Behavior of Marine Fishes Capture Processes and Conservation Challenges Editor
Pingguo He
A John Wiley & Sons, Inc., ...
165 downloads
1638 Views
13MB 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
Behavior of Marine Fishes Capture Processes and Conservation Challenges Editor
Pingguo He
A John Wiley & Sons, Inc., Publication
Behavior of Marine Fishes Capture Processes and Conservation Challenges
Behavior of Marine Fishes Capture Processes and Conservation Challenges Editor
Pingguo He
A John Wiley & Sons, Inc., Publication
Edition first published 2010 © 2010 Blackwell Publishing Ltd. Chapter 8 remains with the U.S. Government. Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical, and Medical business to form Wiley-Blackwell. Editorial Office 2121 State Avenue, Ames, Iowa 50014-8300, USA For details of our global editorial offices, for customer services, and for information about how to apply for permission to reuse the copyright material in this book, please see our website at www.wiley.com/wiley-blackwell. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Blackwell Publishing, provided that the base fee is paid directly to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license by CCC, a separate system of payments has been arranged. The fee codes for users of the Transactional Reporting Service are ISBN-13: 978-0-8138-1536-7/2010. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks
or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Library of Congress Cataloging-in-Publication Data Behavior of marine fishes : capture processes and conservation challenges / editor Pingguo He. – 1st ed. p. cm. Includes bibliographical references and index. ISBN 978-0-8138-1536-7 (hardback : alk. paper) 1. Marine fishes–Behavior. 2. Fishery Management. 3. Marine fishes–Conservation. I. He, Pingguo. QL620.B44 2010 597.177–dc22 2009050959 A catalog record for this book is available from the U.S. Library of Congress. Set in 9.5 on 12 pt Times by Toppan Best-set Premedia Limited Printed in Singapore 1
2010
Contents Contributors
ix
Preface Clement S. Wardle
xi
Introduction Pingguo He
xiii
Part One: Locomotion and Sensory Capabilities in Marine Fish
3
Chapter 1
5
Chapter 2
Chapter 3
Swimming in Marine Fish John J. Videler and Pingguo He 1.1 Introduction 1.2 The Swimming Apparatus 1.3 Swimming-Related Adaptations 1.4 Styles of Swimming 1.5 Interactions between Fish and Water: Fish Wakes 1.6 Energy Required for Swimming 1.7 Swimming Speeds and Endurance 1.8 Concluding Remarks Fish Vision and Its Role in Fish Capture Takafumi Arimoto, Christopher W. Glass, and Xiumei Zhang 2.1 Introduction 2.2 Structure of the Fish Eye 2.3 Visual Function 2.4 Visual Capacity: Visual Acuity, Separable Angle, and Maximum Sighting Distance 2.5 Color and Appearance of Fishing Gear Underwater 2.6 Fish Vision and Its Application in Fish Capture 2.7 Concluding Remarks Hearing in Marine Fish and Its Application in Fisheries Hong Young Yan, Kazuhiko Anraku, and Ricardo P. Babaran 3.1 Introduction 3.2 Properties of Underwater Sound and Vibration 3.3 Underwater Sound Sources and Their Characteristics 3.4 General Morphology and Functions of Inner Ears and Ancillary Structures 3.5 Responses of Fish to Sound and Its Application in Fisheries 3.6 Concluding Remarks
v
5 5 8 12 13 14 18 19 25 25 25 27 32 35 36 40 45 45 45 47 48 53 60
vi
Contents
Part Two: Fish Behavior near Fishing Gears during Capture Processes
65
Chapter 4
67
Chapter 5
Chapter 6
Chapter 7
Chapter 8
Fish Behavior near Bottom Trawls Paul D. Winger, Steve Eayrs, and Christopher W. Glass 4.1 Introduction 4.2 Trawl Gear and Trawl Fisheries 4.3 Fish Behavior in the Pretrawl Zone (Zone 1) 4.4 Fish Behavior between Trawl Doors and in the Net Mouth (Zone 2) 4.5 Fish Behavior inside the Trawl Net and the Codend (Zone 3) 4.6 Factors Influencing Fish Behavior near Trawls 4.7 Concluding Remarks Fish Behavior in Relation to Longlines Svein Løkkeborg, Anders Fernö, and Odd-Børre Humborstad 5.1 Introduction 5.2 Worldwide Longline Fisheries 5.3 Description of the Gear 5.4 Chemoreception and Food Search—The Basis for Bait Fishing 5.5 Interactions between the Fish and the Longline Gear 5.6 Conservation Challenges and Potential Solutions 5.7 Concluding Remarks 5.8 Future Challenges Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues Bjarti Thomsen, Odd-Børre Humborstad, and Dag M. Furevik 6.1 Introduction 6.2 Worldwide Use of Fish Pots 6.3 Fish Behavior in Relation to Pots 6.4 Conservation Challenges and Solutions 6.5 Concluding Remarks Large-scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges Pingguo He and Yoshihiro Inoue 7.1 Introduction 7.2 Trap Fisheries and Trap Designs 7.3 Fish Behavior in and around Traps 7.4 Fish Behavior and Trap Designs 7.5 Size and Species Selectivity and Mortality of Escapees and Discards 7.6 Conservation Issues and Mitigation Measures in Trap Fisheries 7.7 Concluding Remarks Fish Behavior near Gillnets: Capture Processes, and Influencing Factors Pingguo He and Michael Pol 8.1 Introduction 8.2 Capture Mechanisms, Gear Designs, and Fishing Efficiency 8.3 Size Selectivity of Gillnets 8.4 Fish Behavior and Gillnet Fishing 8.5 Measures to Reduce Bycatch and Discards in Gillnets 8.6 Interaction of Marine Mammals, Seabirds, and Sea Turtles with Gillnets 8.7 Derelict Gillnets: Ghost Fishing Problems and Solutions 8.8 Concluding Remarks
67 67 69 75 82 89 95 105 105 106 108 110 114 123 130 132 143 143 143 146 150 154 159 159 159 165 170 172 174 178 183 183 184 187 189 192 195 197 198
Contents Chapter 9
Electric Senses of Fish and Their Application in Marine Fisheries Hans Polet 9.1 Introduction 9.2 Properties of an Electric Field in Water 9.3 The Electric Field 9.4 Application in Marine Fisheries 9.5 Conservation Issues 9.6 Concluding Remarks
Part Three: Contemporary Issues in Capture and Conservation in Marine Fisheries Chapter 10
Chapter 11
Chapter 12
Technical Measures to Reduce Bycatch and Discards in Trawl Fisheries Norman Graham 10.1 Introduction 10.2 Bycatch and Discard in World Fisheries 10.3 Cause of Bycatch and Discard 10.4 Technical Measures to Reduce Bycatch and Discard 10.5 Implementation of Discard Reduction Measures in Trawl Fisheries 10.6 Discussion 10.7 Concluding Remarks Mortality of Animals that Escape Fishing Gears or Are Discarded after Capture: Approaches to Reduce Mortality Petri Suuronen and Daniel L. Erickson 11.1 Introduction 11.2 Mortality of Discards and Escapees 11.3 Assessment of Mortality 11.4 Factors Causing Stress, Injury, and Mortality 11.5 Measures to Improve Survival 11.6 Concluding Remarks Effect of Trawling on the Seabed and Mitigation Measures to Reduce Impact Pingguo He and Paul D. Winger 12.1 Introduction 12.2 Review of Recent Studies on the Seabed Impact of Trawling 12.3 Description of Trawls and Their Operation 12.4 Pelagic and Semipelagic Trawls 12.5 Groundgear Modifications 12.6 Trawl Door Considerations 12.7 Other Trawl Gear Components 12.8 Beam Trawls 12.9 Concluding Remarks
Chapter 13 Measures to Reduce Interactions of Marine Megafauna with Fishing Operations Dominic Rihan 13.1 Introduction 13.2 Species and Fisheries Involved 13.3 Extent of Bycatch 13.4 Nature of the Problem
vii 205 205 206 209 219 228 231
237 239 239 240 242 242 257 258 259 265 265 266 269 276 283 286 295 295 295 296 299 301 305 306 309 310 315 315 316 318 320
viii
Contents 13.5 Regulatory Frameworks 13.6 Potential Mitigation Measures 13.7 Concluding Remarks
321 325 337
Appendix Species Names Mentioned in the Text
343
Index
347
Color Plate Section
Contributors
Norman Graham Marine Institute Rinville, Oranmore Galway, Ireland
Kazuhiko Anraku Faculty of Fisheries, Kagoshima University Kagoshima, Japan Takafumi Arimoto Fish Behavior Section Tokyo University of Marine Science and Technology Tokyo, Japan
Pingguo He University of Massachusetts Dartmouth School for Marine Science and Technology New Bedford, MA, USA
Ricardo P. Babaran College of Fisheries and Ocean Sciences University of the Philippines in the Visayas Miagao, Iloilo, the Philippines
Odd-Børre Humborstad Fish Capture Division Institute of Marine Research Bergen, Norway
Steve Eayrs Gulf of Maine Research Institute Portland, ME, USA
Yoshihiro Inoue Nakano-ku Tokyo, Japan
Daniel L. Erickson Oregon Department of Fish and Wildlife Marine Resources Program Newport, OR, USA
Svein Løkkeborg Fish Capture Division Institute of Marine Research Bergen, Norway
Anders Fernö Department of Biology University of Bergen Bergen, Norway
Michael Pol Massachusetts Division of Marine Fisheries New Bedford, MA, USA Hans Polet Institute for Agricultural and Fisheries Research Oostende, Belgium
Dag M. Furevik Fish Capture Division Institute of Marine Research Bergen, Norway
Dominic Rihan Irish Sea Fisheries Board Dublin, Ireland
Christopher W. Glass Institute for the Study of Earth, Oceans and Space University of New Hampshire Durham, NH, USA
ix
x Petri Suuronen Finnish Game and Fisheries Research Institute Helsinki, Finland Bjarti Thomsen Faroe Marine Research Institute Faroe Islands John J. Videler Dept. of Marine Biology Groningen University The Netherlands Clement S. Wardle Warkbraes, Craigievar Aberdeenshire, UK Paul D. Winger Fisheries and Marine Institute Memorial University of Newfoundland St. John’s, NL, Canada
Contributors Hong Young Yan Institute of Fishery Sciences National Taiwan University Sensory Electrophysiology Laboratory Institute of Cellular and Organismic Biology Academia Sinica (Taiwan National Academy of Science) Jiaoshi, Taiwan Xiumei Zhang College of Fisheries Ocean University of China Qingdao, China
Preface
the different fishing gears have been developed by the fishermen to make use of the details of fish behavior to catch or release the fish. Small fish pass through the open meshes of a trawl because they are soon exhausted when trying to swim with the moving net. Selective trawls work with precise knowledge of this scale effect. They are towed at just above the sustained swimming speed of the size and species of targeted fish. Larger fish, which are not exhausted, swim away when the net is lifted, whereas smaller fish are quickly exhausted and pass through the meshes or over herding ropes and wires. During the past 50 years, large amounts of data on fish behavior and fish capture processes have been collected. Interpretations of these data have generated a wealth of knowledge that has been applied to the catching and conservation of fish. The details of this developing knowledge tend to become scattered in refereed papers in different scientific journals and are often forgotten as time passes. It is important that every few years these papers are reexamined by experts and reinterpreted in relation to the current problems encountered in the world’s marine fisheries. The refereed papers, as collected in this book, give the reader, scientist, or fisherman authoritative views, both retrospective and forward looking, on the issues facing global marine fisheries. Not only does the behavior of target and bycatch species need to be understood, but all species of megafauna, which include many protected species, that may interact with the fish capture processes. The behavioral details must be studied and avoidance procedures developed. It is thus very relevant that an up-to-date collection of
For hundreds of years, nets were considered to sieve or filter the fish from the water. Then in 1952, a breakthrough was made during the first diving observations when fish were filmed near a towed fishing net. Instead of being passively filtered by the net, the fish were orderly swimming ahead of the groundgear of a Danish seine net, being herded through the net funnel and actively escaping through open meshes of the codend. The new technologies and underwater cameras allowed SCUBA divers to observe and record that fish were reacting to nets in many different ways that were related to the biological characteristics of fish (e.g., species, size, sex, and state of maturity), environmental conditions of the fishing ground (e.g., temperature, seasonal changes, day/night length, tide, current, wind, depth, changing color and light levels, clarity of water, and bioluminescence), and operational parameters (e.g., towing speed, colors of twines, gear and ship sounds). With so many variables, two hauls may never be the same, making causes and effects very difficult to investigate with limited tows during experiments at sea. Large aquaria and swimming tunnels were built so that many of these variables could be accurately controlled to allow experiments that measured the physiological limits of fish that were relevant when reacting to fishing gears under a variety of conditions. This research has included issues such as fish swimming capacity, schooling behavior, visual and hearing capability, bait odor distribution and chemoreception, electro senses, and learning and conditioning. Knowing the limits to the response abilities of fish gives us clues to help explain how
xi
xii
Preface
knowledge and a synthesis of fish behaviors, fish capture processes, and conservation issues are presented at a time when conservation is becoming the priority above and beyond the traditional values of marine resources. Currently and in the near future, all harvesting and extraction of our natural resources will need careful studies of the effects on the eco-
system and will involve detailed conservation measures in the capture fishery, including the avoidance or management of subjects like seabed damage, genetic selection due to size-selective fishing, damage to species diversity, and total fishing mortality. Clement S. Wardle
Introduction
stocks while avoiding depleted stocks became one of the major management tools. As a result, the fishing industry is allowed to catch only fish of certain species and sizes and at specific locations and times. Consequently, the aim of understanding fishing gear and fish behavior turned from catching more fish to catching selected fish. Development of selective fishing gears to reduce bycatch and discards became the mainstream of marine capture fishery research. As a result, research on fish behavior, especially fish reaction to fishing gears, has flourished since the 1970s. In a broad sense, fish behavior can be defined as the adaptation of fish to internal and external environments and to natural and artificial stimuli. In a narrower sense, related to fish capture, fish behavior can be considered the reaction of fish to physical and chemical stimuli associated with the fishing gear and its operation and the reaction to environmental changes in the forms of movement and distribution. Recognition of the importance of fish behavior in understanding and improving size and species selectivity and in realizing rational exploitation of the resource has encouraged applied fish behavior studies in the context of fish capture. As one of the few remaining wild resource harvesting activities, fishing operations have increasingly been criticized for collateral effects on other animals and the marine ecosystem. Discards in world marine fisheries represent huge wastes of resources. It has become a moral issue as well as a threat to global food security. Fish population decline and endangerment of some of charismatic
Humans might have started observing and interpreting fish behavior when they first threw rocks at a fish in a stream. Hunting has continued to this day, although increasingly sophisticated, in the wild capture fishery. During the course of fishing for various species under different conditions, a variety of fishing gears and methods have emerged. New materials and technologies have been invented and applied to design and construct gears and to study fish behavior, so as to capture fish more efficiently as well as to conserve fish for future use. Powered fishing boats and hauling equipment, synthetic gear materials, and echo-sounding equipment may be the most influential technologies in commercial marine capture fisheries, while echo-sounders and sonar, underwater cameras, and data storage tags have made great impacts on the study of fish behavior in the field. Modern fishing technologies, coupled with insatiable demands for seafood products, have caused great stress to fishery resources globally, which only a little over a century ago were still considered to be “inexhaustible.” Commercial fishing activity now occurs from the shoreline to depths exceeding several thousand meters, in the remotest of oceans, and even under polar ice caps. Dramatic collapse of some of the world’s major fish stocks, notably Norwegian herring and Chinese greater yellow croaker in the 1970s and Newfoundland northern cod in the early 1990s, stunned many in the fishing industry, fishery research, and management arenas. Because recovery can take a long time, if they do recover (as seen in Norwegian herring), selective fishing for healthy
xiii
xiv
Introduction
megafaunal species related to fishing operations have caused great concern and become a challenge for both the fishing and research communities. This book reviews and summarizes current understanding of fish behavior as it relates to capture processes in world marine fisheries and presents conservation challenges facing the fishing industry and fisheries researchers. It also illustrates potential technical solutions to the issues. The book is divided into three parts: (1) locomotive and sensory modalities relevant to capture; (2) fish behavior near fishing gears; and (3) conservation challenges in marine fisheries. The 13 chapters in the book were written by 22 leading researchers in the field and reviewed by 14 well-known experts, representing a total wisdom of 36 global scientists from 16 countries from Asia, Europe, and North America. The first part of the book, covering the locomotion and sensory capability of fish, consists of three chapters. Chapter 1 (by Videler and He) reviews and summarizes fish swimming mechanisms, styles, and capabilities. Chapter 2 (by Arimoto, Glass, and Zhang) describes fish vision, underwater light and the visual environment, and how and how well a fish can see and visually respond to a moving or stationary object. Chapter 3 (by Anraku, Yan, and Babaran) describes underwater acoustic properties and fish hearing capability, as well as presenting examples of the application of sound and acoustic stimuli in controlling fish behavior. Information presented in these three chapters is frequently been referred to in later chapters when the behavior of fish near fishing gears is described and interpreted. The second part of the book consists of six chapters that describe fish behavior near different fishing gears and the fish capture processes. The major commercial fishing gears included are otter trawl (Chapter 4 by Winger, Eayrs, and Glass), longline (Chapter 5 by Løkkeborg, Fernö, and Humborstad), fish pot (Chapter 6 by Thomsen, Humborstad, and Furevik), large-scale fish traps and setnets (Chapter 7 by He and Inoue), gillnet (Chapter 8 by He and Pol), and the theory and practice of fish catching using electricity (Chapter 9 by Polet). Each chapter starts with a review of the fishery and the fishing gear and describes fishing processes and fish behav-
ior near the fishing gear and then concludes with conservation issues and potential solutions related to the fishing gear type. The third and last part of the book illustrates specific conservation challenges and solutions across fishing gear types and fisheries. The conservation issues reviewed include bycatch and discard in trawl fisheries (Chapter 10 by Graham), mortality of discards and escapees (Chapter 11 by Suuronen and Erickson), seabed effects of trawling (Chapter 12 by He and Winger), and fishery interactions with megafauna species (Chapter 13 by Rihan). We have made great progress during the past 50 years in understanding fish behavior near fishing gears and in applying this knowledge in conservation-oriented fishing gear designs and operations. Yet there are still many things we do not know and we are still facing many challenges. As we enter the second decade of this century, conservation in marine fisheries will be placed in an even more prominent position, as greater attention is focused on the vast ocean, resulting in more rules and regulations to protect and preserve fisheries, ecosystems, and protected species. Further understanding of the behavior of fish and associated animals will be needed to achieve goals of conservation and sustainable utilization of marine fishery resources. This book is intended as a reference for fishery researchers, students, managers, and conservation enthusiasts. It can also be used as a textbook for fishery courses at undergraduate and graduate levels. I am grateful to Dr. Clem Wardle, the worldrenowned fish behaviorist and retired Aberdeen Marine Laboratory researcher who wrote the preface for the book. His insight on the subject is valuable even 10 years after his retirement. I would like to thank all authors for their commitments to writing chapters and their timely completion of the manuscript. I would like particularly to thank the following reviewers for their critical reviews and helpful suggestions: Michael Davis (USA), Steve Eayrs (USA), Michael Fine (USA), Daniel Foster (USA), Emma Jones (New Zealand), Sven-Gunnar Lunnaryd (Sweden), Bob van Marlen (the Netherlands), Henry Milliken II (USA), Barry O’Neill (UK), Michael Pol (USA), Craig Rose (USA), Al Stoner (USA), Anna-Liisa Toivonen
Introduction (Finland), and Timothy Werner (USA). Persistent understanding and encouragement of my commissioning editor at Blackwell Sciences, Justin Jeffryes, are fully appreciated. Last but not least, I am grateful to my wife, Miao, and daughters, Fiona and Joanna, for putting up with
xv
me over the past 3 years while I was writing chapters and editing the book. It would not have been possible without their support and understanding. Pingguo He Raynham, MA, USA
Behavior of Marine Fishes Capture Processes and Conservation Challenges
Part One Locomotion and Sensory Capabilities in Marine Fish
Chapter 1 Swimming in Marine Fish John J. Videler and Pingguo He
1.1 INTRODUCTION After 500 million years of natural selection, fish are extremely well adapted to various constraints set by the aquatic environment in which they live. In the dense fluid medium, they are usually neutrally buoyant and use body movements to induce reactive forces from the water to propel themselves. Animal movements are powered by contracting muscles, and these movements consume energy. The basic principles of fish locomotion are used by approximately 25,000 extant species. The variation in swimming styles, within the limits of these principles, is great. Swimming includes steady swimming at various speeds, accelerating, braking, maneuvering, jumping, diving downward, and swimming upward. Swimming behavior is different for every species and changes in each individual during growth from larva to adult. Speed, agility, and endurance maxima determine the chances for survival to a considerable extent. Peak performance in absolute terms is positively related to temperature and body length. Ultimately, performance affects the evolutionary fitness of each individual and is a significant factor that is directly related to capture by or escape from both active (e.g., trawls and seines) and passive (e.g., gillnet, longlines, and traps) fishing gear. This chapter reviews the current knowledge about fish swimming mechanisms and abilities to provide a background for discussions in later chapters.
fish into two lateral halves, and lateral longitudinal muscles that are segmentally arranged in blocks, or myotomes. The vertebral column is laterally highly flexible and virtually incompressible longitudinally. Consequently, contraction of the muscles on one side of the body bends the fish, and waves of curvature along the body can be generated by series of alternating contractions on the left and the right side (Videler and Wardle 1991). Fish vertebrae are concave fore and aft (amphicoelous) and fitted with a neural arch and spine on the dorsal side. In the abdominal region, lateral projections are connected with the ribs enclosing the abdominal cavity. The vertebrae in the caudal region bear a hemal arch and spine. Neural and hemal spines point obliquely backward. The number of vertebrae varies greatly among species—European eels (Anguilla anguilla) have 114 vertebrae, and the numbers in the large order of Perciformes vary between 23 and 40 (Ford 1937). The number is not necessarily constant within a species. Atlantic herring (Clupea harengus), for example, can have between 54 and 58 vertebrae (Harder 1975a, 1975b). The end of the vertebral column is commonly adapted to accommodate the attachment of the tail fin. Several vertebrae and their arches and spines are partly rudimentary and have changed shape to contribute to the formation of platelike structures providing support for the fin rays of the caudal fin. Most fish species have unpaired dorsal, caudal, and anal fins and paired pectoral and pelvic fins. Each fin is powered by intrinsic musculature. The lateral muscles are usually the main target of the fishing industry. Relatively short lateral muscle
1.2 THE SWIMMING APPARATUS Fish are aquatic vertebrates with a skull, a vertebral column supporting a medial septum that divides the
5
6
Locomotion and Sensory Capabilities in Marine Fish
Figure 1.1. (A) The myotomes and myosepts on the left side of the king salmon. Myotomes have been removed at four places to reveal the complex three-dimensional configuration of the lateral muscles. [Redrawn after Greene and Greene (1913).] (B) A cross section through the upper left quarter of the caudal region of a salmon. Red muscle fibers are situated in the dark area near the outside. The lines represent the myosepts between the complex myotomes. [Based on Shann (1914).]
fibers are packed into myotomes between sheets of collagenous myosepts. The myotomes are cone shaped and stacked in a segmental arrangement on both sides of the median septum (Fig. 1.1A). In cross sections through the caudal region, the muscles are arranged in four compartments. On each side is a dorsal and a ventral compartment; in some groups, they are separated by a horizontal septum. The left and right halves and the dorsal and ventral moieties are mirror images of each other. In cross sections, the myosepts are visible as more or less concentric circles of collagen. The color of the muscle fibers may be red, white, or intermediate in different locations in the myotomes (Fig. 1.1B), which was first described by Lorenzini in 1678 (Bone 1966). Red fibers are usually situated directly under the skin. The deeper white fibers form the bulk of lateral muscles, and in some species intermediately colored pink fibers are found between the two. The red fibers are slow but virtually inexhaust-
ible and their metabolism is aerobic. They react to a single stimulus owing to the high density of nerve terminals on the fibers. The white fibers are fast, exhaust quickly, and use anaerobic metabolic pathways. White fibers are either focally or multiply innervated. Pink fibers are intermediate in most aspects. The red muscles of some large tuna and shark species are positioned well inside the white muscle mass, an arrangement that can increase the muscle temperature by as much as 10°C during swimming (Carey and Teal 1969). The final paragraph describes how this halves the twitch contraction time of the white muscles and doubles the maximum swimming speed. Fish fins consist of two layers of skin, usually supported by fin rays that are connected to supporting skeletal elements inside the main body of the fish. Intrinsic fin muscles find their origin usually on the supporting skeleton and insert on the fin rays. The fins of elasmobranchs (sharks and rays) are
Swimming in Marine Fish
7
Figure 1.2. The structure of a typical teleost fin ray. (A) Dorsal or ventral view: the left and right fin ray halves are each other’s mirror image. (B) Lateral view; the size of the bony elements decreases to the right after each of the bifurcations. The position of the bifurcations in the various branches does not show a geometrically regular pattern. (C) Longitudinal section through the bony elements of the fin ray at a position indicated in (A). Note that a joint with densely packed collagenous fibers connects the elements. The collagenous fibers connecting the fin ray halves have a curly, serpentine appearance. [From Videler (1993).]
permanently extended and rather rigid compared with those of teleosts (bony fish). Elasmobranch fin rays consist of rows of longitudinally connected small pieces of cartilage in a juxtaposed arrangement. Intrinsic muscles on both sides of the rows running from the fin base to the edge bend these fins. Teleost fins can be spread, closed, and folded against the body. There are two kinds of teleost fin rays: spiny, stiff unsegmented rays and flexible segmented ones. Spiny rays stiffen the fin and are commonly used for defense. The flexible rays (Fig. 1.2) play an important role in adjusting the stiffness and camber of the fins during locomotion. They consist of mirror-image halves, each of which has a skeleton of bony elements interconnected by collagenous fibers. Muscles pulling harder on the fin ray head on one side will shift the two halves with respect to each other and bend the ray or increase the stiffness against bending forces from the water. In contrast to elasmobranch fins, there are no muscles on the fin itself. The body shape of fish may vary greatly among species, but the best pelagic swimmers have a common form. Their bodies are streamlined, with gradually increasing thickness from the point of the snout to the thickest part at about one-third of the length. From that point, the thickness gradually decreases toward the narrow caudal peduncle. A
moving body in water encounters friction and pressure drag. Friction drag is proportional to the surface area, and pressure drag is proportional to the area of the largest cross section. A spherical body has the lowest friction for a given volume; a needle-shaped body encounters minimal amounts of pressure drag. An optimally streamlined body is a hybrid between a sphere and a needle and offers the smallest total drag for the largest volume. It has a diameter-to-length ratio between 0.22 and 0.24. The best pelagic swimmers have near-optimal thickness-to-length ratios (Hertel 1966). The mechanically important part of fish skin is the tissue (the stratum compactum) underneath the scales, which consists of layers of parallel collagenous fibers (see Videler 1993 for a review). The fibers in adjacent layers are oriented in different directions, and the angles between the layers vary between 50 and 90 degrees, but the direction in every second layer is the same. The packing of layers resembles the structure of plywood, except that in the fish stratum compactum there are also radial bundles of collagen connecting the layers; the number of layers varies between 10 and 50. In each layer, the fibers follow left- and right-handed helices over the body surface. The angle between the fibers and the longitudinal axis of the fish decreases toward the tail. In some species, the
8
Locomotion and Sensory Capabilities in Marine Fish
stratum compactum is firmly connected to the myosepts in the zone occupied by the red muscle fibers; in other fish, there is no such connection. The strongest fish skins that have been tested are those of eel and shark. Values of Young’s modulus (the force per unit cross-sectional area that would be required to double the length) of up to 0.43 GPa (1 GPa = 109 N/m2) have been measured. This is about one-third of the strength of mammalian tendon, for which values of 1.5 GPa have been measured. Scales are usually found at the interface between fish and water. Several swimming-related functions have been suggested. Scales might serve to prevent transverse folds on the sides of strongly undulating fish, keeping the outer surface smooth. Spines, dents, and tubercles on scales are usually arranged to form grooves in a direction of the flow along the fish. Roughness due to microstructures on scales in general creates small-scale turbulence, which could delay or prevent the development of dragincreasing large-scale turbulence (Aleyev 1977). Fish mucus covering the scales is supposed to reduce friction with the water during swimming. This assumption is based on the idea that mucus shows the “Toms effect” (Lumley 1969), which implies that small amounts of polymers are released that preclude sudden pressure drops in the passing fluid. Measurements of the effects of fish mucus on the flow show contradictory results varying from a drag reduction of almost 66% (Pacific barracuda, Sphyraena argentea) to no effect at all (Pacific bonito, Sarda chiliensis) (Rosen and Cornford 1971). Experiments with rainbow trout showed that mucus increases the thickness of the boundary layer
(Daniel 1981). The layer of water around the fish is affected by the presence of the fish during swimming. Viscosity causes a layer of water close to the fish to travel with it at the speed of the fish. There is a gradient of decreasing water velocities in a direction away from the surface of the fish. The thickness of the boundary layer is defined as the distance from the surface of the fish to where the water is no longer dragged along. A thick boundary layer implies that the gradient is gradual, which reduces viscous friction. However, the penalty for a thicker boundary layer is that the fish has to drag along a larger amount of water. The conclusion might be that the effect of mucus is beneficial during slow-speed cruising but detrimental during fast swimming and acceleration (Videler 1995). 1.3 SWIMMING-RELATED ADAPTATIONS Some fish species are adapted to perform some aspect of locomotion extremely well, whereas others have a more general ability to move about and are specialized for different traits not related to swimming. Generalists can be expected to have bodies that give them moderately good performance in various special functions. Specialists perform exceptionally well in particular skills. Fast accelerating, braking, high-speed cruising, and complex maneuvering are obvious examples. The special swimming adaptations shown in Figure 1.3 are only a few of a wealth of possible examples. A closer study of the swimming habits of a large number of species will show many more specialist groups than the dozen or so described here (e.g., Lindsey 1978).
Figure 1.3. A representation of fish swimming specializations. See text for detailed explanation. (a) pike (Esox lucius), barracuda (Sphyraena argentea); (b) forkbeard (Phycis phycis), saithe (Pollachius virens); (c) bluefin (Thunnus thynnus), porbeagle shark (Lamna nasus); (d) angelfish (Pterophyllum scalare), butterfly fish (Chaetodon sp.); (e) sunfish (Mola mola), opah (Lampris guttatus), louvar (Luvarus imperialis); (f) swordfish (Xiphias gladius), sailfish (Istiophorus platypterus); (g) turbot (Scophthalmus maximus), skate (Raja batis); (h) moray eel (Muraena Helena), eel (Anguilla anguilla); (i) rainbow wrasse (Coris julis), sand eel (Ammodytes tobianus); (j) sea-horse (Hippocampus ramulosus), Atlantic flying fish (Cheilopogon heterurus), hatchet fish (Gasteropelecus sp.); cornet fish (Fistsularia sp.). [Modified from Videler (1993).]
h a
Pikc
Moray
Barracuda
Eel b
Forkbeard
Saithe
i c
Bluefin tuna
Porbeagle Rainbow wrasse
Sandeel
Anglefish d
j Butterflyfish Sea-horse
e Opah
Sunfish Louvar
e
k
Atlantic flying fish
f
Swordfish
Sailfish
Hatched fish
g
Turbot
Skate
l Cornet fish
9
10
Locomotion and Sensory Capabilities in Marine Fish
Specialists in accelerating, such as the pike (Esox lucius) and the barracuda (Fig. 1.3a), are often ambush predators. They remain stationary or swim very slowly until a potential prey is within striking distance. These species have a reasonably streamlined body and large dorsal and anal fins positioned extremely rearward, close to the caudal fin. Acceleration during the strike is achieved by the first two beats of the tail, the effect is enlarged by the rearward position of the dorsal and anal fins. The relative skin mass of the pike is reduced, compared with that of other fish, increasing the relative amount of muscles and decreasing the dead mass that has to be accelerated with the fish at each strike. Maximum acceleration rates measured for pike vary between 40 and 150 m/s2, which is 4 to 15 times the acceleration due to gravity (G = 9.8 m/s2). The highest peak acceleration value reported for pike is 25 times gravity (Harper and Blake 1990, 1991). Braking is difficult while moving in a fluid medium. Fish use the unpaired fins and tail, usually in combination with the pectoral and pelvic fins, for braking. Gadoids with multiple or long unpaired fins are good at braking. Forkbeard (Phycis phycis) swims fast and close to the bottom with elongated pelvic fins extended laterally for the detection of bottom-dwelling shrimp. The fish instantly spreads the long dorsal and anal fins and throws its body into an S-shape when a prey item is touched. In the process, the fin rays of the tail fin are actively bent forward. Braking is so effective that the shrimp has not yet reached the caudal peduncle before the fish has stopped and turned to catch it. The highest deceleration rate measured is 8.7 m/s2 for saithe (Pollachius virens) (Geerlink 1987). The contribution of the pectorals to the braking force is about 30%; the rest comes from the curved body and extended median fins (Geerlink 1987) (Fig. 1.3b). Cruising specialists (Fig. 1.3c) migrate over long distances, swimming continuously at a fair speed. Many are found among scombrids and pelagic sharks, for example. Cruisers have highly streamlined bodies, narrow caudal peduncles with lateral keels, and high-aspect-ratio tails (aspect ratio is the tail height squared divided by tail surface area). The bluefin tuna (Thunnus thynnus), for example, is an extreme long distance swimmer, crossing the
Atlantic twice a year. The body dimensions are very close to the optimum values, with a thickness-tolength ratio near 0.25 (Hertel 1966). Cruising speeds of 3-m-long bluefin tunas measured in large enclosures reached 1.2 L/s (260 km/d), where L equals body length (Wardle et al. 1989). Angelfish (Pterophyllum scalare) and butterfly fish (Chaetodon sp.) (Fig. 1.3d) are maneuvering experts with short bodies and high dorsal and anal fins. Species of this guild live in spatially complex environments. Coral reefs and freshwater systems with dense vegetation require precise maneuvers at low speed. Short, high bodies make very short turning circles possible. Angelfish make turns with a radius of 0.065 L (Domenici and Blake 1991). For comparison, the turning radius of a cruising specialist is in the order of 0.5 L, an order of magnitude larger. Sunfish (Mola mola), opah (Lampris guttatus), and louvar (Luvarus imperialis) are among the most peculiar fish in the ocean (Fig. 1.3e). They look very different from each other but they all have large body sizes. The sunfish reaches 4 m and 1500 kg, the opah may weigh up to 270 kg, and the louvar is relatively small with a maximum length of 1.9 m and weight of 140 kg. While little is known about the mechanics of their locomotion, they all seem to swim slowly over large distances. The opah uses its wing-shaped pectorals predominantly, and the louvar has a narrow caudal peduncle and an elegant high-aspect-ratio tail similar to those of the tunas. The sunfish has no proper tail, but the dorsal and ventral fins together form an extremely highaspect-ratio propeller. Sunfish swim very steadily, moving the dorsal and ventral fins simultaneously to the left side and, half a cycle later, to the right side. The dorsal and ventral fins have a cambered wing profile in cross section with a rounded leading edge and a sharp trailing edge. The intrinsic fin muscles fill the main part of the body and insert on separate fin rays, enabling the sunfish to control the movements, camber, and profile of its fins with great precision. Although there are no measurements to prove this as yet, it as appears that these heavy species specialize in slow steady swimming at low cost. Inertia helps them to keep up a uniform speed, while their well-designed propulsive fins generate just enough thrust to balance the drag as efficiently as possible.
Swimming in Marine Fish Swordfishes (Xiphiidae) and billfishes (Istiophoridae) (Fig. 1.3f) show bodily features unique to these fish—the extensions of the upper jaws, the swords, and the shape of the head. They are probably able to swim briefly at speeds exceeding those of all other nektonic animals, reaching values of well over 100 km/h (Barsukov 1960). The sword of swordfish is dorsoventrally flattened to form a long blade of up to 45% of the body length. The swords of billfish (including sailfish, spearfish, and marlin) are pointed spikes, round in cross section and shorter (between 14% and 30% in adult fish, depending on species) than those of swordfish. All swords have a rough surface, especially close to the point. The roughness decreases toward the head. One other unique bodily feature of the swordbearing fishes is the concave head. At the base of the sword, the thickness of the body increases rapidly with a hollow profile up to the point of greatest thickness of the body. The rough surface on the sword reduces the thickness of the boundary layer of water dragged along with the fish (Hertel 1966). This reduces drag. The concave head probably serves to avoid drag-enhancing large-scale turbulence. The caudal peduncle is dorsoventrally flattened, fitted with keels on both sides. These features and the extremely high-aspect-ratio tail blades with rearward-curved leading edges are hallmarks of very fast swimmers. The shape of the body of flatfish (Pleureonectiformes) and rays (Rajiformes) (Fig. 1.3g) offers the opportunity of hiding in the boundary layer of the seabed where speeds of currents are reduced. There is another possible advantage connected with a flat body shape. Both flatfish and rays can be observed swimming close to the bottom. These fish are negatively buoyant and, like flying animals, must generate lift (a downwash in the flow) at the cost of induced drag to remain “waterborne.” Swimming close to the ground could reduce the drag due to lift generation considerably, depending on the ratio between height off the ground and the span of the “wings” (Anderson and Eberhardt 2001). Only a few species can swim both forward and backward. Eels (Anguillidae), moray eels (Mutaenidae), and conger eels (Congeridae) (Fig. 1.3h) can quickly reverse the direction of the propulsive wave on the body and swim backward. The
11
common feature of these fish is the extremely elongated flexible body without a high tail fin. Swimming is usually not very fast; they prefer to swim close to the bottom and operate in muddy or maze-type environments. Rainbow wrasses (Coris julis) and sand eels (Ammodytes tobianus) (Fig. 1.3i) sleep under a layer of sand—sand eels in daytime and rainbow wrasse during the night. Both species swim headdown into the sand using high-frequency lowamplitude oscillations of the tail. If the layer of sand is thick enough, the speed is not noticeably reduced. Body shapes are similar—that is, slender with a well-developed tail. The wrasses use their pectoral fins for routine swimming and move body and tail fin during escapes and to swim into the sand (Videler 1988). Most neutrally buoyant fish species are capable of hovering in one spot in the water column. Some species can hardly do anything else. Seahorses (Fig. 1.3j) and pipefishes (Syngnathidae) rely on camouflage for protection from predators. They are capable of minute adjustments of the orientation of their body using high-frequency, low-amplitude movements of the pectoral and dorsal fins. Seahorses are the only fish with a prehensile tail. Flying fish (Exocoetidae) (Fig. 1.3k) have exceptionally large pectoral fins to make gliding flights out of the water when chased by predators. Some species are four-winged because they use enlarged pelvic fins as well. The lower lobe of the caudal fin is elongated and remains beating the water during takeoff. Hatchet fishes (Gasteropelecus sp.) actually beat their pectoral fins in powered flight. The pectoral fins have extremely large intrinsic muscles originating on a greatly expanded pectoral girdle. Many additional species occasionally or regularly leap out of the water but they are not specially adapted to fly. Cornet fishes (Fistularia sp.) (Fig. 1.3l) are predators of small fish in the littoral of tropical seas, most often seen above sea grass beds or sandy patches between coral reefs. They seem to have two tails; the first one is formed by a dorsal and anal fin, and the second one is the real tail. Beyond the tail is a long, thin caudal filament. These fish hunt by dashing forward in one straight line without any side movements of the head, using large-amplitude
12
Locomotion and Sensory Capabilities in Marine Fish
strokes of the two tail fins and the trailing filament (J.J.V., personal observations). It looks as though the double–tail fin configuration with the trailing filament serves to allow fast acceleration without recoil movements of the head. Precise kinematic measurements are needed to provide evidence for this assumption. 1.4 STYLES OF SWIMMING Most fish species swim with lateral body undulations running from head to tail; a minority use the movements of appendages to propel themselves. The waves of curvature on the bodies of undulatory swimmers are caused by waves of muscle activations running toward the tail with a 180-degree phase shift between the left and the right side (Videler and Wardle 1991). The muscular waves run faster than the waves of curvature, reflecting the interaction between the fish’s body and the reactive forces from the water. The swimming speed varies between 0.5 and 0.9 times the backward speed of the waves of curvature during steady swimming (Videler and Hess 1984). The wavelength of the body curvature of slender eel-like fish is about 0.6 L, indicating that there is more than one wave on the body at any time. Fast-swimming fish such as mackerel and saithe have almost exactly one complete wave on the body, and on short-bodied fish such as carp and scup, there is less than one wave on the length of the body during steady swimming. The maximum amplitude (defined as half the total lateral excursion) may increase toward the tail linearly, as in eels and lampreys, or according to a power function in other species (Wardle et al. 1995). The increase in maximum amplitude is concentrated in the rearmost part of the body in fast fish like tuna. The maximum amplitude at the tail is usually in the order of 0.1 L with considerable variation around that value. The period of the waves of curvature determines the tail beat frequency, which is normally linearly related to the swimming speed (Bainbridge 1958). The distance covered per tail beat is the “stride length” of a fish. It varies greatly between species but also for each individual fish. Maximum values of more than one body length have been measured for mackerel; the least distance covered per beat of the same individual was reported to be 0.7 L (Videler and Hess 1984; Wardle and He
1988). Many species reach values between 0.5 and 0.6 L during steady swimming bouts. Left–right undulations of the body from head to tail are used by fish varying in body shape from eel to tuna. However, the amplitudes of the undulations and the way in which they are generated greatly differ among species. In eel-like fishes, there is a fairly tight connection between the position of the backward-moving bend and muscle shortening on the concave side of the bend. One-sided waves of muscle contraction run from head to tail, causing the local bending of the body with a slight delay. There is always more than one complete wavelength present on the body of these fish between the head and tail (Wardle et al. 1995). Mackerel-type fishes, on the other hand, have only one wavelength on the body at all times, and the lateral muscles on one side of the body from head to tail are activated simultaneously (Wardle and Videler 1993). Simultaneous muscle activity would result in a C-shaped lateral bending of the body if it took place outside the water. However, during swimming, water reacts more strongly to sideward movements of the high tail than to the movements of the rounded side of the rest of the body. Lateral muscles in the anterior part of the body actually shorten when activated, but muscles in the rear part and especially in the caudal peduncle are active when they are being stretched by the sideward push of the water against the tail. Muscles there produce negative work by resisting being stretched. This type of muscle activity provides higher forces and power per unit crosssectional area. Therefore, the rear part of these fishes can be narrower (to contribute to a streamlined body) and still transmit the high forces from the anterior muscle mass to the tail blade (Wardle and Videler 1993; Wardle et al. 1995). Fish locomotion using paired and median fins was reviewed by Blake (1983). Swimming with appendages includes pectoral fin swimming and median fin propulsion. Pectoral fin movements of, for example, labrids (Labridae), shiner perches (Cymatogaster aggregata), and surfperches (Embiotocidae) make an elegant impression. The beat cycle usually consists of three phases. During the abduction phase, the dorsal rays lead the movement away from the body and downward. The adduction phase brings the fin back to the body surface led by
Swimming in Marine Fish horizontal movement of the dorsal rays. During the third phase, the dorsal rays rotate close to the body back to their initial position. Stride lengths vary with speed and may reach more than one body length at optimal speeds. Undulations of long dorsal and anal fins can propel fish forward and backward and are used in combination with movements of the pectoral fins and the tail. There is usually more than one wave on each fin (e.g., up to 2.5 waves on the long dorsal fin of the African electric eel, Gymnarchus niloticus). 1.5 INTERACTIONS BETWEEN FISH AND WATER: FISH WAKES Every action of the fins or the body of a fish will, according to Newton’s third law, result in an equal
13
but opposite reaction from the surrounding water. A swimming fish produces forces in interaction with the water by changing water velocities locally. The velocity gradients induce vortices, being rotational movements of the fluid. Vortices either may end at the boundary of the fluid or may form closed loops or vortex rings with a jet of water through the center (Videler et al. 2002). Furthermore, vortex rings can merge to form chains (Fig. 1.4). Quantitative flow visualization techniques have been successfully applied to reveal the flow patterns near fish using body undulations to propel themselves (Müller et al. 1997). The interaction between undulating bodies and moving fins and the water results in complex flow patterns along and behind the swimming animals (Videler 1993). A schematic
Figure 1.4. Schematic drawings of vortex ring structures. (A) A single vortex ring. The ring shaped centre of rotation is drawn as a line. The rotations of the vortex ring structure draw a jet of water through the centre of the ring, indicated by the arrow. (B) A 3-D reconstruction of a chain of three connected vortex rings. A resulting jet of fluid undulates through the centre of the vortex rings building the chain. [From Videler et al. (2002).]
14
Locomotion and Sensory Capabilities in Marine Fish
Figure 1.5. Artist’s impression of the flow behind a steadily swimming saithe. The tail blade is moving to the left and in the middle of the stroke. At the end of each half-stroke a column vortex is left behind when the tail blade changes direction. Tail tip vortices are shed dorsally and ventrally when the tail moves from side to side. Together the vortices form a chain of vortex rings (as shown schematically in Fig. 1.4) with a jet of water winding through the centers of the rings in the opposite swimming direction. [From Videler (1993).]
three-dimensional impression of the wake generated by the tail behind a steadily swimming fish is shown in Figure 1.5. This shows the dorsal and ventral tail tip vortices generated during the tail beat as well as the vertical stop–start vortices left behind by the trailing edge of the tail at the end of each half-stroke. During a half-stroke, there is a pressure difference between the leading side of the fin and the trailing side. Dorsal and ventral tip vortices represent the water escaping at the fin tips from the leading side, with high pressure to the trailing side where the pressure is low. At the end of the halfstroke, the tail changes direction and builds up pressure on the opposite side of the fin, leaving the previous pressure difference behind as a vertical vortex column. These vertical, dorsal, and ventral vortex systems form a chain through which a jet of water undulates opposite the swimming direction. If we concentrate on what happens in a mediofrontal plane through the fish and the wake, we expect to see left and right stop–start vortices with an undulating backward jet between them. The rotational sense should be anticlockwise on the right of the fish and clockwise on the left. Visualizations of the flow in the mediofrontal plane of swimming fish
reveal that this picture of the wake is correct (Fig. 1.6) (Müller et al. 1997). Flow patterns around fish using paired and unpaired fins, as well as the tail and those near maneuvering fish, are much more complex systems of jets and vortex rings. Such complex flow patterns have been published for a number of species during recent years (Lauder and Tytell 2006). 1.6 ENERGY REQUIRED FOR SWIMMING Swimming fish use oxygen to burn fuel to power their muscles. Carbohydrates, fat, and proteins are the common substrates. A mixture of these provides about 20 J/ml oxygen used (Videler 1993). Measurements of energy consumption during swimming are mainly based on records of oxygen depletion in a water tunnel respirometer. Respiration increases with swimming speed, body mass, and temperature and varies considerably among species. The highest levels of energy consumption measured in fish are about 4 W/kg (Videler 1993). Fast, streamlined fish can increase their metabolic rates up to 10 times resting levels during swimming at the highest sustainable speeds. Short-burst speeds
Swimming in Marine Fish
15
Figure 1.6. The wake of a continuously swimming mullet. The arrows represent the flow velocity in mm/s scaled relative to the 10 mm/s bar on the bottom. The shaded circles indicate the centers of the column vortices. The picture represents a horizontal cross-section half way down the tail through the wake drawn in Figure 1.5. [Based on Müller et al. (1997).]
powered by anaerobic white muscles can cost as much as 100 times resting rates. Most of the energy during swimming at a constant speed is required to generate sufficient thrust to overcome drag. The drag on a steadily swimming fish is proportional to the square of the swimming speed—the energy required increases as the cube of the speed. In other words, if a fish wants to swim twice as fast, it will have to overcome four times as much drag and use eight times as much energy. A fair comparison of the energy used requires standardization of the speed at which the comparison is made. The energetic cost of swimming is the sum of the resting or standard metabolic rate and the energy required to produce thrust. Expressed in watts (joules per second), it increases as a J-shaped curve with speed in m/s (Fig. 1.7) (Videler 1993). The exact shape of the curve depends mainly on the species, size, temperature, and condition of the fish. Owing to the shape of the curve, there is one optimum speed at which the ratio of metabolic rate over speed reaches a minimum. This ratio represents the amount of work a fish has to do to cover 1 m (J/s divided by m/s). To make fair comparisons possible, the
optimum speed (Uopt), where the amount of energy used per unit distance covered is at a minimum, is used as a benchmark. Series of measurements of oxygen consumption at a range of speeds provide the parameters needed to calculate Uopt and the energy used at that speed. The energy values are normalized by dividing the active metabolic rate at Uopt (in W = J/s = Nm/s) by the weight of the fish (in N) multiplied by Uopt (in m/s), to reach a dimensionless number for the cost of transport. Hence, COT represents the cost to transport one unit of weight over one unit of distance at Uopt. Available data show that Uopt is positively correlated with mass (M) and proportional to M0.17. The value of Uopt decreases, however, with mass if Uopt is expressed in L/s and is proportional to M−0.14. While there is great variation in measured Uopt, 2 L/s can serve as a reasonable first estimate of the optimum speed in fish. At Uopt the COT values are negatively correlated with body mass (M−0.38) (Fig. 1.8). Fish use an average of 0.07 J/N to swim their body length at Uopt. If the weight and the size of the animals are taken into account as well by calculating the energy needed to transport the body weight
16
Locomotion and Sensory Capabilities in Marine Fish
Figure 1.7. A theoretical curve of the rate of work as a function of swimming speed. SMR is the standard or resting metabolic rate at speed 0. The amount of work per unit distance covered (J/m) is at a minimum at Uopt. [From Videler (1993).]
over the length, the amount of energy used to swim at Uopt increases in proportion to M0.93 (Fig. 1.9) (Videler 1993). Burst-and-coast (or kick-and-glide) swimming behavior is commonly used by several species (Weihs 1974; Videler and Weihs 1982). It consists of cyclic bursts of swimming movements followed by a coast phase in which the body is kept motionless and straight. The velocity curve in Figure 1.10 shows how the burst phase starts off at an initial velocity (Ui) lower than the average velocity (Uc). During a burst, the fish accelerates to a final velocity (Uf), higher than Uc. The cycle is completed when velocity Ui is reached at the end of the deceleration during the coast phase. Energy savings in the order of 50% are predicted if burst-and-coast swimming is used during slow and high swimming speeds instead of steady swimming at the same average speed (Videler and Weihs 1982). The model predictions are based on the assumption that there is a three-fold difference in drag between a rigid gliding body and an actively moving fish.
COT
102
1
Undulatory P t l Pectoral Tuna Shark Salmon Larvae
10-2 10-4
10-2
1
Body mass (kg) Figure 1.8. Doubly logarithmic plot of dimensionless COT, being the energy needed to transport one unit of mass over one unit of distance during swimming at Uopt, related to body mass. The connected points indicate series of measurements of animal groups indicated separately; “undulatory” and “pectoral” refer to measurements of fish using body plus tail and pectoral fins respectively, for propulsion. [From Videler (1993).]
Swimming in Marine Fish
17
Figure 1.9. Doubly logarithmic plot of the energy needed by a swimming fish to transport its body weight over its body length as a function of body mass. Symbols as in Figure 1.7. [Based on Videler (1993).]
Figure 1.10. Part of a velocity curve during burst-and-coast swimming of cod. The average speed Uc was 3.2 L/s. The initial speed and the final velocity of the acceleration phase are indicated as Ui and Uf, respectively. [From Videler (1993).]
Schooling behavior probably has energy-saving effects (Weihs 1973). As seen in Figure 1.5, the wake of a steadily swimming fish shows an undulating jet of water in the opposite swimming direction through a chain of vortex rings. Just outside this system, water will move in the swimming direction. Theoretically, following fish could make use of this
forward component to facilitate their propulsive efforts (Weihs 1973). One would expect fish in a school to swim in a distinct three-dimensional spatial configuration in which bearing and distance among school members showed a distinct constant diamond lattice pattern and a fixed phase relationship among tail beat frequencies. This has not,
18
Locomotion and Sensory Capabilities in Marine Fish
however, been confirmed by actual observations. On the other hand, energetic benefits for school members have been confirmed by indirect evidence. It has been observed that the tail beat frequency of schooling Pacific mackerel (Scomber japonicus) is reduced compared with solitary mackerel swimming at the same speed (Fields 1990). In schools of sea bass, trailing individuals used 9% to 14% lower tail beat frequencies than fish in leading position. There is also some evidence showing that fast swimming fish in a school use less oxygen than the same number of individuals would use in total in solitary swimming at the same speed (Herskin and Steffensen 1998). 1.7 SWIMMING SPEEDS AND ENDURANCE The relationship between swimming speed and endurance is not straightforward due to the separate use of red, intermediate, and white muscle. Virtually inexhaustible red muscles drive slow cruising speeds; burst speeds require all-out contraction of
white muscles lasting only a few seconds. Endurance decreases rapidly when speeds are above cruising speeds. Maximum swimming speeds of fish are ecologically important for obvious reasons. However, slower swimming speeds and the stamina at these speeds represent equally important survival values for a fish. Figure 1.11 relates swimming speed, endurance, and the cost of swimming for a 0.18-m sockeye salmon (Oncorhynchus nerka) at 15°C (Videler, 1993). At low speeds, this fish can swim continuously without showing any signs of fatigue. The optimum speed Uopt is between 1 and 2 L/s. Limited endurance can be measured at speeds higher than the maximum sustained speed (Ums) in this case, somewhat less than 3 L/s. For these prolonged speeds, the logarithm of the time to fatigue (endurance) decreases linearly with increasing velocity up to the maximum prolonged speed (Ump) where the endurance is reduced to a fraction of a minute. Along this endurance trajectory, the fish will switch gradually from partly aerobic to totally anaerobic metabolism. The maximum burst
Figure 1.11. The metabolic rate (linear scale) and the endurance (logarithmic scale) of a 0.18-m, 0.05-kg sockeye salmon as functions of swimming speed in L per second. The water temperature was 15°C. The optimum swimming speed (Uopt), the maximum sustained speed (Ums), the maximum prolonged speed (Ump) and an estimate of the maximum burst speed (Umax) are indicated. [From Videler (1993).]
Swimming in Marine Fish speed in this case is in the order of 7 L/s for sockeye salmon (Brett 1964). A comparison of published data for some marine species reveals that values for Ums for fish varying in size between 10 and 49 cm are between 0.9 and 9.9 L/s, with larger fish achieving less Ums in L/s (Table 1.1). For example, 17-cm-long haddock (Melanogrammus aeglefinus) are capable of swimming at 2.6 L/s for longer than 200 min, but 41-cmlong haddock can only swim at 1.5 L/s for the same period of time (Breen et al. 2004). Bottom-dwelling demersal fish living in complex environments usually have shallower endurance curves than pelagic long-distance swimmers, which fatigue more quickly when they break the limit of the maximum sustained speed (Videler 1993). Endurance in fish swimming at prolonged speeds is limited by the oxygen uptake capacity. Higher speeds cause serious oxygen debts. The maximum burst speed in m/s increases with body length (Fig. 1.12). Average relative values for adult fish are between 10 and 20 L/s. Small fish larvae swim at up to 50 L/s during startle response bursts (Fuiman 1986). Speed record holders in m/s are to be found among the largest fish. Unfortunately, reliable measurements are usually not available. The maximum burst speed of fish depends on the fastest twitch contraction time of the white lateral muscles (Wardle 1975). For each tail beat, the muscles on the right and on the left have to contract once. Hence, the maximum tail beat frequency is the inverse of twice the minimum contraction time. The burst speed is found by multiplying the stride length by the maximum tail beat frequency. Muscle twitch contraction times halve for each 10°C temperature rise, and the burst speed doubles (Videler 1993). Larger fish of the same species have slower white muscles than smaller individuals. The burst swimming speed in L/s decreases with size with a factor of on average 0.89 for each 10 cm length increase. Estimates based on muscle twitch contraction times and measured stride length data for 2.26 m long blue fin tuna vary between 15 and 23 m/s (Wardle et al. 1989). Estimates for 3 m long swordfish exceed 30 m/s (Barsukov 1960). Measured values for burst speeds are difficult to find. The maximum swimming speed in terms of L/s ever recorded in captivity is that of a 30 cm mackerel
19
swimming at 18 L/s (or 5.4 m/s, Wardle and He 1988). At that speed, the tail beat frequency was 18 Hz and the stride length was 1 L.
1.8 CONCLUDING REMARKS Understanding fish swimming performance involves studies of the functional morphology of the swimming apparatus and requires insight in swimmingrelated adaptations. Undulations of body and tail propel the majority of species; others predominantly use movements of paired and unpaired fins. Hydrodynamic interactions between the moving fish and water represent the forces required. Visualization of the flow patterns reveals the vortex systems, pressure distributions, and forces. The energy used to swim is obtained by muscles burning fuel and can be as high as 4 W/kg. The metabolic rate increases exponentially with speed. Fair comparisons among species can be made by looking at the most economic swimming performance at the speed where the energy used per unit weight and unit distance is at a minimum. These dimensionless costs of transport decrease with body mass. The amount of work that needs to be done to transport the weight of a body over its length increases with body mass. Schooling behavior may reduce the costs of transport. Pelagic fish commonly use red lateral muscle during steady swimming at low speeds and white lateral muscle during burst swimming. The stride length of these fish is more or less constant and therefore swimming speed can be predicted from the tail beat frequency which in turn is directly related to the contraction times of the lateral the muscles. Swimming speeds can be classified as sustained speeds, prolonged speeds and burst speeds. Endurance of swimming is reduced at higher swimming speeds during prolonged swimming. Temperature has a profound effect on the swimming capacity with both endurance and swimming speed reduced at lower temperatures. Maximum swimming speeds double for every 10°C increase in temperature. The maximum swimming speed of many marine fish species is between 10 and 20 L/s. Swimming performance affects the evolutionary fitness of a species. For each individual, it is a significant factor directly related to capture by or escape from fishing gears.
Table 1.1. Maximum Sustained Swimming Speed (Ums) and Endurance at Prolonged Speeds of Some Marine Fish Species. Ums (cm/s)
Ums (L/s)
40 49
42 45
1.1 0.9
… …
FT
0.8 0.8
He (1991)
Atlantic cod Gadus morhua
36 36
75 90
2.1 2.5
logE = −0.99 • U + 3.99 logE = −1.13 • U + 4.96
FT
5 8
Beamish (1966)
Atlantic herring Clupea harengu
25
102
4.1
logE = −1.43 • U + 8.37
AT
13.5
He and Wardle (1988)
Atlantic mackerel Scomber scombrus
31
110
3.6
logE = −0.96 • U + 5.45
AT
11.7
He and Wardle (1988)
American shad Alosa sapidissima
42
logE = −1.78 • U + 19.02
FT
Haddock Melanogrammus aeglefinus
17
44
2.6
…
AT
24 31 34 41
53 58 57 60
2.2 1.9 1.7 1.5
… … … …
Jack mackerel Trachurus japonicus
14 21
90 90
6.4 4.3
logE = −7.2 • logU + 9.3
FT
19
Xu (1989)
Japanese mackerel scomber japonicus
10
99
9.9
logE = −0.62 • U + 4.38
FT
19
Beamish (1984)
Redfish Sebastes marinus
17 16 16
52 52 52
3.1 3.3 3.3
logE = −0.25 • U + 1.71 logE = −0.23 • U + 1.70 logE = −0.42 • U + 2.94
FT FT FT
5 8 11
Beamish (1966)
Saithe Pollachius virens
25
88
3.5
logE = −1.63 • U + 5.60
AT
14.4
He and Wardle (1988)
34 42 50
100 106 110
2.9 2.5 2.2
logE = −1.52 • U + 5.91 logE = −1.36 • U + 6.16 logE = −1.17 • U +5.95
Species
Length (cm)
Atlantic cod Gadus morhua
Striped bass Morone saxatilis
42–57
E–U Relation
logE = −0.69 • U + 10.65
Method
FT
T (°C)
Source
Castro-Santos (2005) 9.9
Breen et al. (2004)
Castro-Santos (2005)
Note. T, temperature; FT, flume tank; AT, annular tank; E, endurance (in min); U, swimming speed (in m/s).
20
Swimming in Marine Fish
21
Figure 1.12. Maximum swimming speed of some marine species in relation to their body length. Letter symbols, sources, and temperatures (when available) are as follows: AS— American shad, Alosa sapidissima, Castro-santos (2005); AW—Alewife, Alosa pseudoharengus, Castro-santos (2005); BH—blueback herring, Alosa aestivalis, Castro-santos (2005); CD—cod, Gadus morhua, 9.5° to 12°C, Blaxter and Dickson (1959); FH—flathead, platycephalus bassensis, 20°C, Yanase et al. (2007); HD—haddock, Melanogrammus aeglefinus, 12°C, Wardle (1975); HR—Atlantic herring, Clupea harengus, 10° to 15°C, Misund (1989); JK—jack mackerel, Trachurus japonicus, 23°C, Xu (1989); KW—kawakawa, Euthunnus affinis, 25°C, cited in Beamish (1978); MK—Atlantic mackerel, Scomber scombrus, 12°C, Wardle and He (1988); SB—striped bass, Morone saxatilis, Castro-santos (2005); SE—seabass, Dicentrarchus labrax, 20°C, Nelson and Claireaux (2005); SH1—saithe, Pollachius virens, Videler (1993); SH2— saithe,10.8°C, He (1986); SH3—saithe, 14° to 16°C, Blaxter and Dickson (1959); SK1—skipjack tuna, Katsuwonus pelamis, Walters and Fierstine (1964); SK2—skipjack tuna, cited in Magnuson (1978); SP—sprat, Sprattus sprattus, 12°C, Wardle (1975); WA—wahoo, Acanthocybium solandrei, >15°C, Fierstine and Walters (1968); WH—whiting, Gadus merlangus, 9° to 13°C, Blaxter and Dickson (1959); YT—yellowfin tuna, Thunnus albacares, Fierstine and Walters (1968). Lines for 20 L/s (solid), 10 L/s (dashed), and 5 L/s (dotted) are drawn to indicate swimming speed in L/s. (Modified from He, 1993).
REFERENCES Aleyev YG. 1977. Nekton. The Hague: Dr W. Junk. Anderson DF and Eberhardt S. 2001. Understanding Flight. New York: McGraw-Hill. Bainbridge R. 1958. The speed of swimming as related to size and to the frequency and amplitude of the tail beat. J. Exp. Biol. 35: 109–133. Barsukov VV. 1960. The speed of movement of fishes. Priroda 3: 103–104 (in Russian).
Beamish FWH. 1966. Swimming endurance of some northwest Atlantic fishes. J. Fish. Res. Bd. Can. 23: 341–347. Beamish FWH. 1978. Swimming capacity. In: Hoar WS and Randall DJ (eds). Fish Physiology, Vol. 7. Locomotion. pp 101–187. New York and London: Academic Press. Beamish FWH. 1984. Swimming performance of three southwest Pacific fishes. Mar. Biol. 79: 311–313.
22
Locomotion and Sensory Capabilities in Marine Fish
Blake RW. 1983. Median and paired fin propulsion. In: Webb PW and Weihs D (eds). Fish Biomechanics. pp 214–247. New York: Praeger. Blaxter JHS and Dickson W. 1959. Observations of the swimming speeds of fish. J. Cons. Perm. Int. Explor. Mer. 24: 472–479. Bone Q. 1966. On the function of the two types of myotomal muscle fibre in elasmobranch fish. J. Mar. Biol. Assoc. UK. 46: 321–349. Breen M, Dyson J, O’Neill FG, Jones E and Haigh M. 2004. Swimming endurance of haddock (Melanogrammus aeglefinus L.) at prolonged and sustained swimming speeds, and its role in their capture by towed fishing gears. ICES J. Mar. Sci. 61: 1071– 1079. Brett JR. 1964. The respiratory metabolism and swimming performance of young sockeye salmon. J. Fish. Res. Bd. Can. 21: 1183–1226. Carey FG and Teal JM. 1969. Regulation of body temperature by the bluefin tuna. Comp. Biochem. Physiol. 28: 205–214. Castro-Santos T. 2005. Optimal swim speeds for traversing velocity barrier: an analysis of volitional high-speed swimming behavior of migratory fishes. J. Exp. Biol. 208: 421–432. Daniel TL. 1981. Fish mucus: in situ measurements of polymer drag reduction. Biol. Bull. 160: 376–382. Domenici P and Blake RW. 1991 The kinematics and performance of the escape response in the angelfish (Pterophyllum eimekei). J. Exp. Biol. 156: 187– 204. Fields PA. 1990. Decreased swimming effort in groups of pacific mackerel (Scomber japonicus). Am. Soc. Zool. 30: 134A. Fierstine HL and Walters V. 1968. Studies in locomotion and anatomy of Scombroid fishes. Mem. South. Calif. Acad. Sci. 6: 1–31. Ford E. 1937. Vertebral variation in teleost fishes. J. Mar. Biol. Assoc. UK. 22: 1–60. Fuiman LA. 1986. Burst-swimming performance of larval zebra danios and the effect of diel temperature fluctuations. Trans. Am. Fish. Soc. 115: 143–148. Geerlink PJ. 1987. The role of the pectoral fins in braking in mackerel, cod and saithe. Neth. J. Zool. 37: 81–104. Greene CW and Greene CH. 1913. The skeletal musculature of the king salmon. Bull. U.S. Bur. Fish. 33: 21–60. Harder W. 1975a. Anatomy of Fishes. Part I, Text. Stuttgart: E. Schweizerbart’sche. Harder W. 1975b. Anatomy of Fishes. Part II, Figures and Plates. Stuttgart: E. Schweizerbart’sche.
Harper DG and Blake RW. 1990. Fast-start performance of rainbow trout Salmo gairdneri and northern pike Esox lucius. J. Exp. Biol. 150: 321–342. Harper DG and Blake RW. 1991. Prey capture and the fast-start performance of northern pike Esox lucius. J. Exp. Biol. 155: 175–192. He P. 1986. Swimming Performance of Three Species of Marine Fish and Some Aspects of Swimming in Fishing Gears. PhD thesis. University of Aberdeen, Aberdeen, UK. He P. 1991. Swimming endurance of the Atlantic cod, Gadus morhua L. at low temperatures. Fish. Res. 12: 65–73. He P. 1993. Swimming speeds of marine fish in relation to fishing gears. ICES Mar. Sci. Symp. 196: 183–189. He P and Wardle CS. 1988. Endurance at intermediate swimming speeds of Atlantic mackerel, Scomber scombrus L., herring, Clupea harengus L., and saithe, Pollachius virens L. J. Fish Biol. 33: 255–266. Herskin J and Steffensen JF. 1998. Energy savings in sea bass swimming in a school: measurements of tail beat frequency and oxygen consumption at different swimming speeds. J. Fish Biol. 53: 366–376. Hertel H. 1966. Structure-Form-Movement. New York: Reinhold. Lauder GV and Tytell ED. 2006. Hydrodynamics of undulatory propulsion. In: Shadwick RE and Lauder GV(eds).FishPhysiology,Vol.23.FishBiomechanics. pp 425–468. London: Academic Press. Lindsey CC. 1978. Form, function and locomotory habits in fish. In: Hoar WS and Randall DJ (eds). Fish Physiology, Vol. 7. Locomotion. pp 1–100. New York: Academic Press. Lumley JL. 1969. Drag reduction by additives. Annu. Rev. Fluid Mech. 3: 367–384. Müller UK, van den Heuvel BLE, Stamhuis EJ and Videler JJ. 1997. Fish foot prints: morphology and energetics of the wake behind a continuously swimming mullet (Chelon labrosus Risso). J. Exp. Biol. 200: 2893–2906. Misund OA. 1989. Swimming behavior of herring (Clupea harengus L) and mackerel (Scomber scombrus L) in purse seine capture situations. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 541– 546. St. John’s, Newfoundland: Marine Institute. Nelson JA and Claireaux G. 2005. Sprint swimming performance of juvenile European sea bass. Trans. Am. Fish. Soc. 134: 1274–1284. Rosen MW and Cornford NE. 1971. Fluid friction of fish slimes. Nature. 234: 49–51.
Swimming in Marine Fish Shann EW. 1914. On the nature of lateral muscle in teleostei. Proc. Zool. Soc. Lond. 22: 319–337. Videler JJ. 1988. Sleep under sand cover of the labroid fish Coris julis. In: Koella WP, Obál F, Schultz H and Visser P (eds). Sleep ’86. pp 145–147. Stuttgart: Gustav Fischer. Videler JJ. 1993. Fish Swimming. London: Chapman and Hall. Videler JJ. 1995. Body surface adaptations to boundarylayer dynamics. In: Biological Fluid Dynamics. pp 1–20. Cambridge: The Society of Biologists Limited. Videler JJ and Hess F. 1984. Fast continuous swimming of two pelagic predators: saithe (Pollachius virens) and mackerel (Scomber scombrus). A kinematic analysis. J. Exp. Biol. 109: 209–225. Videler JJ, Stamhuis EJ, Müller UK and van Duren LA. 2002. The scaling and structure of aquatic animal wakes. Integr. Comp. Biol. 42: 988–996. Videler JJ and Wardle CS. 1991. Fish swimming stride by stride: speed limits and endurance. Rev. Fish Biol. Fish. 1: 23–40. Videler JJ and Weihs D. 1982. Energetic advantage of burst-and-coast swimming of fish at high speeds. J. Exp. Biol. 97: 169–178. Walters V and Fierstine HL. 1964. Measurements of swimming speeds of yellowfin tuna and wahoo. Nature. 202: 208–209.
23
Wardle CS. 1975. Limit of fish swimming speed. Nature. 255: 725–727. Wardle CS and He P. 1988. Burst swimming speeds of mackerel, Scomber scombrus L. J. Fish Biol. 32: 471–478. Wardle CS, Videler JJ and Altringham JD. 1995. Tuning in to fish swimming waves: body form, swimming mode and muscle function. J. Exp. Biol. 198: 1629–1636. Wardle CS, Videler JJ, Arimoto T, Franco JM and He P. 1989. The muscle twitch and the maximum swimming speed of giant bluefin tuna, Thunnus thynnus L. J. Fish Biol. 35: 129–137. Weihs D. 1973. Hydromechanics of fish schooling. Nature. 245: 48–50. Weihs D. 1974. Energetic advantage of burst swimming of fish. J. Theor. Biol. 48: 215–229. Xu G. 1989. Study on the Fish Swimming Movement and Its Application in Fishing by Trawls. PhD thesis. Tokyo University of Fisheries, Tokyo, Japan (in Japanese with English summary). Yanase K, Eayrs S and Arimoto T. 2007. Influence of water temperature and fish length on the maximum swimming speed of sand flathead, Platycephalus bassensis: implications for trawl selectivity. Fish. Res. 84: 180–188.
24
Locomotion and Sensory Capabilities in Marine Fish
SPECIES MENTIONED IN THE TEXT African electric eel, Gymnarchus niloticus alewife, Alosa pseudoharengus American shad, Alosa sapidissima angelfish, Pterophyllum scalare Atlantic cod, cod, Gadus morhua Atlantic flying fish, Cheilopogon heterurus Atlantic herring, Clupea harengu Atlantic mackerel, Scomber scombrus blueback herring, Alosa aestivalis bluefin tuna, Thunnus thynnus butterfly fish, Chaetodon sp. cornet fishes, Fistsularia sp. European eel, eel, Anguilla anguilla flathead, Platycephalus bassensis forkbeard, Phycis phycis haddock, Melanogrammus aeglefinus hatchet fish, Gasteropelecus sp. jack mackerel, Trachurus japonicus Japanese mackerel, Scomber japonicus kawakawa, Euthunnus affinis louvar, Luvarus imperialis moray eel, Muraena helena opah, Lampris guttatus
Pacific barracuda, barracuda, Sphyraena argentea Pacific bonito, Sarda chiliensis Pacific mackerel, Scomber japonicus pike, Esox lucius porbeagle shark, Lamna nasus rainbow wrasse, Coris julis redfish, Sebastes marinus sailfish, Istiophorus platypterus saithe, Pollachius virens sand eel, Ammodytes tobianus seabass, Dicentrarchus labrax seahorse, Hippocampus ramulosus shiner perch, Cymatogaster aggregata skate, Raja batis skipjack tuna, Katsuwonus pelamis sprat, Sprattus sprattus striped bass, Morone saxatilis sunfish, Mola mola swordfish, Xiphias gladius turbot, Scophthalmus maximus wahoo, Acanthocybium solandrei whiting, Gadus merlangus yellowfin tuna, Thunnus albacares
Chapter 2 Fish Vision and Its Role in Fish Capture Takafumi Arimoto, Christopher W. Glass, and Xiumei Zhang
2.1 INTRODUCTION Vision in vertebrates, including fish, has been extensively studied (see Crescitelli 1977; Douglas and Djamgoz 1990; Guthrie and Muntz 1993; Lythgoe 1979; Nicol 1989). While the structure of the eye is well known and mechanisms of vision have been described for a number of fish, many commercially important marine species have received little attention. Despite many years of research into the visual systems of fish, detailed knowledge and understanding of the role of fish vision in their reaction to fishing gears during capture processes are far from complete. Understanding visual characteristics of fish is an important component in understanding the fish capture process and interactions between fish and fishing gear. This chapter reviews the structure and function of the eyes of marine fishes as well as the underwater visual environment to assist in better understanding the reaction of fish to fishing gear. Some examples of using fish vision in the design and operation of fishing gear are also illustrated. Knowledge gaps are identified for future research in the field of fish vision and visual behavior in relation to fish capture processes.
Lythgoe 1979; Nicol 1989). A short review on the structure and function of fish eyes and their role in the fish capture process is provided here. Although there is great diversity in the details of individual component that reflects the variation in the light environment inhabited by fish, a typical fish eye is illustrated in Figure 2.1. The eye has two main functions: optics and accommodation (Fernald 1990).
2.2 STRUCTURE OF THE FISH EYE The form, function, and structure of the fish eye (Fig. 2.1) have been documented extensively in the past (Atema et al. 1988; Collin and Marshall 2003; Douglas and Djamgoz 1990; Kawamoto 1970;
Accommodation: The Focusing of the Image on the Retina In most fish, the image is focused on the retina by movement of the lens rather than by a change in the shape of the lens, as occurs in other vertebrates.
Optics: The Collection and Formation of an Image Both sensitivity and acuity depend on the brightness of an image reaching the retina, and this is affected by the properties of the eye. The fish pupil is usually immobile, and light control is performed by the retinomotor mechanism involving movement of melanin granules in the retinal pigment cells. Optical resolution depends on lens quality, receptor size, and density. Images are formed by the refractive properties of the lens as the cornea of most fish eyes has a refractive index almost identical to that of water and contributes little to the optics of the eye.
25
26
Locomotion and Sensory Capabilities in Marine Fish
Figure 2.1. Structure of fish eye. (Modified from Kawamoto 1970.)
2.2.1 Lens The lens is the main refracting component of the eye and is usually spherical in shape. Details of the lenses of fish eyes are discussed by Sivak (1990). The spherical lens of fish eyes has been shown to produce good resolution due in part to a refractive index gradient within the lens; a higher refractive index at the center of the lens decreases with radius in all directions (Fernald 1985). Differences in refraction may be obtained by a variation in the composition of the lens as indicated by the ratio of proteins and water in the lens, which varies among species (Nicol 1989). Accommodation in the fish eye is a result of movement of the lens (rather than of a change in its shape) or of altering the depth of the eyeball to change the lens–retina distance in some species. The lens is moved backward to focus an image in teleosts, moved forward in elasmobranches, and pushed backward by flattening the cornea in lampreys (Nicol 1989). Studies on accommodation have shown that the direction of lens movement and the region of highest cone density in the retina are related (Tamura 1957; Tamura and Wisby 1963).
2.2.2 Retina As with the general structure of the eye, retinas of fish display similar characteristics but differ in details, reflecting the unique visual-environmental conditions under which different fish species live, whether in shallow well-lit lagoons or the dark bottom of a deep ocean (see Guthrie and Muntz 1993; Nicol 1989). The generalized fish retina is composed of an outer sheet of pigment epithelium covering a layer of photoreceptors (visual cell layer) and an inner layer of nervous tissue (Fig. 2.2). Teleosts generally possess both rod and cone photoreceptor cells. Rods have only one pigment and are used for scotopic (dark-adapted) vision, whereas cones may have up to four pigments and are used for color or photopic (light-adapted) vision. Cones may be double or single, and some teleosts may have triple or quadruple cones. Where multiple cones exist, cones may be structurally different or indistinguishable and may even contain different pigments. Single and double cones are usually found in specific arrangements or mosaics, which vary among species. The number of cones may also change across the retina within a single species.
Fish Vision and Its Role in Fish Capture
27
Figure 2.2. Retinal structure of Pacific saury (Cololabis saira) indicating layers within the retina. (Hajar 2007.) For color detail, please see color plate section.
The neural layers of the retina are composed of a nuclear layer, a ganglion cell layer, and a plexiform layer. The signal from the retina passes from the photoreceptors to the bipolar cells and then to the ganglion cells. It is successively modified by the horizontal, amacrine, and inner plexiform cells. As with the structure of the eye as a whole, details of the structure of the retinal layer and its component cells are subject to great variations among species (Wagner 1990). 2.3 VISUAL FUNCTION 2.3.1 Color Vision Most fish species can distinguish color by the use of red-, green-, and blue-sensitive cones. At least two types of cones are required for color discrimination, while some freshwater and shallow-living marine species have the capability to detect ultraviolet radiation with a fourth type of cone. Many deep sea species are often referred to as color-blind,
because their retinas are composed of only rods with no cones. Color vision and color discrimination can be determined by behavioral conditioning techniques. However, care is needed in inferring true color discrimination capability from these experiments because characteristics of an apparently simple color can be confounded by subtle aspects such as hue, saturation, and brightness of the targets. Electroretinogram (ERG) is used to monitor the response of retina to stimulation by different wavelengths of light (i.e., color) and to determine spectral sensitivity of fish eyes. Figure 2.3 shows ERG wave patterns of walleye pollock (Theragra chalcogramma) in dark-adapted condition. Figure 2.4 provides the spectral sensitivity curve of walleye pollock showing the Purkinje phenomenon where the wavelength of highest sensitivity is shifted from 540 nanometer (nm) for the light-adapted eye (photopic vision) to 470 nm for the dark-adapted eye (scotopic vision) (Zhang 1992).
28
Locomotion and Sensory Capabilities in Marine Fish (Anthony 1981) or by monitoring schooling performance in obligate schooling species (Glass et al. 1986), and by electrophysiological monitoring of the retinal response to varying light intensities (Kobayashi 1962). Different fishing gears provide a different contrast image according to ambient light conditions, gear type, and the visual sensitivity of the fish. The contrast of an object against the water background appears to be more important than the brightness of the object (Wardle 1987, 1993).
Figure 2.3. Electroretinogram (ERG) amplitude in dark-adapted eyes of walleye pollock (Theragra chalcogramma) under different light intensities. (Zhang 1992.)
2.3.2 Light Vision Photosensitivity is the ability of fish eye to receive light and to get visual information in ambient light conditions. Light intensity varies with water depth, time of day, and transparency or turbidity of water. To adapt to the wide range of light intensities encountered at sea, functional changes between cone and rod cells are made through shifting of positions of visual cells according to the ambient light intensity. Rods are highly sensitive to low light intensities, while cones are used at high light intensities. This allows fish to function visually over a wide range of light intensities in the natural environment. The minimum light intensity threshold for fish to function visually has been determined for a number of fish species by cardiac condition response
2.3.3 Motion Vision A moving image is generally of more importance to an animal than a static image. Detection of movement is dependent on visual acuity and persistence time—the time taken to process the image. The frequency at which flickering images fuse to produce a continuous image is referred to as the flicker fusion frequency (FFF) or critical flicker frequency (CFF) and is dependent on light intensity, temperature, and flash duration. Research on the perception of flicker in fish has been periodically reviewed (Douglas and Hawryshyn 1990; Landis 1954; Nicol 1989; Protasov 1970). The ability of fish to detect a moving or flickering image is affected by the level of illumination. Fish can perceive motion at a wide range of light intensities from 10−7 to 10−4 lux (Protasov 1970). At low light intensities, detection of a target is only possible if sufficient light is received to activate the photoreceptors. Increased sensitivity to low light levels is facilitated by temporal summation, which is a low threshold frequency of flicker fusion. In other words, an image must persist for a longer duration if it is to be perceived under low light levels; therefore, fast-moving images will not likely be detected. As light intensity is increased, the sensitivity to, or detection of, an image is greatly enhanced. The amount of temporal summation necessary to perceive a moving or flickering image decreases as sufficient light is received in a shorter time period to activate the receptors (Douglas and Hawryshyn 1990). Simply put, decreasing light intensity leads to a decrease in capacity for perception of motion. As with spectral sensitivity, FFF can be determined using either behavioral or electrophysiologi-
Fish Vision and Its Role in Fish Capture
29
Figure 2.4. Relative electroretinographic amplitude in lightadapted eyes (open circle) and two dark-adapted eyes of different time, showing the phenomenon of Pukinje shift. (Zhang 1992.) For color detail, please see color plate section.
cal approaches. Values of FFF obtained by electroretinography have been shown to increase with light intensity, although the relationship is not linear (see Blaxter 1970; Loew and McFarland 1990; Protasov 1970; Zhang and Arimoto 1993a). Curves obtained by plotting FFF against log light intensity are generally double branched and the point of inflection is a result of change in receptors from rods to cones. The point of inflection differs among species and the switch in receptor type may be mediated to some degree by photochemical movements of photoreceptors and the pigment epithelium (Douglas and Hawryshyn 1990). Behavioral techniques have also been used to investigate FFF. The optomotor response—that is movement of the eyes, head, curvature of the body or trunk, or movement of the entire animal in response to follow a moving image–has often been used as a determinant of FFF and visual acuity (Harden-Jones 1963; Nicol 1989; Sbikin 1981).
Behavioral results may differ from those obtained by electrophysiological techniques using electroretinography, possibly because the perceived FFF through behavioral techniques is a result of complex visual processing whereas an ERG is obtained in a relatively early stage of the process (Douglas and Hawryshyn 1990). Comparative studies on flicker fusion have often been presented as maximum photopic FFF. Elasmobranches generally have lower FFF than teleosts. Pelagic species living near the surface tend to have higher thresholds than do those dwelling in the bottom of deeper waters. Protasov (1970) was the first to relate FFF to ecology of the fish. For example, motion perception is important in detection of predators and prey; spurt reaction of prey will not be detected by a predator with a lower FFF. It is important to remember that many FFF values were determined from laboratory experiments where the background was usually black. In natural
30
Locomotion and Sensory Capabilities in Marine Fish
conditions, flickering or moving targets are often viewed against a nonuniform and changing background (Nicol 1989). The detection of movement has important implications in how fish react to fishing gears, in particular, trawl gears. The herding and optomotor reactions, by which fish hold station with the gear components such as bobbins, floats, ropes, and meshes until it becomes exhausted, are both a result of these images constantly moving relative to the background and relatively stationary to the fish. The higher FFF of a species, more acute is the sense of motion detection. In addition to FFF, fish have an ability to detect a moving object that is related to the pattern of the visual cells in the retina. A mosaic arrangement of cones confers better detection of moving objects and is found in species such as jack mackerel (Trachurus symmetricus) and chum salmon (Oncorhynchus keta) that rely on superior visual capabilities for predation and schooling (Wagner 1990). A random arrangement that confers lower motion detection capability is found in Japanese eel (Anguilla japonica), which has a lifestyle less reliant on the need for high-definition motion detection. Predatory species feeding on fast-moving preys require high temporal resolution motion vision. Higher FFF can be achieved with higher eye temperatures, which in most fish species is similar to ambient water temperature. For example, an increase of FFF from 5 Hz at 10°C to over 40 Hz at 20° was observed in swordfish (Xiphias gladius) (Fritsches et al. 2005). To increase higher temporal resolution for successful hunting of fast-moving prey, swordfish developed a specialized heating system to keep eyes and brain 10° to 15° above ambient water to increase temporal resolution (Fritsches et al. 2005). The increased capability of motion vision may also exist in other “warmbodied” tunas, which have a unique thermal regulation system (Sharp and Dizon 1978). Sharper vision, as well as faster swimming capacity, makes them superior predators. 2.3.4 Form Vision The lens quality and focal accommodation, together with the retinal resolving power, define the ability
Figure 2.5. Diagram showing binocular vision (1), monocular vision (2), and blind zone (3) in a typical teleost fish.
of fish to perceive details of a visual object. A pair of eyes generally located on opposite sides of the body gives a wide visual field with a monocular visual field on each side and a rather narrow binocular visual field facing forward (Fig. 2.5) (Wardle 1993). Different fish species have a different contour pattern of cone densities, with the highest density area being located in different regions. The position of highest cone density area defines the visual axis of each species (Shiobara and Arimoto 1999; Miyagi et al. 2001). The visual axis is related to the visual response, and in general it defines the direction in which the fish has greatest visual sensitivity. In some species like yellowtail (Seriola quinqueradiata), the concentration of high-density area is not obvious, indicating that this active predatory species has good all-round vision. The determination of the visual axis direction derived from the retinal histology is supported by the direction of movement of the lens during focal accommodation, which can be examined by electrical stimulation of the lens muscle. Visual acuity is the ability of fish to resolve and to see fine details of an object or a pattern. There are three different aspects of visual acuity:
Fish Vision and Its Role in Fish Capture
31
Table 2.1. Conversion Table for Visual Acuity Expressed in Minimum Separable Angle (MSA), Minimum Angle of Resolution (MAR), Snellen Notation, and Decimal Unit. MSD (m) for Target Size of MSA (Radian) 2.91 × 10−3 1.45 × 10−3 0.73 × 10−3 0.58 × 10−3 0.36 × 10−3 0.29 × 10−3 0.15 × 10−3
MAR (min)
Snellen Notation
Decimal Unit
0.5 cm
1 cm
2 cm
10 5.0 2.5 2.0 1.25 1.0 0.5
6/60 6/30 6/15 6/12 6/7.5 6/6 6/3.0
0.10 0.20 0.40 0.50 0.80 1.00 2.00
1.72 3.45 6.85 8.62 13.9 17.2 33.3
3.44 6.90 13.7 17.2 27.8 34.5 66.7
6.87 13.8 27.4 34.5 55.6 69.0 133.3
The corresponding maximum sighting distance (MSD) for different target sizes are estimated.
• Target detection—the ability to form an image on the retina • Gap detection—the ability to distinguish fine details such as the gap of the Landolt C Mark • Target recognition—the ability to recognize letters such as the Snellen Notation used in opticians’ clinics
and pigment index (P), which in turn are calculated from the cone position (c) and the thickness (A) of visual cell layer (Fig. 2.6):
Gap detection and target recognition are compatible with the minimum separable angle (MSA) or the minimum angle of resolution (MAR) (see also Section 2.4.1). The visual acuity can be expressed by several different units, which are derived from the MSA. These units can be converted from each other as shown in Table 2.1. The decimal unit is defined as the reciprocal of MAR in minutes of arc. The smaller MAR represents better visual acuity.
P=p A
2.3.5 Retinomotor Response The retinomotor response is a process by which the relative position of cones and rods in the retina changes in response to changes in ambient light intensity. Figure 2.6 shows cone positions of the retina of jack mackerel at different retinal adaptation stages from light, transitional, and dark adapted eyes. In light-adapted eyes, cones are moved to the surface level close to the outer limiting membrane in visual cell layers by shortening its myoid. There is an associated shift of pigments to cover the rods to protect them from strong light. Adaptation stages of the retina can be described by the cone index (C)
C=c A or from the shifting distance of pigment layer (p):
The retinomotor response follows a circadian rhythm. It is influenced by ambient light and can be modified by artificial light during light fishing. Identification of light/dark adaptation stages using cone index provides important information on visual response of target and nontarget species during their capture by fishing gears. In the case of walleye pollock captured by a trawl at a depth of 270 m during daytime, retina conditions were identified as transitional or dark-adapted stages, indicating a reduced visual response of fish to towed gears in deep waters (Zhang et al. 1993). 2.3.6 Optomotor Response The optomotor response, which is sometimes called the optomotor reaction, refers to the phenomenon that the fish maintains a relatively fixed position of a visual image on its retina. Optomotor response may be one of the most significant behavioral responses governing fish reaction to the surrounding stimuli during fish capture processes. The
32
Locomotion and Sensory Capabilities in Marine Fish
Figure 2.6. Changes in the position of layers in the retina during dark adaptation, transitional stage, and light adaptation. p, shifting distance of pigment layer; c, cone position; and A, thickness of visual cell layer. (Zhang 1992.) For color detail, please see color plate section.
response pattern explains a number of observed behaviors such as maintaining position in flowing water or in a school (Shaw 1965). The optomotor response has been used to control fish behavior for the purpose of measuring swimming speeds (Bainbridge 1958; He and Wardle 1988) and for studying spectral sensitivity using different light patterns as visual targets (Hasegawa 2006). Pavlov (1969) described station-holding swimming of fish at the wing of a trawl as being mediated by the optomotor response and discussed potential for its application in improving capture efficiency of the fishing gear. This was further elaborated in the laboratory (Inoue and Arimoto 1976; Inoue and Kondo
1972) and in the field during fish capture processes using direct underwater observation techniques (Wardle 1987).
2.4 VISUAL CAPACITY: VISUAL ACUITY, SEPARABLE ANGLE, AND MAXIMUM SIGHTING DISTANCE Visual acuity is represented by the MSA (α in radian), as defined by the following equation (Tamura 1957):
α=
1 ⎡ 2 × 0.1 × (1 + .025) ⎤ ⎦⎥ F ⎣⎢ n
Fish Vision and Its Role in Fish Capture
33
Minimum Separable Angle
Cone On
α
F
Off On
Landolt C
where F is the focal length of the lens, which is 2.55 times the radius of the lens, and n is cone density, which is the number of cones in an area of 0.01 mm2. The equation is derived from the concept of optical resolving capability by cones in the retina as shown in Figure 2.7, where at least three cones are required to identify the gap in the Landolt C Mark, with one cone for the gap between a pair of dots or between the two tip points of the Landolt C Mark. The visual information received by visual cells is integrated and modulated at ganglion cells that are located at the surface layer in the lining of retina and transmitted to the optic nerve in the retinal level. Collin and Pettigrew (1989) reasoned that the visual acuity is related to the density of ganglion cells, with the same concept of discriminating a pair of dots as the optical resolving capability. While the approach is theoretically feasible, it is difficult to identify and count the number of ganglion cells during histological examinations. Behavioral approaches to determine the MSA are possible using classic conditioning techniques by training fish to identify the direction of the opening of the Landolt C Mark. After the fish is conditioned with a larger Landolt C Mark, the mark size is systematically reduced until the fish cannot discern the direction of the opening. The MSA derived from behavioral approach is usually smaller than that obtained from histological examinations (Shiobara and Arimoto 2003).
Figure 2.7. Landolt C Mark, cone density, and the minimum separable angle (α). F, Focal length. (Hajar 2007.)
Figure 2.8 shows the MSA for a number of species and sizes of fish. Overall, visual acuity is affected by several optical factors, including ambient light intensity and contrast of visual targets against the background. Larger fish have smaller angles, indicating they have better visual acuity and are more capable of distinguishing smaller or finer details of visual objects. The maximum sighting distance (D) can be estimated from the MSA (α, in radian) and the size of the visual target (l): D=l α assuming ideal optical conditions, which include a high contrast target, in high illumination conditions, and in highly transparent water. Under these ideal optical conditions, the maximum sighting distance of walleye pollock in relation to the size of fish and the size of visual targets is shown in Figure 2.9. In general, larger fish have larger sighting distance and can detect fishing gear components or other underwater objects from farther away (for more details, see Arimoto and Namba 1996). The maximum sighting distance derived from the MSA may not fully represent the true capability of fish to detect a fishing gear or its components. The estimation is solely dependent on target size and cone density. Here the target size of floats, bobbins, or knots of netting is assumed to be the length between the top and bottom or left and right of the
Figure 2.8. Visual acuity as expressed by the minimum separable angle (in min) in relation to fish body length in yellowtail (Theragra chalcogramma) (Miyagi 2001), red sea bream (Pagrus major) (Shiobara et al. 1998), jack mackerel (Trachurus symmetricus), walleye pollock (Theragra chalcogramma) (Zhang 1992), and Pacific saury (Cololabis saira). (Hajar et al. 2008.)
Figure 2.9. Maximum sighting distance of walleye pollock (Theragra chalcogramma) of different body lengths when viewing object of 2- to 10-cm visual target. (Zhang and Arimoto 1993.)
34
Fish Vision and Its Role in Fish Capture
35
2.5 COLOR AND APPEARANCE OF FISHING GEAR UNDERWATER
variations in many factors, but a general classification of coastal and oceanic water types by color has been established by Jerlov (1964). In general, spectral absorption of light in the open oceans produces an effect, as shown in Figure 2.10, where light in the green/blue part of the spectrum transmits deeper into the water column than other wavelengths. This can have a profound effect on behavioral ecology of fish residing in different areas of the ocean, whether in the upper strata of the ocean, in shallow waters near the coastline, or in deep ocean bottoms. It should be noted, however, that much of the world’s commercial fishing operations are conducted in relatively deep waters either in the absence of visible light or in dim monochromatic conditions.
2.5.1 Spectral Properties of Seawater There is a great diversity in the underwater visual environment inhabited by fish. Many factors, including water property, nature of light source, and suspended particles, affect the distribution of light as it passes down and through the water column. The characteristics of underwater visual environment and their influence on fish vision and behavior have been described by a number of authors, notably Guthrie and Muntz (1993); Blaxter (1970, 1988), Lythgoe (1979), Loew and McFarland (1990), Nicol (1989), and Protasov (1970). Underwater natural light comes from either solar radiation or bioluminescence. As solar radiation passes through the water, it is refracted toward the vertical and progressively filtered and diminished due to absorption and scattering effect of water molecules, dissolved pigments, and particulate matter. This results in generally monochromatic conditions at depths at which fishing gears are usually operated, with the light coming more or less from directly above. Upwelling or horizontal transmission of light generated through bioluminescence, where it occurs, is of low intensity (approximately 5% of overall illumination) but has a significant effect on the visibility of fishing gear and may under certain circumstances result in directed behavioral reactions by fish in its vicinity. The spectral transmission curves of waters of oceans and coastal areas can vary greatly due to
2.5.2 Visual Contrast of Fishing Gear Blaxter et al. (1964) summarized the role of visual senses of fish when responding to fishing gears. Wardle (1983) stressed the importance of understanding visual contrast of fishing gear against the background and suggested that it was more important than the brightness of the gear itself. Wardle (1993) described and illustrated the herding effect induced by a high-contrast image presented by the otter boards and the trawl mouth and the entrapping effect induced by the less visible rear section of the net. The visual angle for detecting targets was compared with contrast perception for different parts of the gear as they were viewed from below against the bright surface and from above against the dark background (see Chapter 8, Figure 8.7). There is a complex relationship between color and contrast of gear components, ambient light intensity. and water quality. Kim and his colleagues (Kim 1998; Kim and Wardle 1998a, 1998b) modeled visual stimulus and contrast of fishing gear components and made some interesting predictions on fish behavior near fishing gear under visual conditions. But in general it has been demonstrated that light-colored netting panels are more difficult to detect against a bright background because there is little contrast between the image and its background. The reverse is also true for materials that contrast strongly with the background against which they are viewed. The contrast required to distinguish the target against its background can be defined by the
target, not the gap distance as used in calculating the MSA. Furthermore, the line acuity required for resolving mesh twines or grating resolutions for netting panels tends to be larger than the visual acuity for a point source target. Regardless, the maximum sighting distance calculated from the given formula is a useful tool for predicting visual capability of fish, especially for comparing different species and sizes of fish. It can also be used for estimating visibility of targets and visual range of fish and for manipulating the visual appearance of fishing gears and their components to make them more or less visible.
36
Locomotion and Sensory Capabilities in Marine Fish
Figure 2.10. Spectral absorption of light in the open oceans where light in the green/blue part of the spectrum transmits deeper into the water column than other wavelengths (reprinted from http:// ultramaxincorp. com/?p2=/modules/ ultramax/catalog. jsp&id=23). For color detail, please see color plate section.
following equation and has been referred to as the apparent contrast: C = (I − Ib ) Ib where I and Ib represent target and background radiance (W m−2) or irradiance (W sr−1 m−2). The contrast threshold was determined in several species by observing behavioral responses to food reward, including striped beak-perch (Oplegnathus fasicatus) (Arakawa et al. 2007), red sea bream (Pagrus major) (Miyagi 2001), and Japanese common squid (Tadorodes pacificus) (Siriraksophon et al. 1995; Siriraksophon and Morinaga 1996). The cardiac conditioning technique was also reported for Atlantic cod (Gadus morhua) (Anthony 1981). While the contrast threshold varies among species, target types, background conditions, and ambient light intensity, it generally showed higher contrast sensitivity in dark ambient light conditions due to function of the rod cell in scotopic vision and lower contrast sensitivity in brighter conditions mediated by the cone cell response in photopic vision.
The contrast of a net is determined by the apparent contrast of twine when viewed against the background. However, when viewing a netting panel, the twine of the netting panel comprises only a small portion of the overall area being viewed and this can have an impact on visibility of the netting and visual range of the fish. The projected area ratio, or netting solidity, is the ratio of the area of the mesh twine to the total area of the netting and is related to mesh size, twine thickness, knot type, and hanging ratio. More solid netting (i.e., netting with larger projected area ratios) is more visible and provides fish with a larger visual range. 2.6 FISH VISION AND ITS APPLICATION IN FISH CAPTURE 2.6.1 Herding and Capture of Fish by Trawl under Visual Conditions The importance of fish vision in relation to the capture process of trawl fisheries was first highlighted by Blaxter et al. (1964) and Parrish (1969) through a series of laboratory and field observa-
Fish Vision and Its Role in Fish Capture tions. Since these early studies, advances in underwater observation techniques with visual, photographic, and acoustic tools (Graham et al. 2004; Urquhart and Stewart 1993) have greatly increased our knowledge and understanding of the capture process of trawls—for example, the level of illumination required for fish to form visual images of the fishing gear (Wardle 1983, 1987, 1989, 1993; also see Chapter 4). The herding by the otter boards, trawl warps, and sand cloud is the first stage of the capture process, where the visual stimuli of the trawl gear assists with accumulating fish inside and toward the trawl mouth. Observations of fish avoidance reaction to the wing of the net, the netting of the trawl mouth, and other components have been described by Kim and Wardle (2003). In that study, they quantified optical characteristics such as visual contrast of gear components against water background as viewed by a fish. They modeled fish behavior in response to a suite of visual stimuli. Zhang and Arimoto (1993b) modeled escape distances for fish to avoid a trawl as a function of the maximum sighting distance and fish swimming speed. Despite these efforts, there is disconnect between our knowledge of visual physiology and how it relates to observed visual reactions to fishing gears, but this may be a fertile area for future research. This research may have important implications for developing conservation-oriented fishharvesting measures such as design of nets that avoid bycatch and reduce discard through manipulation of visual images of the gear. 2.6.2 Use of Light in Fishing Fishing with lights is one of the most advanced and successful methods for catching squids and other pelagic species (Ben-Yami 1976, 1988; Inada and Arimoto 2007). The technique has been successfully used to attract and aggregate fish for centuries. Started with fire torches, a wide range of light fishing equipment and methods has been developed for net gears such as purse seines and lift nets, as well as for hook-and-line methods for squid, mackerel, and jack mackerels. The method has been used in small-scale fisheries along the coast of Japan and some Southeast Asian countries, as well as in largescale offshore and oceanic fisheries. A wide variety of technologies and lighting equipment has been
37
developed over time to best match water quality at specific fishing grounds. Specific lighting parameters (e.g., color and intensity) have been applied according to the behavior of target species. While fishing techniques have evolved over time, there is little knowledge of why squids and fish are attracted to light, although many explanations have been offered, such as: • Schooling for feeding under the light • Conditioned responses to light intensity gradients • Curiosity behavior and other social behavior • Positive phototaxis making them orient to the light source • Optimum light intensity for feeding and other activities • Disorientation and immobilization due to localized high light levels in surrounding dark conditions It is understood, however, that attraction is likely due to a combination of the factors and almost certainly varies both temporally and spatially (BenYami 1976, 1988). Light fishing continues to be one of the largest and most productive fishing techniques globally and as such has generated a wide and diverse range of research and development activities. While much of this activity is focused on improving commercial productivity, there are many current initiatives to reduce energy costs associated with light production through use of light emitting diodes (LED) as the primary source of light and to tailor light quality (e.g., intensity and color) to different species aimed at species-selective fishing (Inada and Arimoto 2007). Both topic areas have important implications for conservation of energy and fish stocks for sustainable and profitable exploitation of fishery resources in the future. 2.6.3 Use of Light and Illusion in Guiding and Blocking As outlined earlier, many of the important behavioral reactions of fish to fishing gears are governed or mediated by the visual system. This has long been exploited, often unsuspectingly, by fishermen. For example, there are many opinions regarding the
38
Locomotion and Sensory Capabilities in Marine Fish
best or most appropriate color of fishing twine for commercial success, indicating that fishermen do indeed understand the role that color, contrast, and spectral qualities of water have on the behavior of fish in response to their fishing operations (Jones et al. 2005). Fishing gear may have evolved to exploit the optomotor response in the case of trawls or to exploit the visibility of twine in the case of monofilament gillnets. However, the role of light and manipulation of visual stimuli to modify natural behavior patterns of fish and to guide fish to desired directions or destinations has only recently been explored in new and innovative ways with regard to fish capture in commercial fishing operations. This field of research may yet prove to yield significant advances in improving fishing efficiency and selectivity, reducing bycatch and discard, and reducing the impact of fishing on the habitat. The use of fish behavior to guide animals has long received attention in rivers and waterways in many countries (Coutant 2001). Behavioral systems for guiding fish away from turbines or for facilitating fish migration through fish passages are generally attractive financially for users of a water resource as they can be less costly and easier to implement than structural systems (Coutant 2001). However, the notion of merely hanging a simple device such as a strobe light or other visual cue from an intake to “frighten” fish away or to attract fish along a different path has proved to be oversimplified and often misguided. Nevertheless, behavioral control systems have been studied extensively and are often considered an appropriate first step in problem solving (Coutant 2001). Within the field of marine commercial fisheries, while the importance of vision and the role of visual images of fishing gear components have been recognized for a long time, there have been, with a few notable exceptions, surprisingly few directed attempts to manipulate patterns of fish behavior. In the 1970s and 1980s, Wardle and a team of coworkers in Aberdeen, Scotland (Wardle 1983), developed an experimental, or concept, trawl with black-and-white striped patterns in its wings and main belly of the net (Fig. 2.11). The patterns were designed to maximize the effect of the optomotor response in the wings of the net and to guide exhausted fish more quickly to the codend for sub-
sequent capture. Later, a striped panel constructed from regular dark (blue) twine and bright Glownet material made by Nichimo of Japan was designed to provide visual guiding even when ambient illumination levels dropped below the visual threshold for image formation. Both nets proved successful in concept but were never further pursued toward commercialization. Nevertheless, they helped demonstrate how fish, even in commercial fishing conditions, could have their behavior manipulated by controlling the visual stimuli surrounding them. Subsequent studies by Wardle et al (1991) and Glass et al (1993, 1995) continued this exploration of behavior modification through manipulation of visual stimuli. Wardle et al (1991) demonstrated the importance of spectral quality of water in rendering monofilament gillnets visible or invisible to approaching fish (also see Chapter 8). In this study, they identified that the angle of the twine relative to the viewer was important in determining its visibility and that knots of the netting acted in a fashion similar to lenses, making knots the most visible part of the netting. These critical observations helped explain some of the observed behavioral reactions of fish to gillnet panels and prompted further investigation into the reactions of fish when surrounded by netting panels in the body and codends of towed fishing gears. Observations of fish in fishing gears have consistently shown that fish, under visual conditions, keep clear of the netting and are herded effectively by the panels of the wingends and the front regions of the net (Glass et al. 1993). Fish entering the extension and codend region of a net have also been shown to avoid the netting and do not attempt to pass through the open meshes around them. Different mesh configurations (diamond, square, hexagonal) present a different visual stimulus to fish and this stimulus is further affected by the color or contrast of the twine as viewed against the water background. Glass et al. (1993) also showed that fish reacted differently to component parts of a mesh. Horizontal bars elicited a different behavioral reaction to that shown when the fish were presented with vertically oriented bars. The importance of this study and the studies of Cui et al. (1991) and Wardle et al. (1991) is that they demonstrate that the natural behavior patterns of commercial fish species can be
Fish Vision and Its Role in Fish Capture
39
Figure 2.11. (A, B) A conceptual trawl design using optomotor response of fish to improve catch efficiency. (Courtesy Clem Wardle.)
altered by subtle changes in the surrounding visual stimulus. This led Glass et al. (1993) to postulate that by careful consideration of the behavior patterns of fish to surrounding panels of netting and by identification of the important components of the overall visual stimulus, it may be possible to create a system of visual illusions within a fishing net, which would stimulate the fish to approach and even penetrate meshes more readily. In a series of experiments in the laboratory and then at sea, Glass et al. (1995) and Glass and Wardle (1995) further investigated this concept to alter natural behavior patterns to encourage fish to escape. Throughout the fish capture process in trawls, there is little to encourage fish to escape or show avoidance responses. They postulated that a strong visual stimulus at the back of the net would encourage fish to attempt to escape through the meshes of the netting surrounding them rather than avoiding the meshes altogether. In developing the
strong visual stimulus that might be needed to elicit a response, they investigated the visual appearance of high-speed plankton samplers used to sample larval fish populations (Fig. 2.12). There was evidence of behavioral avoidance in larger-length classes, and this was thought to be induced by the high-contrast image of the approaching sampler, thereby allowing larger larval fish to react and swim out of its path (Glass et al. 1995). By recreating a similar high-contrast visual stimulus in laboratory experiments, Glass et al. (1995) identified that fish could, in fact, be induced to pass through meshes surrounding them when it appeared that alternative clear paths were either blocked or appeared to be blocked by visual patterns. To create the illusion that the path through the net was blocked, Glass and Wardle (1995) recreated a stimulus identical to that developed in laboratory experiments by placing a simple black tunnel in the net (Fig. 2.13). The fish were free to move through the tunnel but, when
40
Locomotion and Sensory Capabilities in Marine Fish
viewed from the front, the illusion of the path being blocked elicited a significant modification in the behavior. Fish were reluctant to pass through the tunnel, swam vigorously ahead of it, and responded by attempting to swim through open meshes just
ahead of the black tunnel. This illustrated how the knowledge of the underwater visual field and the behavior and sensory physiology of fish can be used in an applied manner to manipulate behavior of fish within the context of fishing operations.
Figure 2.12. Aberdeen Gulf III plankton sampler as viewed from the front. The engulfing black hole is believed to have induced larger plankton to escape by swimming out of the pass. (Wardle 1983.)
2.7 CONCLUDING REMARKS While the visual system of fish is well understood, there is a lack of research on its function and role in mediating behavior in commercial marine fish species during capture processes. Visual capacity of fish can be described by visual acuity, spectral sensitivity, and motion detection ability. Fish vision plays an important part in its reaction to fishing gears in catching a prey, and even more so in their attraction to light sources during light fishing. Many fish species have a two-gear system in vision provided by rod and cone cells in their retina, responsible for dark (scotopic) and light (photopic) conditions, respectively. Despite the earlier effort on modification and manipulation of visual stimulus and subsequent behavioral modification of fish within and around a net, this field of research has yet to achieve its full potential. With improvements in technology, there is tremendous scope for
Figure 2.13. Schematic illustration of the black tunnel (A) and underwater photograph of Atlantic cod (Gadus morhua) escaping just ahead of the black tunnel (B). (Crown copyright, reproduced with the permission of Marine Scotland.)
Fish Vision and Its Role in Fish Capture herding, guiding, and manipulating fish within and around a net, by use of their visual systems, in both light and dark conditions. Implications for improved fishing efficiency, conservation of fish stocks through reduction of bycatch and discard, reduction of interaction, and mortality of protected species may potentially revolutionize fishing operations. REFERENCES Anthony PD. 1981. Visual contrast thresholds in the cod Gadus morhua L. J. Fish Biol. 19: 87–104. Arakawa H, Watanabe T and Morikawa Y. 2007. Visual contrast threshold of striped beak-perch Oplegnathus fasciatus. Fish. Sci. 73: 469–471. Arimoto T and Namba K. (eds). 1996. Fish Behavior and Physiology for Fish Capture Technology. Tokyo: Koseisha-Koseikaku. 128 pp. (in Japanese). Atema J, Fay RR, Popper AN and Tavolga WN. 1988. Sensory Biology of Aquatic Animals. New York: Spring-Verlag. 936 pp. Bainbridge R. 1958. The speed of swimming of fish as related to size and to the frequency and the amplitude of the tail beat. J. Exp. Biol. 35: 109–133. Ben-Yami M. 1976. Fishing with Light: FAO Fishing Manuals. Surrey: Fishing News Books. 150 pp. Ben-Yami M. 1988. Attracting Fish with Light. FAO Training Series—14. Rome: FAO. 72 pp. Blaxter JHS. 1970. Light. Animals. Fishes. In: Kinne O (ed). Marine Ecology, Vol.1. Environmental Factors. pp 213–320. London: Wiley. Blaxter JHS. 1988. Sensory performance, behavior, and ecology of fish. In: Atema J, Fay RR, Popper AN and Tavolga WN (eds). Sensory Biology of Aquatic Animals. pp 203–232. New York: Springer-Verlag. Blaxter JHS, Parrish BB and Dickson W. 1964. The importance of vision in reactions of fish to drift nets and trawls. In: Modern Fishing Gear of the World 2. pp 529–536. London: Fishing News Books. Collin SP and Marshall NJ. 2003. Sensory Processing in Aquatic Environments. New York: SpringerVerlag. 446 pp. Collin SP and Pettigrew JD. 1989. Quantitative comparison of the limits on visual spatial resolution set by the ganglion cell layer in twelve species of reef teleosts. Brain Behav. Evol. 34: 184–192. Coutant CC (ed). 2001. Behavioral Technologies for Fish Guidance. Am. Fish. Soc. Symp. 26: 193 p. Crescitelli F (ed). 1977. Handbook of Sensory Physiology, Vol. VII/5: The Visual System in Vertebrates. New York: Springer-Verlag. 813 pp.
41
Cui G, Wardle CS, Glass CW, Johnstone ADF and Mojsiewicz WR. 1991. Light level thresholds for visual reaction of mackerel, Scomber scombrus L., to colored monofilament nylon gillnet materials. Fish. Res. 10: 255–263. Douglas RH and Djamgoz MBA (eds).1990. The Visual System of Fish. Devon: Chapman and Hall. 526 pp. Douglas RH and Hawryshyn CW. 1990. Behavioral studies of fish vision: an analysis of visual capabilities. In: Douglas RH and Djamgoz MBA (eds). The Visual System of Fish. pp 373–418. Devon: Chapman and Hall. Fernald RD. 1985. Growth of the teleost eye: novel solutions to complex constraints. Environ. Biol. Fish. 13: 113–123. Fernald RD. 1990. The optical system of fishes. In: Douglas RH and Djamgoz MBA (eds). The Visual System of Fish. pp 45–61. Devon: Chapman and Hall. Fritsches KA, Brill RW and Warrant EJ. 2005. Warm eyes provide superior vision in swordfishes. Curr. Biol. 15: 55–58. Glass CW and Wardle CS. 1995. Studies on the use of visual stimuli to control fish escape from codends. II. The effect of a black tunnel on the reaction behavior of fish in otter trawl codends. Fish. Res. 23: 165–174. Glass CW, Wardle CS and Gosden SJ. 1993. Behavioral studies of the principles underlying mesh penetration by fish. ICES Mar. Sci. Symp. 196: 92–97. Glass CW, Wardle CS, Gosden SJ and Racey DN. 1995. Studies on the use of visual stimuli to control fish escape from codends. I. Laboratory studies on the effect of a black tunnel on mesh penetration. Fish. Res. 23: 157–164. Glass CW, Wardle CS and Mojsiewicz WR. 1986. A light intensity threshold for schooling in the Atlantic mackerel (Scomber scombrus). J. Fish Biol. 29A: 71–81. Graham N, Jones EG and Reid DG. 2004. Review of technological advances for the study of fish behavior in relation to demersal fishing trawls. ICES J. Mar. Sci. 61: 1036–1043. Guthrie DM and Muntz WRA. 1993. Role of vision in fish behavior. In: Pitcher TJ (ed). Behavior of Teleost Fishes. 2nd ed. pp 89–128. Devon: Chapman and Hall. Hajar MAI. 2007. Visual Physiology of Fish for Understanding the Capture Process of Light Fishing. PhD thesis. Tokyo University of Marine Science and Technology, Tokyo, Japan.
42
Locomotion and Sensory Capabilities in Marine Fish
Hajar MAI, Inada H, Hasobe M and Arimoto T. 2008. Visual acuity of Pacific saury Cololabis saira for understanding capture process. Fish. Sci. 74: 461–468. Harden-Jones FR. 1963. The reaction of fish to moving backgrounds. J. Exp. Biol. 40: 437–446. Hasegawa E. 2006. Comparison of the spectral sensitivity of three species of juvenile salmonids. J. Fish Biol. 68: 1903–1908. He P and Wardle CS. 1988. Endurance at intermediate swimming speeds of Atlantic mackerel, Scomber scombrus L., herring, Clupea harengus L., and saithe, Pollachius virens L. J. Fish Biol. 33: 255–266. Inada H and Arimoto T. 2007. Trends on research and development of fishing light in Japan. J. Illum. Eng. Inst. Jpn. 91: 199–209. Inoue M and Arimoto T. 1976. On the optomotor reaction of fish relevant to fishing method. III. Experiments for fishing purpose. J. Tokyo Univ. Fish. 63: 9–16. Inoue M and Kondo T. 1972. On the optomotor reaction of fish relevant to fishing method. J. Tokyo Univ. Fish. 58: 9–16. Jerlov NG. 1964. Optical classification of ocean water. In: Tyler JE (ed). Physical Aspects of Light in the Sea. pp 45–49. Honolulu: Univ. Hawaii Press. Jones E, Glass C and Milliken H. 2005. The reaction and behavior of fish to visual components of fishing gears and the effect on catchability in survey and commercial situations. ICES CM 2004/B: 05, ref. ACE. Annex 2. pp 68–112. Kawamoto N. 1970. Fish Physiology. Tokyo: Koseisha Koseikaku. 554 pp. (in Japanese). Kim YH. 1998. Modeling on contrast threshold and minimum resolvable angle of fish vision. Bull. Kor. Soc. Fish. Technol. 33(3): 43–51. Kim YH and Wardle CS. 1998a. Modeling the visual stimulus of towed fishing gear. Fish. Res. 34: 165–177. Kim YH and Wardle CS. 1998b. Measuring the brightness contrast of fishing gear, the visual stimulus for fish capture. Fish. Res. 34: 153–166. Kim YH and Wardle CS. 2003. Optomotor response and erratic response: quantitative analysis of fish reaction to towed fishing gears. Fish. Res. 60: 455–470. Kobayashi H. 1962. A comparative study on electroretinogram in fish, with special reference to ecological aspects. J. Shimonoseki Coll. Fish. 11(3): 17–148.
Landis C. 1954. Determinants of the critical flickerfusion threshold. Physiol. Rev. 34: 259–286. Loew ER and McFarland WN. 1990. The underwater visual environment. In: Douglas RH and Djamgoz MBA (eds). The Visual System of Fish. pp 1–43. Devon: Chapman and Hall. Lythgoe JN. 1979. The Ecology of Vision. Oxford: Oxford University Press. 244 pp. Miyagi M. 2001. Study on Visual Physiology of Fish for Application in Net Gears. PhD thesis. Tokyo University of Fisheries, Tokyo, Japan. (In Japanese). Miyagi M, Akiyama S and Arimto T. 2001. The development of visual acuity in yellowtail Seriola quinqueradiata. Bull. Jpn. Soc. Sci. Fish. 67(3): 455–459 (in Japanese with English abstract). Nicol JAC. 1989. The Eyes of Fishes. Oxford: Oxford University Press. 308 pp. Parrish BB. 1969. A review of some experimental studies of fish reactions to stationary and moving objects of relevance to fish capture processes. FAO Fish. Rep. 62: 233–245. Pavlov DS. 1969. The optomotor reaction of fishes. FAO Fish. Rep. 62: 803–808. Protasov VR. 1970. Vision and near orientation of fish. Trans. by M. Raveh for Israel Program for Scientific Translations. Washington, DC: U.S. Dept of Commerce. 175 pp. Sbikin YN. 1981. The optomotor reaction and some characteristics of the vision of young sturgeon. J. Ichthyol. 21: 167–171. Sharp GD and Dizon AE. 1978. The Physiological Ecology of Tunas. New York: Academic Press. 485 pp. Shaw E. 1965. The optomotor response and the schooling of fish. ICNAF Spec. Publications 6: 753–755. Shiobara Y, Akiyama S and Arimoto T. 1998. Developmental changes in the visual acuity of red sea bream Pagrus major. Fish. Sci. 64: 944–947. Shiobara Y and Arimoto T. 1999. Behavioral analysis of feeding experiment on visual axis of red sea bream Pagrus major. Bull. Jpn. Soc. Sci. Fish. 65: 728–731(in Japanese with English abstract). Shiobara Y and Arimoto T. 2003. Change in visual acuity and retinal adaptation according to light intensity for red sea bream Pagrus mojor. Nippon Suisan Gakkaishi. 69: 632–636. Siriraksophon S and Morinaga T. 1996. Effect of background brightness on the visual contrast threshold of the Japanese common squid. Fish. Sci. 62: 534–537. Siriraksophon S, Nakamura Y and Matsuike K. 1995. Study on visual contrast threshold in Japanese
Fish Vision and Its Role in Fish Capture common squid Todarodes pacificus. Fish. Sci. 61: 574–577. Sivak JG. 1990. Optical variability of fish lens. In: Douglas RH and Djamgoz MBA (eds). The Visual System of Fish. pp 63–80. Devon: Chapman and Hall. Tamura T. 1957. A study on visual perception in fish, especially on resolving power and accommodation. Bull. Jpn. Soc. Sci. Fish. 22: 536–557. Tamura T and Wisby WJ. 1963. The visual sense of pelagic fishes especially the visual axis and accommodation. Bull. Mar. Sci. Gulf Caribb. 13: 433–448. Urquhart GG and Stewart PAM. 1993. A review of techniques for the observation of fish behavior in the sea. ICES Mar. Sci. Symp. 196: 135–139. Wagner H-J. 1990. Retinal structure of fishes. In: Douglas RH and Djamgoz MBA (eds). The Visual System of Fish. pp 109–157. Devon: Chapman and Hall. Wardle CS. 1983. Fish reaction to towed fishing gears. In: Macdonald AG and Priede IG (eds). Experimental Biology at Sea. pp 168–195. London: Academic Press. Wardle CS. 1987. Investigating the behavior of fish during capture. In: Bailey RS and Parrish BB (eds). Developments in Fisheries Research in Scotland. pp. 139–155. Surrey: Fishing News Books.
43
Wardle CS. 1989. Understanding fish behavior can lead to more selective fishing gears. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 12– 18. St. John’s, Newfoundland: Marine Institute. Wardle CS. 1993. Fish behavior and fishing gear. In: Pitcher TJ (ed). Behavior of Teleost Fishes. 2nd ed. pp 609–643. London: Chapman & Hall. Wardle CS, Cui G, Mojsiewicz WR and Glass CW. 1991. The effect of color on the appearance of monofilament nylon underwater. Fish. Res. 10: 243–253. Zhang XM. 1992. Study on Visual Physiology of Fish for Applying Trawl Net Operation. PhD thesis. Tokyo University of Fisheries, Tokyo, Japan (in Japanese). Zhang XM, Akiyama S, Arimoto T, Inoue Y and Matsushita Y. 1993. Retinal adaptation of walleye pollock in trawl fishing ground of north Pacific. Bull. Jpn. Soc. Sci. Fish. 59: 481–485 (in Japanese with English abstract). Zhang XM and Arimoto T. 1993a. Electroretinogram critical fusion frequency of jack mackerel Trachurus japonicus by strobe light. J. Tokyo Univ. Fish. 80: 61–67. Zhang XM and Arimoto T. 1993b. Visual physiology of walleye pollock Theragra chalcogramma in relation to capture by trawl nets. ICES Mar. Sci. Symp. 196: 113–116.
44
Locomotion and Sensory Capabilities in Marine Fish
SPECIES MENTIONED IN THE TEXT Atlantic cod, Gadus morhua chum salmon, Oncorhynchus keta jack mackerel, Trachurus symmetricus Japanese common squid, Tadorodes pacificus Japanese eel, Anguilla japonica
Pacific saury, Cololabis saira red seabream, Pagrus major striped beak-perch, Oplegnathus fasicatus swordfish, Xiphias gladius walleye pollock, Theragra chalcogramma yellowtail, Seriola quinqueradiata
Chapter 3 Hearing in Marine Fish and Its Application in Fisheries Hong Young Yan, Kazuhiko Anraku, and Ricardo P. Babaran
ical processes involved in the acousticolateralis system and to make the best use of the knowledge in the designs of fishing gear and operation. This chapter reviews the fundamental knowledge of physical properties of sound and discusses the physiological characteristics of the ear. Methodologies involved in understanding fish hearing in terms of frequency range and hearing threshold are reviewed along with a discussion on how exposure to noise will impact the overall hearing abilities of targeted fishes. The last section of this chapter provides a few examples of practical applications of acoustic signals in fisheries either to attract or to expel fish.
3.1 INTRODUCTION The acousticolateralis system of fish is composed of two major structures—the inner ear and the lateral line. The inner ear is responsible for the balance and detection of acoustic signals, whereas the lateral line detects water-borne vibration signals (Hawkins 1986). These mechanosensory functions are crucial for the survival of fish. In terms of anatomical structure, the functional units of the inner ear are sensory hair cells and are used to detect underwater acoustical signals, whereas the neuromasts of the lateral line detect waterborne lowfrequency vibrations caused by physical as well as biological forces. The sensation of water-borne sound and vibration offers fish a dual detection system to measure mechanical disturbances of their environment. In turn, fish can listen to sounds produced by either conspecifics or heterospecifics, and they can take corresponding actions such as retreating or escalating agonistic behavior or being attracted to the source if the sounds are courtship signals. Likewise, sensation from the lateral line informs recipients of the presence of obstacles, predators or prey. In conjunction with their eyes, the lateral line system also participates in the schooling behavior of fish (Pitcher 1979). The successful operation of fisheries whether at the commercial or subsistence level requires proper designs of fishing gears and methods. In light of how the auditory functions of ears and the sensation abilities of lateral line modulate the behaviors of fish, it would be useful to understand the physiolog-
3.2 PROPERTIES OF UNDERWATER SOUND AND VIBRATION An understanding of two major properties of sound in terms of sound pressure and water particle motion (displacement, velocity, and acceleration) is necessary to understand the responses of the ear and the lateral line system. The propagations of sound pressure and particle motion are complicated because of large differences in the attenuation level related to the type of sound source, its frequency, and the distance from the stimulus. 3.2.1 Sound Source and Sound Field Sound propagation speed in air (340 m/s) and water (1500 m/s) are different because the acoustic impedance values (ρc) of sound (where ρ represents the density of medium and c is the velocity of sound) in the two media are different due to differences in
45
46
Locomotion and Sensory Capabilities in Marine Fish
media density. The acoustic impedance is a measure of the total reaction of a medium to sound transmission (i.e., the easiness of sound passing through the specific medium). The acoustic impedances of air and water are about 39.6 g cm−2 s−1 and 150,000 g cm−2 s−1, respectively. Harris and van Bergeijk (1962) described the propagation of sound pressure and water particle motion generated by two types of sound source—a monopole and a dipole source. A monopole sound source is represented by a pulsating air bubble in the water that changes its volume. Sounds generated from underwater speakers and fish gas bladders are considered monopole sounds. Following the expansion of the air bubble, water particles move along the radial direction relative to the center of the bubble, and this movement is transferred from one particle to the next. Displacement of water particles decreases in an inverse proportion to the square of the distance from the sound source. Furthermore, a compression wave (i.e., pressure wave) is also induced during the propagation process, because water has a slight compressibility. As sound propagates, its pressure decreases inversely with increasing distance from the source. Within this sound field, the magnitude is larger for particle motion within the distance of λ/2π where λ is the wavelength, while the pressure is larger outside the distance of λ/2π. The sound field within the distance of λ/2π from the source is termed the near field and the sound field outside of the distance of λ/2π is termed the far field. In the case of dipole sound source, sound is generated by the motion of the object without changing its volume. Displacement of water particles is the largest along the axis of the motion and decreases with the cube of the distance from the sound source. The magnitude of sound pressure, on the other hand, decreases with the square of the distance from the sound source. Hence, the propagation properties of sound pressure and water particle motion are different depending on the sound source. In the meantime, sound frequency affects the velocity and acceleration of particle motion but not the displacement. For example, the velocity of the particle motion u is expressed as u = 2πfd, where f and d are sound frequency and displacement, respectively. Therefore, velocity of particle motion is
higher in high-frequency sounds. In terms of attenuation of a sound pressure level (SPL), high-frequency sounds lose their energy more rapidly than do low-frequency sounds because of the high rate of absorption—that is, the transformation of sound energy to thermal energy. Hence, low-frequency sounds remain in the medium for longer distances, and this is why low-frequency sounds dominate the underwater world (Hawkins 1986). 3.2.2 Sound Pressure Level SPL in water can be calculated with the following equation: SPL ( dB) = 20 log ( p p0 ) where p is the pressure level of the sound in μPa and p0 is the reference pressure level. For underwater sounds, the reference pressure level is 1 μPa; therefore, underwater sound pressure is usually expressed as “dB re 1 μPa.” For sounds in air, p0 is 20 μPa, which is the hearing threshold for human with 1000-Hz sound stimulus. Prior to 1990, underwater acoustic studies used “dB re 1 μbar” as the measurement unit. To convert “dB re 1 μbar” data to “dB re 1 μPa,” 100 dB is added to the “dB re 1 μbar” data—that is, dB re 1 μPa = dB re 1 μbar + 100. 3.2.3 Variations of Underwater Sound Pressure Levels Characteristics of underwater sound vary with the location in water. Hatakeyama (1996) reviewed underwater sound pressure in relation to various sound sources and approximated auditory thresholds of fishes (Fig. 3.1). One of the distinctive differences of the underwater sound is the noise levels in shallow waters compared with those in deep waters. In coastal marine waters, snapping sounds of the pistol shrimp (Alpheoidea spp.) are usually very loud, but such sounds have not been detected in deep water or freshwater (Urick 1983). Hearing thresholds are much lower in otophysan fishes (Ostariophysi: Otophysi), also known as hearing specialists, which account for 64% of the freshwater species (Nelson 1994), than are nonotophysan fishes, which are called as hearing generalists (details described later). Artificial sounds generated
Hearing in Marine Fish and Its Application in Fisheries
47
Figure 3.1. Underwater sound pressure levels of various types of sound sources. (Redrawn from Hatakeyama 1992.)
from underwater piling drilling and dynamite explosions result in high SPLs. In general, highintensity sound is considered aversive for fishes and can cause damage to fish (e.g., dynamite fishing [high sound pressure in coupling with compression waves]). A 200-dB SPL sound would cause an estimated 1010 μPa (or about 100 gf/cm2) force to the fish. For the lowest threshold level of otophysan fishes is 60 dB, the force is about 103 μPa (or about 0.01 mgf/cm2), whereas the lowest threshold level for nonotophysan fishes is 90-dB SPL, equivalent to 3.16 × 104 μPa (or about 0.3 mgf/cm2). 3.3 UNDERWATER SOUND SOURCES AND THEIR CHARACTERISTICS 3.3.1 Natural Underwater Ambient Sounds Underwater ambient noise covers a wide range of frequencies from 1 Hz up to about 100 kHz (NRC 2003; Urick 1983). Urick (1983) classified sources of ambient noise in the ocean into the following six categories:
• Those resulted from tides and hydrostatic pressure changes of relatively large amplitude and at the low-frequency end of the spectrum • Seismic disturbances that generate noise between 1 and 100 Hz • Oceanic turbulence in the form of irregular random water currents of large or small scales (Wenz 1962)—For instance, steady current at 1 knot can generate noise around 106 dB (re 1 μPa). • Ship traffic that generates frequency in the range of 50 to 500 Hz—Such noise can be detected at distances of 1000 miles or more from the site of measurement. • Surface waves that caused noise in the frequencies between 1 and 50 kHz (NRC 2003)—When below 5 to 10 Hz, the dominant ambient noise source was the nonlinear interaction of oppositely propagating ocean surface waves. • Noise caused by precipitation (rain, hail, and snow)—The spectrum of rain noise, for wind speeds below 1.5 m/s, showed a peak at 13.5 kHz
48
Locomotion and Sensory Capabilities in Marine Fish with a sharp cutoff on the low-frequency side and a gradual fall-off (7 dB per octave) on the high-frequency side.
3.3.2 Sounds Produced by Fishes Many fish produce calls as part of a specific behavioral pattern. Sounds are believed to elicit changes in the behavior of other individuals of the same or different species. Sounds vary in structure and characteristics depending on the mechanism used to produce them. In general, two major sound types are produced by fish. The stridulatory sounds result from rubbing hard parts of the body. For example, members of the grunt family, Pomadasyidae, produce a sharp, vibrant call by grating a dorsal patch of pharyngeal denticles against smaller ventral patches. Some catfish (Family: Siluridae, e.g., channel catfish Ictalurus punctatus) produce a squeak when the enlarged pectoral spines are moved against each other (Hawkins 1986; Fine et al. 1997). The second type of sound, the drumming sound, is produced by contraction of skeletal muscles along the body wall against a gas-holding structure, such as gas bladder. The notable cases are sounds produced in oyster toadfish (Opsanus tau), croaker (Micropogonius undulates) (Fine et al. 1997, 2001), and fish of the Family Gadiade, such as haddock (Melanogrammus aeglefinus), cod (Gadus morhua), pollock (Pollachius pollachius), and Tadpole-fish (Raniceps raniceps) (Hawkins 1986). The major difference between stridulation sound and drumming sound is that the former tends to have wider frequency bandwidth than the latter. Recently, fishermen have use the knowledge and new technology (e.g., hydrophone) to locate aggregation of spawners of sciaenid fish in the coastal area of central Taiwan, and this has resulted in a large-quantity catch of sciaenids. This has raised concerns that the wild stock of these sciaenids could be depleted within a short time if the new technology is not properly managed (Tu et al. 2004). 3.3.3 Sounds Produced by Fishing, Research, and Whale-Watching Vessels Modern fishing and research vessels use diesel engines in conjunction with high-thrust propulsion systems that can generate significant levels of noises that are radiated underwater. Field tests
revealed that a ship with the combination of a diesel engine and generator produced noise in the frequency range of 8 to 6 kHz with SPL between 110 and 140 dB (Mitson and Knudsen 2003). The frequency range and SPL of vessel noise are thus in the hearing range of many commercial important marine fish species such as cod (10–600 Hz, 60– 140 dB) and herring (50–1500 Hz, 55–150 dB) (Astrup and Møhl 1993; Chapman and Hawkins 1973; Blaxter et al. 1981; Enger 1967; Sand and Karlsen 1986; Schwarz and Greer 1984). Winger (2004) reported that Atlantic cod responded to an approaching vessel from as far as 1500 m. Over the past few decades, whale-watching has been promoted in many parts of the world and has become an important tourist industry (Hoyt 2000). A field study conducted in the Juan de Fuca Strait area of southern British Columbia and northwestern Washington, where killer whale watching is a significant business, revealed that SPLs of noise generated by the vessels were in the range of 145 to 169 dB (100–20 kHz). The recorded killer whale call source levels were 105 to 124 dB and the audiogram showed best frequency at 20 kHz (range 100– 100 kHz) with a hearing threshold of 40 dB. These data clearly indicated that noises generated from whale-watching vessels could be perceived by whales and most of the fish in the vicinity of the vessels (Erbe 2002). The long-term consequences of noise exposure generated by whale-watching vessels on dolphins and whales remain to be examined. Bottom trawl is carried out by towing a net over the bottom of the ocean (see Chapters 4 and 12). Such operations result in underwater noise when the fishing gear makes contact with the seabed. Buerkle (1968, 1977) reported that Atlantic cod were able to detect noise generated by a bottom trawl at a range of at least 2.5 km. Response of fish as indicated by a change in behavior due to noises from a combination of the trawler and the trawl it was towing has been reported (Winger 2004). 3.4 GENERAL MORPHOLOGY AND FUNCTIONS OF INNER EARS AND ANCILLARY STRUCTURES The inner ear of fish, including elasmobranches, consists of three semicircular canals (with associ-
Hearing in Marine Fish and Its Application in Fisheries
49
Figure 3.2. Anatomical structure of a typical fish (croaking gourmi [Trichopis pumila]) inner ear. The inner ear was stained with osmium tetraoxide to enhance contrast.
ated cristae ampullaris) and three otolithic end organs: the saccule, utricle, and lagena (Fig. 3.2). Despite some variations in the structure of otolithic organs in fishes, the basic functional morphology is essentially the same among fishes (Popper and Fay 1999). Lining some portions of the wall of the organ is a piece of sensory epithelium that contains sensory hair cells and supporting cells. Above the sensory hair cells, the sac also contains an otolith, a calcium carbonate structure that lies close to the sensory hair cells. Differences in density between the otolith and the adjacent sensory hair cells trigger relative movement between sensory hair cells and the otolith when a sound wave passes through the ear. Because of a denser mass of the otolith, its movement is smaller than that of the sensory hair cells, causing the bending of cilia bundles as well as kinocilia on top of the sensory hair cells. The shearing action between sensory hair cells and the otolith generates evoked potentials, which are then transmitted along ascending auditory neuronal pathways to the central hearing structures. In addi-
tion to hearing end organs, in some fish, some ancillary structures such as gas bladder or otic gas bladders also aid hearing by picking up the pressure component of the sound into the ear through direct or indirect contact with the hearing end organs. 3.4.1 Hearing Abilities of Fish Because of the similar acoustic impedance of the fish body and the surrounding water, the fish body is considered transparent to passing sound waves. Due to its physical nature, only lowfrequency sound with high energy can be perceived by such direct stimulation mechanism of sensory hair cells. Therefore, for most fishes that rely on hearing only through particle stimulation mechanism, their hearing ability is limited to a narrow frequency band (less than 1000 Hz) with high sound pressure threshold (as high as 120 dB at the best frequency). Such fishes are hence termed “hearing generalist” species. Certain species, however, evolved mechanisms to enhance their hearing through gas-containing structures that are coupled
50
Locomotion and Sensory Capabilities in Marine Fish
to the inner ears. The low-density gas that is enclosed inside the gas bladder changes volume when sound waves pass through the fish. It is generally believed that the passing sound waves lead to compression and expansion of the gas inside the gas bladder and omnidirectional sound is generated. The transmission of the resonant sound entering into the ears contributes to the hearing ability of fish. However, there are many studies showing exceptions to this generalization, as reported by Connaughton et al. (1997), Barimo and Fine (1998), Yan et al. (2000), and Fine et al. (2001, 2004), in which the gas bladder of weakfish, toadfish, goby, and gouramis does not contribute to auditory function. However, fishes in the superorder Ostariophysi (e.g., cyprinoids, characoids, and siluroids) have a specialized mechanical coupling structure (i.e., the Weberian ossicles) that connect the gas bladder to the inner ear (Furukawa and Ishii 1967). Hence, vibrations caused by the passing sound to the gas bladder are transmitted to the ears and hearing abilities are enhanced. Because of their extended hearing frequency range (up to 8000 Hz in certain catfish) and low thresholds (60 dB in goldfish), these fishes are called “hearing specialist” species. In addition, some species have either an otic gas bladder that is attached directly to the saccule or have embedded the inner ear adjacent to the suprabranchial chamber where a pocket of air is enclosed to pick up pressure component of the sound (Yan 1998; Yan and Curtsinger 2000, Yan et al. 2000). The radiographs of carp (Cyprinus carpio) (a hearing specialist), red sea bream (Pagrus major) (a hearing generalist), and bastard halibut (Paralichthys olivaceus) (a hearing generalist) show that the former two species have gas bladders and the latter does not (Fig. 3.3.). Not all fish with a gas bladder can be classified as a hearing specialist if they lack a mechanical coupling between the gas bladder and the inner ear (e.g., red sea bream). For these species, the pressure component of the passing sound cannot be picked up and transmitted into the inner ear and hence the hearing ability is limited. Figure 3.4 shows audiograms obtained from several commercially harvested species: bastard halibut (Fujieda et al. 1996), red sea bream (Ishioka et al. 1988), jacopever (Sebastes schlegeli) (Motomatsu et al. 1996), walleye pollock (Theragra
Figure 3.3. Soft radiographs of carp Cyprinus carpio (a hearing specialist) (A), red sea bream Pagrus major (a hearing generalist) (B), and bastard halibut Paralichthys olivaceus (a hearing generalist) (C). S indicates swimbladder.
chalcogramma) (Park and Iida 1998), Japanese jack mackerel (Trachurus japonicus) (Babaran et al. unpublished data), spotlined sardine (Sardinops melanostictus) (Akamatsu et al. 2003), and masu salmon (Oncorhynchus masou) (Kojima et al. 1992). Although sound frequencies were limited to 2000 Hz during tests, the audiograms clearly indicate that the most sensitive frequency range lies between 100 and 1000 Hz. Many other fish species, such as cichlid fish (Astronotus ocellatus) and European eel (Anguilla anguilla), are sensitive only to low-frequency sounds (Jerkøet al. 1989; Yan and Popper 1992). Some commercially important species such as dab (Limanda limanda) have a hearing frequency range of 30 to 200 Hz with the
Hearing in Marine Fish and Its Application in Fisheries
Figure 3.4. Auditory threshold curves (audiograms) of (Po) bastard halibut (Fujieda et al. 1996), (P) red sea bream (Ishioka et al. 1988), (S) jacopever Sebastes schlegeli (Motomatsu et al. 1996), (T) walleye pollock Theragra chalcogramma (Park and Iida 1998), (Tj) Japanese jack mackerel Trachurus japonicus (Babaran et al. unpublished), (Sm) Spotlined sardine Sardinops melanostictus (Akamatsu et al. 2003), and (O) masu salmon Oncorhynchus masou. (Kojima et al. 1992.)
best frequency at 100 Hz, whereas the hearing ability of Atlantic cod is limited to between 30 and 400 Hz with best frequency at 100 Hz. The Atlantic salmon (Salmo salar) has a hearing range of 30 to 300 Hz with the best frequency at 150 Hz (Hawkins 1986). Overall, these three important commercial species have very limited hearing frequency range and high hearing thresholds. Because of their limited hearing ability, they are not easily affected by ambient noises of low magnitudes. Hearing abilities of cetaceans are in the range of 50 Hz to 100 kHz with best hearing frequency around 10 kHz to 50 KHz and thresholds around 30 to 50 dB (NRC 2003). Because the echoes they produce have wavelengths greater than the diameter of the filaments or twines of a gillnet, the sound could pass through the filament without producing
51
an echo. Cetaceans relying on echo location can thus swim into the gillnet, resulting in bycatch and mortalities. Two methods of gear modifications have been tested to reduce small cetacean bycatch. Acoustic deterrents or pingers that emit highfrequency sounds were used to alert marine mammals of the presence of the fishing gear (Kraus et al. 1997; Goodson 1997; Mooney et al. 2007). The practices were applied to many European and North American fisheries and showed some successes. However, major drawbacks of using pingers are high cost, habituation of cetaceans, wide variability of success rates, and the unintended “dinnerbell” effect, which in fact resulted in unintended outcome of luring cetaceans to the nets to take fish caught in nets (Trippel et al. 2003). The second method is the use of an alternative gear materials by modifying the characteristics of the filaments of the gillnet to increase their acoustic reflectivity (i.e., target strength [TS]), making them easier to be detected by echo-locating odontocetes. Net alterations included air-filled monofilament nylon, multifilament nets, weighted filaments woven into nets, and adding a filler (e.g., barium sulfate) to the nylon to increase net density (Trippel et al. 2003). The latest tests showed that barium sulfate– and iron oxide–enhanced nets increased reflectivity compared with control nets, with the barium sulfate nets generating the highest TS values. The tested TS values indicated that dolphins should be able to detect these nets in time to avoid contact and entanglement, but porpoises, with typically lower source levels, may not detect nets at a range great enough to avoid entanglement (Mooney et al. 2007). These researches are ongoing regarding different mammals and in different fisheries. Further discussions on this topic can be found in Chapters 7, 8, and 13. 3.4.2 Research Techniques on Auditory Physiology of Fish Traditionally, studies of fish hearing have used behavioral or electrophysiological methods. Behavioral methods are based on conditioning the fish with acoustic signals in conjunction with either reward or punishment (Fay 1969; Yan 1995; Yan and Popper 1991). These psychoacoustical methods are time consuming, with experimental paces dictated by physical or cognitive conditions of the test
52
Locomotion and Sensory Capabilities in Marine Fish
subject. In addition, behavioral responses near the threshold fluctuated greatly even in a specific individual. Therefore, threshold determinations are made with the aid of mathematical probability paradigms. The electrophysiological methods record neuronal activities either from microphonics of auditory organs (Saidel and Popper 1987) or from single-unit recordings, which register single nerve fiber discharge patterns in response to acoustic signals (Enger and Anderson 1967). These two invasive electrophysiological methods have some common limitations. Preparations are rather complicated and invasive surgery is needed. The placement of electrodes is restricted to specific end organs or fibers; therefore, responses do not accurately represent the whole auditory pathways. Lately, a noninvasive auditory brainstem response (ABR) method was developed to measure responses of the whole auditory pathways in fish (Kenyon et al. 1998). Advantages of the ABR protocol include noninvasiveness, rapid completion, and measurement of relevant neuronal cells participated in the signal processing. These advantages are in stark contrast to qualitative psychoacoustical method or invasive electrophysiological method. The ABR method involves simple sedation of test fish, placement of recording electrodes on the cephalic region, and presentation of acoustic stimuli via either airborne sound or waterborne sound to the subject. Recordings of evoked potentials are through electrodes to a two-stage amplifier, averaged, and displayed on a computer screen. Determinations of hearing threshold (in terms of dB; re 1 μPa) of a particular frequency are made using either the traditional visual inspection method (Kenyon et al. 1998) or statistical method (Yan 1998). The whole process of obtaining an auditory tuning curve can be completed in less than 2 hours. The ABR system has been widely used by auditory research scientists around the world since its first publication in 1998. The ABR method has become the de facto standard method for fish auditory physiology research. 3.4.3 Effects of Noise Exposure on Hearing Ability of Fishes Urick (1983) chronicled noises that come from seismic disturbances, oceanic turbulence, ship traffic, surface waves, thermal noise, and coastal
water wave actions as well as biological sounds (e.g., calls of porpoises, noises of a mass of snapping shrimps). Snapping shrimps can produce sound ranging from 700 to 30,000 Hz with SPL as high as 70 dB (re 1 μPa), and the croaking sound of croakers (family: Sciaenidae) had a frequency range of 100 to 3000 Hz (dominant frequency around 200 Hz) with the highest sound pressure greater than 110 dB (re 1 μPa) (Barimo and Fine 1998; Fine et al. 2004; Urick 1983). The ambient noise caused by rain, as recorded in Long Island Sound, New York, showed a frequency range of 700 to 20,000 Hz with sound pressure as high as 85 dB (re 1 μPa) (Urick 1983). These natural or biological underwater sounds generally exert no harm to fish. However, anthropogenic sounds (i.e., man-made noises) have been increasingly become an issue that could harm the welfare of fish and other marine animals (Richardson and Würsig 1997). In a series of pioneering experiments, Scholik and Yan (2001, 2002a, 2002b) and Scholik et al. (2004) demonstrated that exposure to the noise (band width: 100–6000 Hz, with dominant frequency at 1300 Hz) generated by an 55-horsepower (hp) outboard engine hampered the hearing ability of a hearing specialist, the fathead minnow (Pimephales promelas), and led to elevation of the hearing threshold at 1, 1.5, and 2 kHz for 7.8, 13.5, and 10.5 dB, respectively. Further experiments by exposing fathead minnow to white noise (i.e., different frequencies of sound with equal energy) of 142 dB (band width 0.3–4 KHz) for either 1, 2, 4, 8, and 24 hours showed elevation of thresholds at 0.8, 1, 1.5, and 2 kHz. Even 14 days after exposure, hearing thresholds for 1.5 and 2 kHz were still significantly higher than the baseline data. The threshold shifts were, however, not observed in a hearing generalist species, the bluegill sunfish (Lepomis macrochirus) (Scholik and Yan 2002b). These results indicate that hearing specialist species are more vulnerable to negative impacts of prolonged exposure to noise than are hearing generalist species. A follow-up study by Scholik et al. (2004) showed that the negative effect of noise exposure to fathead minnow can be mitigated by feeding diets added with vitamin E at a dose of 450 mg/kg. The rationale of using vitamin E to offset the deleterious effect of reactive oxygen species (ROS) (i.e.,
Hearing in Marine Fish and Its Application in Fisheries free radicals) from the acoustic trauma is based on the chain-braking antioxidant effect of vitamin E to neutralize ROS (Chow et al. 1999). Following Scholik and Yan’s work, the study of noise effects has become a research topic of interest and a number of laboratories have examined the various effects of noise exposure on fish hearing. For examples, the effect of powerboat races in Alpine lake, exposure to laboratory-induced noises, exposure to seismic airguns, and ship noises (e.g., McCauley et al. 2003) have been examined. These studies provide further evidence that underwater noise exerts a negative impact on fish hearing, communication and can causes behavioral changes. Interestingly, two gobies living under the waterfall areas in Italy evolved hearing ability and produce sounds outside the frequency range of the background noise (Lugli et al. 2003). Yan et al. (2006) found that the levels of noises generated from various types of aerators used in aquaculture ponds were between 119 dB and 154 dB. These noise levels were found higher than hearing thresholds of 15 species of fish and shrimp commonly cultured in Taiwan. The long-term effects of noises generated from aerators remain to be investigated and attention should be paid to improve the design of aerators with less noise so as not to exert a negative effect on the growth and welfare of cultured fish and shrimp. 3.5 RESPONSES OF FISH TO SOUND AND ITS APPLICATION IN FISHERIES 3.5.1 Acoustic Attraction Underwater sound travels at a speed of about 1500m/s and may be used to control fish behavior over a longer distance compared with chemical or visual stimuli. Several applications of the use of sound in fishery operation to attract fish have been reviewed by Hashimoto and Maniwa (1964, 1966), and Maniwa and Hatakeyama (1970, 1975). For example, it was reported that fish schools can be driven into the set-net by the vocal sound of Risso’s dolphin (Grampus griseus) and that the yellowtail (Seriola quinqueradiata) could be attracted to the surface from a deep layer on the fishing ground by the swimming and feeding sounds of conspecifics.
53
One of Japanese traditional fishing methods, the “donburi” or “boko,” uses acoustical signals to attract fish to the desired fishing area. This technique is still being used in Shibushi Township, Kagoshima Prefecture, Japan, to harvest demersal fishes, such as red sea bream (Pagrus major). A donburi used in the Kagoshima area is a device consisting of a conical shaped lead, measuring 126 mm high and 47 mm at the bottom diameter and weighing about 880 g (Fig. 3.5). It has a 22-mm diameter and 48-mm-deep opening at the bottom and is held by a 7-m-long rope. donburi sizes vary among fishermen, but it is generally believed that larger ones are better because they can produce higher-intensity sounds than can smaller ones. The donburi is deployed while a fishing vessel is anchored and its engine is turned off. A fisherman throws the donburi so that it hits the water surface perpendicularly, to generate sound and to form a column of tiny bubbles while sinking. Air bubbles rise to the water surface for longer than 10 s. Nonperpendicular casts result in larger bubbles and a weaker sound. A fisherman casts donburi 10 to 20 times during one fishing operation. Fishing is carried out by using a handline while the donburi is cast. Catch rates increase gradually with fishing depth gradually raising from the bottom to a shallow layer by as much as 10 m. The donburi is effective in fishing grounds of 40 to 50 m in depth for red sea bream (Pagrus major), crimson sea bream (Evynnis japonica), threestripe tigerfish (Terapon jarbua), and sharpnose tigerfish (Rhyncopelates oxyrhynchus). To record sounds generated by the donburi, two replicas of fisherman’s gear shown in Figure 3.5 were made (K. Anraku, unpublished data)—one with an opening on the bottom and the other one without the opening. Sound recordings were made under quiet conditions in Lake Ikeda, Kagoshima Prefecture of Japan, at water depth of 30 m. The donburi produced sounds of the highest intensity when it hit the water surface, with sounds generated by air bubbles following (see sonograph in Fig. 3.6A; impact time at about 2.6 s mark). Power spectrum analyses indicate a broadband sound with frequencies ranging from 1 to greater than 10 kHz (Fig. 3.6B, spectrograph).The SPLs were measured (and averaged) at 1, 5, 10, 20, and 30 m, respectively,
54
Locomotion and Sensory Capabilities in Marine Fish
Figure 3.5. Fish attraction device used by Japanese fishermen, locally called donburi and used in Shibushi town, Kagoshima, Japan.
while the donburi was cast repeatedly (Fig. 3.5B). Even at 30-m water depth (Fig. 3.6C), both types of donburi produce sounds greater than 100 dB. Sounds produced by the donburi with a hole in the bottom of it (Fig. 3.6C) were 4.3 dB higher than sounds for the one without a hole. It appears that function of the bottom hole in a donburi is to produce louder sounds. 3.5.2 Use of Sound in Fish Guidance Devices The release of artificially raised fish seedlings into coastal waters has been an integral part of sea ranching operations in Japan for more than 40 years (Shishidou 2002). Proper introduction of seedlings into reef areas is crucial for their initial survival. The Research Institute of Oita Prefectural Government in Japan first applied the sound conditioning method in their marine ranching projects to prevent fish from dispersing and to enhance the recapture rate of released fish (Kamijyo 1998). Here we describe an active fish guidance method that uses an acoustic conditioning technique to transport fish over a long distance (Anraku et al. 2006a,
2006b). The method allows the transport of fish seedlings to a desired location in the sea without physical handling so to minimize physiological stress and physical injury. It combines acoustic, visual, and feeding stimuli to control fish during field transportation. The research involved red sea bream, which is one of the major sea ranching species in Japan. The following descriptions are mainly from work by Anraku et al. (2006a, 2006b). Conditioning About 50,000 juvenile red sea bream of about 50 mm in body length were held in a net cage for conditioning use. They were fed with artificial pellets while a sinusoidal tone burst (pulse width and pulse interval were both 0.75 s) with a frequency of 300 Hz was played to condition the fish with sound and food. By repeating such food and acoustic coupling conditioning, fish learned to associate feeding with acoustic stimuli. Once they were conditioned, the fish would respond to sound stimuli by swimming rapidly close to the water surface even before feed pellets were given. A visual target
Hearing in Marine Fish and Its Application in Fisheries
55
Figure 3.6. (A) Sound wave and (B) sound spectrograph recorded before and after a donburi is cast to sea (impact time at 2.6 s). (C) The mean sound pressure levels recorded at the different water depths, which were generated by a donburi with a hole. Error bar indicates standard deviation of means of 10 measurements. (K. Anraku, unpublished data.)
made of strings of blue plastic ribbons (each string was 15 cm in length) served as the aggregation point for fish. The underwater speaker was suspended in water at all times while the visual target was suspended only during conditioning. Feed pellets were given 0.5 to 1 min after sound was broadcasted. Each training session lasted between 5 and 10 min and a total of 20 to 30 training sessions were conducted each day for 9 days. The fishes’ response to the sound stimuli clearly changed within the first day but at least 3 days of training were required before dense aggregation around the visual target was observed. Once the fish were trained, fish would immediately respond to sound stimuli by increasing their swimming speed and swimming around the visual target in a schooling formation. The fish school followed the target even when its position was changed.
The Guiding Device The fish guidance device (FGD; Zeni Lite Buoy Co. Ltd., Tokyo, Japan) was made from an aluminum frame and equipped with a feeding device and an underwater speaker (Fig. 3.7). The feeding device was made of metal frames (2.5 m in length), PVC pipes, plastic hoses, and a water pump. This feeding apparatus dispenses feed pellets (particle size φ = 1.1–1.3 mm) through the holes (4 mm in diameter) drilled on the PVC pipes using a water pump. The blue plastic ribbons serving as visual targets for conditioning in the cage training stage were attached to the frame to aggregate the fish. Three underwater TV cameras were installed on the FGD to monitor fish behavior during the guidance experiments. An underwater speaker was attached to the towing boat to broadcast specific acoustic signals to the fish.
56
Locomotion and Sensory Capabilities in Marine Fish
Figure 3.7. (A) Structure of the fish guidance device (FGD). (B) The school of fish following the FDG. (C) Underwater camera view of the school following the FDG.
Field Trials in the Open Sea Fish guiding experiments with the use of FGD (Fig. 3.7) were conducted with acoustically conditioned juvenile red sea bream (55-mm mean body length), which were released into Kawashiri Port, Kagoshima, in Japan. Three separate guiding trials were conducted. The first two trials used 1000 conditioned fish, while the third trial used 700 individuals. In the first trial, the guiding distance was not prescribed and food was not given after leaving the port. In the following two trials, the guiding distances were prescribed to 1000 m and 3000 m, respectively, and food was given during the guidance trials. The feeding rate B (g/s) was calculated according to the following formula (Lovell 1977): B = 0.05 × N × W × V D
where N is the number of individuals following the FGD, W (g) is the average body weight of fish, V (m/s) is the towing speed of FGD, and D is the guiding distance (m). Earlier findings showed that satiated fish would not aggregate under the FGD and would not follow a moving FGD (K. Anraku, unpublished data). Daily satiation level of feeding was estimated as 5% of the fish body weight. The fish guiding experiment was terminated when either the distance traveled exceeded the prescribed guiding distance or the majority of the fish left the FGD. Figure 3.7B shows that fish actively swam toward feeding pipes and actively fed on pellets discharged. Fish formed a tightly packed school behind the pipes (Fig. 3.7C). Fish feeding at the front of the school were usually overtaken and then returned to
Hearing in Marine Fish and Its Application in Fisheries the trailing end of the school. Smaller fish with poor swimming ability usually fed on remains of pellets and were seldom seen at the front of the school. The number of fish remained with the FGD after leaving the port is shown in Figure 3.8. The guidance rate is the ratio of the number of fish with the FGD to the initial number. In the first trial (Figure 3.8A) during which no pellets were given, the initial total number of fish with the FGD was only 170 even
Figure 3.8. Changes in the numbers of fish (left y-axis) following the FGD (x-axis, in min) and changes in guidance rates (right y-axis) during guiding experiments in the open sea. (A) Guiding without the use of food pellets; (B) 1000-m guiding trial with food pellets given; (C) 3000-m guiding trial with food pellets given.
57
though a total of 1000 fish were released. This finding indicates that feeding remains an important incentive for fish to follow the FGD during the guidance. The guidance rate rapidly decreased with time elapsed and distance traveled (Fig. 3.8A). In trials B and C, a large number of fish remained with the FGD even after the targeted distance 1000 m was reached. Guidance rates rapidly decreased at the distance between 1000 and 2000 m, during which the towing boat was turning to change direction. The drop in guidance rate over the distance was likely the outcome of many fish that broke away from the FGD due to fluctuations in towing speed. The total distance traveled by the fish was very short in the first trial. Most fish were seen to swim down toward the bottom and formed various smaller schools. However, it was also observed that fish could be guided to a distance as far as 3000 m with a guidance rate as high as 40%. These guidance experiments showed that a large proportion of fish can reliably be guided for about 1000 m when combining acoustic, visual, and feeding stimuli. However, further experiments are needed to sort out the respective roles of acoustic, visual, and feeding cues in fish guiding.
3.5.3 Role of Sound and Hearing in Fish Aggregation Devices A traditional fish aggregation device (FAD) termed “payao” has long been used in certain parts of Philippine waters. A payao is an anchored FAD that is made up of a bamboo raft, anchoring rope, and a cement anchor weight (Fig. 3.9). The bamboo raft provides the buoyancy to the payao structure, serving as a marker to indicate its position and as the attachment for the series of palm fronds that play a role in fish aggregation. The anchor line is made of 14- to 16-mm φ polypropylene rope and tethers the raft to a 500-kg cement anchor. Similar structures have been used in Malaysia and Indonesia, where they are called unjam and rumpon, respectively (FAO; http://www.fao.org/fishery/equipment/fad/ en). Globally, the use of various kinds of FADs is widespread and covers many parts of the world’s oceans (Fonteneau et al. 1999; Kamijyo 1998). Fishing with various forms of FADs has long been practiced in Japan and in Mediterranean
58
Locomotion and Sensory Capabilities in Marine Fish
Figure 3.9. Schematic diagram of a fish aggregation device (payao in the Philippines).
regions, although most of the anchored structures are usually deployed in coastal waters. The practice of tuna fishing with anchored FADs started in the late 1960s when field tests were conducted with the deployment of drifting types of payaos in oceanic waters in the Philippines (Floyd 1985). The successful completion of that experiment led to the proliferation of payaos and a progressive increase in tuna production in the Philippines waters over the past three decades starting in the 1970s. Since then, similar structures were also deployed in Okinawa, Japan, Hawaii, and several other areas in the Pacific, Atlantic, and Indian Oceans. With the use of FADs, large tuna species (yellowfin tuna, Thunnus albacares; bigeye tuna, Thunnus obesus; and skipjack tuna, Katsuwonus pelamis), large pelagic species (dolphinfish, Coryphaena hippurus; blue marlin, Makaira mazarra; and sailfish Istioporus inducus), small tunas (bullet tuna, Auxis rochei; frigate tuna,
Auxis thazard; and kawakawa, Euthynnus affinis), and other small pelagic fish species (scads, Decapterus spp., and bigeye scad, Selar crumenophthalmus) have became regular targeted species (Fonteneau et al., 1999). There seems to be many incentives why fish prefer to associate with FADs. These include the search for food, the use of the FAD as shelter from predators, as a schooling companion, alternative environment, cleaning station, spatial reference, and meeting point of individuals to form schools, and the requirement of some fish to converge in biologically rich environments (Freon and Misund 1999). However, none of these hypotheses can adequately explain how and why fish are attracted to the FADs or how they remain aggregated within the effective area of the FADs. Because the understanding behind these processes is not yet clear, scientists are increasingly looking at the possibility that sound generated from FADs may provide an important acoustic sensory cue to explain the attraction or aggregation of fish near FADs like a payao. Westenberg (1953) previously suggested that the vibrations of the palm leaves of a rumpon in Indonesia may be responsible for the attraction of fish, and called for more research in this area of study. Until recently, however, there were no publications of the sounds generated by anchored FADs like a payao (Dempster and Tarquet 2004). Indeed, in payaos, surface waves and water currents acting on its parts may be responsible for the sounds generated with the structure itself (Dempster and Tarquet 2004). There are at least two possible sources of the sounds generated from the payao. One source is the raft itself, and the other is the anchor line. The mechanisms of sound generation by these two parts of the payao are different. The raft of a payao oscillates with the surface waves in the open sea, generating highly audible sounds. The sound is partly generated when the front end of the raft plows into incoming waves or when the rear end of the raft dips into the water with each passing wave. As it moves with the waves, the entire raft vibrates especially when it splashes down hard on the sea surface. However, field observations reveal that the ability of the raft to generate sound is not necessarily coupled to the big waves; even small waves acting
Hearing in Marine Fish and Its Application in Fisheries head-on against the hollow ends of the bamboo could generate a loud sound that is perceptible to the human ear at distances of greater than 50 m (authors’ personal observation, unpublished data). Meanwhile, the action of tidal currents on the anchor line of the payao also generates sound similar to that of the Aeolian tunes of telephone lines when blown by a consistently strong wind. The sound could also be generated by the anchor line resulting from its lateral vibrations due to the shedding of vortices when acted on by passing water currents. The first field recording of underwater sound ever made near an anchored FAD was made near a payao in the Philippines (Babaran et al., 2008). Payao-generated sounds were recorded with a hydrophone at various depths and distances downstream from the payao. The dominant peak depended on the conditions in the field at the time of recording. For example, during rough weather, the payao generated sound at a level as high as 145 dB close to the raft with a frequency centered around 63 Hz (Babaran et al., 2008). This sound, which was higher than the background noise by about 20 dB, had a limited range and attenuated rapidly with both increasing distances from the raft and increasing depths. The frequency of the gener-
59
ated sound matched with those of the raft’s pitching and rolling motions as recorded by accelerometers. Between these two motions, however, energy due to the former was much more dominant, apparently because of the tendency of the raft to orientate itself parallel to the wind’s direction. The sounds generated by the vibrations of the payao’s anchor line when acted by strong water current were distinct from the sounds generated by the raft’s motion. Measurements of underwater sound generated by the anchor line revealed a dominant peak at 49 Hz with SPLs ranging from 91 to 101 dB as measured 3 m from the anchor line at a depth of 15 m (Fig. 3.10). Moreover, within a range of 40 m from the raft’s position, the recorded SPL of this low-frequency sound did not vary with increasing distance from the payao but decreased with depth (Fig. 3.8) (Babaran et al. 2008). This finding was important because sound resulting from the vibration of the anchor line, which seemed more dominant than raft-generated sound, appeared to be the stimulus used by fish, particularly tuna, when they navigated between anchored FADs (Holland et al. 1990; Oohta and Kakuma 2005). It is interesting to note that the SPLs generated by a payao falls within the hearing range of the fish (as seen in Fig. 3.4).
Figure 3.10. Power spectrum of the sound generated by a fish aggregation device (payao in the Philippines). Sounds are recorded near the payao, at 3-m distance horizontally and 15-m depth, which is set in Panay Bay, Philippines.
60
Locomotion and Sensory Capabilities in Marine Fish
3.5.4 Aversive Sound to Reduce Fish Entrapment in the Cooling Water Intakes Much research to develop techniques to control fish behavior to prevent their entry into water intakes of the power plants has been carried out worldwide (Carlson and Popper 1997). Conventional coalburning and nuclear power plants have to draw in a large volume of water to cool either steam turbines or reactors. Inevitably, some fish are either impinged or entrapped (Grimes 1975; Hanson et al. 1977; Stanford et al. 1982). The compositions of species and economic losses due to the impingement and entrapment of fishes have been assessed, and a case study in Taiwan showed that US$2.5 million yearly economic loss could be due to the intake in the Nuclear Power Plant I (Shao et al. 1990). Additionally, many fish deformities caused by the thermal plume of cooling water discharge have been observed in Taiwan. Four species of marine fish—Jarbua tarpon (Terapon jarbua), large scale mullet (Liza macrolepis), milkfish (Chanos chanos), and grey mullet (Mugil cephalus)—were attracted to the thermal plume at the discharge outlet site of a Nuclear Power Plant II in northern Taiwan, which caused extensive lordosis of vertebral columns. Prolonged exposure to high water temperature caused deficiency of ascorbic acid in the fish’s muscle, which then led to mismatched growth between vertebrates and therefore muscles that resulted in lordosis formation (Shao et al. 1990). Various methods, including mechanical screening devices, electric barriers, strobe light, and sound (EPRI 1992, Humbles 1993), have been deployed to reduce either impingement or entrapment (see summary in Ross and Dunning 1996). Work by Ross and Dunning (1996) demonstrated that by broadcasting high-frequency sound (122– 128 KHz) at a pressure level of 190 dB, alewives (Alosa pseudoharengus) could be driven from the water intake of the James A. FitzPatrick nuclear power plant in Ontario, Canada. In Taiwan, a study using aversive underwater sound to drive fish from intake areas has been undertaken since 2006. First, the audiograms of 11 species of the most frequently entrapped fish were obtained with the aforementioned ABR protocol. Based on their best hearing frequencies and threshold data, randomized pulsed low-frequency sound
(100–2000 Hz interval; 100 Hz duration; 1 s in each frequency) was digitized with a function generator and amplified through an Industrial Power Amplifier (IPA 300T) and broadcast with a underwater speaker (Lubell Labs LL9162; sound pressure of 187 dB, measured with an Okidata SW-1030 hydrophone placed 10 m from the speaker) at the water intake of Nuclear Power Plant II in northern Taiwan coast (Wu et al. 2009). A total of 17 field tests were conducted from November 13, 2006, to February 26, 2008, each with sound on or off for a period of 24 h. During periods when the sound was off, a total of 17 species (1076 individuals) were entrapped. During the periods when sound was on, 10 species (572 individuals) were entrapped. The results indicated that sound significantly reduced the entrapment rate by almost 50% (Wu et al. 2009). The promising results prompted the planning and execution of the long-term use of underwater aversive sound to repel fish from the cooling water intakes and discharge sites of nuclear power plants in Taiwan. 3.6 CONCLUDING REMARKS The underwater world is full of sound and vibration. Fish evolved to have a mechanosensory system to detect both sound and vibration. Hearing generalist fish have a narrower hearing frequency range (less than 1500 Hz) and higher hearing threshold (above 100 dB, re 1 μPa) than do hearing specialist fish (up to 8 kHz and down to 60 dB re 1 μPa). For communication purposes, some fish evolved to produce sound. Unwanted noises generated by fishing, research, and whale-watching vessels can inevitably affect commercially important fish species. Prolonged exposure to noise results in reduced hearing abilities of fish. Fish can be conditioned by coupling the sound with food and visual cues. Devices utilizing this knowledge can be used to guide the fish over a long distance for underwater transport purposes to minimize physiological stress. A donburi fishing method that uses sound generated underwater to attract fish has been used in the Kagoshima area of Japan. Many FADs using sound as the main attraction feature have been deployed by fishermen around the world. Artificial underwater noise has been widely used by power plant operators to drive fish away from cooling water
Hearing in Marine Fish and Its Application in Fisheries intake areas to reduce unwanted impingement and entrapment. REFERENCES Akamatsu T, Nanami A and Yan HY. 2003. Spotlined sardine Sardinops melanostictus listens to 1-kHz sound by using its gas bladder. Fish. Sci. 69: 348–354. Anraku K, Kawamura G, Nakahara M, Shigesato N and Archdale MV. 2006a. Fish behavior controlmethods in marine ranching in Japan-I. Development of conditioning method on the basis of hearingability and auditory behavior. INOC-UMS/BMRI, ICCOSMA. pp 391–397. Anraku K, Makino T, Okawa F, Watanabe K, Masu S, Ozono H, Takeshita H, Kawamura G and Archdale MV. 2006b. Fish behavior control methods in marine ranching in Japan–III. Development of fish guidance device. INOC-UMS/BMRI, ICCOSMA. pp 405–413. Astrup J and Møhl B. 1993. Detection of intense ultrasound by the cod (Gadus morhua L.). J. Exp. Biol. 182: 31–42. Babaran RP, Anraku K, Ishizaki M, Watanabe K, Matsuoka T and Shirai H. 2008. Sound generated by a payao and comparison with auditory sensitivity of jack mackerel Trachurus japonicus. Fish. Sci. 74: 1207–1214. Barimo JF and Fine ML. 1998. Relationship of swimbladder shape to the directionality pattern of underwater sound in the oyster toadfish. Can. J. Zool. 76: 134–143. Blaxter JHS, Gray JAB and Denton EJ. 1981. Sound and startle responses in herring shoals. J. Mar. Biol. Assoc. UK. 61: 851–869. Buerkle U. 1968. Relation of pure tone thresholds to background noise level in the Atlantic cod (Gadus morhua). J. Fish. Res. Bd. Can. 25: 1155–1160. Buerkle U. 1977. Detection of trawling noise by Atlantic cod (Gadus morphua L.). Mar. Freshw. Behav. Physiol. 4: 233–242. Carlson TJ and Popper AN. 1997. Using sound to modify fish behavior at power-production and water-control facilities: A workshop. December 12–13, 1995. Portland State University, Portland, OR. Phase II: Final Report to Bonneville Power Administration, Contract No. 1986BP62611, Project No. 199207101, 360 pp. (BPA Report DOE/ BP-62611-11.) Chapman CJ and Hawkins AD. 1973. A field study of hearing in the cod (Gadus morhua L.). J. Comp. Physiol. 85: 147–167.
61
Chow CK, Ibrahim W, Wei Z and Chan AC. 1999. Vitamin E regulates mitochondrial hydrogen peroxide generation. Free Rad. Biol. Med. 27: 580– 587. Connaughton M, Fine ML and Taylor MH. 1997. The effects of seasonal hypertrophy and atrophy on fiber morphology, metabolic substrate concentration and sound characteristics of the weak fish sonic muscle. J. Exp. Biol. 200: 2449–2457. Dempster T and Tarquet M. 2004. Fish aggregation device (FAD) research: gaps in current knowledge and future directions for ecological studies. Rev. Fish Biol. Fish. 14: 21–42. Enger PS. 1967. Hearing in herring. Comp. Biochem. Physiol. 22: 527–538. Enger PS and Anderson R. 1967. An electrophysiological field study of hearing in fish. Comp. Biochem. Physiol. 22: 517–525. EPRI. 1992. Evaluation of strobe lights for fish diversion at the York Haven hydroelectric project. Electric Power Research Institute EPRI-TR-101703. Final Report. Boston, MA. 120 pp. Erbe C. 2002. Underwater noise of whale-watching boats and potential effects on killer whales (Orcinus orca), based on an acoustic impact model. Mar. Mam. Sci. 18: 394–418. Fay RR. 1969. Behavioral audiogram for the goldfish. J. Audit. Res. 9: 112–121. Fine ML, Friel JP, McElroy D, King CB, Loesser KE and Newton S. 1997. Pectoral spinelocking and sound production in the Channel Catfish Ictalurus punctatus. Copeia. 1997: 777–790 Fine ML, Malloy KL, King CB, Mitchell SL and Cameron TM. 2001. Movement and sound generation by the toadfish swimbladder. J. Comp. Physiol. (A).187: 371–379. Fine ML, Schrinel J and Cameron TM. 2004. The effect of loading on disturbance sounds of the Atlantic croaker Micropogonius undulates: air vs. water. J. Acoust. Soc. Am. 116: 1271–1275. Floyd JM. 1985. Development of the Philippine tuna industry. Pacific Islands Development Program, East-West Center, Honolulu, Hawaii. 60 pp. Fonteneau A, Pallares P and Pianet R. 1999. A worldwide review of purse seine fisheries on FADs. In: Tuna Fishing and Fish Aggregating Devices Symposium. pp.13–25. Caribbean, Martinique, October 15–19, 1999. Freon P and Misund OA. 1999. Dynamics of Pelagic Fish Distribution and Behavior: Effects on Fisheries and Stock Assessment. Cambridge: Fishing News Books. 360 pp.
62
Locomotion and Sensory Capabilities in Marine Fish
Fujieda S, Matsuno Y and Yamanaka Y. 1996. The auditory threshold of the bastard halibut Paralichthys olivaceus. Nippon Suisan Gakkaishi. 62: 201–204. Furukawa T and Ishii Y. 1967. Neurophysiological studies on hearing in goldfish. J. Neurophysiol. 30: 1377–1403. Goodson AD. 1997. Developing deterrent devices designed to reduce the mortality of small cetaceans in commercial fishing nets. Mar. Freshw. Behav. Physiol. 29: 211–236. Grimes CB. 1975. Entrapment of fishes on intake water screens at a stream electric generating station. Chesapeake Sci. 16:172–177. Hanson CH, White JR and Li WH. 1977. Entrapment and impingement of fishes by power plant coolingwater intakes. Mar. Fish. Rev. 1266: 7–17. Harris GG and van Bergeijk WA. 1962. Evidence that the lateral-line organ responds to near-field displacements of sound sources in water. J. Acoust. Soc. Am. 34: 1831–1841. Hashimoto T and Maniwa Y. 1964. Research on the luring of fish schools by utilizing underwater acoustical equipment (1). Tech. Rep. Fish. Boat. 19: 1–12. Hashimoto T and Maniwa Y. 1966. Research on the luring of fish schools by utilizing underwater acoustical equipment (2). Tech. Rep. Fish. Boat. 20: 1–5. Hatakeyama Y. 1992. Hearing abilities of fish. Fish. Eng. 28: 111–119. (In Japanese) Hatakeyama Y. 1996. Hearing abilities of fish and its responses to underwater sound (II). J. Mar. Acoust. Soc. Jpn. 23: 132–139 (in Japanese). Hawkins AD. 1986. Underwater sound and fish behavior. In: Pitcher T (ed). The Behavior of Teleost Fishes. pp 114–151. Baltimore, MD: The Johns Hopkins University Press. Holland KN, Brill RW and Chang RKC. 1990. Horizontal and vertical movements of yellowfin and bigeye tuna associated with fish aggregating devices. Fish. Bull. 88:493–507. Hoyt E. 2000. Whale watching 2000: worldwide tourism numbers, expenditures, and expanding socioeconomic benefits. Yarmouthport, MA: International Fund for Animals Welfare. 157 pp. Humbles G. 1993. Evaluation of Smith-Root type electric fish barrier. Yakima Project, Yakima, WA. 84 pp. Ishioka H., Hatakeyama Y and Sakaguchi S. 1988. The hearing ability of the red sea bream Pagrus major. Nippon Suisan Gakkaishi. 54: 947–951. Jerkø K, Turunen-Rise I, Enger PS and Sand O. 1989. Hearing in the eel (Anguilla anguilla). J. Comp. Physiol. (A).165:455–459.
Kamijyo Y. 1998. Example of marine ranching system (Kaiyou Bokuzyou no Jitsurei). In: Soeda H, Hatakeyama Y and Kawamura G (eds). Hearing Physiology in Fishes (Gyorui no Choukaku Seiri). pp 359–367. Tokyo: Kouseisya Kouseikaku. Kenyon TN, Ladich F and Yan HY. 1998. A comparative study of hearing ability in fishes: the auditorybrainstem response approach. J. Comp. Physiol. (A).182: 307–318. Kojima T, Shimamura T, Yoza K, Okumoto N, Hatakeyama Y and Soeda H. 1992. W-shaped auditory threshold curves of masu salmon Oncorhynchus masou. Nippon Suisan Gakkaishi. 58: 1447–1452. Kraus SD, Read AJ, Solow A, Baldwin K, Spradlin T, Anderson E and Williamson J. 1997. Acoustic alarms reduce porpoise mortality. Nature. 388: 525. Lovell RT. 1977. Fish nutrition: energy. Comm. Fish Farmer. 2: 29–30. Lugli M, Yan HY and Fine ML. 2003. Acoustic communication in two freshwater gobies: the relationship between ambient noise, hearing thresholds and sound spectrum. J. Comp. Physiol. (A).189: 309–320. Maniwa Y and Hatakeyama Y. 1970. Research on the luring of fish schools by utilizing underwater acoustical equipment (3). Tech. Rep. Fish. Boat. 24: 1–5. Maniwa Y and Hatakeyama Y. 1975. Research on the luring and driving away of fish schools by utilizing underwater acoustical equipment (4)–Experiments and practical use in the squid fishing. Tech. Rep. Fish. Boat. 28: 1–22. McCauley RD, Fewrell J and Popper AN. 2003. High intensity anthropogenic sound damages fish ear. J. Acoust. Soc. Am. 113: 638–642. Mitson RB and Knudsen HP. 2003. Causes and effects of underwater noise on fish abundance estimation. Aquat. Living Resour. 16: 225–263. Mooney TA, Au WWL, Nachtigall PE and Trippel EA. 2007. Acoustic and stiffness properties of gillnets as they relate to small cetacean bycatch. ICES J. Mar. Sci. 64: 1324–1332. Motomatsu K, Hiraishi T, Yamamoto K and Nashimoto K. 1996. Auditory threshold and critical ratio of black rockfish Sebastes schlegeli. Nippon Suisan Gakkaishi. 62: 785–790. Nelson JS. 1994. Fishes of the World, 3rd. ed. New York: Wiley and Sons. 624 pp. NRC. 2003. Ocean noise and marine mammals. Washington, DC: National Research Council. 204 pp. Oohta I and Kakuma S. 2005. Periodic behavior and residence time of yellowfin and bigeye tuna associated with fish aggregating devices around Okinawa
Hearing in Marine Fish and Its Application in Fisheries Islands, as identified with automated listening stations. Mar. Biol. 146:581–594. Park Y and Iida K. 1998. Walleye pollock, In: Soeda H, Hatakeyama Y and Kawamura G (eds). Hearing Physiology in Fishes (Gyorui no Choukaku Seiri). pp 223–233. Tokyo: Kouseisha-Kouseikaku. Pitcher TJ. 1979. Sensory information and the organization of behavior in a shoaling cyprinid. Anim. Behav. 27: 126–149. Popper AN and Fay RR. 1999. The auditory periphery in fishes. In: Fay RR and Popper AN (eds). Comparative Hearing: Fish and Amphibians. pp 43–100. New York: Springer-Verlag. Richardson WJ and Würsig B. 1997.Influences of man-made noise and other human actions on cetacean behavior. Mar. Freshw. Behav. Physiol. 29: 183–209. Ross QE and Dunning DJ. 1996. Reducing impingement of alewives with high frequency sound at a power plant in Ontario. N. Am. J. Fish. Manag. 16: 548–559. Saidel WM and Popper AN. 1987. Sound perception in two anabantoid fishes. Comp. Biochem. Physiol. 88: 37–44. Sand O and Karlsen HE. 1986. Detection of infrasound by the Atlantic cod. J. Exp. Biol. 125: 197–204. Scholik AR and Yan HY. 2001. Effects of underwater noise on auditory sensitivity of a cyprinid fish. Hear. Res. 152: 17–24. Scholik AR and Yan HY. 2002a. Effects of boat engine noise on the auditory sensitivity of the fathead minnow, Pimephales promelas. Environ. Biol. Fish. 63: 203–209. Scholik AR and Yan HY. 2002b.The effects of noise on the auditory sensitivity of the bluegill sunfish, Lepomis macrochirus. Comp. Biochem. Physiol. (A).133: 43–52. Scholik AR, Lee US, Chow CK and Yan HY. 2004. Dietary vitamin E protects the fathead minnow, Pimephales promelas, against noise exposure. Comp. Biochem. Physiol. (C).137: 313–323. Schwarz AL and Greer GL. 1984. Responses of Pacific herring, Clupea harengus pallasi, to some underwater sounds. Can. J. Fish. Aquat. Sci. 41: 1183–1192. Shao KT, Lin CP, Ho LT and Lin PL. 1990. Study on the fish communities from northern and southern waters of Taiwan by analyzing the impingement data. J. Fish. Soc. Taiwan 17:73–90. Shishidou H. 2002. Stocking effectiveness of red sea bream, Pagrus major, in Kagoshima Bay, Japan. Fish. Sci. 68(suppl. 1): 904–907. Stanford RM, Jordan SW, Talhelm DR, Liston CR, Korson C and Steinmuller MH. 1982. The bioeco-
63
nomic impact of impingement and entrapment on yellow perch in Lake Erie. N. Am. J. Fish. Manag. 2: 285–293. Trippel EA, Holy NL, Palka DL, Shepard TD, Melvin GD and Terhyne JM. 2003. Nylonbarium sulphate gillnet reduces porpoise and seabird mortality. Mar. Mam. Sci. 19: 240–243. Tu C, Wei RC and Chan HS. 2004. Passive acoustic localization for sciaenid habitat in coastal water of Taiwan. J. Acoust. Soc. Am. 115: 2474. Urick RJ. 1983. Principles of underwater sound. Los Altos, CA: Peninsula Publishing. 423 pp. Wenz GM. 1962. Acoustic ambient noise in the ocean: spectra and sources. J. Acoust. Soc. Am. 34: 1936–1956. Westenberg J. 1953. Acoustical properties of some Indonesian fisheries. J. Mar. Sci. 18:311–325. Winger P. 2004. Effect of Environmental Conditions on the Natural Activity Rhythms and Bottom Trawl Catchability of Atlantic Cod (Gadus morhua). PhD thesis, Memorial University of Newfoundland. 151 pp. Wu YH, Yu HY, Shao IT, Lee ZC, Lin ST,Yan HY, Hsu CH, Lee CP, and Jiang HH. 2009. The method using underwater sound in reducing fish entrainment and impingement at the cooling water inlets of nuclear power plants in northem Taiwan. TaiPower Eng. Month. 733: 108–117. Yan HY. 1995. Investigations of fish hearing ability using an automated reward method. In: Klump GM, Dooling RJ, Fay RR and Stebbins WC (eds). Methods in Comparative Psychoacoustics. pp 263– 276. Basel, Switzerland: Birkhauser Verlag. Yan HY. 1998. Auditory role of the suprabranchial chamber in gourami fish. J. Comp. Physiol. (A).183: 325–333. Yan HY and Curtsinger WS. 2000. The otic gas bladder as an ancillary auditory structure in a mormyridfish. J. Comp. Physiol. (A).186: 595–602. Yan HY and Popper AN. 1991. An automated positive reward method for measuring acoustic sensitivity in fish. Behav. Res. Method. Instru. Compu. 23: 351–356. Yan HY and Popper AN. 1992. Auditory sensitivity of the cichlid fish Astronotus ocellatus (Cuvier). J. Comp. Physiol. (A).171: 105–109. Yan HY, Fine ML, Horn NS and Colon WE. 2000. Variability in the role of the gas bladder in fish audition. J. Comp. Physiol. A. 186: 435–445. Yan HY, Lin SH, Yu HY and Lee WJ. 2006. Analyses on noises generated by aerators and its possible impacts on cultured fish and shrimp. Proc. TaiwanFrance Symp. Marine Biodiver. Sustain. Fish. Maricult. pp. 95–108. Kao-Hsiung, Taiwan.
64
Locomotion and Sensory Capabilities in Marine Fish
SPECIES MENTIONED IN THE TEXT alewife, Alosa pseudoharengus Atlantic salmon, Salmo salar bastard halibut, Paralichthys olivaceus bigeye scad, Selar crumenophthalmus bigeye tuna, Thunnus obesus blue marlin, Makaira mazarra bluegill sunfish, Lepomis macrochirus bullet tuna, Auxis rochei carp, Cyprinus carpio channel catfish, Ictalurus punctatus cichlid fish, Astronotus ocellatus cod, Gadus morhua crimson sea bream, Evynnis japonica croaker, Micropogonius undulates dab, Limanda limanda dolphinfish, Coryphaena hippurus European eel, Anguilla anguilla fathead minnow, Pimephales promelas frigate tuna, Auxis thazard grey mullet, Mugil cephalus jacopever, Sebastes schlegeli
Japanese jack mackerel, Trachurus japonicus Jarbua tarpon, Terapon jarbua kawakawa, Euthynnus affinis large scale mullet, Liza macrolepis masu salmon, Oncorhynchus masou milkfish, Chanos chanos oyster toadfish, Opsanus tau pistol shrimp, Alpheoidea spp. pollock, Pollachius pollachius red sea bream, Pagrus major Risso’s dolphin, Grampus griseus sailfish, Istioporus inducus scads, Decapterus spp. sharpnose tigerfish, Rhyncopelates oxyrhynchus skipjack tuna, Katsuwonus pelamis spotlined sardine, Sardinops melanostictus tadpole-fish, Raniceps raniceps threestripe tigerfish, Terapon jarbua walleye pollock, Theragra chalcogramma yellowfin tuna, Thunnus albacares yellowtail, Seriola quinqueradiata
Part Two Fish Behavior near Fishing Gears during Capture Processes
Chapter 4 Fish Behavior near Bottom Trawls Paul D. Winger, Steve Eayrs, and Christopher W. Glass
4.1 INTRODUCTION A bottom trawl is a towed fishing gear that is designed to catch fish, shrimp, or other target species that live on or in close proximity to the seafloor (Fig. 4.1). The process by which fish enter and are retained involves a complex sequence of fish behaviors in response to the fishing vessel and the various components of the trawl. Observing and understanding these behavior patterns represent a critical step in the effective design of mobile trawling systems. Underwater observations of fish behavior in relation to trawls began as early as the 1960s by researchers in Canada, Scotland, and Russia (Beamish 1966a, 1969; Martyshevskii and Korotkov 1968; Parrish et al. 1969) and soon the elaborate relationship between trawl design and fish behavior began to be articulated (Okonski 1969) together with mathematical models of fish behavior (Foster 1969). Nearly 40 years later, the field continues to grow with much of our current understanding of fish behavior near bottom trawls coming about through the technological advancement of various observational techniques including scuba diving, towed vehicles, underwater cameras, telemetry, and hydroacoustics. For a review of these techniques, see Urquhart and Stewart (1993), Godø (1998), and Graham et al. (2004). In this chapter, we review the current knowledge of fish behavior in relation to bottom trawls, building on the earlier valuable reviews by Wardle (1983, 1986, 1993), Laevastu and Favorite (1988), Engås (1994), Godø (1994), and Glass and Wardle (1995a). We attempt to distil more than 100 studies since the 1960s on fish behavior in response to
visual and auditory stimuli produced by the vessel, doors, sand clouds, sweeps, footgear, and trawl netting. We review the typical patterns of behavior as well as several extrinsic and intrinsic factors known to influence behavior. We also equally emphasize where possible the high degree of between-individual variability in behavioral expression that is often observed. We interpret this variability as differences in tradeoffs (Fernö 1993) at the individual level that minimize costs and maximize benefits. Finally, we broadly extend the application of the economic hypothesis of antipredator behavior (Ydenberg and Dill 1986) initially introduced to the field of fish capture by Fernö and Huse (2003). 4.2 TRAWL GEAR AND TRAWL FISHERIES The historical development of a bottom trawl can be traced back to the early use of beam trawls during the fourteenth century with a series of later technological strides during the Industrial Revolution and post–World War II years (see Graham 2006 for review). It is the principal technique by which most demersal fish and shrimp are captured, accounting for approximately 22% of the world’s fish production (Kelleher 2005). Trawl fishing is practiced by nearly all of the world’s coastal states and can be found in estuaries, coastal regions, and the high seas to depths of 2000 m or more. In many regions of the world, the types of bottom trawls used and their operational techniques have not changed much over the past 50 years. But for
67
68
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.1. Schematic drawing of a complete bottom trawl fishing system. Design, shape, and size of individual components vary depending on the fishery and operation
other regions, particularly those of developed countries, the trawls used today barely resemble those of even 20 years ago (Walsh et al. 2002). Modern designs are more advanced and sophisticated as a result of increasing fuel costs, the need for species and size selectivity, stringent bycatch restrictions, and the necessity to minimize the impact on the environment. Meeting these challenges has led to significant improvements in the way bottom trawls are designed and tested, including advances in computer design, simulation, and physical modeling (Winger et al. 2006). A bottom trawl is designed and engineered as a system of parts that work together with predictable geometry and performance (Fig. 4.1). Depending on the trawl design and targeted species, towing speeds range from 2.0 to 5.0 knots (1.0 m/s to 2.5 m/s). The otter boards or doors provide horizontal spreading force through a combination of hydrodynamic lift and frictional shear with the seabed. They also help keep the trawl net on the seabed and generate a sand cloud to assist herding of fish into the net. Wire sweeps and bridles connect the doors to the wingends of the trawl net and vary in length depending on the fishery, often short (less than 40 m) or non-
existent for shrimp fisheries and long (approximately 350 m) for flatfish fisheries. The trawl body consists of a series of netting panels selvaged together. The design, shape, and dimensions of these panels help define the overall shape and performance of the net. The vertical opening of the mouth of the net is usually achieved by attaching positive buoyancy (e.g., floats) or hydrodynamic kites to the headline, although in some shrimp fisheries direct attachment of the headline to the otterboard means the height of the net is equivalent to the height of the otterboard. To maintain seabed contact, chain, rubber discs, or steel bobbins are attached to the fishing line and/or bolsh line of the net. The process by which fish enter and are retained by a net involves a complex sequence of fish behaviors. A helpful approach for understanding and describing this process is to compartmentalize it into three zones as shown in Figure 4.2. Zone 1 describes fish behavior in the pretrawl zone, including their initial detection and reaction to the lowfrequency noise produced by the vessel, as well as avoidance behavior in response to the trawl warps. Zone 2 describes fish behavior in response to the trawl doors, sweeps, and net mouth. And, finally,
Fish Behavior near Bottom Trawls
69
Figure 4.2. Schematic drawing of a bottom trawl illustrating the different zones that fish may encounter during the capture process. (Adapted from Walsh 1996.) Hatched areas represent the “sweep zones”—fish in these zones must be herded or guided into the net path to become vulnerable to capture.
Zone 3 describes fish behavior once inside the trawl body. 4.3 FISH BEHAVIOR IN THE PRETRAWL ZONE (ZONE 1) 4.3.1 Underwater Radiated Noise The fish capture process begins well ahead of the vessel, where fish initially detect and respond to low-frequency noise produced by the vessel, warps, doors, and trawl. The combination of these sounds produces an underwater-radiated noise signature that is highly specific to each vessel and trawling operation. In most cases, this noise signature is dominated by low-frequency tones (10–10,000 Hz) produced mainly by the vessel. In fact, detailed measurements of vessel-radiated noise alone have shown extreme variation in the frequency spectrum and pressure levels between vessels (Mitson 1993; Mitson and Knudsen 2003). Factors that determine a vessel’s noise signature include cylinder firing rate, engine load, gearbox configuration, and propeller pitch (see Anon 1995 in press for review). Additional noise is also produced by the vibrations of trawl warps, door contact with the seabed, and
trawl components as they pass through the water (Buerkle 1977; Handegard et al. 2003; Handegard and Tjøstheim 2005; Mitson 1993), all of which contribute to the overall noise signature. For many species of fish, these low-frequency sounds occur directly within their hearing range. Over the past few decades, audiograms of hearing sensitivity have been collected for a number of fish species (see Anon 1995 in press; Fréon and Misund 1999; Popper 2003 for reviews). Figure 4.3 illustrates a few examples of the relationship that exists between hearing threshold (dB) and sound frequency (Hz). Popper et al. (2004) recently argued that fish can be grouped as hearing “specialists” or hearing “nonspecialists” (i.e., generalists). The hearing generalists detect sounds up to 1500 Hz and have lower sensitivity (higher thresholds). Hearing specialists, by comparison, detect sounds of 3000 Hz or above and have better sensitivity (lower thresholds). While this species-specific variation in hearing capability is probably the result of variation in selection pressure over time (Manley et al. 2004), it has immediate and important relevance to how well fish detect approaching fishing gear, particularly vessels and trawls.
70
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.3. Audiograms of fish hearing sensitivity for several species, including Atlantic cod (Gadus morhua), Atlantic salmon (Salmo salar), American shad (Alosa sapidissima), herring (Clupea harengus), North Sea plaice (Pleuronectes platessa), common dab (Limanda limanda), North Sea pollack (Pollachius pollachius), and goldfish (Carassius auratus auratus). (Data from Mitson 1993; Popper 2003.)
Once the dB–frequency relationship is known for a species (e.g., Fig. 4.3), it is possible to calculate the theoretical distance that the species can detect vessel/trawl noise (see Mitson 1993). For example, Atlantic cod (Gadus morhua) have one of the lowest thresholds (Fig. 4.3) and are therefore capable of relatively long-range detection (Sand and Karlsen 1986) for certain frequencies. It has been estimated that cod can detect trawling noise from a minimum distance of 3.2 km during summer, reducing to 2.5 km during winter due to increased ambient noise in the sea at that time of year (Buerkle 1977). The main question remains, how well can fish discriminate the direction and distance from which a vessel/trawl is approaching? This cannot be an easy task given the diffuse and complex sounds emitted from an approaching vessel and trawl. Several studies have attempted to investigate different aspects of sound source localization in haddock (Melanogrammus aeglefinus), pollack (Pollachius
pollachius), ling (Molva molva), walleye pollock (Theragra chalcogramma), and cod (Gadus morhua) (Chapman 1973; Chapman and Johnstone 1974; Hawkins and Sand 1977; Mann et al. 2009; Olsen 1969; Schuijf 1975). Using elegant behavioral or electrophysiological methods, these studies have demonstrated rather convincingly the capacity for sound source localization in many species, in both the near- and far-field. But these have mainly been in response to pure-tones under laboratory conditions, not the diffuse and complex sounds of a vessel and bottom trawl. Fay (2005) reviews several theories as to “how” fish collect and process sounds from their environment and concludes that our knowledge of sound source localization in fish remains incomplete, thus identifying a key field for future investigation. 4.3.2 Reaction Distance A significant body of evidence suggests there is large difference between the distance at which fish
Fish Behavior near Bottom Trawls first “detect” approaching sounds and the distance at which they “react” to approaching sounds. In other words, their detection thresholds and reaction thresholds are very different. Detection thresholds are determined by the physics of the inner ear and are likely to be predictable within a species, size class, ambient noise in the water, and physical property of the water. Response thresholds by comparison, are the outcome of a behavioral tradeoff. They are the behavioral expression made by individual fish that are attempting to minimize costs and maximize benefits in response to an approaching threat (Fernö 1993; Godin 1997). It was recently predicted (Fernö and Huse 2003) that the timing of reaction by fish to an approaching vessel and trawl should follow similar economics to prey fleeing from predators. If this is true, it opens the tantalizing opportunity to apply basic principles from established predator–prey theory and, in particular, the growing field of “optimal escape theory,” initiated by Ydenberg and Dill (1986). Since their influential model was published more than 20 years ago, dozens of behavioral studies have found the model to be robust in its ability to predict relative reaction distances to
71
threats across species and taxa (see reviews by Cooper and Frederick 2007; Stankowich and Blumstein 2005). In general, when a fish detects a threat, usually an attacking predator, it must sequentially decide (1) whether and when to flee, (2) in which direction to flee, (3) how fast to flee, and (4) how far to flee. In Ydenberg and Dill’s (1986) economic (costbenefit) model, a fish under threat of a predator continually chooses between two behavioral options, staying where it is (and perhaps continuing with an ongoing activity) or fleeing, as the distance between it and the predator shrinks. The distance at which the fish flees (i.e., reaction distance) is determined by a balance between the costs of these two options (Fig. 4.4). The costs of fleeing (F) are assumed to increase linearly and the costs (or risk) of remaining (R) to decrease proportionately, with increasing distance to the predator. Timing is critical as fleeing too early may result in lost opportunities (high F) and remaining too long may result in injury or death (high R). A fish under threat is assumed to reassess its choice of action moment by moment as the distance between itself and the threat changes and should opt for the behavior with the
Figure 4.4. Ydenberg and Dill’s (1986) economic model of reaction distance (D) for fish under the threat of a predator. (Adapted from Godin 1997.) We broadly extend this model to help develop predictions about response thresholds and corresponding optimal reaction distances for fish in response to a bottom trawl. See text for details.
72
Fish Behavior near Fishing Gears during Capture Processes
lowest cost. The animal should opt for remaining when F > R and opt for fleeing when R > F. The optimal reaction distance (D*) is defined as the intersection of the two curves (see Fig. 4.4). It is the point where the costs of remaining just balance the costs of fleeing (i.e., escaping). The model predicts that the optimal reaction distance should increase with increasing cost of remaining (D2* → D3*) and decrease with increasing cost of fleeing (D2* → D1*). As a trawl approaches an aggregation of fish in the wild, fish in the pretrawl zone (Zone 1) are faced with the important decision of whether and when to react. The optimal reaction distance (D*) should be determined by a balance between the costs of fleeing (F) and remaining (R). Factors that define the R cost curve will be related to the “apparent risk” of the vessel and trawl (e.g., type, age, speed, engine load, and propeller pitch, trawl rigging, performance, and operation). The F cost curve will be defined by lost opportunities (e.g., foraging and spawning) and energy expenditure. Layered over this, of course, are then the several extrinsic and intrinsic factors that are expected to modify behavioral expression (see Section 4.6). It is quite plausible to expect that each individual fish could have a unique intersection of the curves and therefore a unique optimal reaction distance (D*). Several studies have attempted to empirically measure the reaction distance of demersal species in response to approaching vessels and bottom trawls. While most of the early research used traditional echo-sounders to monitor the response of whole aggregations (e.g., Olsen et al. 1983a; Ona and Godø 1990), recent approaches now monitor the movements of individual fish using either splitbeam echo-sounders (e.g., De Robertis et al. 2008; Handegard et al. 2003; Handegard and Tjøstheim 2005; Hjellvik et al. 2003; Jørgensen et al. 2004; McQuinn and Winger 2003; Ona et al. 2007) or acoustic telemetry of individual fish tagged with acoustic transmitters (Engås et al. 1998; HardenJones et al. 1977; Winger 2004). Both have proven highly effective at estimating reaction distances and determining avoidance patterns. The individual reaction distances of acoustically tagged cod were studied by Engås et al. (1998) and Winger (2004) in response to an approaching
vessel and bottom trawl. Both studies used an acoustic positioning system (Voegeli et al. 2001) to monitor the behavior of individual fish for the periods before, during, and after an encounter (Fig. 4.5). Beginning at a distance of 1 to 3 km the array, a vessel course was plotted to take the vessel and trawl directly over the fish with the closest possible precision. The vessel then exited the opposite end of the array and took up position at a distance of 1 to 3 km. Engås et al. (1998) found that cod were capable of initiating avoidance responses at distances ranging from 470 to 1470 m. These results were the first of their kind, demonstrating that reaction distances in some cases can occur at distances greater than 1 km. The results were verified in a similar study by Winger (2004), which found reaction distances between 206 and 1512 m. Figure 4.6 illustrates various changes in swimming speed observed in response to an approaching vessel at different speeds, as well as a vessel towing a trawl. The corresponding right axis shows the distance between the fish and the approaching threat when the change in swimming speed occurred (i.e., reaction distance). Together, these studies indicate that cod are capable of detecting and reacting to an approaching trawler from considerable distances and that the optimal reaction distance can be highly variable depending on the behavioral tradeoff chosen by individual fish. Further discussions on extrinsic and intrinsic factors that are known to affect reaction distance are given in Section 4.6.
4.3.3 Avoidance Patterns Once a fish has detected and decided to react at some optimal distance to the threat of a fishing vessel (see earlier), it must decide in which direction and how fast to swim. The traditional view has long been held that fish react with a graduated response, beginning first with a slow adjustment in swimming direction away from the approaching stimulus (Olsen et al. 1983a). Several studies have provided empirical data to support the theory, including those of Ona and Godø (1990) and Nunnalle (1991), which both documented a density draining phenomenon in advance of an approaching vessel, indicative of horizontal avoidance. However,
Fish Behavior near Bottom Trawls
73
Figure 4.5. Illustration of the experimental setup used by Engås et al. (1998) and Winger (2004) to investigate the individual reaction distances of acoustically tagged cod in response to an approaching vessel and trawl.
recent evidence suggests that for some vessels, certain species may actually swim in toward the vessel path as it approaches (Handegard and Tjøstheim 2005). For the individual fish, the decision on which direction to swim in response to an approaching vessel/trawl is complicated by the nonuniform propagation of noise intensity. As a vessel and trawl move along a given trajectory, high and low areas of noise intensity systematically form and collapse in response to the moving hull, warps, and trawl. The hull’s ability to shadow propeller
cavitations produces large lobes of high-intensity noise on the vessel’s port and starboard sides. The resulting “butterfly pattern” (see Anon in press; Misund 1994) has been shown in some studies to induce intense outward horizontal movement toward the port and starboard (Engås 1994; Soria et al. 1996), to attract fish inward toward the vessel track in other studies (Handegard and Tjøstheim 2005), and even to trap fish into swimming in the forward direction in line with the vessel’s track, known as the pursuit effect (Misund 1994; Misund et al. 1996).
74
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.6. Examples of Atlantic cod behavior in response to an approaching vessel (top and middle) and vessel plus trawl (bottom). Top, Increase in swimming speed in response to a vessel approaching at 10.0 knots. Middle, Decrease in swimming speed in response to a vessel approaching at 4.5 knots. Bottom, Increase in swimming speed in response to a vessel and trawl approaching at 3.0 knots.
In addition to horizontal movement, fish occurring in the pelagic zone may also dive vertically toward the seabed in response to vessel-radiated noise. This has been documented for several species including capelin, herring, anchovy, common sardine, haddock, and cod (Gerlotto et al. 2004; Handegard et al. 2003; Hjellvik et al. 2003; Olsen et al. 1983b; Ona and Godø 1990; Michalsen et al. 1999; Vabø et al. 2002). In a recent study, Handegard and Tjøstheim (2005) used a free-floating splitbeam echo-sounder system to examine the simultaneous horizontal and vertical displacements of individual fish in response to an approaching vessel and trawl. The authors provide an impressive threedimensional model of the directional behavior of gadoids built using multiple years of experimental data. One of the key findings was that vertical displacements of fish (diving) tended to occur at the start of the tow when vessel noise drops markedly. They also noted the strongest and sharpest responses toward the warps. These findings contradict earlier indications that the timing of diving occurs in response to the gradual rise in vessel noise caused by an approaching vessel.
Finally, avoidance patterns in the pretrawl zone are often characterized by a change in swimming speed. According to the avoidance model proposed by Olsen et al. (1983a), changes in swimming speed should occur when a change in swimming direction alone has been insufficient in reducing the approaching threat. Field studies investigating the swimming speeds of fish in response to an approaching vessel/ trawl have commonly reported increases in swimming speed (e.g., Handegard et al. 2003; Handegard and Tjøstheim 2005; McQuinn and Winger 2003; Michalsen et al. 1999; Olsen et al. 1983a; Winger 2004) but also in rare occasions observed decreases in speed (Engås et al. 1998; Winger 2004). Figure 4.6 illustrates a few examples of individual cod modifying their swimming speed in response to an approaching vessel at different speeds, as well as a vessel towing a trawl. Figure 4.6, A and C, reveals a more than doubling of swimming speed, whereas Figure 4.6B illustrates an example of a fish that reduced its swimming speed from roughly 20 cm/s to near zero, well in advance of the approaching threat (approximately 1500 m). Both of these avoidance patterns are considered to be adaptive from a
Fish Behavior near Bottom Trawls predator-evasion point of view (Lima and Dill 1990) and are probably reflective of differences in distance to shelter/cover for these particular fish, hence affecting the steepness of the F-cost curve, and their likely reaction distance (Fig. 4.4).
4.4 FISH BEHAVIOR BETWEEN TRAWL DOORS AND IN THE NET MOUTH (ZONE 2) As a trawl begins to move through an aggregation of fish, a proportion of these fish will enter between the doors. Fish located in this zone will be either directly in the path of the net itself or in one of the two sweep zones between the wings of the net and trawl doors (see Fig. 4.2). The net path is defined as the area swept between the wingends of the trawl net. Fish located in this zone are directly available for capture by the trawl net. However, fish in the
75
two sweep zones must be herded (or guided) into the net path to become available for capture. 4.4.1 Herding Patterns: Roundfish Roundfish that are located near the seafloor in advance of an approaching trawl tend to respond to trawl doors as soon as their presence is sensed visually. Naturally, this is dependent on both the visual field and the visual range of the species. Observations have shown that fish tend to choose the path of least resistance in this situation. That is, they swim in a manner that maintains the threat at the edge of their visual range, keeping at least one eye on the door at all times as it passes by. This behavior results in what is known as the “fountain maneuver” (Fig. 4.7). The process unwittingly guides many of the fish directly into the path of the net, increasing their vulnerability to capture (Hall et al. 1986; but see Wardle 1993 for detailed description).
Figure 4.7. Typical fountain maneuver of roundfish in response to trawl doors and their subsequent herding behavior into the trawl mouth. Dotted lines represent the limit of the visual range of fish, which varies across species and, most important, with the underwater light field. This behavioral model is based on a series of observations from several studies conducted over time and illustrates common behavior patterns observed at high light intensities.
76
Fish Behavior near Fishing Gears during Capture Processes
Fish that swim around the outside of the doors effectively escape capture and are unlikely to reenter the capture zones. Fish that swim around the inside of the doors tend to immediately enter the path of the net and begin swimming toward the net mouth, maintaining position between the sweeps and sand clouds. In many cases, the visual stimuli of the netting, floats, and footgear are still outside the visual range, presenting what appears to be a clear route to safety. It is not until these components enter the visual range of the fish (denoted by the second dotted line in Fig. 4.7) that fish tend to alter course and begin swimming in the direction of tow in the net mouth. Factors affecting the fountain maneuver as well as the degree of variation within and between species are not fully understood. Experiments have demonstrated that the herding efficiency of some species can be improved by increasing the length of sweep between the door and wingend (i.e., sweep length). Engås and Godø (1989a) found that catch rates for large cod and haddock generally increased with increasing sweep length from 20 to 120 m. They also found that smaller fish (less than 29.5-cm length) were generally underrepresented with increasing sweep length, indicating the process can also be size selective for roundfish (see discussion by Engås 1994). Somerton (2004), by contrast, was unable to duplicate this finding for either Pacific cod (Gadus macrocephalus) or walleye pollock and concluded that these species have weak herding responses. Co-related to sweep length is the angle of attack of the sweep to the direction of tow (sweep angle), which is typically between 10 and 20 degrees. Strange (1984) found that the catching efficiency for cod and haddock was reduced at sweep angles greater than 20 degrees. At such large angles of attack, the sand clouds are likely to deviate from the sweep and trail well outside of the net. This leaves only the relatively inconspicuous sweep to herd fish toward the wing-end, reducing overall herding effectiveness. 4.4.2 Herding Patterns: Benthic Species For benthic species such as flatfish, monkfish, and skates, herding is typically induced after direct or near contact with the doors, sand clouds, and sweeps (see review by Ryer 2008). Reaction distances are
generally short, and in most cases fish are seen swimming in close proximity to the sweep (Fig. 4.8). Under sufficient light conditions, animals will typically select a swimming trajectory that is approximately perpendicular (90 degrees) to the advancing sweeps. This results in the fish slipping along the sweep toward the path of the net (Fig. 4.9A), increasing their vulnerability to capture. However, not all fish will behave this way, and swimming trajectory can be somewhat variable among individuals and among species (S.J. Walsh and P.D. Winger, unpublished observations). Individuals may select trajectories that do not direct them into the path of the net, thus evading capture (Fig. 4.9B). With the exception of Reid et al. (2007), empirical estimates of swimming trajectory and its variability have not been adequately estimated for many species, highlighting another key area for future investigation. Behavioral observations have also revealed that the choice of swimming behavior can be somewhat variable (Harden-Jones et al. 1977; Hemmings 1969, 1973; High 1969; Main and Sangster 1981; Reid et al. 2007; Ryer and Barnett 2006). Once disturbed from the seafloor, individuals may choose to (1) swim slower than the speed of the advancing sweep, in which case they would be overtaken and escape from the gear, (2) swim continuously at the same speed as the sweep, “keeping station” a certain distance ahead of the stimulus, or (3) swim at a speed greater than the sweep for a period of time and then slow down or settle onto the seafloor. Only in the latter two cases do fish progressively slip along the sweep toward the mouth of the trawl, increasing their vulnerability to capture as they move toward the net path. Several authors have attempted to mathematically model these different behavior patterns (Foster 1969; Foster et al. 1981; Fuwa 1989; Fuwa et al. 1988; Reid et al. 2007; Tanaka et al. 1991). The models demonstrate that herding efficiency is highly sensitive to subtle changes in behavior (i.e., swimming continuously versus periodic settling), indicating that the choice of swimming behavior probably determines the likelihood of capture. Functional explanations for the three behavior patterns are still unresolved and the degree of variation within and between species is unknown, although recent laboratory observa-
Fish Behavior near Bottom Trawls
77
Figure 4.8. Typical herding behavior of benthic species such as flatfish and skates in response to the sweep of a bottom trawl. Reaction distances are often low and most animals are seen swimming approximately perpendicular (90 degrees) to the sweep.
tions suggest species-specific differences are likely and that extrinsic factors such as ambient light level and water temperature are important proximate factors (Ryer and Barnett 2006). Finally, for herding to be effective, these fish must also have sufficient endurance to reach the net path. Assuming most fish swim perpendicular (90 degrees) to the advancing sweep, the “herding speed” will be determined by the forward towing speed of the trawl and the sweep angle (i.e., angle of attack of the sweep to the direction of tow). Depending on trawl rigging and operation, herding speeds usually range between 0.2 and 0.6 m/s. Given that all fish are stimulated to swim at the
same speed, fish of different sizes will operate at different levels within their performance range and exhibit different gaits (Winger et al. 2004). The distance required to swim to reach the net path is determined by the sweep angle and the position along the sweep where the fish initially encounter the gear. The probability of successfully swimming this distance is thought to be size and temperature dependent (see Section 4.6). 4.4.3 In the Trawl Mouth Video and photographic observations have repeatedly shown that behavioral diversity is greatest in the trawl mouth. Different species and sizes of fish
78
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.9. Under sufficient light conditions, most flatfish and skates will select a swimming trajectory that is perpendicular (90 degrees) to the advancing sweeps. This results in the fish slipping along the sweep toward the path of the net (A). In some cases, individuals may select swimming trajectories that never increase their vulnerability to capture (B). Unless these individuals change their swimming trajectory, they might never transition into the path of the net.
all arrive here in varying physiological conditions depending on their avoidance behavior in Zone One, as well as their subsequent herding behavior in Zone Two. The trawl mouth is the melting pot where everything comes together at high speed and quick decisions have deliberate outcomes. Because of the high variability in behavior in this region of the gear, several authors have attempted to analyze the behavior by coding or classifying patterns of behavior into meaningful categories (Albert et al. 2003; DeAlteris et al. 1992; Eayrs and Piasente 2006; Piasente et al. 2004; Walsh and Hickey 1993). Once this is done, the frequency and duration of the patterns can be quantified, making interpretation considerably easier. For many species, the most common response is to orient in the direction of tow and keep station with the advancing trawl (Fig. 4.10). This behavior
occurs at high light intensities and appears to be an optomotor reflex in response to the visual cues (contrast) produced by the surrounding footgear and netting panels (see Glass and Wardle 1989; Main and Sangster 1981; Walsh and Hickey 1993; Wardle 1993; also see Chapter 2). This behavior is similar to that observed in the laboratory where fish can be induced to follow moving stripes and patterns projected on the floor and walls of a tank (Breen et al. 2004; Harden-Jones 1963; He and Wardle 1988). For some species, this reflex produces a strong motivation to maintain position (or station) within the mouth of a moving trawl, whereas for other species, the response is less apparent (or nonexistent), producing a more erratic behavior in the trawl mouth (Kim and Wardle 2003). As light intensity drops, ordered patterns of behavior eventually cease. Glass and Wardle (1989)
Fish Behavior near Bottom Trawls
79
Figure 4.10. (A, B) Typical behavior of fish in the mouth of a bottom trawl at high light intensities. Fish are seen swimming forward in the direction of tow, immediately ahead of the footgear. (Photograph from Glass and Wardle 1989.)
and later Walsh and Hickey (1993) observed several demersal species at various angles to the approaching gear, showing little evidence of reaction at low light intensities (Fig. 4.11). Under these conditions, reaction thresholds are relatively high and the corresponding reaction distances short, in some cases even resulting in collision with the gear (Fig. 4.12). See Section 4.6 for a detailed discussion on the role of light level, color, and contrast. Naturally, not all species are created equal in their swimming capability and performance. The period of time that fish are capable of swimming in the trawl mouth, and hence their vulnerability to
capture, is heavily dependent on the towing speed of the trawl, which is carefully chosen to match the sustained swimming speed range of the targeted species (see Chapter 1). Several studies have managed to visually observe the endurance of different species swimming in the mouth of a trawl under different contexts. Reported values range from as low as a few seconds for jack mackerel (Trachurus japonicus) (Martyshevskii and Korotkov 1968), 3 to 4 s for skates (Raja sp.) (DeAlteris et al. 1992), 5 to 6 s for Spanish sardine (Sardinella anchovia) (Korotkov 1970), 1 to 10 s for Greenland halibut (Reinhardtius hippoglossoides) (Albert
80
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.11. (A, B) Typical behavior of fish in the mouth of a bottom trawl at low light intensities. Fish are oriented at various angles to the approaching gear, showing no evidence of reaction. (Photograph from Glass and Wardle 1989.)
et al. 2003), 2 to 12 s for various flatfish in the Gulf of Alaska/Bering Sea (Bublitz 1996), up to 10 s for blue grenadier (Macruronus novaezelandiae) (Piasente et al. 2004), up to 1.0 min for tiger flathead (Neoplatycephalus richardsoni) (Piasente et al. 2004), up to 1.5 min for dogfish (Sqaulus acanthias) (DeAlteris et al. 1992), up to 2.0 min for sand flathead (Platycephalus bassensis) (Yanase et al. 2009), 2.5 min for haddock (Main and Sangster 1983), 3.0 min for squid (Loligo pealeii) (Glass et al. 1999), up to 8.0 min for Pacific halibut
(Hippoglossus stenolepis) (Rose 1995), and up to 15 min for saithe (Pollachius virens) (Main and Sangster 1983). However, in situ estimates of swimming endurance in the trawl mouth are often costly to obtain and difficult to interpret. With this in mind, several studies have investigated the endurance of commercially targeted species under controlled laboratory conditions for the direct application to trawling (Beamish 1966b; Breen et al. 2004; Chandler 1967; Dogˇanyilmaz-Özbilgin et al. 2006; He 1991; He
Fish Behavior near Bottom Trawls
81
Figure 4.12. (A, B) Collision and escapement of fish under the footgear of a bottom trawl. The process is known to be species and size selective as well as dependent on ambient light intensity. (Photograph from Walsh and Hickey 1993.)
and Wardle 1988; Winger et al. 1999, 2000). Using swimming flumes or large annular tanks, researchers have investigated the period of time that fish are capable of swimming at specific speeds before fatiguing. Once the endurance relationships are defined, the probability of swimming a given period of time can be predicted for various swimming/ towing speeds. Figure 4.13 illustrates various endurance probability curves for Atlantic cod at speeds comparable to those experienced in the trawl mouth. These curves show that the probability of cod achieving a given endurance decreases rapidly with increasing swimming speed. For example, the
probability that cod has endurance greater than 2 min in the trawl mouth is 85% at 1.1 m/s, dropping to 35% at 1.3 m/s, and 0% at speeds equal to or greater than 1.5 m/s. The findings demonstrate that for this species, endurance is expected to be highly sensitive to changes in towing speed, and that subtle changes in the speed of the trawl through the water could have a significant effect on the turn-over rate of cod in the trawl mouth (see Winger et al. 2000 for more details). The point at which fish cease swimming in the trawl mouth is triggered by a behavioral decision. In most cases, we assume it is coupled with the
82
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.13. Estimated probability curves for the swimming endurance of Atlantic cod at swimming speeds comparable to those experienced in the trawl mouth. (Data from Winger et al. 2000.)
onset of metabolic exhaustion as it is often evidenced by a change in gait from steady to unsteady swimming just prior to falling back into the trawl (e.g., Main and Sangster 1983). However, recent studies have clearly shown that “exhaustion” and “fatigue” may not necessarily be one and the same (see Peake and Farrell 2006). Exhaustion refers to the state in which a fish has fully depleted its stored energy, whereas fatigue is a behavioral decision that can occur in advance of energy depletion. Breen et al. (2004) found under laboratory conditions that seemingly exhausted haddock were in fact not exhausted at all, just “unwilling” to continue swimming. This suggests that the mechanism associated with the point at which the fatigue decision is invoked while swimming in the trawl mouth is undoubtedly complex and cannot be entirely explained by metabolic exhaustion alone. This is, of course, supported by the fact that fish are often observed swimming actively once they have fallen back into the trawl (e.g., Eayrs and Piasente 2006; Grimaldo et al. 2007; He et al. 2008; Jones et al. 2008; Piasente et al. 2004). For further discussion, see Ryer (2008). Once the decision to cease swimming occurs, the majority of fish will turn and fall back into the trawl
(Zone 3), but some will in fact escape over the headline or under the footgear. Those that escape under the footgear, such as flatfish, skates, and cod (Wardle 1984), may do so through active escapeseeking or simply through accidental collision with the footgear. Underwater observations and experiments using mini-sampling nets behind the footgear have together demonstrated that escapement under the trawl is often species specific and size dependent (Albert et al. 2003; DeAlteris et al. 1992; Engås and Godø 1989b; Godø and Walsh 1992; Ingólfsson and Jørgensen 2006; Korotkov 1970; Langeland 2005; Walsh 1992; Weinberg and Munro 1999). 4.5 FISH BEHAVIOR INSIDE THE TRAWL NET AND THE CODEND (ZONE 3) 4.5.1 Entry and Orientation The entry of fish into the trawl net is highly variable within and between species and dependent on a host of extrinsic and intrinsic factors (see Section 4.6). Attempts to characterize entry behavior have ranged from the highly descriptive (e.g., Main and Sangster 1981) to the highly quantitative (e.g., Albert et al. 2003; Bublitz 1996; Castro et al. 1992).
Fish Behavior near Bottom Trawls For most species of fish, entry into the trawl net marks a transition. They were, just moments ago, swimming in the net mouth under the influence of the optomotor reflex. Suddenly a behavioral change has occurred, triggered by either fatigue, the onset of metabolic exhaustion, collision with another fish, social facilitation, or some combination of these factors. Some species, such as sandeels (Ammodytidae) and haddock (Main and Sangster 1981; Wardle 1993), rise upward as they enter the trawl net and swim directly and purposely toward the codend; depending on ambient light levels and net length, these fish see a clear passage surrounded by netting disappearing into the gloom (Wardle 1993; Watson 1988). Cod, flatfish, and saithe, on the other hand, typically turn toward the trawl net and may even attempt escape between the discs of the footgear (e.g., Engås 1994; Wardle 1993) or dive through large meshes if available (e.g., Beutel et al. 2008; Milliken and DeAlteris 2004). Some species make burst-swimming maneuvers in seemingly random directions, striking netting and other fish as they pass through the trawl mouth, while others with limited swimming capability, such as New Zealand dory (Cyttus novaezelandiae) (Piasente et al. 2004), remain motionless and do not respond unless they collide with netting or other fish. The height at which fish rise over the footgear and enter the trawl net is species dependent and can range from a few centimeters to several meters. Benthic species such as skates and flatfish typically enter very low (e.g., Bublitz 1996; Rose 1995), whereas gadoids, small pelagics, and squid tend to rise higher and enter at greater heights (e.g., Glass et al. 1999; Main and Sangster 1981, Thomsen 1993). Several examples exist where gear technologists have successfully used these species-specific behavioral differences to reduce bycatch of nontargeted species (see reviews by Glass 2000; Graham 2006). On reaching the relatively confined funnel section of netting immediately ahead of the codend (sometimes called the extension or intermediate section), some species, including haddock and cod, will turn and swim in the towing direction (Grimaldo et al. 2007; He et al. 2008). As exhaustion sets in, each individual will slow and eventually enter the exten-
83
sion, where increased crowding may disrupt their orderly behavior and elicit randomly orientated burst-swimming behavior. This behavior is likely to cause collision with netting or other fish, or even the escape of small individuals through the side and upper meshes of the trawl net. It is in this location where square-mesh panels or exit windows may be inserted to increase rates of fish escape (e.g., Broadhurst et al. 2002; Graham et al. 2003; Krag et al. 2008). The vertical location of fish during passage through the extension of a trawl net is often highly variable and sometimes species specific. Main and Sangster (1981) observed whiting (Merlangius merlangus) swimming directly toward the codend close to the bottom of the extension and Atlantic mackerel (Scomber scombrus) doing likewise before turning and swimming out of the trawl mouth. Blue grenadier were observed by Piasente et al. (2004) passing through the extension at all heights, burst-swimming in random directions and often colliding with netting or other fish (these collisions usually elicited further bouts of highspeed swimming and additional collisions). Atlantic bumper (Chlorscombrus chrysurus) and anchovies (Ancoa hepstus) passed toward the codend through the upper half of the extension, while red snapper (Lutjanus campechanus) passed through the lower half, with smallest individuals closest to the bottom of the trawl net (D. Foster, National Marine Fisheries Service, Mississippi Labs, personal communication). The behavior of flatfish in the extension can also be highly variable and may range from dynamic to passive. Sand flathead may swim in any direction, often at high speed, or lie motionless on the lower panels of netting (Yanase et al. 2009), particularly on inclined panels where hydrodynamic forces hold them in position (S. Eayrs, personal observations). These fish may remain in place for several minutes or longer, until altering hydrodynamics, collision with passing fish, or another stimulus causes them to rise upward and swim toward the codend. In contrast, He et al. (2008) observed that over 80% of flounders (mainly yellowtail flounder [Limanda ferruginea]) swam slowly close to the bottom of the extension and that most of these individuals were facing the codend. Tiger flathead (Neoplatycephalus
84
Fish Behavior near Fishing Gears during Capture Processes
richardsoni) pass through the net cruise-swimming and enter the codend either facing toward the codend or trawl mouth (Piasente et al. 2004), while rock sole (Pleuronectes bilineatus) show little ability to maneuver or slow their progress toward the codend (Rose 1995). 4.5.2 Fish Behavior inside a Codend Most fish that enter a codend are considered to be exhausted and have limited capability for sustained swimming, having endured to a greater or lesser degree forced swimming to minimize or avoid contact with the trawl gear and other fish. The passage of fish from the net mouth to the codend may take several seconds or longer, depending on fish condition (stamina and available reserves of energy), swimming direction, towing speed, and length of trawl net. Using a multicamera system, Yanase et al. (2009) noted that motionless sand flathead took 7 s on average to pass from the mouth of a fish trawl to the codend, while individuals swimming vigorously in the towing direction took 17 s on average to reach the codend. In a shrimp trawl, juvenile red snapper took 15 to 20 s to swim from the net mouth to a turtle excluder device, a distance of 9 m (Engås et al. 1999). Stronger swimming fish can be expected to take a similar duration to these reported times, or even longer, to enter a codend under similar circumstances. Upon reaching the codend, individuals may become part of the accumulated catch, or turn and swim ahead of the catch for a period of time (Watson 1988; Wardle 1992; O’Neill et al. 2003). Small fish may be passively filtered through the netting, whereas those still capable of active swimming may be able to detect, orientate, and swim through open meshes of the codend and escape. Species with poor swimming capability, such as whiptails and New Zealand dory, appear motionless as they travel from the net mouth to the codend, although contact with other fish or netting often results in a burst-swimming response (Eayrs and Piasente 2006). Strong swimming species, such as spotted warehou (Seriolella punctata), can maintain cruiseswimming behavior within the trawl net for long periods before gradually falling back into the codend, although some of these individuals are also able to swim the entire length of the trawl net during
haulback and escape through the trawl mouth (Piasente et al. 2004). The construction of the codend and the amount of accumulated catch are dominant factors affecting the swimming duration of fish and their likelihood of escape from this part of the trawl net. As the codend is towed through the water, netting gathered at the rear of the codend generates water turbulence and eddies, some of which is pushed forward while some is displaced laterally through open meshes in the codend. This turbulence, together with vesselinduced pulsing movement of the codend, may assist fish swimming by reducing the speed required to maintain station within the codend (Broadhurst et al. 1999; Rose 1995) and serve to provide fish an additional opportunity to escape through the codend meshes (Jones et al. 2008; O’Neill et al. 2003). Main and Sangster (1991) compared the swimming duration of fish in codends constructed from diamond-mesh or square-mesh netting and found a three-fold increase in swimming duration in diamond-mesh codends. The open-square meshes resulted in a 25% increase in relative water flow through the codend and, despite having higher filtering potential—due to consistency of mesh opening—fish with poor swimming capability may struggle to swim against this higher flow and may need to make repeated escape attempts before successfully orientating themselves and swimming through a square mesh. Thus, while the use of square-mesh codends seems intuitively a more appropriate option to allow certain species to escape, higher water flow may to some extent be counterproductive and hamper or delay fish escape from these codends. Further influencing water movement in the codend is catch-induced turbulence. As the catch accumulates in the codend, fish block the open meshes and further dam water movement through the codend. Additional water is now displaced forward and will pass laterally through the codend meshes ahead of the catch. Using a flume tank, Wakeford (2004) characterized the flow field within a codend fitted with a bycatch reduction device (Fig. 4.14). Turbulent flow was particularly apparent ahead of a simulated catch, including a reversal in flow immediately ahead of the catch along the upper section of the codend. There was also a sig-
Fish Behavior near Bottom Trawls
85
Figure 4.14. Typical flow field observed throughout a standard extension/codend containing a bycatch reduction device. (Data from Wakeford 2004.) Variation in water speed through the trawl is known to modify fish behavior. Under certain conditions, turbulent flow and eddies are expected, creating areas of flow reversal as shown here near the top of the codend. For color detail, please see color plate section.
nificant reduction in flow velocity extending approximately 3 m ahead of the catch. Placing a bycatch reduction device (BRD) in this region would have a high likelihood of success because displaced water aids fish escapement through the device (Broadhurst et al. 1999; Eayrs 2007). However, if the BRD is located too far ahead of the catch, fish will be unable to swim forward and reach the device and escape. Bycatch reduction will therefore be comprised and consist mainly of individuals with the capability to respond and escape through the device as they initially pass through the trawl extension. Further hampering determination of the optimum location of a BRD is inconsistent catch volume within and between individual tows; the efficacy of these devices is therefore variable and their precise location in a codend is difficult to determine. In shrimp-trawl fisheries, where a BRD that is located too close to the accumulated catch will result in substantial shrimp loss, fishermen usually take a conservative approach and locate the device as far from the codend drawstrings as is legally permitted (Eayrs 2007). An option to overcome this problem is to stimulate fish to seek escape before they reach the codend. Efforts to achieve this have usually
been based on partial blockage of the codend and include tapered funnels (Watson 1988), hummer bars (Brewer et al. 1998; Watson 1992), deflector grids (Watson 1988), netting cones (Engås et al. 1999) or floats, or the use of contrasting materials such as colored selvedges, dyed netting, or black cylinders/tunnels (e.g., Glass and Wardle 1995b; He et al. 2008; Madsen et al. 1998; Wardle 1992). To date, there has only been limited uptake of these devices by fishermen, suggesting varying and/or limited operational performance. 4.5.3 Fish Behavior in Response to Bycatch Reduction Devices Numerous BRDs1 have been developed over the past few decades for different trawl fisheries around the world (see reviews by Eayrs 2007; Glass 2000; Watson et al. 1999). See Figure 4.15 for some common examples. In some instances, the behavior
1
In the broadest sense a bycatch reduction device is any addition or modification to a trawl net designed to allow the escape of unwanted animals. In many fisheries, including shrimp-trawl fisheries, this term specifically refers to devices designed to allow the escape of fin-fish and other small animals such as crabs and jellyfish.
86
Fish Behavior near Fishing Gears during Capture Processes
Figure 4.15. Examples of bycatch reduction devices (BRDs), including (A) the downward excluding Super Shooter TED, (B) Radial Escape Section (RES) for escape of strong swimming fish from the codend, (C) square-mesh codend for escape of small fish from the codend, and (D) square-mesh window (or escape panel) for the escape of small fish from the codend. (Images from Eayrs 2007.)
of fish in response to these devices has been thoroughly evaluated and is known to be in be influenced by a variety of intrinsic factors such as physiological condition, motivational state, fish size, and visual ability, as well as extrinsic factors such as ambient light levels, water temperature, and BRD design, position, and orientation. This, however, is the exception rather than the rule, and a thorough evaluation of the behavior of most species in relation to BRDs awaits further research. Inclined separator grids are primarily designed to mechanically filter and separate the catch by size (e.g., Fig. 4.15a), although the behavior of some species in response to a grid can also aid their escape from the trawl. Several factors are known to affect the hydrodynamic performance of these devices (Riedel and DeAlteris 1995), their catching performance, and associated fish behavior (see a review by Eayrs 2007). As fish pass through the extension of a trawl net toward a grid, they are faced with the option of either turning and maintaining station within the extension, swimming through the escape opening located ahead of the grid, or attempting passage through the bars of the grid. Variation in behavior is often high and many nontarget species
seem particularly tenacious at seeking areas of reduced flow and resisting escape from the trawl net. Engås et al. (1999) reported that red snapper turned to swim with the trawl net and avoid passage through a netting funnel located immediately ahead of a turtle excluder device (TED) and that this response was presumably due to visual stimuli presented by the device. As these fish tired, they passed through the funnel tail first and then encountered the TED. Most of these fish passed through the TED and continued toward the codend; however, many took up station either behind the TED or swam forward (back through the TED) and took up station adjacent the funnel where it was attached to the extension of the trawl net. These fish swam with low tail-beat frequencies, suggesting reduced water flow in this region of the trawl net, and apparently were not inclined to leave for the remainder of the tow. Given the high variation in behavior reported by these authors, it would seem that new or different strategies are required to facilitate the escape of this species from the trawl net. Currently, the relationship between fish escape and the orientation of a grid or TED is not well understood. Anecdotal evidence indicates that
Fish Behavior near Bottom Trawls upward excluding grids improve fish exclusion rates because down welling light is reflected from the bars of the grid and this increases the distance at which the grid first becomes visible. This reflected light may also improve the visibility of the escape opening and further contribute to fish escape. In contrast, the bars of a downward excluding grid will be in shadow to an approaching fish and the escape opening may be less visible. This explanation is certainly plausible; however, the superior performance of upward excluding grids may simply reflect the preferred direction of fish escape, and these conflicting explanations highlight a need to better understand the interplay between fish response to visual stimuli and their natural escape direction. A dramatic example of the importance of grid orientation was recently demonstrated in a southeastern U.S. shrimp fishery. Aware that several nontargeted finfish species such as red snapper resisted vertical displacement by an upward excluding grid, the researchers tested an innovative sideopening grid and observed an estimated 80% increase in the escapement of nontargeted fish. In response to the horizontal bars of the grid, most red snapper turned to face the trawl mouth orientated perpendicular to the grid, then passed along the face of the grid before swimming through the escape opening (D. Foster, National Marine Fisheries Service, Mississippi Labs, personal communication). The orientation and size of a fish may also play a role in their response as they approach an inclined separator grid. Hannah et al. (2003) reported that fish facing an inclined grid often swam up the grid and through the escape opening in a very short period, whereas fish facing away from the grid often took longer to escape. Large fish in particular often responded by maintaining their position a short distance ahead of the grid where turbulent water induced by the funnel or guiding panel reduced the effort required to swim. As these fish became exhausted, many drifted back, then moved upward to avoid the grid and swam through the escape opening (Hannah et al. 2003). Some large fish have also been observed passing through a grid and entering the codend, only to reemerge later and escape through the trawl mouth (S. Eayrs, personal observations). In contrast, small fish, flatfish, and
87
others with relatively poor swimming capability are sometimes held hydrodynamically against the bars of a grid. These fish are usually dislodged after a short period by sudden grid movement or by contact with passing fish (Eayrs 2007). Small fish have also been observed attempting to maintain station immediately behind the bars of a grid (Engås et al. 1999; Eayrs 2007) or even deliberately lodging themselves behind the bars where they are attached to the trawl netting (S. Eayrs, personal observations). This behavior is considered adaptive in that it reduces energy expenditure; in this region, the swimming speed required to maintain station with the grid may be reduced by 50% or more (Fig. 4.14). This behavior may continue for several minutes until the individual is overcome with exhaustion or contacted by other passing fish. As flatfish pass through a grid, they may attempt to rest on netting distorted by the circumference of the grid. In this position, these fish are orientated at an angle to the apparent water movement and thus held in place against the netting; these fish may take several minutes to reach the codend (Eayrs 2007). The escape of fish through a BRD relies on their ability to detect, orientate, and swim through the escape openings of the device. Özbilgin and Wardle (2002) reported that haddock typically used one of two methods to escape through a square-mesh panel (e.g., Fig. 4.15d)—either slowly working their way through a mesh with a single tail-beat followed by motionless behavior or with rigorous acceleration through a mesh followed by continuous high-speed swimming away from the trawl net. Grimaldo et al. (2007) observed that loose or slack rigging of such panels caused them to “flutter,” and this helped provoke undersized haddock to actively escape and improve escape success. This suggests that both the visual capability of fish and ambient light level are important parameters influencing fish escape. In the absence of vision, fish may not be able to respond to a BRD and only passive filtration is possible, and then only if correctly orientated (Gabr et al. 2007). For fish swimming ahead of accumulated catch in the codend (having missed the opportunity to escape as they initially passed through the extension and codend), reaching a BRD requires swimming at a speed greater than the apparent speed of water in
88
Fish Behavior near Fishing Gears during Capture Processes
the codend. While this will be less than the towing speed, many individuals by this point may simply not have the metabolic reserves to power their escape. Parsons and Foster (2007) argue that the ideal water flow in and around a BRD should be little more than 0.4 m/s to permit successful escape of certain bycatch species. In many tropical shrimp trawl fisheries, bycatch is dominated by small fish that are similar in size or larger than the shrimp. These fish generally have superior swimming capabilities compared with shrimp, and BRDs designed to allow fish to escape by swimming through one or more escape openings are required to reduce catches of these fish. Many of these BRDs are designed to be placed in close proximity to a region of turbulent water flow or generate turbulent water flow themselves (Eayrs 2007). Fish are able to detect turbulent water through their lateral line system and will attempt to remain within an area of turbulent flow, thus increasing their likelihood of encountering a BRD and escaping from the trawl net. Several BRDs have proved particularly successful and have been mandated in certain fisheries, including the fisheye, extended funnel, and Jones/Davis device (see reviews by Watson 2007; Watson et al. 1999). Admittedly, however, none of the BRDs in use today are completely successful in eliminating fish bycatch. In many cases, fish can be guided toward the escape opening of a BRD but make no attempt to escape through the device (Eayrs 2007; Parsons and Foster 2007). The precise reason for this is unknown, but it is thought to be linked to the optomotor reflex and a desire to maintain station within the trawl net. Escapement of these more resistive individuals often occurs only during the haulback procedure, when reduced speed, pulsing codend movements, and slack netting together overcome this reflex and enable fish to swim more readily through the escape openings of the device (Broadhurst et al. 1996; Eayrs 2007; Engås et al. 1999; Madsen et al. 2008; Watson et al. 1999). The key to improved BRD designs therefore seems in the first instance to be linked to deliberate generation of water turbulence to entice fish toward a BRD and, second, by developing methods to overcome the optomotor reflex so fish voluntarily depart from the trawl net throughout the entire tow.
However, efforts to manipulate water movement and improve BRD design have usually been based on a trial-and-error approach to testing and development, and this has meant that the design of many BRDs has remained largely unchanged for several decades. A relatively recent innovation, the fishbox BRD, uses one or more metal plates (or foils) to induce water turbulence adjacent to escape openings. Despite having been tested successfully in shrimp fisheries in the United States and Australia (Eayrs 2007) and outperforming most other BRDs, particularly during daytime, an intimate understanding of fish behavior to variations in foil design and orientation awaits to be fully realized. Overcoming the optomotor reflex to facilitate fish escapement has to date not been effectively achieved and also requires further research. Reducing the visual stimulus (contrast) of the trawl extension, codend, and BRD is an obvious first step but requires a greater understanding of the visual capabilities of fish under a range of ambient light levels encountered in the fishery and their response under these various conditions. Meanwhile, until this is achieved, the influence of the optomotor reflex will dominate fish behavior and BRD performance will not be optimized. The immediate future of BRD research clearly rests in greater knowledge of fish behavior and their response to trawl stimuli, particularly in low-light conditions. The need for divers or cameras to observe fish behavior in close proximity to BRDs remains a significant limitation. The use of flume tanks to assess water flow, test ways to generate turbulence, and develop new concepts is an important component of BRD research (Brewer et al. 1998). These facilities provide a controlled testing environment and have the necessary instrumentation to quantify water velocity and direction in and around these devices and codends (Winger et al. 2006). This knowledge can be used to develop new or improved designs that can be applied to at-sea studies. Watson (1988) argues that deliberately altering water flow is the most promising technique for developing selective shrimp trawls, in combination with knowledge of fish response to trawl stimuli. Wakeford (2004) successfully used a flume tank to assess the effect of a fisheye BRD on water flow in and around a full-sized shrimp-trawl codend,
Fish Behavior near Bottom Trawls and despite only being able to record water flow in two dimensions, the author reported that the device reduced flow in excess of 80% near the upper panel of the codend. Just how this knowledge might be used to improve the performance of the fisheye remains unclear, but it does provide a first step in stimulating new ideas and ways to elicit fish movement toward such devices and escapement from a trawl net. 4.6 FACTORS INFLUENCING FISH BEHAVIOR NEAR TRAWLS 4.6.1 Extrinsic Factors Variation in bottom trawl catchability is known to be associated with spatial and temporal changes in environmental conditions. Ambient light intensity, water temperature, and fish density are generally considered to be three of the most influential extrinsic factors governing fish behavior and fish capture; each is described next. Light Level, Contrast, and Color It is well known that most species of fish have welldeveloped and efficient visual systems that are particularly well adapted to detect very small differences in contrast in the generally monochromatic underwater environment in which they exist (Jones et al. 2004). Many species, particularly fast-swimming pelagic ones, have been demonstrated to have excellent visual acuity and low light sensitivity. In addition, many commercially important fish species can form visual images at light intensities well below those at which human visual systems cease to perform. In some species these minimum light intensity thresholds occur at extremely low levels (less than 10−6 lux) (Glass and Wardle 1989; Walsh and Hickey 1993) and can be difficult to detect and measure reliably. Consequently, environmental conditions, such as underwater light field, can have a significant effect on fish availability (Jones et al. 2004), catchability (Wileman et al. 1996), and reaction behavior (Ryer and Barnett 2006) within and around the net. Observations on the reaction behavior of fish to towed fishing gears have been conducted for more than 60 years and have shown that when light is present, fish are not sieved from the sea but react in quite subtle ways to the complex
89
visual stimuli presented by components of the gear. However, despite the advanced performance of fish visual systems, there will clearly be times and/or depths at which their visual systems cease to perform and this will manifest itself in a different suite of behavioral responses. To date, this aspect has received little attention, perhaps due to the difficulty of “observing” behavior under these dark conditions. The visibility of an object such as a netting panel, trawl float, door, or other component of a net depends largely on its contrast with the background and is therefore dependent on the background against which it is viewed, the properties of the water, as well the direction and intensity of the illumination (Kim and Wardle 1998a, 1998b). All of this combines to produce a unique visual background against which a fish must view its world. The first and perhaps most obvious characteristic about each unique visual background is its visible range. Where water clarity is good and light intensity is high, fish may see an approaching net from afar and may react to its approach by rising in the water column and allowing the net to pass underneath, thus avoiding capture. Conversely, in dark or turbid conditions, visible range may be very short and fish may have little time to react to the approaching net. Under these circumstances, response thresholds are generally expected to be higher, reaction distances shorter, and collisions with the gear more likely (Glass and Wardle 1989; Walsh and Hickey 1993). However, under most circumstances where the fish are able to form visual images, they will react in a predictable manner to the visual stimuli presented by the gear and its components. The headrope and its floats will be seen as a high-contrast silhouette, as will the sweeps and footgear as viewed against the sediment underneath. The netting surrounding the fish appears as a moving pattern, and this induces an optomotor response, causing the fish to turn and swim in the direction of the tow at the same speed as the moving netting (see Section 4.4). Aside from the visual background against which an object is viewed, the nature of the object itself is also important, particularly its color. Netting panels used in bottom trawls may be dyed any number of colors (green, blue, red, orange, yellow), and harvesters rarely agree on the combination of
90
Fish Behavior near Fishing Gears during Capture Processes
panel colors necessary to achieve optimal catch rates and bycatch reduction. Certain panels may be colored to increase their visibility and assist with herding, whereas other panels may be altered to reduce their conspicuous to assist with escapement of nontargeted species or sizes (for review, see Glass 2000; Jones et al. 2004; Wardle 1993). Visual illusions may also be used to encourage fish to actively select certain behavior patterns. Laboratory experiments have demonstrated that fish will choose to penetrate meshes when alternative routes are blocked or appear blocked due to visual illusion (Glass et al. 1995). In this case, the visual stimulus was a black tunnel that was believed to appear like the looming open mouth of a large predator. Subsequent sea trials were carried out using black PVC-coated canvas laced into the extension of a trawl to create a black tunnel effect. Video observations of fish showed a strong aversion to entering the tunnel and vigorous attempts to penetrate square-mesh escapement panels located immediately ahead of the perceived “threat” (Glass and Wardle 1995b). Trials in the Gulf of Maine have also demonstrated promising results of black tunnels at separating target from nontarget species (He et al. 2008), although other attempts to duplicate the findings have been less successful (Brewer et al. 1998). Water Temperature With the exception of certain scombrids and sharks, water temperature is known to be an important environmental constraint on the behavior of most species of fish. It affects almost every aspect of an individual’s activities, including metabolism, muscle performance, growth rate, activity level, foraging routine, vigilance to predators, and even behavior toward fishing gear. It is generally thought that temperature influences the catchability of a bottom trawl in two ways: (1) by affecting response threshold and (2) by affecting swimming capability. In Section 4.3, it was noted that an individual’s response threshold toward an approaching threat is the outcome of a behavioral tradeoff that minimizes costs and maximizes benefits. If we assume the costs of fleeing (F) are related in some way to water temperature (probably through energy expendi-
ture), we would expect individuals to adjust their optimal reaction distance (D*) so as to minimize these costs (see Fig. 4.4). In other words, we expect that at unfavorable water temperatures (either too warm or cold), the F-cost curve should rise more steeply, shifting the intersection of the curves to the left, resulting in a reduced reaction distance (D*) to the threat. Field data collected by Winger (2004) support this argument. Atlantic cod exhibited a reduced likelihood of responding to an approaching trawler in winter compared with summer, as well as a corresponding reduction in reaction distance at the lower water temperatures. Under environmentally extreme conditions, we would expect an even higher response threshold, near-zero reaction distance, and an inability for the trawl to elicit a behavioral response whatsoever. It is possible that the passive drifting of walleye pollock into a trawl net under low water temperatures, recorded by Inoue et al. (1993), may be representative of this kind of situation. The second way that temperature affects trawl capture is through its influence on swimming capability. At the most basic level, it alters the rate at which muscle fiber is physically capable of contracting. This limits muscle power and in turn the swimming speeds at which fish are capable of swimming, essentially shrinking the performance range. Although empirical field observations are still largely lacking, laboratory evidence suggests water temperature has a profound effect on the fish capture process, including limiting the herding efficiency of trawl sweeps (Ryer and Barnett 2006; Winger et al. 1999), limiting swimming speed and endurance in the trawl mouth (He 1991, 1993; Yanase et al. 2007), and limiting the ability to escape once inside the net (Özbilgin 2002; Özbilgin and Wardle 2002). For a detailed review, see Chapter 1. Fish Density Variation in distribution, patch size, and fish density are common. Some species maintain aggregations almost continuously, while others do so infrequently or only on occasion. At low densities, the behavior of individual fish is largely independent of neighbors, while at higher densities social interactions with conspecifics are likely to occur. This
Fish Behavior near Bottom Trawls variation in the local density of fish is known to influence aspects of fish capture behavior from the net mouth to the codend. Godø et al. (1999) provides evidence for densitydependent behavior in the net mouth based on a comparative analysis of Norwegian and Canadian underwater video observations. The authors report qualitative differences in behavior at varying densities for cod, haddock, and American plaice. At low densities (one or two fish), cod and haddock displayed “loner behavior,” characterized as kick-andglide swimming while criss-crossing the net mouth. Turnover rates (number of fish entering the net per unit time) were high and escapement under the footgear was common. As fish density increased (more than five fish), cod and haddock typically exhibited “schooling behavior.” This was characterized as a more-uniform, less-frightened behavioral response, resulting in a lower turnover rate and reduced escapement under the footgear. In the case of plaice, low densities were characterized by zigzag behavior and increased escapement under the footgear. At higher densities, the zigzag behavior was often disrupted and collisions with other fish were common, resulting in reduced escapement under the footgear. Evidence that variation in fish density in the net mouth actually corresponds to differences in trawl efficiency has been examined by Walsh (1996) and Godø et al. (1999). Both studies examined the catch of fish in the trawl net compared with fish caught in smaller bag nets that were attached underneath the main trawl, designed to catch downward escaping fish that pass through the footgear. Albeit highly variable, both studies showed a general tendency toward lower efficiency of the footgear (i.e., more fish escaping underneath) at lower densities and greater efficiency (i.e., less fish escaping underneath) at higher densities. Fish density is also expected to affect the performance of BRDs installed within the net. When large pulses of fish are encountered, devices such as selection windows, sorting grids, or separator panels may be temporarily masked by neighboring conspecifics. This reduces the probability of fish encountering the devices and thus reduces the potential sorting efficiency. For example, a BRD that is designed to reduce the capture of small fish
91
may sustain reduced performance at high catch rates (i.e., high densities), resulting in a larger proportion of small fish in the catch, lowering the overall L50 for the haul. Evidence of a negative correlation between L50 and catch rate has been documented by Kvamme and Isaksen (2004) and Jørgensen et al. (2006). Finally, recent evidence suggests that the degree of accumulated catch and subsequent density of fish in the codend affects the duration of station keeping as well as the likelihood of active escape seeking behavior in some species. Jones et al. (2008) collected underwater video observations of haddock in the codend of a commercial whitefish trawl. The authors found that at low densities, haddock exhibited a reduced preference for station keeping, resulting in an increased rate of contact with the netting and subsequent escape. As the catch accumulated in the codend and the density increased, station keeping behavior increased and the rate of contact with the netting decreased. The findings are consistent with similar observations in the net mouth (Godø et al. 1999), suggesting an increased preference in some species to maintain the status quo, or togetherness, when surrounded with the company of other fish. 4.6.2 Intrinsic Factors In addition to the many extrinsic factors that modify fish behavior during capture, there are also a number of intrinsic factors—that is, conditions or states wholly belonging to the individual—that are also known to modify behavioral expression and therefore trawl catchability. The best understood of these is fish size, but there are other lesser studied factors, including an individual’s motivational state, physiological condition, learning, and experience. Fish Size One of the most obvious and well-studied factors affecting fish capture by trawls is fish size. The effect is manifested primarily as length-dependent swimming capability. Both the maximum swimming speed that an animal can achieve and its endurance are closely related to its body length (see Chapter 1). This is particularly important during the herding of benthic species located in the sweep zone as well as for all fish swimming in the trawl
92
Fish Behavior near Fishing Gears during Capture Processes
mouth (see Section 4.4). In both cases, every fish, regardless of size, is stimulated to swim at a certain speed. As a result, fish of different sizes must operate at different levels within their performance range. Small fish are required to swim vigorously with high tail-beat frequencies and will operate at the upper end of their performance range, analogous to sprinting. Larger fish, by comparison, are capable of higher swimming speeds and therefore operate at a lower point in their performance range, analogous to jogging. To effectively swim at these different levels, teleost fish exhibit more than one gait (Alexander 1989; Peake and Farrell 2004; Winger et al. 2004). Gaits represent a multigear system (synonymous with walking, jogging, sprinting) for the fractionation of the swimming performance range (Webb 1994a). Similar to terrestrial animals, individual gaits in fish work over a part of the performance range and a series of gaits together cover the entire performance range. Each gait is defined on the basis of its muscle use, propulsor type, and propulsor kinematics (Webb 1994b). Looking at the trawl mouth as an example, fish of different sizes are observed using different gaits in an effort to produce sufficient thrust to power their swimming and avoid falling back into the trawl. Small fish are characterized by high tail-beat frequencies coupled with burst-and-coast motions. These sprintlike bursts rely heavily on the fast contraction rates of white anaerobic muscle fibers to produce sufficient mechanical power. Medium-sized fish, by comparison, are usually characterized by continuous swimming punctuated with occasional large-amplitude kicks of the tail for extra thrust. As fish size continues to increase, a transition from unsteady to steady swimming occurs, with the largest fish swimming effortlessly in the mouth of the trawl. This final gait transition coincides with a change from prolonged to sustained swimming activity. Fish at these sizes have sufficient cross-sectional muscle area to produce the power for steady cruising and do not recruit white muscle fiber to the extent that smaller fish must (see reviews by Videler 1993; Videler and Wardle 1991; Webb 1994ab). Motivational State Ecological theory tells us that animals are routinely (if not constantly) balancing conflicting demands.
Assuming fish cannot maximize all things all the time, the situation inescapably arises where two or more demands cannot be satisfied simultaneously, and a tradeoff must be made. Usually, this includes things such as balancing the need to eat against the need to avoid being eaten (e.g., Nøttestad et al. 1996). Albeit relatively modern from an evolutionary perspective, mobile trawls also represent a tangible threat, and fish are expected to make behavioral tradeoffs at the individual level that minimize costs and maximize benefits. One of the best ways to observe this tradeoff is to witness changes in response threshold (i.e., reaction distance) under different scenarios (see Section 4.3). One particularly interesting tradeoff scenario is the harvesting of spawning aggregations. Under these conditions, the threat of capture is pitted against maximizing reproductive fitness/output. In this situation, we would predict the cost of fleeing (F) to be high (that is a steeper line than usual) because of lost reproductive opportunities. Hence, vigilance toward approaching threats should be low, raising the response threshold, and lowering the optimal reaction distance (see Fig. 4.4). Empirical evidence to support this argument is limited but not entirely lacking. Olsen (1971) described capelin (Mallotus villosus Müll.) as insensitive to vessel noise during the spawning season. Mohr (1971) compared the avoidance behavior of prespawning aggregations of Atlantic herring to midwater trawls against those of spawning aggregations. He described prespawning herring to be “skittish” but, as the spawning time approached, herring became “sluggish and indolent.” At peak spawning season, almost no reaction to the vessel and trawl could be observed whatsoever. More recently, Skaret et al. (2005) also reported no avoidance of spawning aggregations of herring toward the noise of a large research vessel in shallow water (30–40 m). These observations are atypical during the rest of the year when herring are typically riskaverse and react strongly to such threats (e.g., Misund 1994; Vabø et al. 2002), supporting the argument that motivational state can modify behavioral expression. In general, motivational state is poorly understood and our ability to accurately observe and record tradeoffs under different motivational condi-
Fish Behavior near Bottom Trawls tions is limited. In many cases, such things cannot be adequately teased apart under field conditions due to confounding differences with other extrinsic factors (e.g., water temperature, light intensity, depth, predator–prey field, etc.), which may produce unexpected results at times (e.g., Skaret et al. 2006). Further research is required to illuminate this field. Laboratory experiments, albeit limited in their application, may be the only way to investigate variation in behavioral expression under different motivational states. Physiological Condition The condition of fish is known to vary seasonally in association with things such as feeding, migrating, and reproductive cycles. It can also vary with environmental conditions and food availability. For comparative purposes, it is often assessed on the basis of morphometrics or somatic indices or with the use of biochemical indicators (e.g., Dutil et al. 1998; Martínez et al. 1999). Although empirical evidence is limited, changes in fish condition are suspected to affect trawl catchability in at least three ways. First, physiological condition may affect trawl capture through its influence on response threshold and motivational state (see earlier discussions). Clearly, individuals with poor nutritional status will make different tradeoffs as they balance the conflicting demands of fleeing in response to an approaching trawl or continuing with some existing activity. Even more complicating would be if they are actively engaged in something important such as a feeding opportunity, migration, or spawning. Under this scenario, there would be serious lost opportunities associated with fleeing. This should push the F-cost curve upward, shifting the intersection of the curves to the left, resulting in a decreased reaction distance (D*) to the threat (Fig. 4.4). This is, of course, speculative; empirical evidence to support this hypothesis is completely lacking. Second, physiological condition affects trawl capture through its influence on swimming capability. As fish condition deteriorates, lipid and glycogen reserves are typically mobilized first, but eventually so are white muscle proteins. This changes white muscle composition, reduces muscle mass, limits power output, and shrinks the perfor-
93
mance range. Laboratory studies have demonstrated that starved cod suffer both reduced sprint speeds (Martínez et al. 2002) and reduced swimming endurance (Martínez et al. 2003) compared with their fed counterparts. Although field data are lacking, we would expect this reduced performance to limit the herding efficiency of trawl sweeps, limit swimming speed and endurance in the trawl mouth, and limit escape ability once inside the trawl. Finally, physiological condition is expected to affect trawl capture through its influence on mesh penetration capability. Özbilgin et al. (2006) and Ferro et al. (2008) investigated temporal variation in codend size selection for North Sea haddock over a period of 1 year. The authors found seasonal variation in retention characteristics of the trawl based on differences in fish girth. Interestingly, for fish of a given length, an increase in girth lead to increased retention in the winter, whereas it led to a decrease in the fall. They explain this difference on the basis that the increased girth in winter was associated with gonad development, whereas in the fall it was associated with muscle volume. They speculate that the latter probably had better swimming performance (for a given length and girth) resulting in increased capacity to penetrate and escape open meshes in the codend. Learning and Experience Models of fish reaction behavior often assume that all fish react in a similar manner and that all fish are naïve when they encounter a bottom trawl. These are both erroneous assumptions. In Section 4.6, we explored how motivational state can have a significant effect on the reaction behavior of fish but it is also reasonable to assume that, in many cases, fish may encounter fishing gear on more than one occasion and that their behavioral reactions may be modified through a process of learning from past experience. More than 50 years ago, Golenchenko (1955) raised the possibility that fish could have conditioned responses to active fishing gear. Using the Russian made hydrostat, Kiselev (1968) documented some of the first field observations of fish behavior in response to the repeated sounds of a trawler. He reported that cod were initially “agitated” but, after a few repetitions of the sound, soon
94
Fish Behavior near Fishing Gears during Capture Processes
habituated and showed “no response.” Direct observations of cod and haddock in response to repeated trawl hauls were later conducted during the 1980s using the towed submersible “Tetis” (Zaferman and Serebrov 1989). The authors documented reduced catchability of conditioned fish, especially smaller fish that had escaped the trawl during previous encounters. Using acoustic telemetry, Pyanov (1993) demonstrated that experienced fish will actually move out of the path of an approaching trawler and that the learned behavior could be observed as much as 9 days after the initial encounter. How rapid can fish learn? While this is likely to vary between species and individuals (Hunter and Wisby 1964; Zhuykov and Pyanov 1993), evidence suggests that fish have the capacity to learn quickly, particularly in response to aversive stimuli such as fishing gear. This is because learning about fishing gear is a risky business, and there is little room for mistakes or extensive training. Laboratory studies have demonstrated rather convincingly that fish can learn to avoid approaching nets after just one or two experiences (Hunter and Wisby 1964; Pyanov 1993; Pyanov and Zhuykov 1993) and that this can lead to noticeable reductions in catching efficiency during repeated tows (for a review of the Russian literature, see Pavlov and Kasumyan 1995). However, despite multiple demonstrations of the ability of fish to avoid approaching trawlers, fishing gears continue to capture fish even in heavily fished areas where there is a likelihood that individual fish may encounter and pass through the meshes on more than one occasion. Özbilgin and Glass (2004) demonstrated in a laboratory setting that haddock were reluctant to pass through a mesh panel but learned quickly to penetrate the meshes when a food reward was presented as a conditioned stimulus. Once this response was learned, the fish penetrated the meshes with ease when presented with the conditioned stimulus while still reluctant to do so in the absence of the stimulus. They concluded that haddock were capable of modifying their behavior based on prior experience and argued that fish that encounter a trawl gear on more than one occasion may be capable of “escaping” more effectively on subsequent encounters with a net or that, at the very least, their behavior may be modified in some way.
Can learned behavior be transmitted to naïve individuals? In formal terms, this is referred to as “social learning” and is defined as the process by which individuals acquire new behavior or information about their environment via the observation or interaction with other animals (see review by Brown and Laland 2003). Two remarkable studies demonstrated social learning among fish in response to approaching nets under laboratory conditions (Brown and Warburton 1999; Hunter and Wisby 1964). In these experiments, nets were towed along a tank and the ability of fish to locate and escape through a small hole was evaluated. Failure to locate the hole resulted in a negative/punitive experience (trapped) and success resulted in a positive experience (escape). Both studies showed that groups of fish solved the problem faster than isolates or smaller groups and that escape latency decreased with repeated exposure. The findings demonstrate the potential value of social interactions in avoiding active fishing gear. In these examples, fish were more likely to make the “right” decision sooner and with greater accuracy when they were able to monitor the behavior of several fellow conspecifics, benefiting from a discovery made by any one of them. Other supporting evidence for social learning in fish has been documented for herring (Soria et al. 1993) as well as haddock (Özbilgin and Glass 2004). In the latter study, naïve haddock learned to penetrate mesh more quickly when in the company of experienced fish. One clear implication of such behavioral modification through learning is that the efficiency and selectivity of a net may not be the same for all individuals that encounter the net and that “experienced” fish may have a higher probability of escapement, and hence potential survivability, than naïve fish. This may have important implications for surveys and other scientific studies such as those that seek to determine selective efficiency of fishing gears where the assumption is that all fish react in a similar manner to the net. The effect of learned avoidance behavior in fishing gears has also been discussed in detail by Fréon et al. (1993) and Soria et al. (1993), who concluded that there could be a long-term decrease in catchability in exploited stocks where fish are relatively more experienced than for fish of unexploited stocks.
Fish Behavior near Bottom Trawls Although a body of evidence exists in support of the role of learning and experience in behavioral modification of reactions to fishing gears, this remains an area that may yet yield important insights through further research. 4.7 CONCLUDING REMARKS This chapter reviewed the current knowledge of fish behavior in relation to bottom trawls, with discussions starting in the pretrawl zone ahead of the vessel and finishing with behavior in the codend. The entire capture process has been discussed along with key extrinsic and intrinsic factors that are known to influence behavior. We have described typical patterns of behavior but equally emphasized the high degree of variability in behavioral expression that is often observed. We have extended the application of the economic hypothesis of antipredator behavior (Ydenberg and Dill 1986) initially introduced to the field of trawl capture by Fernö and Huse (2003) and found that, in many cases, it is helpful in formulating predictions of fish behavior in response to vessels and bottom trawls. ACKNOWLEDGMENTS The authors are grateful to Emma Jones (NIWA, New Zealand) and Daniel Foster (NMFS, Mississippi, USA) for their helpful comments and suggestions on the manuscript. REFERENCES Albert OT, Harbitz A and Høines ÅS. 2003. Greenland halibut observed by video in front of survey trawl: behavior, escapement, and spatial pattern. J. Sea Res. 50: 117–127. Alexander RM. 1989. Optimization and gaits in the locomotion of vertebrates. Physiol. Rev. 69: 1199–1227. Anon. 1995. Underwater noise of research vessels: review and recommendations. Ed. by RB Mitson. ICES Coop. Res. Rep. 209: 61 pp. Anon. In press. Causes and consequences of fish reactions to fisheries research vessels. Ed. by F. Gerlotto, ICES Coop. Res. Rep. Beamish FWH. 1966a. Reaction of fish to otter trawls. Fish. Can. 19(5): 19–21. Beamish FWH. 1966b. Swimming endurance of some Northwest Atlantic fishes. J. Fish. Res. Bd. Can. 23: 341–347.
95
Beamish FWH. 1969. Photographic observations on reactions of fish ahead of otter trawls. FAO Fish. Rep. 62(3): 511–521 Beutel D, Skrobe L, Castro K, Ruhle P Sr, Ruhle P Jr, O’Grady J and Knight J. 2008. Bycatch reduction in the Northeast USA directed haddock bottom trawl fishery. Fish. Res. 94: 190–198. Breen M, Dyson J, O’Neill FG, Jones E and Haigh M. 2004. Swimming endurance of haddock (Melanogrammus aeglefinus L.) at prolonged and sustained swimming speeds, and its role in their capture by towed fishing gears. ICES J. Mar. Sci. 61: 1071–1079. Brewer D, Rawlinson N, Eayrs S and Burridge C. 1998. An assessment of bycatch reduction devices in a tropical Australian prawn trawl fishery. Fish. Res. 36: 195–215. Broadhurst MK, Kennelly SJ and Eayrs S. 1999. Flowrelated effects in prawn trawl codends: potential for increasing the escape of unwanted fish through square-mesh panels. Fish. Bull. 97: 1–8. Broadhurst MK, Kennelly SJ and Gray CA. 2002. Optimal positioning and design of behavioral-type by-catch reduction devices involving square-mesh panels in penaeid prawn-trawl codends. Mar. Freshw. Res. 53: 813–823. Broadhurst MK, Kennelly SJ and O’Doherty G. 1996. Effects of square-mesh panels in codends and of haulback delay on bycatch reduction in the oceanic prawn-trawl fishery of New South Wales. Aust. Fish. Bull. 94: 412–422. Brown C and Laland KN. 2003. Social learning in fishes: a review. Fish Fish 4: 280–288. Brown C and Warburton K. 1999. Social mechanisms enhance escape responses in shoals of rainbowfish, Melanotaenia duboulayi. Environ. Biol. Fish 56: 455–459. Bublitz CG. 1996. Quantitative evaluation of flatfish behavior during the capture by trawl gear. Fish. Res. 25: 293–304. Buerkle U. 1977. Detection of trawling noise by Atlantic cod (Gadus morhua L.). Mar. Behav. Physiol. 4: 233–242. Castro KM, DeAlteris JT and Milliken HO. 1992. The application of a methodology to quantify fish behavior in the vicinity of demersal trawls in the Northwest Atlantic, USA. Mar. Technol. Soc. Conf. Proc. 1992: 310–315. Chandler RA. 1967. Swimming endurance of haddock. Fish. Res. Bd. Can. Manuscr. Rep. Ser. 930: 14 pp. Chapman CJ. 1973. Field studies of hearing in teleost fish. Helgoländer wiss. Meeresunters 24: 371–390.
96
Fish Behavior near Fishing Gears during Capture Processes
Chapman CJ and Johnstone ADF. 1974. Some auditory discrimination experiments on marine fish. J. Exp. Biol. 61: 521–528. Cooper WE Jr and Frederick WG. 2007. Optimal flight initiation distance. J. Theor. Biol. 244: 59–67. DeAlteris JT, Castro KM and Milliken HO. 1992. Development of an underwater video camera and recording system for observing fish behavior in the vicinity of a bottom trawl and a methodology to quantitatively analyze the resulting data. Final Report, Rhode Island Sea Gant Program, DOC Award No. NA-89AA-D-56082, 20 pp. De Robertis A, Hjellvik V, Williamson NJ and Wilson CD. 2008. Silent ships do not always encounter more fish: comparison of acoustic backscatter recorded by a noise-reduced and a conventional research vessel. ICES J. Mar. Sci. 65: 623–635. Dogˇanyilmaz-Özbilgin Y, Özbilgin H and Bas¸aran F. 2006. Relationship between critical and maximum swimming speeds. ICES WGFTFB Report, pp 145–153. Dutil J-D, Lambert Y, Guderley H, Blier PU, Pelletier D and Desroches M. 1998. Nucleic acids and enzymes in Atlantic cod (Gadus morhua) differing in condition and growth rate trajectories. Can. J. Fish. Aquat. Sci. 55: 788–795. Eayrs S. 2007. A guide to bycatch reduction in tropical shrimp-trawl fisheries. Revised edition. Rome: FAO. 105 pp. Eayrs S and Piasente M. 2006. In situ examinination of fish behavior in response to a demersal trawl net. Presented at ICES Symposium on Fishing Technology in the 21st Century, Oct 30-Nov 3, 2006. Boston, MA. Engås A. 1994. The effects of trawl performance and fish behavior on the catching efficiency of demersal sampling trawls. In: Fernöö A and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 45–68. Oxford: Fishing News Books Ltd. Engås A and Godø OR. 1989a. The effect of different sweep lengths on the length composition of bottomsampling trawl catches. J. Cons. Int. Explor. Mer. 45: 263–268. Engås A and Godø OR. 1989b. Escape of fish under the fishing line of a Norwegian sampling trawl and its influence on survey results. J. Cons. Int. Explor. Mer. 45: 269–276. Engås A, Foster D, Hataway BD, Watson JW and Workman I. 1999. The behavioral response of juvenile red snapper (Lutjanus campechanus) to
shrimp trawls that utilize water flow modifications to induce escapement. Mar. Technol. Soc. J. 33(2): 43–50. Engås A, Haugland EK and Øvredal JT. 1998. Reactions of cod (Gadus morhua) in the pre-vessel zone to an approaching trawler under different light conditions: preliminary results. Hydrobiologia 371/372: 199–206. Fay RR. 2005. Sound source localization by fishes. In: Popper AN and Fay RR (eds). Sound Source Localization. pp 36–66. New York: Springer. Fernö A. 1993. Advances in understanding of basic behavior: consequences for fish capture studies. ICES Mar. Sci. Symp. 196: 5–11. Fernö A and Huse I. 2003. Fish avoidance of survey vessels and gear: can predictions based on the response of fish to predators explain the observed variations? Presented at the ICES Symposium on Fish Behavior in Exploited Ecosystems, June 23– 26, 2003. Bergen, Norway. Ferro RST, Özbilgin H and Breen M. 2008. The potential for optimizing yield from a haddock trawl fishery using seasonal changes in selectivity, population structure, and fish condition. Fish. Res. 94: 151–159. Foster JJ. 1969. The influence of fish behavior on trawl design with special reference to mathematical interpretations of observations on the swimming speeds of fish and results of C.F. experiments. FAO Fish. Rep. 62(3): 731–773. Foster JJ, Campbell CM and Sabin GCW. 1981. The fish catching process relevant to trawls. Can. Spec. Publ. Fish. Aquat. Sci. 58: 229–246. Fréon P, Gerlotto F and Misund OA. 1993. Consequences of fish behavior for stock assessment. ICES Mar. Sci. Symp. 196: 190–195. Fréon P and Misund OA. 1999. Dynamics of Pelagic Fish Distribution and Behavior: Effects on Fisheries and Stock Assessment. Oxford: Fishing News Books. 348 pp. Fuwa S. 1989. Fish herding model by ground ropes considering reaction of fish. Nippon Suisan Gakkaishi. 55: 1767–1771. Fuwa S, Sato O, Nashimoto K and Higo N. 1988. Fish herding model by ground rope. Nippon Suisan Gakkaishi 54: 1155–1159. Gabr M, Fujimori Y, Shimizu S and Miura T. 2007. Behavior analysis of undersized fish escaping through square meshes and separating grids in simulated trawling experiment. Fish. Res. 85: 112–121. Gerlotto F, Castillo J, Saavedra A, Barbieri MA, Espejo M and Cotel P. 2004. Three-dimensional
Fish Behavior near Bottom Trawls structure and avoidance behavior of anchovy and common sardine schools in central southern Chile. ICES J. Mar. Sci. 61: 1120–1126. Glass CW. 2000. Conservation of fish stocks through bycatch reduction: a review. Northeast Naturalist 7: 395–410. Glass CW and Wardle CS. 1989. Comparison of the reactions of fish to a trawl gear, at high and low light intensities. Fish. Res. 7: 249–266. Glass CW and Wardle CS. 1995a. A review o fish behavior in relation to species separation and bycatch reduction in mixed fisheries. In: Solving Bycatch: Considerations for Today, and Tomorrow. pp 243–250. Alaska Sea Grant College Program Report No. 96–03, University of Alaska Fairbanks. Glass CW and Wardle CS. 1995b. Studies on the use of visual stimuli to control fish escape from codends. II. The effect of a black tunnel on the reaction behavior of fish in otter trawl codends. Fish. Res. 23: 165–174. Glass CW, Sarno B, Milliken HO, Morris GD and Carr HA. 1999. Bycatch reduction in Massachusetts inshore squid (Loligo pealeii) trawl fisheries. Mar. Technol. Soc. J. 33(2): 35–42. Glass CW, Wardle CS, Gosden SJ and Racey DN. 1995. Studies on the use of visual stimuli to control fish escape from codends. I. Laboratory studies on the effect of a black tunnel on mesh penetration. Fish. Res. 23: 157–164. Godin J-G. 1997. Evading predators. In: Godin J-G (ed). Behavioral Ecology of Teleost Fishes. pp 191– 236. Oxford: University Press. Godø OR. 1994. Factors affecting the reliability of groundfish abundance estimates from bottom trawl surveys. In: Fernö A and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 169–199. Oxford: Fishing News Books. Godø OR. 1998. What can technology offer the future fisheries scientist—possibilities for obtaining better estimates of stock abundance by direct observations. J. Northw. Atl. Fish. Aquat. Sci. 23: 105–131. Godø OR and Walsh SJ. 1992. Escapement of fish during bottom trawl sampling—implications for resource assessment. Fish. Res. 13: 281–292. Godø OR, Walsh SJ and Engås A. 1999. Investigating density-dependent catchability in bottom trawl surveys. ICES J. Mar. Sci. 56: 292–298. Golenchenko AP. 1955. Speech. In: Proceedings of the Conference on Fish Behavior and Searching. pp 53–54. Ichthyological Commission, 5, Academy of Science, USSR, Moscow. 236 pp (in Russian).
97
Graham N. 2006. Trawling: historic development, current status and future challenges. Mar. Technol. Soc. J. 40: 20–24. Graham N, Jones EG and Reid DG. 2004. Review of technological advances for the study of fish behavior in relation to demersal fishing trawls. ICES J. Mar. Sci. 61: 1036–1043. Graham N, Kynoch RJ and Fryer RJ. 2003. Square mesh panels in demersal trawls: further data relating haddock and whiting selectivity to panel position. Fish. Res. 62: 361–375. Grimaldo E, Larsen RB and Holst R. 2007. Exit windows as an alternative selective system for the Barents Sea demersal fishery for cod and haddock. Fish. Res. 85: 295–305. Hall SJ, Wardle CS and MacLennan DN. 1986. Predator evasion in a fish school: test of a model for the fountain effect. Mar. Biol. 91: 143–148. Handegard NO, Michalsen K and Tjøstheim D. 2003. Avoidance behavior in cod (Gadus morhua) to a bottom-trawling vessel. Aquat. Liv. Resour. 16: 265–270. Handegard NO and Tjøstheim D. 2005. When fish meet a trawling vessel: examining the behavior of gadoids using a free-floating buoy and acoustic split-beam tracking. Can. J. Fish. Aquat. Sci. 62: 2409–2422. Hannah WR, Jones SA and Matteson KM. 2003. Observations of fish and shrimp behavior in ocean shrimp (Pandalus jordani) trawls. Newport, OR: OR Dept. Fish. Wildl. Mar. Res. Prog. 28 pp. Harden-Jones FR. 1963. The reaction of fish to moving backgrounds. J. Exp. Biol. 40: 437–446. Harden-Jones FR, Margetts AR, Greer Walker G and Arnold GP. 1977. The efficiency of the Granton otter trawl determined by sector-scanning sonar and acoustic transponding tags. Rapp. P.-v. Réun. Const. Int. Explor. Mer. 170: 45–51. Hawkins AD and Sand O. 1977. Directional hearing in the median vertical plane by cod. J. Comp. Physiol. 122: 1–8. He P. 1991. Swimming endurance of Atlantic cod, Gadus morhua L., at low temperatures. Fish. Res. 12: 65–73. He P. 1993. Swimming speeds of marine fish in relation to fishing gears. ICES Mar. Sci. Symp. 196: 183–189. He P and Wardle CS. 1988. Endurance at intermediate swimming speeds of Atlantic mackerel, Scomber scombrus L., herring, Clupea harengus L., and saithe, Pollachius virens L. J. Fish Biol. 35: 255–266.
98
Fish Behavior near Fishing Gears during Capture Processes
He P, Smith T and Bouchard C. 2008. Fish behavior and species separation for the Gulf of Maine multispecies trawls. J. Ocean Technol. 3(2): 59–77. Hemmings CC. 1969. Observations on the behavior of fish during capture by the Danish seine net, and their relation to herding by trawl bridles. FAO Fish. Rep. 62(3): 645–655. Hemmings CC. 1973. Direct observation of the behavior of fish in relation to fishing gear. Helgoländer wiss. Meeresunters 24: 348–360. High WL. 1969. SCUBA diving, a valuable tool for the investigating the behavior of fish within the influence of fishing gear. FAO Fish. Rep. 62(2): 253–667. Hjellvik V, Michalsen K, Aglen A and Nakken O. 2003. An attempt at estimating the effective fishing height of the bottom trawl using acoustic survey recordings. ICES J. Mar. Sci. 60: 967–979. Hunter JR and Wisby WJ. 1964. Net avoidance behavior of carp and other species of fish. J. Fish. Res. Bd. Can. 21: 613–633. Ingólfsson OA and Jørgensen T. 2006. Escapement of gadoid fish beneath a commercial bottom trawl: relevance to the overall trawl selectivity. Fish. Res. 79: 303–312. Inoue Y, Matsushita Y and Arimoto T. 1993. The reaction behavior of walleye pollock (Theragra chalcogramma) in a deep/low-temperature trawl fishing ground. ICES J. Mar. Sci. 196: 77–79. Jones EG, Glass CW and Milliken HO. 2004. The reaction and behavior of fish to visual components of fishing gears and its effect on catchability in survey and commercial situations. ICES FTFB Working Group meeting, Gdynia, Poland. April 20–23, 2004. 58 pp. Jones EG, Summerbell K and O’Neil F. 2008. The influence of towing speed and fish density on the behavior of haddock in a trawl cod-end. Fish. Res. 94: 166–174. Jørgensen R, Handegard NO, Gjøsæter H and Slotte A. 2004. Possible vessel avoidance behavior of capelin in a feeding area and on a spawning ground. Fish. Res. 69: 251–261. Jørgensen T, Ingólfsson OA, Graham N and Isaksen B. 2006. Size selection of cod by rigid grids. Is anything gained compared with diamond mesh codends only? Fish. Res. 79: 337–348. Kelleher K. 2005. Discards in the world’s marine fisheries. An update. FAO Fish. Tech. Pap. 470: 131 pp. Kim Y-H and Wardle CS. 1998a. Measuring the brightness contrast of fishing gear, the visual stimulus for fish capture. Fish. Res. 34: 151–164.
Kim Y-H and Wardle CS. 1998b. Modelling the visual stimulus of towed fishing gear. Fish. Res. 34: 165–177. Kim Y-H and Wardle CS. 2003. Optomotor response and erratic response: quantitative analysis of fish reaction to towed fishing gears. Fish. Res. 60: 455–470. Kiselev ON. 1968. Visual observations of fish behavior in natural conditions. pp. 18–22. In: Fish Behavior and Fishing Techniques. Ministry of Fisheries of the USSR, Murmansk. Korotkov VK. 1970. The speed and endurance of fishes escaping from a trawl. J. Ichthyol. 10: 832–836. Krag LA, Frandsen RP and Madsen N. 2008. Evaluation of a simple means to reduce discard in the Kattegat-Skagerrak Nephrops (Nephrops norvegicus) fishery: commercial testing of different codends and square-mesh panels. Fish. Res. 91: 175–186. Kvamme C and Isaksen B. 2004. Total selectivity of a commercial cod trawl with and without a grid mounted: grid and codend selectivity of north-east Arctic cod. Fish. Res. 68: 305–318. Laevastu T and Favorite F. 1988. Fishing and Stock Fluctuations. Farnham: Fishing News Books. 239 pp. Langeland MR. 2005. Escapement of Fish under a Survey Trawl: The Effect of Ground Gear Configuration. MSc thesis, University of Bergen, Norway. 77 pp. Lima SL and Dill LM. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can. J. Zool. 68: 619–640. Madsen N, Moth-Poulsen T and Lowry N. 1998. Selectivity experiments with window codends fished in the Baltic Sea cod (Gadus morhua) fishery. Fish. Res. 36: 1–14. Madsen N, Skeide R, Breen M, Krag LA, Huse I and Soldal AV. 2008. Selectivity in a trawl codend during haul-back operation: an overlooked phenomenon. Fish. Res. 91: 168–174. Main J and Sangster GI. 1981. A study of the fish capture process in a bottom trawl by direct observations from a towed underwater vehicle. Scot. Fish. Res. Rep. 23: 1–23. Main J and Sangster GI. 1983. Fish reactions to trawl gear: a study comparing light and heavy ground gear. Scot. Fish. Res. Rep. 27: 1–17. Main J and Sangster GI. 1991. A study of haddock (Melanogrammus aeglefinus (L.)) behavior in diamond and square mesh codends. Scot. Fish. Work. Pap. 19/91.
Fish Behavior near Bottom Trawls Manley GA, Popper AN and Fay RR. 2004. Evolution of the Vertebrate Auditory System. New York: Springer. 415 pp. Mann DA, Wilson DW, Song J and Popper AN. 2009. Hearing sensitivity of the walleye pollock. Trans. Am. Fish. Soc. 138: 1000–1008. Marténez M, Couture P and Guderley H. 1999. Temporal changes in tissue metabolic capacities of wild Atlantic cod Gadus morhua (L.), from Newfoundland. Fish Physiol. Biochem. 20: 181–191. Marténez M, Guderley H, Nelson JA, Webber D and Dutil J-D. 2002. Once a fast cod, always a fast cod: maintenance of performance hierarchies despite changing food availability in cod (Gadus morhua). Physiol. Biochem. Zool. 75: 90–100 Marténez M, Guderley H, Dutil J-D, Winger PD, He P and Walsh SJ. 2003. Condition, prolonged swimming performance and muscle metabolic capacities of cod Gadus morhua. J. Exp. Biol. 206: 503–511. Martyshevskii VN and Korotkov VK. 1968. Behavior of some fish in the zone of action of the trawl. In: Fish Behavior and Fishing Techniques. pp. 73–79. Murmansk: Ministry of Fisheries of the USSR. McQuinn IH and Winger PD. 2003. Tilt angle and target strength: target tracking of Atlantic cod (Gadus morhua) during trawling. ICES J. Mar. Sci. 60: 575–583. Michalsen K, Aglen A, Somerton D, Svellingen I and Øvredal JT. 1999. Quantifying the amount of fish unavailable to a bottom trawl by use of an upward looking transducer. ICES CM. 1999/J:08. 19 pp. Milliken HO and DeAlteris JT. 2004. Evaluation of a large-mesh panel to reduce the flatfish bycatch in the small-mesh bottom trawls used in the New England silver hake fishery. N. Am. J. Fish. Manag. 24: 20–32. Misund OA. 1994. Swimming behavior of fish schools in connection with capture by purse seine and pelagic trawl. In: Fernö A and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 84–106. Oxford: Fishing News Books. Misund OA, Øvredal JT and Hafsteinsson MT. 1996. Reactions of herring schools to the sound field of a survey vessel. Aquat. Liv. Resour. 9: 5–11. Mitson RB. 1993. Underwater noise radiated by research vessels. ICES Mar. Sci. Symp. 196: 147–152. Mitson RB and Knudsen HP. 2003. Causes and effects of underwater noise on fish abundance estimation. Aquat. Liv. Resour. 16: 255–263. Mohr H. 1971. Behavior patterns of different herring stocks in relation to ship and midwater trawl. In:
99
Kristjonsson M. (ed). Modern Fishing Gear of the World, Vol. 3. pp 368–371. London: Fishing News Books. Nøttestad L, Aksland M, Beltestad A, Fernö A, Johannessen A and Misund OA. 1996. Schooling dynamics of Norwegian spring spawning herring (Clupea harengus L.) in a coastal spawning area. Sarsia 80: 277–284. Nunnalle EP. 1991. An investigation of the avoidance reactions of Pacific whiting (Merluccius productus) to demersal midwater trawl gear. ICES CM 1991/B:5. Okonski S. 1969. The influence of fish behavior on the choice of the trawl net shape and size. FAO Fish. Rep. 62(2): 389–407. Olsen K. 1969. A comparison of acoustic threshold in cod with recordings of ship-noise. FAO Fish. Rep. 62(2): 431–438. Olsen K. 1971. Influence of vessel noise on behavior of herring. In: Kristjonsson M (ed). Modern Fishing Gear of the World, Vol. 3. pp 291–294. London: Fishing News Books. Olsen K, Angell J and Løvik A. 1983b. Quantitative estimations of the influence of fish behavior on acoustically determined fish abundance. FAO Fish. Rep. 300: 139–149. Olsen K, Angell J, Pettersen F and Løvik A. 1983a. Observed fish reactions to a surveying vessel with special reference to herring, cod, capelin, and polar cod. FAO Fish. Rep. 300: 131–138. Ona E and Godø OR. 1990. Fish reaction to trawling noise: the significance for trawl sampling. Rapp. P.-v. Réun. Const. Int. Explor. Mer. 189: 159–166. Ona E, Godø OR, Handegard NO, Hjellvik V, Patel R and Pedersen G. 2007. Silent research vessels are not quiet. J. Acoust. Soc. Am. 121: 145–150. O’Neill FG, McKay SJ, Ward JN, Strickland A, Kynoch RJ and Zuur AF. 2003. An investigation of the relationship between sea state induced vessel motion and cod-end selection. Fish. Res. 60: 107–130. Özbilgin H. 2002. Effect of temperature change on maximum swimming speed of whiting, Merlangius merlangus (Linnaeus, 1758). Turk. J. Zool. 26: 255–262. Özbilgin H and Glass CW. 2004. Role of learning in mesh penetration behavior of haddock (Melanogrammus aeglefinus). ICES J. Mar. Sci. 61: 1190–1194. Özbilgin H and Wardle CS. 2002. Effect of seasonal temperature changes on the escape behavior of haddock, Melanogrammus aeglefinus, from the codend. Fish. Res. 58: 323–331.
100
Fish Behavior near Fishing Gears during Capture Processes
Özbilgin H, Ferro RST, Robertson JHB, Holtrop G and Kynoch RJ. 2006. Seasonal variation in trawl codend selection of northern North Sea haddock. ICES J. Mar. Sci. 63: 737–748. Parrish BB, Blaxter JHS, Pope JA and Osborn RH. 1969. Underwater photography of fish behavior in response to trawls. FAO Fish. Rep. 62(3): 873–884. Parsons GR and Foster DG. 2007. Swimming performance and behavior of red snapper: its application to bycatch reduction. Am. Fish. Soc. Symp. 60: 59–75. Pavlov DS and Kasumyan AO. 1995. A review of Russian studies on the behavior and sensory systems of fishes. J. Ichthyol. 35(1): 124–138. Peake SJ and Farrell AP. 2004. Locomotory behavior and post-exercise physiology in relation to swimming speed, gait transition and metabolism in freeswimming smallmouth bass (Micropterus dolomieu). J. Exp. Biol. 207: 1563–1575. Peake SJ and Farrell AP. 2006. Fatigue is a behavioral response in respirometer-confined smallmouth bass. J. Fish Biol. 68: 1742–1755. Piasente M, Knuckey IA, Eayrs S and McShane PE. 2004. In situ examination of the behavior of fish in response to demersal trawl nets in an Australian trawl fishery. Mar. Freshw. Res. 55: 825–835. Popper AN. 2003. Effects of anthropogenic sounds on fishes. Fisheries 28(10): 24–31. Popper AN, Plachta DTT, Mann DA and Higgs D. 2004. Response of clupeid fish to ultrasound: a review. ICES J. Mar. Sci. 61: 1057–1061. Pyanov AI. 1993. Fish learning in response to trawl fishing. ICES Mar. Sci. Symp. 196: 12–16. Pyanov AI and Zhuykov AY. 1993. Conditioned reflex in fishes for avoidance of active fishing gear. J. Ichthyol. 33(8): 40–50. Reid DG, Allen VJ, Bova DJ, Jones EG, Kynoch RJ, Peach KJ, Fernandes PG and Turrell WR. 2007. Anglerfish catchability for swept-area abundance estimates in a new survey trawl. ICES J. Mar. Sci. 64: 1503–1511. Riedel R and DeAlteris JT. 1995. Factors affecting hydrodynamic performance of the Nordmøre grate system: a bycatch reduction device used in the Gulf of Maine shrimp fishery. Fish. Res. 24: 181–198. Rose CS. 1995. Behavior of North Pacific groundfish encountering trawls: applications to reduce bycatch. pp. 235–241. In: Solving Bycatch: Considerations for Today, and Tomorrow. Alaska Sea Grant College Program Report No. 96–03, University of Alaska Fairbanks.
Ryer CH. 2008. A review of flatfish behavior relative to trawls. Fish. Res. 90: 138–146. Ryer CH and Barnett LAK. 2006. Influence of illumination and temperature upon flatfish reactivity and herding behavior: potential implications for trawl capture efficiency. Fish. Res. 81: 242–250. Sand O and Karlsen HE. 1986. Detection of infrasound by the Atlantic cod. J. Exp. Biol. 125: 197–204. Schuijf A. 1975. Directional hearing of cod (Gadus morhua) under approximate free field conditions. J. Comp. Physiol. 98: 307–332. Skaret G, Axelsen BE, Nøttestad L, Fernö A and Johannessen A. 2005. The behavior of spawning herring in relation to a survey vessel. ICES J. Mar. Sci. 62: 1061–1064. Skaret G, Slotte A, Handegard NO, Axelsen BE and Jørgensen R. 2006. Pre-spawning herring in a protected area showed only moderate reaction to a surveying vessel. Fish. Res. 78: 359–367. Somerton DA. 2004. Do Pacific cod (Gadus macrocephalus) and walleye pollock (Theregra chalcogramma) lack a herding response to the doors, bridles, and mudclouds of survey trawls? ICES J. Mar. Sci. 61: 1186–1189. Soria M, Fréon P and Gerlotto F. 1996. Analysis of vessel influence on spatial behavior of fish schools using a multi-beam sonar and consequences for biomass estimates by echo-sounder. ICES J. Mar. Sci. 53: 453–458. Soria M, Gerlotto F and Fréon P. 1993. Study of learning capabilities of tropical clupeoids using an artificial stimulus. ICES Mar. Sci. Symp. 196: 17–20. Stankowich T and Blumstein DT. 2005. Fear in animals: a meta-analysis and review of risk assessment. Proc. R. Soc. (B) 272: 2627–2634. Strange ES. 1984. Review of the fishing trials with Granton and Saro deep sea trawl gear 1963–1967. Scot. Fish. Work. Pap. 8/84. Tanaka E, Matuda K and Hirayama N. 1991. A simulation model of gear efficiencies of trawlers for flatfish. Nippon Suisan Gakkaishi 57: 1019–1028. Thomsen B. 1993. Selective flatfish trawling. ICES Mar. Sci. Symp. 196: 161–164. Urquhart GG and Stewart PAM. 1993. A review of techniques for the observation of fish behavior in the sea. ICES Mar. Sci. Symp. 196: 135–139. Vabø R, Olsen K and Huse I. 2002. The effect of vessel avoidance of wintering Norwegian spring spawning herring. Fish. Res. 58: 59–77. Videler JJ. 1993. Fish Swimming. London: Chapman and Hall. 260 p.
Fish Behavior near Bottom Trawls Videler JJ and Wardle CS. 1991. Fish swimming stride by stride: speed limits and endurance. Rev. Fish Biol. Fish. 1: 23–40. Voegeli FA, Smale MJ, Webber DM, Andrade Y and O’Dor RK. 2001. Ultrasonic telemetry, tracking and automated technology for sharks. Environ. Biol. Fish. 60: 267–281. Wakeford J. 2004. An investigation into the influence of the Super-Shooted TED and several types of bycatch reduction devices on water flow through a prawn trawl codend. In: Brewer DT, Heales DS, Eayrs SJ, Taylor BR, Day G, Sen S, et al. (eds). Assessment and improvement of TEDs and BRDs in the NPF: a co-operative approach between fishers, scientists, fisheries technologists, economists and conservationists. Final Report on FRDC Project 2000/173. CSIRO Cleveland. 412 pp. Walsh SJ. 1992. Size-dependent selection at the footgear of a groundfish survey trawl. N. Am. J. Fish. Manag. 12: 625–633. Walsh SJ. 1996. Efficiency of bottom sampling trawls in deriving survey abundance indices. NAFO Sci. Coun. Stud. 28: 9–24. Walsh SJ and Hickey WM. 1993. Behavioral reactions of demersal fish to bottom trawls at various light conditions. ICES Mar. Sci. Symp. 196: 68–76. Walsh SJ, Engås A, Ferro R, Fonteyne R and van Marlen B. 2002. To catch or conserve more fish: the evolution of fishing technology in fisheries science. ICES Mar. Sci. Symp. 215: 493–503. Wardle CS. 1983. Fish reaction to towed fishing gears. In: MacDonald AG and Priede IG (eds). Experimental Biology at Sea. pp 167–196. London: Academic Press. Wardle CS. 1984. Fish behavior. Trawl efficiency and energy saving strategies. FIIT Working Paper. Rome: FAO Fishing Technology Service. Sept. 1984. Wardle CS. 1986. Fish behavior and fishing gear. In: Pitcher TJ (ed). Behavior of Teleost Fishes. pp 463–495. London: Chapman and Hall. Wardle CS. 1992. Studies of fish behavior leading to a reduction of bycatch. In: Jones RP (ed). Proceedings of an International Conference on Shrimp Bycatch. May 24–27, 1992, Lake Buena Vista, Florida. pp 163–184. Tallahassee, FL: Southeastern Fisheries Association. Wardle CS. 1993. Fish behavior and fishing gear. In: Pitcher TJ (ed). Behavior of Teleost Fishes. 2nd ed. pp 609–643. London: Chapman and Hall. Watson JW. 1988. Fish behavior and trawl design: Potential for selective trawl development. Proc.
101
World Symp. Fish. Gear and Fish. Vessel Design. pp 25–29. St. John’s, Newfoundland: Marine Institute. Watson JW. 1992. Status of knowledge in the United States relating fish behavior to the reduction of bycatch. In: Jones RP (ed). Proceedings of an International Conference on Shrimp Bycatch. May 24–27, 1992, Lake Buena Vista, Florida. pp 185– 196. Tallahassee, FL: Southeastern Fisheries Association. Watson JW. 2007. Reconciling fisheries with conservation through programs to develop improved fishing technologies in the Unites States. In: Kennelly SJ (ed). By-catch Reduction in the World’s Fisheries. pp 23–36. The Netherlands: Springer. Watson JW, Foster D, Nichols S, Shah A, ScottDenton E and Nance J. 1999. The development of bycatch reduction technology in the Southeastern United States shrimp fishery. Mar. Technol. Soc. J. 33(2): 51–56. Webb PW. 1994a. Exercise performance of fish. In: Jones JH (ed). Comparative Vertebrate Exercise Physiology: Phyletic Adaptations. pp 1–49. San Diego: Academic Press. Webb PW. 1994b. The biology of fish swimming. In: Maddock L, Bone Q and Rayner JMV (eds). Mechanics and Physiology of Animal Swimming. pp 45–62. Cambridge: Cambridge University Press. Weinberg KL and Munro PT. 1999. The effect of artificial light on escapement beneath a survey trawl. ICES J. Mar. Sci. 56: 266–274. Wileman DA, Ferro RST, Fontaine R and Millar RB. 1996. Manual of methods of measuring the selectivity of towed fishing gears. ICES Coop. Res. Rep. 215: 126 pp. Winger PD. 2004. Effect of Environmental Conditions on the Natural Activity Rhythms and Bottom Trawl Catchability of Atlantic Cod (Gadus morhua). PhD thesis, Memorial University of Newfoundland. 151 pp. Winger PD, DeLouche H and Legge G. 2006. Designing and testing new fishing gears: the value of a flume tank. Mar. Technol. Soc. J. 40(3): 44–49. Winger PD, He P and Walsh SJ. 1999. Swimming endurance of American plaice (Hippoglossoides platessoides) and its role in fish capture. ICES J. Mar. Sci. 56: 252–265. Winger PD, He P and Walsh SJ. 2000. Factors affecting the swimming endurance and catchability of Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 57: 1200–1207.
102
Fish Behavior near Fishing Gears during Capture Processes
Winger PD, Walsh SJ, He P and Brown JA. 2004. Simulating trawl herding in flatfish: the role of fish length on behavior and swimming characteristics. ICES J. Mar. Sci. 61: 1179–1185. Yanase K, Eayrs S and Arimoto T. 2007. Influence of water temperature and fish length on the maximum swimming speed of sand flathead, Platycephalus bassensis: implications for trawl selectivity. Fish. Res. 84: 180–188. Yanase K, Eayrs S and Arimoto T. 2009. Quantitative analysis of the behavior of the flatheads (Platycephalidae) during the trawl capture process
as determined by real-time multiple observations. Fish. Res. 95: 28–39. Ydenberg RC and Dill LM. 1986. The economics of fleeing from predators. Adv. Study Behav. 16: 229–249. Zaferman ML and Serebrov LI. 1989. On fish injuring when escaping through the trawl mesh. ICES CM.1989/B:18. 14 pp. Zhuykov AY and Pyanov AI. 1993. Differences in behavior of fish with different learning ability as demonstrated with a model of a trap net. J. Ichthyol. 33(9): 141–147.
Fish Behavior near Bottom Trawls SPECIES MENTIONED IN THE TEXT American shad, Alosa sapidissima Atlantic bumper, Chlorscombrus chrysurus Atlantic cod, cod, Gadus morhua Atlantic mackerel, Scomber scombrus Atlantic salmon, Salmo salar anchovies, Ancoa hepstus blue grenadier, Macruronus novaezelandiae capelin, Mallotus villosus Müll. common dab, Limanda limanda dogfish, Sqaulus acanthias goldfish, Carassius auratus auratus Greenland halibut, Reinhardtius hippoglossoides haddock, Melanogrammus aeglefinus, herring, Clupea harengus jack mackerel, Trachurus japonicus ling, Molva molva New Zealand dory Cyttus novaezelandiae
103
North Sea plaice, Pleuronectes platessa, North Sea pollack, pollock, Pollachius pollachius, Pacific cod, Gadus macrocephalus Pacific halibut, Hippoglossus stenolepis red snapper, Lutjanus campechanus rock sole, Pleuronectes bilineatus saithe, Pollachius virens sand flathead, Platycephalus bassensis sandeels, Ammodytidae sp. skates, Raja sp. Spanish sardine, Sardinella anchovia spotted warehou, Seriolella punctata squid, Loligo pealeii Tiger flathead, Neoplatycephalus richardsoni whiting, Merlangius merlangus walleye pollock, Theragra chalcogramma yellowtail flounder, Limanda ferruginea
Chapter 5 Fish Behavior in Relation to Longlines Svein Løkkeborg, Anders Fernö, and Odd-Børre Humborstad 5.1 INTRODUCTION Fishing by hook and line, including handlining, rod and reel fishing, jigging, trolling, and longlining, is a method used all over the world with a very wide range of vessels, from small artisanal boats to large mechanized longliners. This traditional fishing method has been one of the most important fish capture techniques used since the Stone Age (Bjordal and Løkkeborg 1996). Archaeological excavations show that various materials such as stone, bone, sea shells, and horn have been used for hooks throughout history, and some of these hooks, used on a handline, are believed to be more than 4000 years old. Longlining is a more recent development that was not used on a large scale until industrialization made large numbers of hooks available at a reasonable cost. The use of longlines in Norwegian waters can be dated back to the early 1700s, and Spanish sources indicate that the use of longline gear in the Mediterranean dates back to at least the 1500s (Bjordal and Løkkeborg 1996). A longline is a passive and stationary form of fishing gear. The principle of longline fishing is to attract fish to ingest hooks by using odor-releasing bait that entices the fish to ingest the hook-bait combination. Although longline gear is thus a simple fishing method, there are wide variations in gear construction and mode of operation. Moreover, as the gear can be set either on the bottom or drifting in the water column, a large variety of species can be targeted, from bottom-dwelling flatfish to highly migratory tunas. Longline fishing is therefore a very versatile fishing method.
Understanding the principle behind longline fishing requires knowledge of how fish search for and capture food. The literature on feeding behavior and interactions between fish and longline gear parameters has been synthesized in earlier reviews (Bjordal and Løkkeborg 1996; Løkkeborg 1994). In the course of the past decade, however, a number of behavioral studies have provided new knowledge of great relevance to the performance of longlining, especially regarding food search, feeding motivation, and effects of environmental variables. This literature is reviewed here with the aim of providing a more comprehensive understanding of the interactions between foraging fish and baited hooks. Until recently, concerns about sustainable exploitation of marine resources have focused mainly on the proper management of commercially valuable fish stocks, and longlining has been considered as an environmentally friendly fishing method. However, fishing causes incidental mortality of many nontarget species, and of particular concern are long-lived species such as seabirds, sea turtles, elasmobranches, and marine mammals (Heppell et al. 2005). Thus, gradually more attention is being paid to the general impact of fishing activities on the marine ecosystem (Gislason, 1994). With regard to longline fishing, most attention has been paid to bycatch of seabirds and the development of mitigation measures to solve this problem. The failure of sea turtle populations to recover (six of the seven species are listed as endangered) has also been attributed in part to incidental capture by fishing gear, including longlines. A third important
105
106
Fish Behavior near Fishing Gears during Capture Processes
conservation issue related to longline fishing is shark management [see the U.N. World Food and Agriculture Organization (FAO)’s International Plan of Action for the Conservation and Management of Sharks]. With regard to the principles of ecologically sustainable fishing, this chapter thus focuses on interactions with seabirds, sea turtles, and sharks and discusses how our current knowledge of behavior and gear operation may help to mitigate these impacts. 5.2 WORLDWIDE LONGLINE FISHERIES Baited longlines are used in all oceans and seas, and large proportions (15% to 90%) of several important fish resources are caught by longlines (Bjordal and Løkkeborg 1996). Demersal (bottom) longline fishing takes place on the relatively shallow waters of the continental shelves and slopes down to depths of 3000 m. Many demersal longline fishing operations are carried out in cold waters at relatively high latitudes and target a large variety of fish species. Pelagic (drifting) longlining takes place in deep waters, generally off continental shelves. This fishing method is operated in all oceans from temperate to tropical waters and targets mainly tuna and swordfish (billfish) species. The most important longline fisheries in terms of tonnage of landed catches are described below and summarized in Table 5.1. More detailed information on effort and landings has been provided by Bjordal and Løkkeborg (1996) and Brothers et al. (1999a). 5.2.1 Northeastern Atlantic Ocean Demersal Fisheries Nearly all longline fisheries in the northeast Atlantic region target demersal species. The Norwegian, Icelandic, and Faeroese fleets dominate the longline fishery in this region. The vessels move seasonally between fishing grounds targeting different species, mainly Atlantic cod (Gadus morhua), haddock (Melanogrammus aeglefinus), tusk (Brosme brosme), ling (Molva molva), wolffish (Anarhichas lupus), and Greenland halibut (Reinhardtius hippoglossoides). The longline fleets operate on the shelf and shelf edge on both coastal and high-sea
Table 5.1. Regions and Main Target Species for the World’s Most Important (in Terms of Landings) Longline Fisheries. Region Demersal longlining: Northeastern Atlantic Northwestern Atlantic Northeastern Pacific Northwestern Pacific Southern Ocean Pelagic longlining: Pacific Ocean Atlantic Ocean Indian Ocean
Main Target Species
Cod, haddock, tusk, ling Cod, haddock, spiny dogfish, white hake Cod, halibut, sablefish Walleye pollock, cod Patagonian toothfish Bigeye tuna, yellowfin tuna, albacore Bigeye tuna, albacore, swordfish Yellowfin tuna, bigeye tuna
fishing grounds from west of the British Isles to the Barents Sea. Most coastal vessels are relatively small, and their longlines are baited onshore by hand and coiled into tubs (baskets). Larger vessels using autoline systems operate farther offshore and on high seas fishing grounds. These vessels remain at sea for several weeks and may set more than 30,000 hooks per day. 5.2.2 Northwestern Atlantic Ocean Demersal Fisheries Demersal longlining in this region is carried out by Canadian vessels fishing off Nova Scotia, Newfoundland, and Labrador and in the Gulf of St. Lawrence and by U.S. vessels fishing in the Gulf of Maine and on the Georges Bank. Groundfish species caught include Atlantic cod, haddock, spiny dogfish (Squalus acanthias), Greenland halibut, tusk (cusk), American plaice (Hippoglossoides platessoides), saithe (Pollachius virens), and white hake (Urophycis tenuis). In 1992 a moratorium on groundfish (primarily for cod) came into force in
Fish Behavior in Relation to Longlines Canadian Atlantic waters, but in 1997 a limited longline fishery for cod recommenced around Newfoundland. The Canadian and U.S. longline fleets are dominated by small vessels (less than 45 feet) that bait their lines onshore by hand. The larger vessels (which are small compared with Norwegian and Icelandic vessels) are not mechanized but may remain at sea for several days, baiting their lines by hand.
5.2.3 Northeastern Pacific Ocean Demersal Fisheries This region is divided into three areas—Alaska (Gulf of Alaska and Bering Sea), Canada (British Colombia), and Washington-Oregon-California— and the longline fleet consists entirely of domestic vessels from the United States and Canada. Fishing activity is highest in the Gulf of Alaska and the Bering Sea, where the most important fishing grounds are located. Landings from this area are dominated by Pacific cod (Gadus macrocephalus), Pacific halibut (Hippoglossus stenolepis), and sablefish (blackcod) (Anoplopoma fimbria). Alaskan fisheries account for all of the cod landings and approximately 80% of halibut landings from the northeastern Pacific region. The vessels range in size from relatively small boats to large factory longliners using mechanized longline systems. The Pacific halibut fishery in Alaska was an open access fishery until 1995, when an individual vessel quota system was put into effect. As a result, the number of vessels fell, and the active fishing period changed from a few days to several months.
5.2.4 Northwestern Pacific Ocean Demersal Fisheries Demersal longlining in the northwest Pacific region is carried out by vessels from Japan, South Korea, China, and Taiwan. The vessels operate on both coastal and high-sea fishing grounds, and both demersal longlines and longlines floated off the seabed are used. The longline fleet is dominated by small boats that operate on the coastal fishing grounds. The most important species in this region
107
are walleye pollock (Theragra chalcogrammus), Pacific cod, and Pacific halibut. 5.2.5 Southern Ocean Demersal Fisheries Demersal longline fishing in the Southern Ocean commenced in 1988 and is primarily targeting Patagonian toothfish (Dissostichus eleginoides). This valuable species has a circumpolar distribution south of 55 degrees S, and longline fishing is carried out off the southern part of South America (Argentina, Falkland Islands, and Chile) and around many of the islands of the Southern Ocean, especially those of southern Atlantic and Indian Oceans (e.g., South Georgia, Prince Edward, Crozet, and Kerguelen Islands). The fishery is carried out in very deep waters (down to 3000 m) by large mechanized vessels from several countries. Annual regional quotas are set by the Commission for the Conservation on the Antarctic Marine Living Resources (CCAMLR), but illegal, unreported, and unregulated (IUU) fishing is a problem. 5.2.6 Pelagic Longline Fisheries A large fleet of tuna vessels operates with pelagic longlines in the Indian Ocean and in the central and southern Atlantic and Pacific Oceans. The home ports of most of these vessels are in Japan, Taiwan, and South Korea, which are the major operators in the tuna longline fisheries and land the great majority of the catches. The most important species by weight is bigeye tuna (Thunnus obesus), followed by yellowfin tuna (T. albacares) and albacore (T. alalunga). Catches of bluefin tuna (T. thynnus) are relatively small, but the prices for this species on the Japanese sashimi (raw fish) market are much higher than for the other species. There is also an important pelagic longline fishery targeting swordfish, but this fishery is much smaller than the tuna fisheries (about 10%). The Atlantic Ocean is the most important region for the swordfish longline fishery, and the European and North American fleets are the dominant operators. The tuna longline fisheries have changed over time, driven by economic and market forces. Traditional tuna longlines were set at a maximum depth of about 170 m, and the Japanese longline fleet targeted mainly yellowfin tuna and albacore to supply the U.S. canning industry. In the 1970s,
108
Fish Behavior near Fishing Gears during Capture Processes
a “deep longline” was developed to target bigeye tuna, which is distributed deeper in the water column (more than 200 m) than yellowfin tuna and albacore (less than 200 m). Bigeye tuna fetch a higher price on the sashimi market, and the fleet gradually switched to target this species. The substantial increase in fuel and labor costs in the 1970s forced the Japanese fleet to reduce their export to the low-value canning market and concentrate on high-value products for the sashimi market in Japan. Later the same market-driven shift was seen in the South Korean tuna fleet. 5.3 DESCRIPTION OF THE GEAR Longlining is a versatile fishing method, and there are numerous ways of rigging longlines according to target species and fishing conditions (e.g., bottom type and topography, depth, current). However, all longline gear used in the worldwide fishery is based on the same basic unit, which consists of four parts (Fig. 5.1): • The mainline (groundline); a longline (rope), which has given the gear its name and to which the snood and hook are attached at intervals • The snood (gangion, branchline); a thinner line attached to the main line at one end and to the hook at the other • The hook (large varieties in shape and size) • The bait (cut in pieces or whole finfish, shellfish, or squid/octopus) The unit in practical rigging and use of longlines, however, is the basket, tub, skate, or, in mechanized longlining, magazine. A certain length of mainline with a given number of hooks is baited and coiled into a tub (basket) prior to setting the longline. In mechanized longlining, this unit is replaced by a magazine holding the hooks with the mainline hanging in coils. During setting, several tubs/ baskets or magazines are linked to make a fleet of longlines, which may vary in length from a few hundred meters in small-scale coastal fishing to more than 50 km in high seas mechanized longline fisheries. The mainline and snood vary in length, dimension, and type of material according to the type of fishery. Variations are also found in spacing
Figure 5.1. The basic unit of longline gear consists of four parts: mainline, snood, hook, and bait.
between neighboring snoods/hooks (hook spacing) and in the way in which the snood is attached to the mainline. The widest variations in longline gear characteristics are seen in hook types, in terms of both shape and size. The type of bait used varies according to the food preference of the target
Fish Behavior in Relation to Longlines species, availability, and price, and different baits may be used on the same longline in multispecies fisheries. The most common materials used for mainline and snood are polyamide (nylon) and polyester, both of which have a higher density than seawater and thus sink. More recently, gear manufacturers have developed ropes made from a mixture of different types of materials to obtain longlines with improved properties—in particular, higher breaking strength. Multifilament and monofilament lines have different properties and are used accordingly. Due to their high resistance to chafing, multifilament lines are commonly used for demersal longlines, but monofilament lines may be used for snoods. Monofilament lines are preferred in pelagic and semipelagic longlining because they have higher catching efficiency, probably because they are less visible. The length of the snood varies from relatively short (less than 1 m) in demersal longlining to very long in tuna longlining (up to 30–40 m). The snood may be attached to the mainline in three different ways. It has traditionally been tied to the mainline by a simple knot. More recently, swivels have been used, first on monofilament longlines and later also on demersal multifilament longlines. Swivels prevent twisting of the snood and have been shown to improve catching efficiency. The third way of attachment is by a metal clip that holds the snood and is manually attached to the mainline as it is deployed. This method is commonly used in pelagic tuna longlining and in the fishery for Pacific halibut. With this method, the hook spacing needs to be relatively wide, as the snood is attached during the setting operation. There are wide variations in the types of hook used in longline fisheries. The J-shaped hook has traditionally been most common and entirely dominated the fisheries for groundfish until the mid1980s. During the 1980s, several new hook designs were developed and tested in comparative fishing experiments. These hooks yielded substantial increases in catch rates for many demersal species compared with traditional J-hooks, and marked shifts in hook types were seen in several longline fisheries. For examples, an almost total conversion from J-hooks to circle and EZ-baiter hooks occurred,
109
respectively, in the Pacific halibut and Northeast Atlantic autoline fisheries. Furthermore, using circle hooks instead of traditional J-hooks has proved to be an efficient mitigation measure to reduce incidental bycatch of sea turtles in pelagic longlining (see Section 5.3). Bait type is an important factor affecting the species selectivity of longlining. This is explained by the fact that feeding attractants and stimulants are species specific. There are thus large variations in bait types among different longline fisheries, and fishermen have learned from experience which bait types are most effective for a particular species. Price and availability are other factors that determine the fishermen’s choice of bait type. The bait specimens are usually cut into pieces in demersal longlining, whereas whole specimens are used to target large pelagic species such as tunas. Much effort has been put into the development of artificial baits, but to date no alternative bait has been widely adopted by commercial vessels. There are three main methods of setting longlines, with a wide range of variation in each. Bottom-set or demersal longlines are the traditional and most common setting method, in which the baited lines are laid on the seabed. Demersal longlines are set either as a single mainline or using the double line system (also called the Spanish system). The latter system has a second safety or mother line. This line is unweighted and floats above the mainline to which it is attached at several points. The double-line system is designed to be used in areas with rough bottoms and strong currents and allows the mainline to be retrieved even if it breaks. Bottom-set longlines are used to target demersal species such as cod, halibut, hake (Merluccius spp.), ling, and Patagonian toothfish. Pelagic longlines are suspended from floats at the surface and drift freely in the sea as they are not anchored to the bottom. This setting method is most frequently used in high-seas fisheries targeting tunas, swordfish, and sharks. The fishing depth of the baited hooks is adjusted by varying the length of the float lines according to the depth inhabited by the target species. Surface longlines targeting swordfish may be set very shallow (10–20 m), whereas pelagic lines targeting big eye tuna are set at a few hundred meters depth.
110
Fish Behavior near Fishing Gears during Capture Processes
The third longline setting method, semipelagic longlining, is an intermediate between demersal and pelagic longlining. The mainline is anchored to the seabed but floated off the bottom. Usually, there are floats and sinkers alternating at specific intervals along the mainline. The sinkers are attached to the mainline by a vertical line, and the length of this line is adjusted to ensure that the baited hooks are suspended at the depth inhabited by the target species. Semipelagic longlining is used for hake and other demersal species that seasonally display semipelagic distribution, such as cod, Pollock, and haddock. 5.4 CHEMORECEPTION AND FOOD SEARCH—THE BASIS FOR BAIT FISHING 5.4.1 The Chemosensory Senses and Chemical Stimuli Properties Fishing with bait exploits the feeding behavior of fish and is based on one of the most fundamental activities in an animal’s life—searching for and capturing food (Løkkeborg 1994). Thus, the effectiveness of baited fishing gear ultimately depends on the behavior of the target species and their activity rhythms, feeding motivation, and locomotory abilities (Stoner 2003). Studying the behavior of foraging fish may therefore assist in better understanding the interactions between the target species and the baited gear. However, vulnerability to baited gears varies because changes in activity, feeding, and locomotion are all affected by environmental variables such as temperature, light level, current velocity, ambient prey density, and the presence of potential bait competitors (Stoner 2004). While many aspects of foraging behavior, such as olfaction and taste, feeding stimulants and thresholds, activity rhythms, feeding motivation, foodsearch strategies, and ability to locate prey, have been well studied in some species (e.g., Caprio 1978; Carr 1982; Hara 1992; Ishida and Kobayashi 1992; Johnstone 1980; Løkkeborg et al. 1989; Løkkeborg and Fernö 1999; Stoner 2003; Sutterlin 1975), research on how feeding biology is affected by environmental variables is still in its infancy (Stoner 2004). A starting point for explaining the feeding behavior in fish is optimal foraging theory (Krebs 1978), which assumes that feeding behavior
is determined by the balance between costs and benefits. Fish detect chemical stimuli through at least two different channels of chemoreception: olfaction (smell) and gustation (taste) (Hara 1992, 1993). The olfactory organ in fish is basically paired structures situated in the snout, each of which has an anterior inlet and posterior outlet for a current of water. The receptor cells of the olfactory epithelium or mucosa are located in the floor of the nasal cavity, where they are arranged in folds or lamella to form a rosette (Hara 1992, 1993; Laberge and Hara 2001). Neuron receptor cells carry external information directly through cranial nerve I (olfactory) to the olfactory bulb of the brain, where they terminate and make synaptic contact with second-order bulbar neurons in the form of glomeruli. Onion- or pearshaped taste buds make up the structural basis for the gustatory organ and consist of gustatory microvilliated receptor and supporting and basal cells. Dependent on the species, taste buds can be found on barbels, fins, the oral cavity, pharynx, and gills and over the entire body surface. Chemical information detected by epithelial receptor cells is transmitted by neurons of cranial nerve VI (facial), IX (glossopharyngeal), or X (vagus) to the central nervous system (Marui and Caprio 1992). The olfactory and gustatory organs of fishes exhibit considerable diversity, reflecting adaptations to different environmental conditions and ecological strategies. The vital function, however, is the same—to extract information about the chemical environment, to control behavior. Molecules dissolved in water mediate both senses, and it can be difficult to identify the specific role that each system plays in any particular behavior. However, there are distinct differences in the anatomy of the olfactory and gustatory organs in fish, and these distinctions are also evident in the functional characteristics of these chemosensory systems (Kasumyan and Døving 2003). The gustatory system tends to be more selective than the olfactory system (i.e., a chemical substance that induces feeding behavior need not possess high palatability), and the gustatory spectra of effective taste substances are more species specific than the olfactory spectra. Because of its lower detection thresholds (Ishida and Kobayashi 1992; Johnstone
Fish Behavior in Relation to Longlines 1980), it is believed that the olfactory system initiates the feeding process (Hara 1993), whereas the gustatory system provides the final sensory evaluation of potential food items (Kasumyan and Døving 2003). Nearly all fish use olfaction for distant prey detection, and fish have the capability of locating prey from a distance well beyond their visual range by means of their chemosensory system (Atema 1980). Chemical stimuli are different from visual and acoustic stimuli with regard to two properties that are crucial for their use in attraction of fish to fishing gear. First, a chemical stimulus has long range and can be detected from very long distances (several hundred meters; Løkkeborg 1998), whereas visual and acoustic (except low frequency sounds) stimuli attenuate rapidly. Second, a chemical stimulus lasts for a long period of time (several hours), whereas the two other types of stimuli fade rapidly after being transmitted. These two properties form the basis of a capture principle that has a long range in both time and space and thus make baits and other types of chemical attractants unique when used to attract target species to fishing gears. The spatial range of chemical attractants means that baits can attract fish from a long distance. The area in which a fish will detect and respond to attractants released from an odor source is called the active space (Bossert and Wilson 1963; Elkinton et al. 1984; McQuinn et al. 1988). Because the rate of diffusion in water is very low, the current is the most important agent for dispersion of chemicals in seawater. Turbulence dilutes attractants released from baits, and the concentration tends to decrease with increasing distance from the odor source (but see Section 5.5.3). The distance over which an odor source may attract fish is thus determined by the initial release rate of feeding attractants, the rate of dilution (i.e., turbulence), and the chemosensory capability and response thresholds to feeding attractants of the target fish. The temporal range of an odor source is determined by the change over time in release rate. The rate of release of feeding attractants from traditional mackerel bait decreases rapidly within the first 1.5 h of immersion in flowing seawater and after 2 to 3 h is only one third of the initial rate (Løkkeborg 1990a). Due to this decrease in release rate over
111
time, the length of active space has been estimated to be halved at 24 h compared with 1 h (Løkkeborg et al. 1995). Thus, the development of a system that extends the period of time over which the attractants are released at high concentrations (i.e., longlasting baits) has great potential in the longline fishery. 5.4.2 Mechanisms to Locate an Olfactory Source Aquatic animals that are chemically stimulated need to move upstream to locate an odor source. Olfactory arousal is therefore often followed by upstream swimming toward the chemical source (Atema 1980). There are two possible mechanisms that can be used to locate an odor source. During rheotaxis, the chemical stimulation releases upstream swimming. Although orientation via rheotaxis is not very accurate (Atema 1980), the animal will gradually approach the source. During a gradient search, the animal reacts to a concentration gradient. Given that the concentration increases as the distance to the odor source is reduced and the animal is able to detect the differences in concentration and orient in relation to these, searching based on gradients should be more accurate. However, the gradients are often weak and inconsistent as turbulence can break up the regular pattern, making it very difficult for animals to locate an odor plume by equalizing the sensory input from bilateral olfactory receptors (Webster et al. 2001, Webster and Weissburg 2001). Fish might alternatively sequentially compare concentrations at different points of space based on time-averaged sampling, but the large spatial and temporal variations in concentrations within chemical odor plumes may make it difficult to detect concentration differences even after long periods of sampling (Webster and Weissburg 2001).Chemically stimulated rheotaxis is therefore regarded as the most likely mechanism used by fish and crustaceans to locate odor sources. 5.4.3 Feeding Attractants Many studies have been carried out to identify the chemical nature of feeding attractants for fish. Amino acids comprise the most important group of compounds identified as feeding attractants and stimulants for fish and crustaceans (see reviews by
112
Fish Behavior near Fishing Gears during Capture Processes
Atema 1980; Carr 1982; Carr et al. 1996; Hara 1975; Jones 1992; Mackie 1982). Nearly all studies of chemically stimulated feeding behavior have shown that food extracts lose their stimulatory capacities when their amino acids have been eliminated (Carr and Derby 1986). Ishida and Kobayashi (1992) concluded that the simple and neutral amino acids alanine and glycine play a role as feeding attractants in several fish species. Of a variety of compounds examined electrophysiologically, amino acids also stand out as a highly stimulatory class of compounds, and this sensitivity to amino acids in fish may have evolved as a food-finding mechanism (Ishida and Kobayashi 1992; Sutterlin 1975). However, other groups of compounds also elicit food search. Carr and Derby (1986) state that the major stimulants in tissue extracts of preys that elicit feeding behavior in marine fish and crustaceans are “common metabolites of low molecular weight that include amino acids, quaternary ammonium compounds, nucleosides and nucleotides, and organic acids.” Polyamines have been shown to be potent olfactory stimulants for goldfish (Carassius auratus), and behavioral assays have indicated that polyamines elicit feeding behavior similar to that elicited by amino acids (Rolen et al. 2003). Interestingly, behavioral observations showed that juvenile sablefish responded to squid odor at dilutions of amino acids far below ambient concentrations, indicating that substances other than amino acids are the primary feeding stimulants in this species (Davis et al. 2006). Studies of the role of specific chemicals as feeding stimulants have demonstrated that a mixture of substances acting in concert rather than a single dominant substance is required to yield extracts with stimulatory capacities similar to those of the total tissue extracts of preys (Carr and Derby 1986; Ellingsen and Døving 1986; Johnstone and Mackie 1990; Jones 1992). Furthermore, reviews of the characteristics of substances that act as feeding stimulants have revealed that different species respond to different substances in food extracts (e.g., Carr and Derby 1986; Mackie 1982). Studies using the same squid extract showed that the major stimulants of feeding behavior were different for turbot (Scophthalmus maximus), rainbow trout
(Oncorhynchus mykiss), and plaice (Pleuronectes platessa) (Adron and Mackie 1978; Mackie 1982; Mackie and Adron 1978). Carr (1982) compared the stimulatory capacity of extracts from different organisms and found that their relative effectiveness differed in pinfish (Lagodon rhomboides) and pigfish (Orthopristis chrysopterus). Similarly, field studies showed that four species of marine fish were attracted by different amino acids (Sutterlin 1975). Fishing experiments in commercial longlining and experiences of fishermen have also shown species-selective effects of baits. When several bait types were compared in fishing trials, squid was found to be the most effective bait for capturing cod and hake (Merluccius sp.), whereas mackerel appeared to be more effective for haddock (Martin and McCracken 1954). Bjordal (1983) found that squid bait caught twice as many ling as mackerel but only 9% more tusk. Comprehensive studies of Japanese tuna longlining have shown the speciesspecific effects of bait type on captures of tuna and marlin (Imai 1972; Imai and Shirakawa 1972; Shimada and Tsurudome 1971). Experiments with artificial baits have also demonstrated the effect of bait types on species selectivity (e.g., Løkkeborg 1991; Yamaguchi et al. 1983). For the commercially important cod, several studies have aimed at identifying the primary stimuli that elicit the search for food. Bottom food search behavior in cod was used to determine the feeding stimulants present in shrimp (Ellingsen and Døving 1986). The amino acid glycine was the most potent single component, followed by alanine. There was a synergistic effect among glycine, alanine, proline, and arginine, and a mixture of these substances was more efficient than the total amino acid mixture in the shrimp extract. Two unidentified substances were also found to be highly potent. Laboratory investigations were carried out to identify the feeding stimulants in squid for juvenile cod (Johnstone and Mackie 1990). The study confirmed that amino acids are the major feeding stimulants for cod. A mixture of non–amino acid components was inactive on its own, but there was a synergistic interaction between these components and the amino acid mixture. These results were in general agreement with those obtained in a later
Fish Behavior in Relation to Longlines study based on choice experiments to distinguish preference between feeding stimulants (Franco et al. 1991). Studies aimed at identifying the chemical nature of feeding attractants thus indicate that there is great potential for using baits or extract mixtures to attract specific species to an odor source, to develop species-selective fishing methods. Moreover, as mentioned in Section 4.1, the efficiency of baited gears may be substantially improved by developing long-lasting baits that release feeding attractants at high concentrations over a long period of time. 5.4.4 Chemosensory Thresholds Several laboratory experiments have attempted to determine detection and response thresholds in fish. Electrophysiological techniques have been employed to obtain information on detection thresholds in many freshwater and marine fish species (e.g., Belghaug and Døving 1977; Goh et al. 1979; Nikonov et al. 1990; Sutterlin and Sutterlin 1971), and behavioral methods have been used to study response thresholds (e.g., Løkkeborg et al. 1995; Pawson 1977). Olfaction is generally the more sensitive distant chemoreceptor in fish (Ishida and Kobayashi 1992), but for some species, in particular the genus Ictalurus, the gustatory receptors may be a more acute chemical sense for certain substances (Bardach et al. 1967; Caprio 1978). Johnstone (1980) presented a summary of some comparable detection threshold determinations for freshwater and marine species. Electrophysiological studies of salmon (Salmo salar) demonstrated that the most effective olfactory stimulus tested, alanine, gave a detection threshold of 3.2 × 10−9 M (Sutterlin and Sutterlin 1971). Belghaug and Døving (1977) determined the electrophysiological thresholds to amino acids in char (Salvelinus alpinus) and found that arginine had the highest stimulatory capacity, with a threshold of 2.5 × 10−8 M. Similar studies on coho salmon (Oncorhynchus kisutch), rainbow trout, and white catfish (Ictalurus catus) demonstrated detection thresholds of 10−8 to 10−6 M to several individual amino acids (Hara 1972, 1973; Suzuki and Tucker 1971). Electrophysiological studies of chemosensory capacity have also been carried out on marine fish.
113
The detection thresholds in Atlantic stingray (Dasyatis sabina) and sea catfish (Arius felis) to alanine and cysteine, respectively, were determined at between 10−8 and 10−6.5 M (Silver et al. 1976). Among a range of amino acids, the electrophysiological thresholds to glutamine were estimated at 10−7 M in red sea bream (Chrysophyrys sp.) and below 10−8 to 10−9 M in conger eel (Conger sp.) (Goh et al. 1979). Nikonov et al. (1990) estimated extremely low threshold values of 10−12 to 10−14 M to the most stimulatory amino acids in Black Sea skate (Raja clavata). The detection thresholds in the algivorous rabbitfish (Siganus fuscescens) to 19 amino acids ranged from 10−10 to 10−5 M, with the lowest detection threshold for alanine (Ishida and Kobayashi 1992). A classic conditioning method was used to determine the sensitivity of cod to several amino acids (Johnstone 1980). The amino acids with lowest thresholds for detection, in order of effectiveness, were tyrosine, cysteine, phenylalanine, and glycine, with mean threshold levels ranging from 2.5 × 10−8 M to 6.7 × 10−8 M. This study also investigated the effect of raising the background level of glycine on the response threshold level for glycine. These results indicated that to detect a specific amino acid against a background level of the same substance, the difference in level for detection needs to be proportionally greater with higher background concentrations. These electrophysiological and conditioning studies indicate the threshold concentrations at which the animal is capable of detecting the olfactory stimulus tested. However, attracting fish to an odor source relies on the response of the whole animal (the response threshold), and the threshold at which an animal detects a food extract differs from the threshold for food-searching behavior (Pearson and Olla 1977; Zimmer-Faust and Case 1983). In red hake (Urophycis chuss), the threshold for detection of clam extract was shown to be an order of magnitude lower than the threshold for food searching (Pearson et al. 1980). Furthermore, response thresholds are affected by feeding motivation, and increasing food deprivation and higher temperature have been shown to increase responsiveness to bait odor in sablefish (Løkkeborg et al. 1995; Stoner 2004; Stoner and Sturm 2004). Studies
114
Fish Behavior near Fishing Gears during Capture Processes
on response thresholds are therefore more relevant than those on detection thresholds for various aspects of bait attraction. Unfortunately, few behavioral studies have been carried out to determine response thresholds to food odors. Behavioral response thresholds to glycine were determined to be below 10−7 M in cod and whiting (Merlangius merlangus) (Pawson 1977). Herring (Clupea harengus) larvae were shown to respond to glutamic acid at a threshold concentration of 1.5 × 10−6 M (Dempsey 1978). Response thresholds to the mixture of amino acids and other compounds released from squid bait have been determined in sablefish kept on three different feeding regimes (Løkkeborg et al. 1995). The response thresholds to total dissolved free amino acids (DFAAs) were found to range from 4.4 × 10−8 M in fish fed to satiation to 1.4 × 10−11 M in fish tested after 4 days of food deprivation—the response threshold fell by a factor of 3000 due to food deprivation. These observations suggest that fish deprived of food are capable of responding to baited gears from considerably longer distances than are satiated fish (see Behavior before Stimulation in Section 5.4.1). 5.5 INTERACTIONS BETWEEN THE FISH AND THE LONGLINE GEAR 5.5.1 The Capture Process—A Multitude of Stimuli and Responses The behavior that leads to hooking has traditionally been described as a simple sequence of events, in the course of which a fish moves into the active space, swims against the current in the odor plume toward the longline, and eventually attacks a baited hook and becomes hooked. However, we now know more about the activity and search patterns of fish before odor stimulation and how these patterns influence the probability of entering the active space, the mechanisms of location of an odor source, and the behavior involved after contact with the gear. This new knowledge suggests that the interactions between the fish and the different kinds of stimulus emitted from a baited longline are fairly complex, with the outcome influenced by a great number of physical and biological variables. Some of these variables can be tuned to optimize the gear.
The increase in the number of species studied permits species comparisons, which enhance understanding of the behavior of individual target species. Longlining is primarily based on chemical stimulation involving both the olfactory and gustatory senses, but other types of stimulus are also involved in the catch process. Visual stimuli provided by the line and baits as well as by hooked and unhooked fish play an important role. Movements of baits and struggling fish may also be sensed by the lateral line organ. The relative importance of the sensory modalities differs among species. For instance, on the basis of prey preferences, the relative size of the sensory organs, brain anatomy, and the diel activity rhythm and time of day of responding to baits, visual stimuli seem to be more important in ling than in cod (Fernö et al. 2006; Kotrschal et al. 1998; Løkkeborg et al. 2000; Løkkeborg and Fernö 1999). The influence of different stimuli may also vary within a species. Presoaked baits with reduced release rate of attractants resulted in poorer catch rates for bottom-set longlines but not for pelagic longlines that targeted cod migrating to the spawning grounds (Løkkeborg and Johannessen 1992). In the latter situation, the fish seem to react more to visual than to chemical stimulation. Fish are not always exposed to different stimuli from the gear in a strictly sequential way but may simultaneously encounter olfactory and gustatory stimuli as well as chemical and visual stimuli. The fish must then integrate stimuli from different sensory channels. Different stimuli presumably have an additive effect, but in some cases the stimuli may compete, with the fish trading off their effects. In addition, the order in which an animal encounters different stimuli from the gear may determine the outcome. King crabs approaching a baited pot upstream and then presumably attracted by chemical stimuli were trapped in the odor plume, butting against the net of the pot when they encountered a side without an entrance (Stiansen 2004). In contrast, king crabs approaching the pot down- or across-current and presumably attracted by visual or auditory stimuli showed more flexible behavior. These crabs searched around, and they more often located the entrance and entered the pot. Similar observations of ling and cod have showed that fish that encountered pots upstream when the current
Fish Behavior in Relation to Longlines was perpendicular to the entrance stayed within the odor plume and did not find the entrance area (unpublished data). The response of fish to a longline may also be influenced by which stimulus they initially encounter. The sequence of events leading to hooking can be divided into the following phases: Behavior before Stimulation The spatial and temporary dynamics of fish movements influence the efficiency of longlines. Actively swimming fish will have a higher probability of encountering the bait odor plume than will resting fish, and high swimming speed results in more encounters than slow swimming. Swimming capability varies among species and sizes of fish and is affected by ambient temperature. The path of fish that are searching for food also plays a role. In an individual-based behavioral model, moving at an angle to the current resulted in a higher probability of contact with an odor plume than moving straight into the current or a “random walk” strategy (Vabø et al. 2004). Furthermore, this study showed that cod tracked in a fjord often swim in a zigzag pattern against the current. Diel activity rhythms also influence the probability of encountering the bait odor plume (see Section 5.3). Arousal to the Presence of Bait and Stimulus Categorization Arousal to the presence of food when the concentration of attractants exceeds the thresholds of the chemosensory organs is a necessary first step in the hooking process of baited longlines. However, a fish stimulated by the vicinity of a food item does not automatically respond by approaching the odor source. Defined in this way, arousal is determined by the physiological detection threshold and is not influenced by motivational state and other modifiers. To react further, the fish must classify the stimuli from a source to be of relevance to its survival or reproduction and react accordingly. To respond to a longline by approaching the odor source, the fish must categorize the perceived chemical stimuli as a potential prey. Locating the Bait Odor Source The distance from which an odor source from a baited gear attracts fish (i.e., the active space) can
115
be determined by models of odor dispersion and by tracking fish in the field. Using the dispersion model by Sainte-Marie and Hargrave (1987) and the response thresholds to bait odor, maximum dimensions of the active space within which sablefish would search for food were found to vary from 10 m to several kilometers, depending on state of food deprivation, rate of attractant release from the bait, and current velocity (Løkkeborg et al. 1995). Of these variables, food deprivation had the greatest influence on the size of the active space. After 4 days of food deprivation, the attraction distance for sablefish increased by a factor of 57 over that of fish fed to satiation. Under conditions of low prey availability and great demand for food, therefore, baited gears are likely to attract fish within much larger areas. Field studies have confirmed that fish are capable of locating food odor sources at long distances (Løkkeborg 1998; Løkkeborg and Fernö 1999). In these studies, cod tagged with acoustic transmitters were tracked in their natural environment, and their responses to baited longlines set at various distances were observed (Fig. 5.2). At as far as 700 m from the baits, fish would respond to the bait odor and locate the odor source. Interestingly, this distance is similar to the distances between parallel fleets of longlines as set in commercial fishing in Norway, which suggests that fishermen have managed to optimize an important parameter. Ling have been observed to react at shorter distances (about 100 m; Skajaa 1997). It is important to identify the mechanisms that fish use to locate baited gear to optimize the gear. Upstream movements in the direction of bait have been observed in several field studies (Fernö et al. 1986; Kaimmer 1999; Løkkeborg et al. 1989; Løkkeborg and Fernö 1999; Wilson and Smith 1984). We might initially believe that fish locate the baited gear by swimming toward higher concentrations of attractants using a gradient search, but orientation in a chemical gradient is associated with many problems. Furthermore, in a realistic fishing situation, a gradient search would not always bring the fish into contact with the baited gear. Due to the rapid initial decrease in release of feeding attractants from baits over time (Løkkeborg 1990a), a reverse gradient may well be created, with the
116
Fish Behavior near Fishing Gears during Capture Processes
Baited longline Current direction
100 m
highest concentrations at long distances from the gear (Sigler 2000). In a simulation study of bait location, countercurrent strategies performed much better than strategies based on gradient search (Vabø et al. 2004). Fish orienting by rheotaxis, however, may encounter problems at the edge of the odor plume, and in the simulations, fish that had lost contact with the plume were allowed to swim in progressively larger loops until they were on track again. In such situations, fish might to a certain extent use gradients by comparing different points in space, although the large spatial and tem-
Figure 5.2. Tracks of three cod showing chemically mediated responses to baited longline.
poral variations in concentrations would make this difficult. Hence, although we cannot exclude the possibility that fish use gradients at the edge of the odor plume (Webster and Weissburg 2001), and close to the odor source (Atema 1980), it is safe to assume that rheotaxis is the main mechanism behind upstream swimming and location of a baited longline. This has important practical consequences. If gradient search is the main mechanism, much would be gained by adjusting the release rate of attractants from the baits in such as way as to create
Fish Behavior in Relation to Longlines as strong gradients as possible. If rheotaxis is the most important mechanism, we ought to select or develop baits that initially release attractants at high concentrations to attract fish from long distances and then maintain a release rate such that the fish will encounter above-threshold concentrations all the way to the bait. Sablefish seem to have a great ability to locate baits on a longline even when only a few baits remain (Sigler 1993), and cod can search along a line and localize baits (Løkkeborg and Fernö 1999). However, it may not always be straightforward for a fish to localize an olfactory source. One cod attracted from a large distance was observed to make stops under way, apparently losing the track and searching around (Løkkeborg 1998). The fish may then have entered an area with low concentrations of attractants generated by turbulence. In such situations, the fish may perform search patterns that increase the probability of reencountering the plume (Vabø et al. 2004). Visual Object Categorization, Bait Ingestion, and Hooking Behavior Fish in close contact with a baited hook display a number of behavioral patterns of different intensities. Based on the visual, taste, and mechanical stimuli it perceives, the fish may categorize the baited hook as either edible or inedible and respond accordingly. The behavior patterns have been defined and their occurrences recorded both in the laboratory and in the field (Fernö et al. 1986; Fernö and Huse 1983; Kaimmer 1999; Løkkeborg et al. 1989). Laboratory studies are appropriate for the initial stages of describing and defining different behavior patterns and investigating how the response of individual fish develops over time, but field studies are required to obtain realistic estimates of the frequency and intensity of the reactions (Løkkeborg et al. 1993). A fish makes an Approach toward the bait when it comes into close proximity to the bait but terminates the response without physical contact. During a Taste, the fish is in contact with the bait with the mouth or barbel. Incomplete bite, in contrast to Complete bite, is when the fish only takes a part of the bait into the mouth or does not close its mouth. Jerk is a rapid sideways movement with the head, and a Jerk series is several jerks
117
in rapid succession. Pulling the snood and Chewing the bait are made with the bait in the mouth. Finally, Rush is a rapid burst of swimming with the bait in the mouth, when the fish may either become hooked or escape and swim away. Some of these responses can reflect reactions toward natural prey—for example, a Jerk series could represent the behavior when a fish tries to shake a mussel out of its shell (Brawn 1969). Other responses, such as Rush, may be an escape reaction. Responses to hooks can be described by the frequencies of occurrence and the sequence of the different patterns of behavior. A flow diagram (see Huse and Fernö 1990; Kaimmer 1999) gives an overview of the behavior. A sequence analysis reveals more detail about which behavior patterns tend to precede and follow each other (Kaimmer 1999). In cod, transitions between complete bite and pulling, chewing, and rush are overrepresented (Fig. 5.3). Rushes lead in turn to hooking or the bait being pulled out of the mouth. Incomplete bites, on the other hand, seldom lead to rush or hooking. By identifying the behavior associated with hooking and relating the probability of hooking to this behavior, we can obtain an idea of the efficiency of a particular combination of bait and hook. Rush is strongly associated with hooking, although rushes in some instances can be the result and not the cause of hooking. The probability of hooking can thus be estimated as the number of hooked fish divided by the number of rushes or, alternatively, as the number of hookings relative to the number of all strong responses. The probability based on rushes ranges from 0.08 to 0.52 using different hook-and-bait combinations in laboratory and field studies of various species (Fernö et al. 1986; Huse and Fernö 1990; Kaimmer 1999; Løkkeborg et al. 1989). Many hook types thus appear to be relatively inefficient, and observations made with a highfrequency imaging sonar showed that a very low percentage of sablefish and Pacific halibut (Hippoglossus stenolepis) attracted to longlines and pots were captured (Rose et al. 2005). In one study, only 1.2% of cod observed on video were captured (He 1996). Because a rushing fish should generate sufficient power for the hook to penetrate the mouth when the movement is suddenly stopped by the resistance of the snood, the low hooking probability
118
Fish Behavior near Fishing Gears during Capture Processes
Figure 5.3. Sequence analysis of transitions between different responses of cod towards baited hooks based on a laboratory study by Fernö and Huse (1983). Combinations of behavior patterns that occur more or less frequently than expected by chance are indicated (chisquare test, P < .05). Transitions where the observed and expected numbers of transitions are not significantly different are not marked. Free swimming means that the fish leaves the near field of the baited hook.
indicates that the point of the hook often does not come into contact with the mouth. When it visually detects the longline in the nearfield, the fish may categorize the gear as a prey (resulting in a further approach), a predator (resulting in escape), or an object of no relevance (no response). Several properties of the bait and hook influence the behavior and thereby the outcome of the interaction. Cod and haddock approaching a large bait more often turned away before physical contact than did fish approaching smaller baits (Johannessen et al. 1993), and large baits were also less effective in comparative fishing experiments (Johannessen 1983). Furthermore, large baits have been shown to catch fewer small fish than smaller baits; there is a size-selective effect of bait size (Johannessen 1983; Løkkeborg 1990b; McCracken 1963). In a conflict situation between accepting and
rejecting a baited hook as a potential prey (Fernö and Huse 1983), fish seem to more often reject large baits. The shape of the bait may also affect responsiveness to a baited hook. Lower catch rates of small cod for rectangular shaped artificial baits than when natural shrimp bait was used were explained by a restrained response toward a novel prey item (Løkkeborg 1990b). The hooking probability of a circle hook has been shown to be higher than a J-shaped hook, presumably due to a higher probability of penetrating the mouth during a rush (Huse and Fernö 1990). Circle hooks were also more effective in catching cod in comparative fishing experiments. Similar studies that compared the circle hook and the J-hook also showed the superiority of the circle hook for catching Pacific halibut and hake (Peeling and Rodgers 1985; Quinn et al. 1985) but not for catching sword-
Fish Behavior in Relation to Longlines fish (Xiphias gladius) (Watson et al. 2005). Double and treble hooks, which increase the probability of contact, have been shown to be more effective than single hooks (Fernö et al. 1986). Hooking probability is species specific. For instance, attacks on baited hooks in cod are more often based on a complete bite and are of higher intensity than in haddock, and cod are caught more than twice as often after a rush (Løkkeborg et al. 1989). The different response intensities can be explained by differences in natural foraging behavior, with haddock feeding more on small stationary benthic organisms and cod more on mobile prey, which demand higher-intensity responses. The intensive attacks of cod on prey can therefore be expected to be preceded by more careful and selective behavior, and more cod than haddock in fact terminated their responses to baited hooks without physical contact (Løkkeborg et al. 1989). Haddock more often take only a part of the bait in their mouth and can then pull the bait from the hook, and this species has a reputation among fishermen as a bait stealer. There are generally large individual differences in fish behavior (Magurran et al. 1993), and this should also be expected during interactions with fishing gear (Fernö 1993). In the laboratory, some cod made only a few strong responses toward a baited hook, whereas other individuals made several hundred low-intensity responses (Fernö and Huse 1983). A similar difference between individual fish was also seen in haddock observed in their natural environment (Løkkeborg et al. 1989). The frequency of hooking recorded during behavioral observations cannot be used to directly estimate the efficiency of a given combination of bait and hook. One reason is that a test line differs from an actual longline in many respects, such as in the tension that determines the resistance the fish experience when pulling the snood. Furthermore, fish cannot be assumed to always make only one attack on a baited hook. With a greater number of attempts, the difference between an efficient and a less-efficient hook will decrease and, in theory, both hooks will eventually capture the fish after repeated attacks. With only one attack of each fish, two hook types with hooking probabilities of 0.4 (Hook A) and 0.2 (Hook B) will result in twice as many
119
hooked fish on Hook A, resulting in a 100% difference in catch rate between longlines with Hooks A and B. If the fish make five attempts, the difference will decrease to 37% (1 − 0.65 divided by 1 − 0.85). This may be the reason that the difference between two hooks tends to be smaller in comparative fishing experiments than is suggested by the difference in hooking frequency observed in the laboratory (Huse and Fernö 1990). Nevertheless, behavioral studies of hooking frequency show which hook type is the most efficient. 5.5.2 Internal Factors Food deprivation and motivational state have been shown to affect food searching behavior and responsiveness to prey in several fish species (Atema 1980; Hart 1993; Hart and Connellan 1984; Pearson et al. 1980). Accordingly, hunger has been shown to affect food searching behavior and response intensity to baits. Sablefish responded to lower bait odor concentrations when tested under conditions of lower food rations and longer duration of food deprivation (Løkkeborg et al. 1995). The intensity of behavioral responses to bait odor (swimming speed and turning rate) and response duration were also shown to increase in sablefish with increasing food deprivation (Løkkeborg et al. 1995; Stoner and Sturm 2004). These behavioral changes will affect the distance from which sablefish will start a food search, its search pattern, the location time, and the time spent searching for the odor source. Similarly, Pacific halibut were shown to locate more bait with increasing food deprivation, and fish deprived of food located the baits more rapidly (Stoner 2003). These observations indicate that when food demand is higher, fish intensify their search for food and therefore increase the probability of locating a food odor source, whereas satiated fish will probably not display active food searching behavior. Feeding motivation has also been shown to influence hooking behavior. Whiting and cod with low liver weight, an indication of starvation and increased food demand, tended to swallow the hook and be caught in the stomach, whereas fish with higher liver weight were usually caught in the mouth (Fernö et al. 1986; Johannessen 1983). Seasonal changes in the feeding motivation relative to other motivational systems (e.g., reproduction)
120
Fish Behavior near Fishing Gears during Capture Processes
can also lead to variations in the response. The probability of a cod being hooked was higher in September/October than in December (Løkkeborg et al. 1989), and similarly, the hooking probability of whiting was higher in October than during May– July (the spawning season) (Fernö et al. 1986). Prey abundance affects the level of hunger, and evidence exists for the effects of prey density on responsiveness to baited hooks. Longline fishermen have experienced extremely low catches of cod in the Barents Sea in early spring when the cod are preying on dense schools of capelin (Engås and Løkkeborg 1994). Fishermen from the Faeroe Islands have also observed lower catch rates in years with high prey density in the sea (Petur Steingrund, Faroese Fisheries Laboratory, personal communication). Similarly, longline catch rates of tunas in the tropical Pacific were low in areas where there were large, high-density prey patches (Bertrand et al. 2002). 5.5.3 The External Environment (Light, Temperature, Current) Behavioral observations have shown that cod and other species exhibit diel rhythms in swimming and feeding activities (e.g., Løkkeborg et al. 1989; Løkkeborg and Fernö 1999; Thorpe 1978). The
swimming speed of acoustically tagged adult cod in the autumn increased at dawn, remained high during the day, then gradually decreased during the evening, remaining relatively low throughout the night (Løkkeborg and Fernö 1999). The recorded swimming activity of cod in this study appeared to be related to light level and was regarded as reflecting food search behavior. In a similar study in late spring, with higher light levels at night, a similar but less pronounced rhythm was observed (Løkkeborg 1998) (Fig. 5.4). The locomotory activity of Pacific halibut in a laboratory study was three times as high at high than at lower light levels (Stoner 2003). These observations suggest that activity decreases below a critical light level. This is supported by a study of responses of cod to baited hooks, in which the fish displayed higher activity during the day in both September and December, but the increase in activity in the morning was observed later in December in connection with the changing time of sunrise (Løkkeborg et al. 1989). Similar seasonal changes in diel rhythms of responses to bait have been demonstrated in whiting (Fernö et al. 1986). These findings indicate that the success of fishing with baited gears is influenced by the time of day when the gear is set. More cod were thus observed
Figure 5.4. Diel rhythm in the mean swimming speed of cod in May (open circles) and September (closed circles) observed in a fjord in northern Norway. Arrows indicate sunrise and sunset in September, whereas May has midnight sun at this latitude.
Fish Behavior in Relation to Longlines to locate baits during periods of high (daytime) than low (nighttime) activity, and the time that elapsed until the fish located baits was 50% shorter during the day (Løkkeborg and Fernö 1999). Fishing experiments conducted in commercial fishing showed that longlines set before dawn caught twice as much haddock as longlines set later in the day (Løkkeborg and Pina 1997). Thus, time of day is an important factor to consider when fishing with baited gears. Furthermore, there is little doubt that light level has a direct effect on locomotion in fish, independent of diel rhythms (Stoner 2004). The higher probability of locating baits during the day may also be related to increased visibility. In a laboratory study, Pacific halibut located a larger proportion of the baits offered as the light level increased, and the time taken to locate baits was shorter in the light conditions than in darkness (Stoner 2003). Several fish species have been shown to be relatively unsuccessful in locating and attacking baits in darkness, although the lower light thresholds for feeding were low (e.g., McMahon and Holanov 1995; Ryer and Olla 1999; Stoner 2003). Interestingly, night setting has been shown to be an effective mitigation measure to reduce incidental bycatches of seabirds in longlining, and more seabirds were caught on night-set hooks set in bright moonlight conditions than when there was no moonlight (see Section 5.2). Temperature is another external factor that can influence how fish interact with longlines. Up to a certain limit, increasing temperature increases the scope of activity and swimming activity (Castonguay and Cyr 1998; He 2003; Stoner and Sturm 2004), leading to more encounters with the odor plume and the gear. In addition, temperature will affect the metabolic rate and gastric evacuation (Fry 1971). In most cases, food consumption will increase in line with temperature, leading to stronger responses to baited gear and a higher probability of being hooked. Sablefish at a low temperature swam slowly and attacked and consumed fewer baits (Stoner and Sturm 2004). For a thorough review of the effect of temperature, see Stoner (2004). Current is another external factor that affects food-searching behavior both through its effect on bait odor dispersion and on fish activity. Currents will increase the active space and permit rheotactic
121
responses to the odor source, and more whiting were attracted to bait in the presence of a current (Fernö et al. 1986). However, responses in fish to bait have been shown to decrease when current velocity is high. When current speeds were less than 18 cm s−1, the number of cod and haddock in the vicinity of baits were two or three times as high as in periods with stronger currents (Løkkeborg et al. 1989). However, variations in current velocity below 18 cm s−1 did not appear to influence activity. This could be explained in terms of energy optimization. As food-searching fish swim predominantly upstream to the odor source, it would be energetically advantageous to be active during periods of moderate or low current velocity and to remain in shelter when the current is strong (see Weihs 1987). The current can also influence the probability of contact with a baited gear by influencing the vertical distribution of fish (Michalsen et al. 1996). It is clear that the motivational and physiological state of the target species as well as its sensory capabilities have a major impact on food searching behavior and the likelihood of locating baits. Environmental conditions such as light, temperature, and current also affect locomotion, the probability of bait location, and responsiveness. These internal factors and environmental variables thus have obvious implications for the outcome of fishing operations using baited gears. 5.5.4 Intraspecific and Interspecific Interactions Food-searching behavior and responses to baits are likely to be affected by fish size because larger fish have greater swimming capability and are at less risk of predation than are smaller fish. According to optimal foraging theory, larger fish use a larger foraging area due to their higher optimal swimming speeds (Hart 1993), and constraints on feeding activity are less pronounced due to the lower risk of predation (Milinski 1993), resulting in exploitative competition. Interference competition among fish attracted to baits is also related to size. Both intraspecific (cod, whiting, ling, tusk, halibut) and interspecific (codhaddock, wolffish-cod/haddock) competition for baits has been observed, with the largest individuals emerging as the most successful competitors (Allen
122
Fish Behavior near Fishing Gears during Capture Processes
1963; Bertrand 1988; Fernö et al. 1986; Godø et al. 1997; Løkkeborg and Bjordal 1992; Stoner and Ottmar 2004). Rodgveller et al. (2008) found that longline catch rates of sablefish were negatively correlated with giant grenadier (Albatrossia pectoralis) and rougheye rockfish (Sebastes aleutianus) catch rates, indicating competition for baited hooks in the area studied (Alaskan waters). No negative correlations were found for trawl catches taken in the same area, indicating that the negative correlations for longlines were not due to differing habitat preferences among these species. The presence of other fish in the vicinity of a longline does not always have a negative effect on catchability but may in fact stimulate responses in fish. Groups of fish are less at risk to predators than are single fish, with the balance between feeding and antipredator behavior thus shifted toward the former. Social facilitation and copying the behavior of other fish also stimulate responses (Ryer and Olla 1992), as does competition for food. A laboratory study by Stoner and Ottmar (2004) found that several halibut located and consumed baits more quickly than did single fish. In the field, we observed a haddock that swam slowly back and forth for a long time, nibbling on the baits on the test line (Løkkeborg et al. 1989). Eventually the haddock was hooked and fought to get free, and within a few seconds 10 to 15 large cod entered the field of observation. One cod vigorously attacked and tried to swallow the hooked haddock, while other cod attacked the neighboring hooks and became hooked. Similarly, more whiting became hooked on a test line when there were hooked fish present (Fernö et al. 1986). However, fright responses released by hooked fish have never been observed, and fish of the species studied do not seem to react negatively to trapped fish—a situation that presumably does not take place under natural situations. The stimulatory effect of competitors (social facilitation), hooked fish, and the movements of neighboring baits helps to explain the observation that fish are often found clustered along a longline (Johannessen 1983; Sigler 2000). 5.5.5 Learning The literature on learning in fish has expanded enormously during the past decade, and it is now well
documented that fish have highly developed learning skills and cognitive abilities (Brown et al. 2006). Although the ease with which fish can form an association between a stimulus and a reward or punishment is limited by conceptual learning schemes, learning processes should be expected to modify how the longline is categorized and the sequence of responses toward the gear. The behavior of any given species toward baited hooks does not seem to be fixed, indicating flexibility and learning. In the field, fish are often observed to swim away from the gear after few responses, supporting the idea that they modify their behavior over time. Before a fish encounters the gear, its feeding behavior and its responses to gear-related stimuli can be molded by its recent experience with prey. How fish search for and catch prey may influence both search patterns and which sensory channel fish primarily use. Visual stimuli thus seem to be more important than chemical stimuli for cod feeding on capelin (see Section 5.5.2). This can be used in longline fishing by tuning the type of bait to local natural prey items. Along the same lines, the lower efficiency of artificial baits for small cod than for large cod may be explained by the fact that small cod have less experience of a range of prey organisms, making them more reluctant to attack a novel prey (Løkkeborg 1990b). A clear example of learning can be found in observations of cod that were tagged in situ by allowing them to ingest acoustic transmitters wrapped in mackerel bait (personal observations). A fish that had ingested one tag followed the research vessel to the next tagging station and took another transmitter shortly after the tagging rig was put in place on the bottom. Eventually the fish had taken four valuable tags (US $280 each), intended for tracking other individuals. However, the perturbed researchers hit back, and at its fifth attempt the saboteur was surprised by the baited and wrapped transmitter being replaced by a baited hook. The end of the story is shown in Figure 5.5. This fish responded fast and before the odor plume had dispersed more than a few meters. It was presumably attracted by auditory stimuli from the vessel and learned to associate this sound with the presence of food.
Fish Behavior in Relation to Longlines
123
Figure 5.5. Learning/ conditioning: this cod was able to locate and ingest four acoustic tags wrapped in mackerel bait by associating the sound of the research vessel with the presence of food. For color detail, please see color plate section.
Physical contact with a baited hook has been shown to modify the response. Cod in the laboratory typically terminated their response sequence after a single strong response toward a hook and did not make a new attempt for some time (Fernö and Huse 1983). Over time, the fish apparently came into a conflict situation between feeding and avoiding pain, and some individuals made long series of approaches and retreats without touching the bait. When this article was first submitted for publication, one of the referees questioned whether fish can experience pain, but with our current understanding of fish learning and cognition, it is now generally accepted that fish perceive pain and modify their behavior accordingly (Chandroo et al. 2004). In contrast to cod, haddock typically take small pieces of the bait during incomplete bites without contact with the hook and are thereby rewarded, stimulating further responses. Acoustically tagged cod that encountered a bait bag that was too large to ingest also seem to experience a conflict and were often observed to leave the baited gillnet (Kallayil et al. 2003). After some time, they returned from distances of several hundred meters. This indicates that the output from the reward-evaluating mechanisms in the brain assessing an experience shows a temporal dynamics, with the effect of a negative experience fading away and the tendency to respond to food again
becoming dominant. In some cases, however, the modification can persist for a long period. Carp (Cyprinus carpio) with experience of hooks were still difficult to catch after 1 year (Beukema 1970). If what a fish experiences after terminating a strong response toward a hook is capable of modifying its behavior, there may also be an effect of experience as soon as the fish takes the baited hook into its mouth and manipulates and chews on it. In that case, the efficiency of a hook can be partly dependent on how much it reduces the strength of further responses. Circle hooks are more efficient than J-hooks (see Section 5.5.1), and this is believed to be primarily due to the higher probability of curved hooks penetrating the mouth during a rush. However, it is also possible that the point of a circle hook that is bending inward will come less often into contact with the mouth when a fish is chewing on the bait, and the escalation of the response is therefore not inhibited. 5.6 CONSERVATION CHALLENGES AND POTENTIAL SOLUTIONS 5.6.1 The Main Problems Concerns about the numbers of seabirds that are incidentally killed in various types of fisheries are growing. Most attention has been given to bycatch of seabirds in longlining, especially albatrosses
124
Fish Behavior near Fishing Gears during Capture Processes
taken in the Southern Ocean longline fisheries (Brothers 1991; Cherel et al. 1996; Weimerskirch et al. 1997). Many southern albatross populations are in decline, and longline-induced mortality is believed to be an important factor contributing to this situation (Croxall et al. 1990; Moloney et al. 1994; Poncet et al. 2006; Prince et al. 1994; Weimerskirch and Jouventin 1987). The second conservation issue related to longline fisheries concerns interactions with sea turtles. There are seven species of sea turtles living in the world’s oceans, of which six are listed as endangered (http://www.redlist.org/). The failure of sea turtle populations to recover is attributed in part to incidental capture by fishing gears, and although most concerns have been raised about bycatch of sea turtles in trawl fisheries (Magnuson et al. 1990), pelagic longlines have been implicated as a major source of anthropogenic mortality for loggerhead (Caretta caretta) and leatherback (Dermochelys coricea) sea turtles (Lewison et al. 2004). The final conservation-oriented issue covered here is bycatch of nontarget fish. This problem relates both to discards of specimens below minimum landing size (i.e., juveniles of the target species) and to nontarget species. Bycatch of juveniles can be reduced by improving longline size selectivity, and factors affecting size selection are discussed in Sections 5.1, 5.4, and 5.5 and have been reviewed by Løkkeborg and Bjordal (1992). The effect of fishing on shark stocks has become the focus of considerable international concern (Megalofonou et al. 2005), and the FAO has developed an International Plan of Action for the Conservation and Management of Sharks. Thus, with regard to incidental catch of nontarget fish, we focus on efforts to monitor and reduce shark bycatch in pelagic longlining. It is often difficult to attribute population declines to a specific factor, as the marine environment is subject to much natural variation and thus provides a noisy background for observing changes that can be directly attributed to fishing activities (Gislason 1994). Some catch statistics exist for sharks, but accurate information on the numbers of seabirds and sea turtles killed is difficult to obtain (Lewison et al. 2004; Wienecke and Robertson 2002). Estimates of annual fishing-induced mortality of
these species are poor because captures are rare and observations are few (Pradhan and Leung 2006). Regardless of the actual number of nontarget individuals caught in a fishery and the consequent population-level effects, it is not consistent with the principles of ecologically sustainable management for fisheries to take large numbers of nontarget species (Løkkeborg and Robertson 2002). Furthermore, incidental bycatches of nontarget species are a twofold problem, especially with regard to seabirds. Most seabirds that attack baited hooks manage to seize the bait without becoming hooked, and it is likely that sea turtles and sharks also scavenge baits from longline hooks. Such interactions reduce gear efficiency and profitability due to the associated loss of baits. It should therefore be in the interest of fishermen to reduced interactions with individuals of nontarget species, as reducing bait losses is likely to result in higher target catch rates. A 32% increase in target catch rates was obtained for longlines that were set using a bird-scaring line, compared with longlines set without any mitigation measure (Løkkeborg 2001). 5.6.2 Mitigation Measures Aimed at Reducing Incidental Catches of Seabirds When longlines are set, baited hooks are available to foraging seabirds because they float on the surface for a short while before they start sinking. Seabirds are killed during the line-setting operation when they seize baited hooks, become hooked in the bill or body, and are drawn underwater by the sinking longline. Some seabird species are also capable of diving to depths of several meters and may thus attack baited hooks during the first part of the sinking phase. Birds occasionally become hooked during line hauling, but with careful handling they can be released alive. Most efforts should thus be put into developing measures that will prevent seabirds from seizing baited hooks during the setting operation. Mitigation measures should not only be efficient in minimizing bird capture but also practical and easy to implement in commercial fishing, enforceable, cause no loss of target catch, and offer fishermen incentives to use them (Gilman et al. 2003, 2005). Several mitigation measures capable of reducing the likelihood of seabird bycatch have
Fish Behavior in Relation to Longlines been described (Brothers et al. 1999a; Bull 2007), and they can be divided into four main categories: 1. Avoid peak areas and periods of bird foraging (night setting, area and seasonal closures). 2. Hinder access to baited hooks (underwater setting funnel, weighted lines, thawed bait, line shooter, bait-casting machines, side-setting). 3. Deter birds from taking baited hooks (streamer [bird-scaring] lines, acoustic deterrents, water cannon). 4. Reduce the attractiveness or visibility of the baited hooks (dumping of offal, artificial baits, blue-dyed bait). The most promising and widely tested mitigation measure in demersal longlining is the streamer line (bird-scaring line, tori line, Fig. 5.6a), which is a line with streamers that is towed behind the vessel and deters seabirds from attacking baited hooks while longlines are set. This device has been shown to virtually eliminate seabird bycatch in the longline fisheries in Alaska and in the northeast Atlantic where interactions with the northern fulmar (Fulmarus glacialis) are most common. Experimental studies carried out in the former fishery showed that paired streamer lines reduced seabird bycatch by 88% to 100% compared with control lines with no mitigation measures installed (Melvin
125
et al. 2001). Similar studies conducted in the northeast Atlantic demonstrated that a single streamer line reduced the bycatch by 98% to 100% (Løkkeborg 2003). Mitigation measures such as setting funnel and weighted lines were also shown to significantly reduce seabird bycatch in these fisheries but were not as efficient as the streamer line. Streamer lines are likely to be less efficient in reducing bycatch of diving seabirds as birds may still reach baited hooks beyond the aerial portion of streamer lines. This deficit may be solved or at least significantly reduced by using weighted longlines in combination with streamer lines. In the cod fishery in the Bering Sea, paired streamer lines in combination with integrated weight lines were shown to reduce bycatch of short-tailed shearwaters by 97% compared with control lines with no mitigation measure (Dietrich et al., 2008). Setting longlines at night is the most widely tested mitigation measure in the Patagonian toothfish (Dissosticus eleginoides) fishery in the Southern Hemisphere (Fig. 5.6b). In an experiment conducted in the South Indian Ocean, mortality rates were reduced by 62% when lines were set at night compared with when lines were set during daylight hours (Cherel et al. 1996). Furthermore, for the lines set at night, the mortality rate was 75% lower when the powerful deck lights were turned off. Observer data showed that night setting reduced
Figure 5.6. The bird-scaring (streamer) line is the most efficient mitigation measure tested in demersal fisheries in the northern hemisphere (a), and night-setting when there is no moonlight has proved to be efficient in fisheries in the southern hemisphere (b).
126
Fish Behavior near Fishing Gears during Capture Processes
mortality of white-chinned petrels (Procellaria aequinoctialis) by 81%, and only 1 of a total of 78 albatrosses were caught at night (Weimerskirch et al. 2000). Several other studies have also demonstrated the efficacy of this measure (Ashford et al. 1995; Nel et al. 2002; Reid et al. 2004; Ryan and Watkins 2002). Other efficacious mitigation measures for reducing seabird catches in demersal longlining in the Southern Oceans include dumping offal during line setting (98% and 54%, respectively, in Cherel et al. 1996 and Weimerskirch et al. 2000), weighted longlines (80% in Agnew et al. 2000; 61% to 99% in Robertson et al. 2006), streamer lines (Moreno et al. 1996), and setting funnel (Ryan and Watkins 2002). Night setting has been shown to be an efficient mitigation measure also in pelagic longlining with the probability of taking bycatch reduced by up to 85% (Brothers et al. 1999b; Klaer and Polacheck 1998; Murray et al. 1993). Seabirds were also more likely to be caught on night-set hooks set in bright moonlight condition than on those set when there was no moonlight. In Hawaii, the underwater setting chute was shown to eliminate captures of albatrosses in tuna longlining (Gilman et al. 2003), and in the swordfish fishery, branch lines with added weights and bait that was dyed blue (making it less apparent) reduced the number of contacts with albatrosses by about 90%, and the use of a streamer line reduced contact by about 70% (Boggs 2001). Bait thawing, use of a bait thrower, area fished, and season are other factors that significantly affect seabird mortality in pelagic longlining. There is no single solution of the problem of incidental seabird mortality in longline fisheries, as the efficiency of any given mitigation measure will be influenced by the seabird species assemblage at the particular fishing ground as well as the type of longline gear used (Løkkeborg 2008). Where the northern fulmar is the predominant seabird captured (i.e., northern Atlantic and Pacific Oceans), streamer lines have proved to be very effective in demersal fisheries. In pelagic longlining and in the Southern Hemisphere, where albatrosses and petrels are dominant, night setting is an effective measure, although it should be used in combination with other mitigation devices (e.g., streamer lines and weighted lines) in areas inhabited by nocturnal seabirds and
in bright moonlight conditions. Seabird mortality rates in all longline fisheries can therefore be significantly reduced by appropriate and effective mitigation measures. Considerable reductions in seabird bycatch rates have thus been obtained in many fisheries (Agnew et al. 2000; Gilman et al. 2003; Murray et al. 1993; Reid et al. 2004). For example, from 1993 to 2003, the bycatch rate was reduced from 0.66 to 0.0003 bird per 1000 hooks in the toothfish fishery around South Georgia as a result of the implementation of the CCAMLR conservation measure 25-02 (Reid et al. 2004). 5.6.2 Mitigation Measures Aimed at Reducing Incidental Catches of Sea Turtles Unlike seabirds, sea turtles are seldom caught during the setting operation when the baited hooks are floating at the surface. Sea turtles are able to dive to much greater depths than most seabirds and are caught mainly on pelagic longlines after the gear has sunk to the fishing depth. Thus, different mitigation measures have to be developed to reduce sea turtle bycatch. Due to their limited diving depth, sea turtles are primarily caught by shallow-set longlines and not by deep pelagic or bottom-set longlines (Polovina et al. 2003). Sea turtles caught on pelagic longlines set close to the surface can swim up to the surface to breathe. Nearly all sea turtles caught on shallowset longlines are therefore most likely alive at gear retrieval. Observations in the Spanish surface longline fishery in the western Mediterranean (1999– 2004) showed that less than 2% of the incidental catches of loggerhead turtles were dead (46 of 3480 turtles; Caminas et al. 2006), and all loggerhead turtles (188) caught in a study carried out in the Ionian Sea (Italy) were alive (Deflorio et al. 2005). Thus, it is important to identify and develop the best practices to handle and release captured turtles in such a way as to minimize injury. There is little empirical data on rates of posthooking mortality in sea turtles released from longlines. Findings strongly suggest low rates of postrelease mortality in lightly hooked Olive Ridley sea turtles (Lepidochelys olivacea) caught by shallow-set longlines (Swimmer et al. 2006). Sasso and Epperly (2007) used pop-up archival tags to estimate survival rates, and their results also suggested
Fish Behavior in Relation to Longlines that lightly hooked loggerhead turtles did not suffer any additional mortality relative to control turtles captured by dip net. Casale et al. (2008), however, observed high mortality rates for loggerhead turtles that had swallowed the hook. Thus, mortality rates are likely to differ greatly with regard to hooking position. Measures tested to reduce sea turtle mortality include hook types and bait types that are less efficient in catching turtles, setting at depths beyond the diving depth of turtles, and night setting to prevent turtles from seeing the baited hook. Gilman et al. (2005) have reviewed published studies on the subject as well as others that are planned or in progress. The use of circle hooks to reduce the mortality of sea turtles has been reviewed by Read (2007). Because most research has begun only recently and has not yet been peer-reviewed or published, our knowledge of the efficacy of the proposed mitigation measures is still limited. The most comprehensive work on development of mitigation measures to reduce sea turtle bycatch in longlining is a study of a U.S. pelagic swordfish fishery in the northwestern Atlantic (Watson et al. 2005). Traditionally, J-hooks baited with squid were used in this fishery, until it was closed in 2001 due to interactions with loggerhead and leatherback sea turtles. Watson et al. (2005) evaluated the effectiveness of circle hooks and mackerel bait with respect to reducing interactions with sea turtles and maintaining swordfish catch rates. Circle hooks with mackerel bait reduced bycatch of loggerhead turtle by 90%. Circle hooks with squid bait and J-hooks with mackerel bait significantly reduced loggerhead bycatch (by 86% and 71%, respectively). Mackerel bait, irrespective of hook type, reduced the bycatch of leatherback turtle by about 65%, and circle hooks with squid bait reduced leatherback catch by 57%. The catch rate of swordfish rose when mackerel was used as bait (63% with J-hooks and 30% with circle hooks), whereas the swordfish catch rate was reduced by about 30% using circle hooks with squid bait. In the tuna longline fishery in Hawaii, the use of saury as bait significantly reduced sea turtle bycatch compared with the use of squid as bait (Pradhan and Leung 2006). In conclusion, sea turtle interactions associated with the western Atlantic pelagic swordfish longline
127
fishery can be significantly reduced by using circle hooks or mackerel bait, and when used in combination, a reduction in bycatch of 90% for loggerheads and 65% for leatherbacks can be obtained with an increase in target catch rate. Implementation of this fishing technique made it possible to reopen both this fishery and the Hawaii-based pelagic longline fishery in 2004. Furthermore, few loggerheads caught on circle hooks swallowed the hooks (27%), whereas the majority of the loggerheads caught on J-hooks ingested the hooks (69%) (Watson et al. 2005). Also, Brazner and McMillan (2008) observed that lower proportions of circle hooks were swallowed by loggerheads compared with J-hooks. Swimmer et al. (2006) showed that lightly hooked Olive Ridley turtles (nine turtles) all survived their encounter with shallow-set longline gear using circle hooks. Thus, postrelease mortality is likely to be lower for individuals caught on circle hooks than for those caught on J-hooks. Leatherback turtles were usually hooked externally or entangled in the lines. Read (2007) concluded that circle hooks would significantly reduce sea turtle mortality. In an experiment studying sea turtle bycatch in the fishery targeting swordfish in the western Mediterranean Sea, 93% of the loggerhead specimens were caught on longlines during daytime soak, while swordfish captures were independent of retrieval time (Baez et al. 2007). Watson et al. (2005) found that loggerhead turtle catches increased with increasing soak time. As swordfish longlines were set at sunset and retrieved in the morning, increased soak time implies an increase in daylight soak time. Although the effect of daylight soak time was inconclusive, the authors suspected that there was a confounding effect between total soak time and daylight soak time. Also, in the swordfish-targeted longline fishery in Hawaii, incidental catches of loggerhead turtles increased with increasing soak time (Pradhan and Leung 2006). There was no effect of soak time on leatherback catch rates in these studies, indicating a difference in the time of day at which these two sea turtle species are most likely to interact with baited longlines. Night setting could thus be an efficient way of reducing loggerhead bycatch. Sea surface temperature also affects sea turtle bycatch (Watson et al. 2005). The loggerhead catch
128
Fish Behavior near Fishing Gears during Capture Processes
rate increased 200% to 350% with every 2.8°C increase in surface temperature, with a somewhat lower increase for leatherback. Fishing cooler water did not seem to affect swordfish catch rates when mackerel was used as bait. Analysis of observer and landings data from the Canadian pelagic longline fishery indicated that bycatch of loggerheads was concentrated above 22°C and catch of target species (tuna and swordfish) peaked between 17° and 18°C (Brazner and McMillan 2008). Dying the bait blue has been proposed and tested as a means of mitigating interactions with sea turtles. However, Swimmer et al. (2005) found no differences in rates of interactions with Olive Ridley turtle (8.4 and 8.1 individuals per 1000 hooks) when using untreated versus blue-dyed squid baits. On the basis of these results and other studies, the authors concluded that dying bait blue, which has been shown to reduce interactions with seabirds (see Section 5.2), is not effective in reducing sea turtle bycatch. Setting pelagic longlines deeper has been shown to decrease sea turtle bycatch (Brazner and McMillan 2008; Pradhan and Leung 2006). Analysis of observer data from the Canadian pelagic longline fishery showed that no loggerheads were captured when hooks were set at depths greater than 40 m (Brazner and McMillan 2008). Observer data collected by the Secretariat of the Pacific Community indicated that deep-set longlines reduced sea turtles catches by one order of magnitude compared with shallow-set gear, and turtles caught on deep-set gear were taken on the shallowest hooks (SPREP 2001). Similarly, observer data from Hawaii showed that loggerhead turtles were caught only by shallow longlines targeting swordfish and not by longlines set deep to target bigeye tuna (Polovina et al. 2003). In pelagic tuna longlining, the mainline takes the shape of a catenary curve, which gives a wide range in depth between the shallowest hooks close to the float lines and the deepest hooks in the middle of a basket (Suzuki et al. 1977). Shiode et al. (2005) developed a new method to set all branch lines at almost the same depth, to reduce sea turtle bycatch. The mainline was kept almost horizontal by using midwater floats to lift the section of the mainline between adjacent main floats. This setting method, with only 5-m difference in depth between the shal-
lowest and deepest hooks, may prove to be effective in reducing sea turtle bycatch when the mainline is set deeper and allows a more precise adjustment of the depth of the baited hooks to the swimming depth of the target fish (Shiode et al. 2005). Recently, Shiga et al. (2008) further developed the midwater float system by using two (double) midwater floats. An alternative approach to ensure that all hooks fished at depths greater than 100 m has been developed by Beverly et al. (2009), who suspended the fishing portion of the mainline on long, weighted float lines. 5.6.4 Mitigation Measures Aimed at Reducing Incidental Catches of Sharks The slow growth, late maturity, and low fecundity of sharks make them extremely vulnerable to even modest levels of fishing. Although many regions lack a pelagic fishery that specifically targets sharks (Yokota et al. 2006), other longline fisheries may be a great threat to shark stocks because species with higher production rates, such as swordfish and tuna, continue to support the fishery (Megalofonou et al. 2005). Most studies that deal with shark stock conservation have focused on monitoring and analyzing incidental catch and discards in pelagic fisheries (e.g., Francis et al. 2001; Marin et al. 1998; Megalofonou et al. 2005), whereas to date, few efforts have been made to develop and test measures to prevent shark capture. The most important gear parameters that determine species and size selection in longline fishing are bait type and hook type, of which bait type is the more important factor affecting species selectivity (Løkkeborg and Bjordal 1992). An artificial bait using squid liver (a waste product of the industry) was developed for tuna longlining and tested off the Hawaiian Islands (Januma et al. 2003). This bait significantly reduced shark bycatch, with catch rates that on average were 67% lower than with traditional squid bait. Although catches of tunas were also somewhat lower with the artificial bait, the difference in target catch rates between the two bait types was not significant. In the study carried out in the U.S. Atlantic swordfish fishery, mackerel bait, which proved efficient in reducing sea turtles bycatch (see Section 5.3), was also shown to reduce the catch of blue sharks (30% to 40%) compared
Fish Behavior in Relation to Longlines with squid bait (Watson et al. 2005). In the Hawaiibased swordfish fishery, regulations requiring vessels to switch from using a J-shaped hook with squid bait to a wider circle-shaped hook with fish bait led to a 36% decline in shark catch rate (Gilman et al. 2007). These studies thus confirm the speciesselective effect of bait type, suggesting that there is great potential for reducing shark bycatch by using alternative bait types. Elasmobranches are the most important bycatch species in the semipelagic longline fishery for hake on the coast of the Algarve (southern Portugal), and the effects of removing the lower hooks (i.e., those near the bottom) were evaluated in terms of bycatch reductions and target catch rates (Coelho et al. 2003). Most sharks were caught on hooks near the bottom. These hooks caught very few hake, which were mainly taken by the middle range of the hooks. These results thus indicate that the removal of the lower hooks would result in a significant reduction in shark bycatch with only a small reduction in target catch. Furthermore, the lower hooks often become entangled in the bottom substrate, so that removing them will also offer benefits in terms of handling, gear loss, and bait costs (Coelho et al. 2003). Experiments conducted off northeastern Australia on commercial pelagic longline vessels targeting tuna and billfish (Istiophoridae and Xiiphidae) found lower catch rates of sharks on nylon than on wire leaders (Ward et al. 2008). Sharks may bite through the nylon leaders and escape. Higher catch rates of tuna on nylon than on wire leaders were explained by higher visibility of wire leaders. Although shark caught on pelagic longlines tend to be in good condition and predicted postrelease survival may be high (Francis et al. 2001; Moyes et al. 2006; Ward et al. 2008), this mitigation method has potential to inflict injury and unaccounted mortality in sharks. Information on survivorship is difficult to evaluate (Skomal 2007), and postrelease mortality is evident and survival enhancement in even the more resilient shark species is advocated (Mandelman et al. 2008). Soak time may affect both catch rate and survival rate. Blue shark catch rates were found to increase with increasing soak time (Ward et al. 2004), and increased soak times led to increases in the propor-
129
tion of individuals retrieved dead (Diaz and Serafy 2005). Although shortening longline soak times might be a way of reducing blue shark mortality, this measure would probably be unacceptable to the fishermen and thus difficult to implement because the catches of swordfish would also be lowered with shorter soak times (Ward et al. 2004). The implementation of mitigation measures in longline fisheries (e.g., circle hooks to reduce sea turtle bycatch in the U.S. swordfish fishery; see Section 5.3) requires examination of the effects of such measures on catching efficiency for other nontarget species. Yokota et al. (2006) conducted fishing experiments off the coast of Japan to test the effects of circle hooks on blue shark (Prionace glauca) catches in pelagic longlining and found no significant differences in catch rates or proportion of dead individuals between circle hooks and conventional tuna hooks. In the north-western Atlantic, blue shark catch rates were 8% to 9% higher using circle hooks than when using J-hooks (Watson et al. 2005). Using circle hooks therefore seems to have no effect on reducing blue shark bycatch. However, circle hooks significantly reduced gut hooking in blue sharks compared with J-hooks; presumably post-hooking survival rates also increased with circle hooks (Watson et al. 2005). In most longline fisheries, pelagic sharks are primarily nontarget species and are discarded (Yokota et al. 2006), and it is therefore important to develop practices to minimize injury and postrelease mortality. We may thus conclude that there are few comprehensive studies aimed at the development of effective measures to reduce shark bycatch. More research is needed before appropriate measures can be implemented in the management of longline fisheries where shark bycatch is a problem. Elasmobranches make up a different group of animals than the species targeted in these fisheries, and they have a different feeding behavior with regard to sensory modalities involved, search strategy, prey preference, and diel rhythm. Several species of sharks are capable of detecting magnetic fields (Kalmijn 1971), and the 2006 winner of the “Smart Gear” competition suggested taking advantage of this unique sensory modality in sharks to deter them from taking baited hooks (see http://www.smartgear.org). A first step would be to study in depth
130
Fish Behavior near Fishing Gears during Capture Processes
how these animals search for and capture their food. 5.7 CONCLUDING REMARKS The responses of fish to baited hooks, and how their behavior is influenced by physical and biological variables, affect capture efficiency and size and species selectivity in longline fishing. An overview of the interactions between fish behavior and environmental and gear-related variables is shown in Figure 5.7. Understanding these processes is a prerequisite for improving the accuracy of stock size estimates when catch data from baited gears in resource assessment work are used (Løkkeborg
et al. 1995; Stoner 2004). Also, these processes affect the outcome of commercial fishing operations, and knowledge of how fish react to baited gears is therefore also of great importance for the sustainable harvesting and management of fish stocks. Finally, such information may also help develop mitigation measures for reducing incidental catches of seabirds, sea turtles, and nontarget fish without loss of target catch. Longlining is regarded as a size-selective fishing method, and several behavioral aspects explain why baited gears are more selective than, for example, trawling. Large fish are capable of exploiting food resources more efficiently than small fish, a situa-
Light INTERNAL AND EXTERNAL VARIABLES Ch. 5.2 - 5.5
- Hunger state - Reproductive status - Diel rhytms - Previous experienses
Temperature Current Prey density
STIMULI Ch. 4
VISUAL
OLFACTORY
GUSTATORY / MECHANICAL taste size
type GEAR VARIABLES Ch. 3
Bait texture size
Bait
Attractant concentration
shape
Hook design
prey
OBJECT CATEGORIZATION Ch. 5.1
Object
Object nonedible
predator prey nonedible
prey Object nonedible Escap (compe e titi
Escap e
) on
Search (olfactory rheotaxis) Ig
Attack Ig
est
Ing
Hooked
re no
re no
w
im
Locate
ay
Detect
aw
RESPONSES Ch. 5
S
Figure 5.7. Fish behavior to baited hooks. Variables (internal and external) that affect fish behavior but are not related to the baited gear are shown above the broken line. Stimuli, gear variables, and object categorization and the corresponding behavioral responses are shown at different distances down-current of the baited hooks. Grey and color shades illustrate odor plume and visual range, respectively.
Fish Behavior in Relation to Longlines tion that leads to exploitative competition for baits on a longline (Stoner 2004; Section 5.4). Thus, before the fish come into contact with the gear, a selection process has already taken place that exposes a high proportion of large individuals to the gear (Løkkeborg and Bjordal 1992). Interestingly, seismic air-gun noise has been shown to cause greater reductions in longline catches of large than small cod (Engås et al. 1996b). The stronger response of larger fish was explained by size-dependent swimming capability. There is also competition among fish that encounter baited longlines (i.e., interference competition) (Stoner 2004). In the seismic experiment by Engås et al. (1996b), longline catches of small cod were even shown to increase during seismic shooting, indicating that small fish were more successful in taking the available baits when their larger conspecifics had left the area. Furthermore, the characteristics and visual appearance of the gear may affect small and large individuals differently (see Section 5.1). Baited hooks are a novel food item, and because smaller fish have less experience concerning prey types and more limited diet breathe, their restraint in attacking a novel prey may be more pronounced (Løkkeborg 1990b). There is also a relationship between predator and prey size (Hart 1993; Werner 1974). Last, the hooking probability may be length dependent (Kaimmer 1999). Selective capture of large over small individuals on baited hooks has been demonstrated in several fishing experiments (Bertrand 1988; Engås et al. 1996a; Hamley and Skud 1978; Hovgård and Riget 1992; Huse et al. 1999; 2000). The factors that affect size selectivity are also likely to affect species selection. Large species swim faster and are more successful competitors than small species, and Pacific halibut have been shown to be more efficient than other species in competing for available baits (Skud 1978). Interspecific competition for baits has also been observed in other field studies (Fernö et al. 1986; Godø et al. 1997; Løkkeborg and Bjordal 1992). Perhaps the most important factor that affects species selectivity is bait type (Løkkeborg and Bjordal 1992). There is a very large body of literature demonstrating species-specific preference for feeding attractants, and comparative fishing experiments with different baits have demonstrated clear
131
effects on species composition (see Section 5.3). Fishermen use saithe when targeting wolffish in the Barents Sea and have experienced negligible catches when their longlines have happened to be set on fishing grounds with a low abundance of wolfish, indicating that other species (e.g., cod and haddock) show low preferences for saithe bait. Prey-size preferences may differ among species, and smaller baits resulted in significant increases in catch rates for haddock (120%) but no increase for large cod (Johannessen 1983). In addition to the factors that affect selectivity, longline catching efficiency is affected by gear parameters such as size and type of hook, swivel (versus traditional snood attachment), and mainline and snood materials, as well as operational factors such as setting time, fishing depth (pelagic lines), season, and setting direction relative to current direction. As these factors have pronounced effects on catch rates, they can and must be standardized when using catch per unit effort (CPUE) data from longline surveys as an index of stock size. In the fishery for Pacific halibut, there has been an almost total conversion from J-hooks to circle hooks, and this required adjustments to the CPUE data used for stock assessment as the circle hook is 2.2 times as efficient (Quinn et al. 1985). However, some environmental variables affecting catchability are not easy to measure or standardize. Stoner (2004) provides an interesting review of how physical and biological conditions in the environment influence fish activity, feeding motivation, searching behavior, and bait location and concludes that temperature, light level, current velocity, and prey density have the greatest effects on longline catchability, potentially affecting variation in CPUE by a factor of 10. Target species can occupy wide ranges of these environmental conditions over time and space, and longline survey data used for stock assessments may thus simply reflect variations in fish behavior and catchability rather than trends in fish abundance (Stoner 2004). During the past few years, promising results and important technical advances have made longline fishing more conservation oriented, particularly regarding incidental catches of seabirds. Mitigation measures such as streamer lines, weighted lines, and night setting have nearly eliminated seabird
132
Fish Behavior near Fishing Gears during Capture Processes
bycatch in important demersal fisheries and greatly reduced incidental catches in most pelagic fisheries. Thus, the implementation of appropriate and efficient measures and their compliance have the potential to contribute to negligible longlineinduced mortality in most seabird populations. Measures to reduce incidental captures of sea turtles, such as the circle hook, using finfish as bait, and deep setting, have proved efficient in protecting these endangered species. However, there is no single solution to the problem of incidental catches of sea turtles, and there are great variations in gear construction and mode of operating pelagic longlines. Reported experiments have been carried out only in a few fisheries, and our knowledge of the effects of proposed mitigation measures on sea turtle bycatch and target species catch is too limited to draw firm conclusions except in the case of a few fisheries. Information on how to reduce shark bycatch is very scarce. In general, bait type is the most important gear parameter affecting longline species selectivity (Løkkeborg and Bjordal 1992), and findings suggest that this also applies to sharks (Watson et al. 2005). There are also indications that magnetic fields could be used to deter sharks from baited hooks.
5.8 FUTURE CHALLENGES Ecologically sustainable management for fisheries and conservation-oriented harvesting practices is a topic of growing interest, and consumers are becoming more concerned about how their food has been produced. Several factors that affect species and size selectivity in longlining were discussed in this chapter. Identifying the gear parameters and operational strategies with the greatest potential for selective fishing should be given priority. Combining the most effective factors should improve longline selectivity and thus make this fishing method more sustainable. Gear research carried out in the course of the past few decades has considerably improved the catching efficiency of longlines. One example is the conversion from J-hooks to circle hooks in demersal fisheries. These two hook types are very different in design, and there is probably potential for further improvement through minor changes in hook
design. Choice of bait type is mainly based on fishermen’s experiences, availability, and price, and few comparative fishing experiments have tested the catch efficiency and species selectivity of different natural bait types. Thus, experiments designed to compare different bait types are likely to reveal a potential for improved efficiency and selectivity. This is supported by findings from the study carried out in the western Atlantic swordfish fishery, where mackerel bait was tested as a mitigation measure for sea turtles. Mackerel bait produced 63% higher catch rates of swordfish compared with traditional squid bait in the experiments with J-hooks (Watson et al. 2005). More research should also be aimed at the development of artificial baits based on surplus products or waste materials, as most baits used today are made from resources that could be better used for human consumption. Attention should be paid to developing long-lasting baits, as the release rate of attractants from natural baits decreases rapidly (Løkkeborg 1990a). Furthermore, research should aim to optimize aspects of operational fishing strategy, such as time of setting in relation to diel and tidal cycles, soak time, setting direction relative to the current, and fishing depth. The main conservation issue related to longline fishing is incidental captures of threatened and endangered species such as seabirds, turtles, and sharks. Efficient mitigation measures to reduce incidental captures have been developed for many fisheries, and promising results have been obtained for others. With respect to seabirds, future research should use an experimental approach to fine-tuning the most promising mitigation measures for each specific fishery as there is no single solution to reducing incidental mortality in longline fisheries. These measures include streamer lines, night setting, weighted longlines, and side-setting. Research aimed at reducing sea turtle mortality in longlining is in its infancy, although promising results have already been obtained in the western Atlantic swordfish fishery. Studies should be carried out to test these findings (e.g., circle hooks, finfish as bait, weighted leaders, deep sets, reduced daylight soak time, avoidance of areas of high turtle densities) in other fisheries where interactions with sea turtles are a problem. New experiments are
Fish Behavior in Relation to Longlines essential because, as with seabird interactions, there are differences between fishing grounds in species assemblages, longline gear design, and operational strategies, and best fishing practices to reduce incidental captures will differ accordingly. Although studies aimed at reducing shark bycatch are few, much can be learned from our general knowledge of longline species selectivity. Modern statistical methods (e.g., generalized additive models) and commercial fisheries data can be used to examine and estimate the relative influence of various factors on catch rates, and these models may be used to identify the spatiotemporal, environmental, and operational variables that have the greatest potential for reducing shark bycatch while maintaining profitable catch rates of target species (see Bigelow et al 1999; Walsh and Kleiber 2001). A better understanding of how fish respond to baited hooks is essential for improving the selectivity, efficiency, and conservation aspects of longline fishing. There are a few fisheries, such as the Barents Sea pelagic fishery for haddock, that take high proportions of undersized fish (Huse and Soldal 2000; Løkkeborg and Bjordal 1995), and better understanding of fish behavior may aid in solving this problem. Improved stock size estimation models are perhaps the issue that could gain most from more research on fish behavior in response to baited gear. Stoner (2004) concludes that several environmental variables can have a significant impact on the feeding behavior of fish and produce serious biases in stock assessment surveys carried out using baited gear. To produce unbiased stock size estimates, it is essential to understand relationships between the environment and feeding biology and how physical and biological variables affect fish activity and responsiveness to bait. Such information can be obtained by linking continuous recordings of environmental variables with behavioral observations using acoustic and archival tags, optical and acoustic cameras, acoustic surveys, and laboratory facilities (Stoner 2004). ACKNOWLEDGMENTS We are grateful to Anne-Britt Skar Tysseland for preparing the figures. We also thank Alan W. Stoner for valuable comments that greatly improved the manuscript.
133
REFERENCES Adron JW and Mackie AM. 1978. Studies on the chemical nature of feeding stimulants for rainbow trout, Salmo gairdneri Richardson. J. Fish Biol. 12: 303–310. Agnew DJ, Black AD, Croxall JP and Parkes GB. 2000. Experimental evaluation of the effectiveness of weighting regimes in reducing seabird by-catch in the longline toothfish fishery around South Georgia. CCAMLR Sci. 7: 119–131. Allen KR. 1963. The influence of behavior on the capture of fishes with baits. In: The Selectivity of Fishing Gear. Int. Comm. Northwest Atl. Fish., Spec. Publ. 5: 5–7. Dartmouth, NS, Canada. Ashford JR, Croxall JP, Rubilar PS and Moreno CA. 1995. Seabird interactions with longlining operations for Dissosticus eleginoides around South Georgia, April to May 1994. CCAMLR Sci. 2: 111–121. Atema J. 1980. Chemical senses, chemical signals and feeding behavior in fishes. In: Bardach JE, Magnuson JJ, May RC and Reinhart JM (eds). Fish Behavior and its Use in the Capture and Culture of Fishes. pp 57–101. Manila: International Center for Living Aquatic Resources Management. Baez JC, Real R, Caminas JA and Juan A. 2007. Differential distribution within longline transects of loggerhead turtles and swordfish captured by the Spanish Mediterranean surface longline fishery. J. Mar. Biol. Assoc. U.K. 87: 801–803. Bardach JE,Todd JH and Crickner R. 1967. Orientation by taste in fish of the genus Ictalurus. Science. 155: 1276–1278. Belghaug R and Døving KB. 1977. Odor threshold determined by studies of the induced waves in the olfactory bulb of the char (Salmo alpinus L.). Comp. Biochem. Physiol. 57(A): 327–330. Bertrand J. 1988. Selectivity of hooks in the handline fishery of the Saya de Malha banks (Indian Ocean). Fish. Res. 6: 249–255. Bertrand A, Josse E, Bach P, Gros P and Dagorn L. 2002. Hydrological and trophic characteristics of tuna habitat: consequences on tuna distribution and longline catchability. Can. J. Fish. Aquat. Sci. 59: 1002–1013. Beukema JJ. 1970. Angling experiments with carp: decreased catchability through one trial learning. Neth. J. Zool. 20: 81–92. Beverly S, Curran D, Musyl M and Molony B. 2009. Effects of eliminating shallow hooks from tuna longline sets on target and non-target species in the Hawaii-based pelagic tuna fishery. Fish. Res. 96: 281–288.
134
Fish Behavior near Fishing Gears during Capture Processes
Bigelow KA, Boggs CH and He X. 1999. Environmental effects on swordfish and blue shark catch rates in the US North Pacific longline fishery. Fish. Oceanogr. 8(3): 178–198. Bjordal Å. 1983. Effect of different long-line baits (mackerel, squid) on catch rates and selectivity for tusk and ling. ICES CM/B: 31. 9 pp. Bjordal Å and Løkkeborg S. 1996. Longlining. Oxford: Fishing News Books. 156 pp. Boggs CH 2001. Deterring albatrosses from contacting baits during swordfish longline sets. In: Melvin EF and Parrish JK (eds). Seabird Bycatch: Trends, Roadblocks, and Solutions. pp 79–94. AK-SG01-01. Fairbanks, AK: University of Alaska Sea Grant. Bossert WH and Wilson EO 1963. The analysis of olfactory communication among animals. J. Theor. Biol. 5: 443–469. Brawn VM. 1969. Feeding behavior of Cod (Gadus morhua). J.Fish. Res. Bd Can. 26: 583–596. Brazner JC and McMillan J. 2008. Loggerhead turtle (Caretta caretta) bycatch in the western North Atlantic and opportunities for mitigation. Fish. Res. 91: 310–324. Brothers N. 1991. Albatross mortality and associated bait loss in Japanese longline fishery in the Southern Ocean. Biol. Conserv. 55: 255–268. Brothers NP, Cooper J and Løkkeborg S. 1999a. The incidental catch of seabirds by longline fisheries: worldwide review and technical guidelines for mitigation. FAO Fish. Circ. 937: 100 pp. Brothers N, Gales R and Reid T. 1999b. The influence of environmental variables and mitigation measures on seabird catch rates in the Japanese tuna longline fishery within the Australian Fishing Zone, 1991– 1995. Biol. Conserv. 88: 85–101. Brown C, Laland K and Krause J. 2006. Fish Cognition and Behavior. Fish and Aquatic Resources Series 11. Oxford: Blackwell Publishing. 328 pp. Bull LS. 2007. Reducing seabird bycatch in longline, trawl and gillnet fisheries. Fish & Fish. 8: 31–56. Caminas JA, Baez JC, Valeiras X and Real R. 2006. Differential loggerhead by-catch and direct mortality due to surface longlines according to boat strata and gear type. Sci. Mar. 70: 661–665. Caprio J. 1978. Olfaction and taste in the channel catfish: an electrophysiological study of the responses to amino acids and derivatives. J. Comp. Physiol. 123: 357–371. Carr WES. 1982. Chemical stimulation of feeding behavior. In: Hara TJ (ed). Chemoreception in Fishes. pp 259–273. Amsterdam: Elsevier.
Carr WES and Derby CD. 1986. Chemically stimulated feeding behavior in marine animals. J. Chem. Ecol. 12: 989–1011. Carr WES, Netherton JC III, Gleeeson RA and Derby CD. 1996. Stimulants of feeding behavior in fish: analysis of tissue of diverse marine organisms. Biol. Bull. 190: 149–160. Casale P, Freggi D and Rocco M. 2008. Mortality induced by drifting longline hooks and branchlines in loggerhead sea turtles through observation in captivity. Aquat. Conserv. Mar. Freshw. Ecosys. 18: 945–954. Castonguay M and Cyr DG. 1998. Effects on temperature on spontaneous and thyroxine-stimulated locomotor activity of Atlantic cod. J. Fish Biol. 53: 303–313. Chandroo KP, Duncan IJH and Moccia RD. 2004. Can fish suffer? Perspectives on sentience, pain, fear and stress. App. Anim. Behav. Sci. 86: 225– 250. Cherel Y, Weimerskirch H and Duhamel G. 1996. Interaction between longline vessels and seabirds in Kerguelen waters and a method to reduce seabird mortality. Biol. Conserv. 75: 63–70. Coelho R, Bentes L, Goncalves JMS, Lino PG, Ribeiro J and Erzini K. 2003. Reduction of elasmobranch by-catch in the hake semipelagic near-bottom longline fishery in the Algarve (Southern Portugal). Fish. Sci. 69: 293–299. Croxall JP, Rothery P, Pickering S and Prince PA. 1990. Reproductive performance, recruitment and survival of Wandering Albatrosses Diomedea exulans at Bird Isand, South Georgia. J. Anim. Ecol. 59: 775–796. Davis MW, Spencer ML and Ottmar ML. 2006. Behavioral responses to food odor in juvenile marine fish: acuity varies with species and fish size. J. Exp. Mar. Biol. Ecol. 328: 1–9. Deflorio M, Aprea A, Corriero A, Santamaria N and Metrio G. 2005. Incidental captures of sea turtles by swordfish and albacore longlines in the Ionian sea. Fish. Sci. 71: 1010–1018. Dempsey CH. 1978. Chemical stimuli as a factor in feeding and intraspecific behavior of herring larvae. J. Mar. Biol. Assoc. U.K. 58: 739–747. Diaz GA and Serafy JE. 2005. Longline-caught blue shark (Prionace glauca): factors affecting the numbers available for live release. Fish. Bull. 103: 720–724. Dietrich KS, Melvin EF and Conquest L. 2008. Integrated weight longlines with paired streamer lines- Best practice to prevent seabird bycatch in
Fish Behavior in Relation to Longlines demersal longline fisheries. Biol. Conserv. 141: 1793–1805. Elkinton JS, Cardé RT and Mason CJ. 1984 Evaluation of time-average dispersion models for estimating pheromone concentration in a deciduous forest. J. Chem. Ecol. 10: 1081–1108. Ellingsen OF and Døving KB 1986. Chemical fractionation of shrimp extracts inducing bottom food search behavior in cod (Gadus morhua L.). J. Chem. Ecol. 12: 155–168. Engås A and Løkkeborg S. 1994. Abundance estimation using bottom gillnet and longline—the role of fish behavior. In: Fernö A. and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 134–165. Oxford: Fishing News Books. Engås A, Løkkeborg S, Ona E and Soldal AV. 1996b. Effects of seismic shooting on local abundance and catch rates of cod (Gadus morhua) and haddock (Melanogrammus aeglefinus). Can. J. Fish. Aquat. Sci. 53: 2238–2249. Engås A, Løkkeborg S, Soldal AV and Ona E. 1996a. Comparative fishing trials for cod and haddock using commercial trawl and longline at two different stock levels. J. Northw. Atl. Fish. Sci. 19: 83– 90. Fernö A. 1993 Advances in understanding of basic behavior: consequences for fish capture studies. ICES Mar. Sci. Symp. 196: 5–11. Fernö A and Huse I. 1983. The effect of experience on the behavior of cod (Gadus morhua L.) toward a baited hook. Fish. Res. 2: 19–28. Fernö A, Huse G, Jakobsen PJ and Kristiansen TS. 2006. The role of fish learning skills on fisheries and aquaculture. In: Brown C, Laland K and Krause J (eds). Fish Cognition and Behavior. pp 278–310. Oxford: Blackwell Publishing. Fernö A, Solemdal P and Tilseth S. 1986. Field studies on the behavior of whiting (Gadus merlangus L.) toward baited hooks. Fisk Dir. Skr. Ser. Hav Unders. 18: 83–95. Francis MP, Griggs LH and Baird SJ. 2001. Pelagic shark bycatch in the New Zealand tuna longline fishery. Mar. Fresh. Res. 52: 165–178. Franco J, Johnstone ADF and Mackie AM. 1991. Studies of bait preference in the cod, Gadus morhua L.: characterization of feeding stimulants using an operant conditioning technique. Fish. Res. 10: 229–242. Fry FEJ. 1971. The effect of environmental factors on the physiology of fish. In: Hoar WS and Randall DJ (eds). Fish Physiology, Vol. 6. Environmental
135
Relations and Behavior. pp 1–98. San Diego, CA: Academic Press. Gilman E, Boggs C and Brothers N. 2003. Performance assessment of an underwater setting chute to mitigate seabird bycatch in the Hawaii pelagic longline tuna fishery. Ocean Coast. Manag. 46: 985–1010. Gilman E, Brothers N and Kobayashi DR. 2005. Principles and approaches to abate seabird by-catch in longline fisheries. Fish Fish. 6: 35–49. Gilman E, Kobayashi D, Swenarton T, Brothers N, Dalzell P and Kinan-Kelly I. 2007. Reducing sea turtle interactions in the Hawaii-based longline swordfish fishery. Biol. Conserv. 139: 19–28. Gislason H. 1994. Ecosystem effects of fishing activities in the North Sea. Mar. Pollut. Bull. 29: 520–527. Godø OR, Huse I and Michalsen K. 1997. Short communication: Bait defence behavior of wolffish and its impact on long-line catch rates. ICES J. Mar. Sci. 54: 272–275. Goh Y, Tamura T and Kobayashi H. 1979. Olfactory responses to amino acids in marine teleosts. Comp. Biochem. Physiol. 62(A): 863–868. Hamley JM and Skud BE. 1978. Factors affecting longline catch and effort: II. Hook spacing. Int. Pac. Halib. Comm. Sci. Rep. 64: 15–24. Hara TJ. 1972. Electrical responses of the olfactory bulb of Pacific salmon, Oncorhynchus nerka and Oncorhynchus kisutch. J. Fish. Res. Bd. Can. 29: 1351–1355. Hara TJ. 1973. Olfactory responses to amino acids in rainbow trout, Salmo gairdneri. Comp. Biochem. Physiol. 44A: 407–416. Hara TJ. 1975. Olfaction in fish. Prog. Neurobiol. 5: 271–335. Hara TJ. 1993. Role of olfaction in fish behavior. In: Pitcher TJ (ed). Behavior of Teleost Fishes. pp 171–199. London: Chapman and Hall. Hara TJ. 1992. Fish Chemoreception. London: Chapman and Hall. 373 pp. Hart PJB. 1993. Teleost foraging: facts and theories. In: Pitcher TJ (ed). Behavior of Teleost Fishes. pp 253–284. London: Chapman and Hall. Hart PJB and Connellan B. 1984. The cost of prey capture, growth rate and ration size in pike (Esox lucius) as functions of prey weight. J. Fish Biol. 25: 279–291. He P. 1996. Bait loss from bottom-set longlines as determined by underwater observations and comparative fishing trials. Fish. Res. 27: 29–36. He P. 2003. Swimming behaviour of winter flounder (Pleuronectes americanus) on natural fishing
136
Fish Behavior near Fishing Gears during Capture Processes
grounds as observed by underwater video camera. Fish. Res. 60: 507–514. Heppell SS, Heppell SA, Read AJ and Crowder LB. 2005. Effects of fishing on long-lived marine organisms. In: Norse E and Crowder L (eds). Marine Conservation Biology: The Science of Maintaining the Sea’s Biodiversity. pp 211–231. Washington, DC: Island Press. Hovgård H and Riget FF. 1992. Comparison of longline and trawl selectivity in cod surveys off West Greenland. Fish. Res. 13: 323–333. Huse I and Fernö A. 1990. Fish behavior studies as an aid to improved longline hook design. Fish. Res. 9: 287–297. Huse I and Soldal AV. 2000. An attempt to improve size selection in pelagic longline fisheries for haddock. Fish. Res. 48: 43–54. Huse I, Gundersen AC and Nedreaas KH. 1999. Relative selectivity of Greenland halibut (Reinhardtius hippoglossoides, Walbaum) by trawls, longlines and gillnets. Fish. Res. 44: 75–93. Huse I, Løkkeborg S and Soldal A. 2000. Relative selectivity in trawl, longline and gillnet fisheries for cod and haddock. ICES J. Mar. Sci. 57: 1271–1282. Imai T. 1972. Studies on the several raw fish baits in a tuna long-line. Fishing-I. Mem. Fac. Fish., Kagoshima Univ. 21: 45–50. Imai T and Shirakawa O. 1972. Studies on the several raw fish baits in tuna long-line. Fishing-II. Mem. Fac. Fish., Kagoshima Univ. 21: 51–62. Ishida Y and Kobayashi H. 1992. Stimulatory effectiveness of amino acids on the olfactory response in an algivorous marine teleost, the rabbitfish Siganus fuscescens Houttuyn. J. Fish Biol. 41: 737–748. Januma S, Miyajima K and Abe T. 2003. Development and comparative test of squid liver artificial bait for tuna longline. Fish. Sci. 69: 288–292. Johannessen T. 1983. Influence of hook and bait size on catch efficiency and length selection in longlining for cod (Gadus morhua L.) and haddock (Melanogrammus aeglefinus L.). MSc. thesis, University of Bergen (in Norwegian). Johannessen T, Fernö A and Løkkeborg S. 1993. Behavior of cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) in relation to various sizes of long-line bait. ICES Mar. Sci. Symp. 196: 47–50. Johnstone ADF. 1980. The detection of dissolved amino acids by the Atlantic cod, Gadus morhua L. J. Fish Biol. 17: 219–230.
Johnstone ADF and Mackie AM. 1990. Laboratory investigations of bait acceptance by the cod, Gadus morhua L.: identification of feeding stimulants. Fish. Res. 9: 219–230. Jones KA. 1992. Food search behavior in fish and the use of chemical lures in commercial and sports fishing. In: Hara TJ (ed). Fish Chemoreception. pp 288–320. London: Chapman and Hall. Kaimmer SM. 1999. Direct observations on the hooking behavior of Pacific halibut, Hippoglossus stenolepis. Fish. Bull. 97: 873–883. Kalmijn AJ. 1971. The electric sense of sharks and rays. J. Exp. Biol. 55: 371–383. Kallayil JK, Jorgensen T, Engås A and Fernö A. 2003. Baiting gill nets—how is fish behavior affected? Fish. Res. 61: 125–133. Kasumyan AO and Døving KB. 2003. Taste preferences in fishes. Fish Fish. 4: 289–347. Klaer N and Polacheck T. 1998. The influence of environmental factors and mitigation measures on bycatch rates of seabirds by Japanese longline fishing vessels in the Australian region. Emu. 98: 305–316. Kotrschal K, van Staaden MJ and Huber R. 1998 Fish brains: evolution and environmental relationships. Rev. Fish Biol. Fish. 8: 373–408. Krebs JR. 1978. Optimal foraging: Decision rules for predators. In: Krebs JR and Davis NB (eds). Behavioral Ecology: An Evolutionary Approach. pp. 23–63. Oxford: Blackwell Scientific Publications. Laberge F and Hara TJ. 2001. Neurobiology of fish olfaction: a review. Brain Res. Rev. 36: 46–59. Lewison RL. Freeman SA and Crowder LB. 2004. Quantifying the effects of fisheries on threatened species: the impact of pelagic longlines on loggerhead and leatherback sea turtles. Ecol. Lett. 7: 221–231. Løkkeborg S. 1990a. Rate of release of potential feeding attractants from natural and artificial bait. Fish. Res. 8: 253–261. Løkkeborg S. 1990b. Reduced catch of under-sized cod (Gadus morhua) in longlining by using artificial bait. Can. J. Fish. Aquat. Sci. 47: 1112–1115. Løkkeborg S. 1991. Fishing experiments with an alternative longline bait using surplus fish products. Fish. Res. 12: 43–56. Løkkeborg S. 1994. Fish behavior and longlining. In: Fernö A and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 9–27. Oxford: Fishing News Books. Løkkeborg S. 1998. Feeding behavior of cod, Gadus morhua: activity rhythm and chemically mediated food search. Anim. Behav. 56: 371–378.
Fish Behavior in Relation to Longlines Løkkeborg S. 2001. Reducing seabird bycatch in longline fisheries by means of bird-scaring lines and underwater setting. In: Melvin EF and Parrish JK (eds). Seabird Bycatch: Trends, Roadblocks, and Solutions. pp 33–41. AK-SG-01-01. Fairbanks, AK: University of Alaska Sea Grant. Løkkeborg S. 2003. Review and evaluation of three mitigation measures—bird-scaring line, underwater setting and line shooter—to reduce seabird bycatch in the north Atlantic longline fishery. Fish. Res. 60: 11–16. Løkkeborg S. 2008. Review and assessment of mitigation measures to reduce incidental catch of seabirds in longline, trawl and gillnet fisheries. FAO Fish. Circ. 1040: 24 pp. Løkkeborg S and Bjordal Å. 1992. Species and size selectivity in longline fishing: a review. Fish. Res. 13: 311–322. Løkkeborg S and Bjordal Å. 1995. Size-selective effects of increasing bait size by using an inedible body on longline hooks. Fish. Res. 24: 273–279. Løkkeborg S and Fernö A. 1999. Diel activity pattern and food search behavior in cod, Gadus morhua. Environ. Biol. Fish. 54: 345–353. Løkkeborg S and Johannessen T. 1992. The importance of chemical stimuli in bait fishing— fishing trials with presoaked baits. Fish. Res. 14: 21–29. Løkkeborg S and Pina T. 1997. Effects of setting time, setting direction and soak time on longline catch rates. Fish. Res. 32: 213–222. Løkkeborg S and Robertson G. 2002. Seabird and longline interactions: effects of a bird-scaring streamer line and line shooter on the incidental capture of northern fulmars (Fulmarus glacialis). Biol. Conserv. 106: 359–364. Løkkeborg S, Bjordal Å and Fernö A. 1989. Responses of cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) to baited hooks in the natural environment. Can. J. Fish. Aquat. Sci. 46: 1478–1483. Løkkeborg S, Bjordal Å and Fernö A. 1993. The reliability and value of studies of fish behavior in longline gear research. ICES Mar. Sci. Symp. 196: 41–46. Løkkeborg S, Olla BL, Pearson WH and Davis MW. 1995. Behavioral responses of sablefish, Anoplopoma fimbria, to bait odor. J. Fish Biol. 46: 142–155. Løkkeborg S, Skajaa K and Fernö A. 2000. Foodsearch strategy in ling (Molva molva L.): crepuscular activity and use of space J. Exp. Mar. Biol. Ecol. 247: 195–208.
137
Mackie AM. 1982. Identification of the gustatory feeding stimulants. In: Hara TJ (ed). Chemoreception in Fishes. pp 275–291. Amsterdam: Elsevier. Mackie AM and Adron JW. 1978. Identification of inosine and inosine-5’-monophosphate as the gustatory feeding stimulants for the turbot, Scophthalmus maximus. Comp. Biochem. Physiol. 60(A): 79–83. Magnuson JJ, Bjorndal KA, Dupaul WA, Graham GL, Owens FW, Peterson CH, Pritchard PCH, Richardson JI, Saul GE and West CW. 1990. Decline of the Sea Turtles: Causes and Prevention. Washington, DC: National Academies Press. Magurran A, Seghers BH, Carbalho GR and Shaw PW. 1993. Evolution of adaptive variation in antipredator behavior. Mar. Behav. Physiol. 23: 29–44. Mandelman JW, Cooper PW, Werner TB and Lagueux KM. 2008. Shark bycatch and depredation in the U.S. Atlantic pelagic longline fishery. Rev. Fish Biol. Fish. 18: 427–442. Marin YH, Brum F, Barea LC and Chocca JF. 1998. Incidental catch associated with swordfish longline fisheries in the south-west Atlantic Ocean. Mar. Fresh. Res. 49: 633–639. Martin WR and McCracken FD. 1954. Relative efficiency of baits for groundfish. Progr. Rep. Atl. Biol. Sta. 126: 17–20. Marui T and Caprio J 1992. Teleost gustation. In: Hara TJ (ed). Fish Chemoreception. pp 171–198. London: Chapman and Hall. McCracken FD. 1963. Seasonal movements of the winter flounder, Pseudopleuronectes americanus, (Walbaum) on the Atlantic coast. J. Fish. Res. Bd Can. 20: 551–586. McMahon TE and Holanov SH. 1995. Foraging success of largemouth bass Micropterus salmoides at different light intensities: implications for time and depth of feeding. J. Fish Biol. 46: 759–767. McQuinn IH, Gendron L and Himmelman JH. 1988. Area of attraction and effective area fished by a whelk (Buccinum undatum) trap under variable conditions. Can. J. Fish. Aquat. Sci. 44: 2054–2060. Megalofonou P, Constantinos Y, Damalas D, De Metrio G, Deflorio M, de la Serna JM and Macias D. 2005. Incidental catch and estimated discards of pelagic sharks from the swordfish and tuna fisheries in the Mediterranean Sea. Fish. Bull. 103: 620–634. Melvin EF, Parrish JK, Dietrich KS and Hamel OS. 2001. Solutions to seabird bycatch in Alaska’s demersal longline fisheries. Washington Sea Grant Program. Project A/FP-7. WSG-AS 01-01.
138
Fish Behavior near Fishing Gears during Capture Processes
Michalsen K, Godø OR and Fernö A. 1996. Diel variation in the catchability of gadoids and its influence on the reliability of abundance indices. ICES J. Mar. Sci. 53: 389–395. Milinski M. 1993. Predation risk and feeding behavior. In: Pitcher TJ (ed). Behavior of Teleost Fishes. pp 285–305. London: Chapman and Hall. Moloney CL, Cooper J, Ryan PG and Siegfried WR. 1994. Use of population model to assess the impact of longline fishing on Wandering Albatrosses Diomedea exulans populations Biol. Conserv. 70: 195–203. Moreno CA, Rubilar PS, Marschoff E and Benzaquen L. 1996. Factors affecting the incidental mortality of seabirds in the Dissostichus eleginoides fishery in the southwest Atlantic (Subarea 48.3, 1995 season). CCAMLR Sci. 3: 79–91. Moyes CD, Fragoso N, Musyl MK and Brill RW. 2006. Predictiong postrelease survival in large pelagic fish. Trans. Am. Fish. Soc. 135: 1389–1397. Murray TE, Bartel JA, Kalish R and Taylor PR. 1993. Incidental capture of seabirds by Japanese bluefin tuna longline vessels in New Zealand waters, 1988– 1992. Biol. Conserv. 3: 181–210. Nel DC, Ryan PG and Watkins BP. 2002. Seabird mortality in the Patagonian toothfish longline fishery around Prince Edward Islands. Antarct. Sci. 14(2): 151–161. Nikonov AA, Ilyin YN, Zherelova OM and Fesenko EE. 1990. Odor thresholds of the Black Sea skate (Raja clavata). Electrophysiological study. Comp. Biochem. Physiol. 95(A): 325–328. Pawson MG. 1977. Analysis of a natural chemical attractant for whiting, Merlangius merlangus L. and cod, Gadus morhua L. using a behavioral bioassay. Comp. Biochem. Physiol. 56(A): 129–135. Pearson WH and Olla BL. 1977. Chemoreception in the blue crab, Callinectes sapidus. Biol. Bull. 153: 346–354. Pearson WH, Miller SE and Olla BL. 1980. Chemoreception in the food-searching and feeding behavior of the red hake, Urophycis chuss (Walbaum). J. Exp. Mar. Biol. Ecol. 48: 139–150. Peeling D and Rodgers DM. 1985. Circle hook comparison study. Fisheries and Oceans, Scotia-Fundy region, Halifax, Canada. Project Rep. 78: 16 pp. Polovina JJ, Howell E, Parker DM and Balazs GH. 2003. Dive-depth distribution of loggerhead (Caretta caretta) and olive ridley (Lepidochelys olivacea) sea turtles in the central North Pacific: might deep longline sets catch fewer turtles? Fish. Bull. 101: 189–193.
Poncet S, Robertson G, Phillips RA, Lawton K, Phalan B, Trathan PN and Croxall JP. 2006. Status and distribution of wandering, black-browed and greyheaded albatrosses breeding at South Georgia. Polar Biol. 29: 772–781. Pradhan NC and Leung P. 2006. Incorporating sea turtle interactions in a multi-objective programming model for Hawaii’s longline fishery. Ecol. Econ. 60: 216–227. Prince PA, Rothery P, Croxall JP and Wood AG. 1994. Population dynamics of Black-browed and Greyheaded Albatrosses Diomedea melanophrys and D. chrysostoma at Bird Island, South Georgia. Ibis 136: 50–71. Quinn TJ, Deriso RB and Hoag SH. 1985. Methods of Population Assessment of Pacific Halibut. Sci. Rep. IPHC. 72: 52 pp. Read AJ. 2007. Do circle hooks reduce the mortality of sea turtles in pelagic longlines? A review of recent experiments. Biol. Conserv. 135: 155– 169. Reid TA, Sullivan BJ, Pompert J, Enticott JW and Black AD. 2004. Seabird mortality associated with Patagonian Toothfish (Dissostichus eleginoides) longliners in Falkland Islands waters. Emu 104: 317–325. Robertson G, McNeill M, Smith N, Wienecke B, Candy S and Olivier F. 2006. The effectiveness of integrated weight (fast sinking) longlines in reducing the mortality of white-chinned petrels (Procellaria aequinoctialis) and sooty shearwaters (Puffinus griseus) in demersal longline fisheries. Biol. Conserv. 132: 458–471. Rodgveller CJ, Lunsford CR and Fujioka JT. 2008. Evidence of hook competition in longline surveys. Fish. Bull. 106(6): 364–374. Rolen SH, Sorensen PW, Mattson D and Caprio J. 2003. Polyamines as olfactory stimuli in the goldfish Carassius auratus. J. Exp. Biol. 206: 1683–1696. Rose CS, Stoner AW and Matteson K. 2005. Use of high-frequency imaging sonar to observe fish behavior near baited fishing gear. Fish. Res. 76: 291–304. Ryan PG and Watkins BP. 2002. Reducing incidental mortality of seabirds with an underwater longline setting funnel. Biol. Conserv. 104: 127–131. Ryer CH and Olla BL. 1992. Social mechanisms facilitating exploitation of spatially variable ephemeral food patches in a pelagic marine fish. Anim. Behav. 44: 69–74. Ryer CH and Olla BL. 1999. Light-induced changes in the prey consumption and behavior of two juve-
Fish Behavior in Relation to Longlines nile planktivorous fish. Mar. Ecol. Prog. Ser. 181: 41–51. Sainte-Marie B and Hargrave BT. 1987. Estimation of scavenger abundance and distrances of attraction to bait. Mar. Biol. 94: 431–443. Sasso CR and Epperly SP. 2007. Survival of pelagic juvenile loggerhead turtles in the open ocean. J. Wildl. Manag. 71: 1830–1835. Shiga M, Shiode D, Hayashi S, Tokai T and Hu F. 2008. Method for estimating buoyancy of midwater float required to standardize hook depth in pelagic longline. Fish. Sci. 74: 479–487. Shimada K and Tsurudome M. 1971. On the bait for tuna long-line-II. On the saury, mackerel and mackerel scad baits for tuna fishing. Mem. Fac. Fish. Kagoshima Univ. 20: 119–130. Shiode D, Hu F, Shiga M, Yokota K and Tokai T. 2005. Midwater float system for standardizing hook depths on tuna longlines to reduce sea turtle by-catch. Fish. Sci. 71: 1182–1184. Sigler MF. 1993. Stock Assessment and Management of Sablefish, Anoplopoma fimbria, in the Gulf of Alaska. PhD thesis, University of Washington. Sigler MF. 2000. Abundance estimation and capture of sablefish (Anoplopoma fimbria) by longline gear. Can. J. Fish. Aquat. Sci. 57: 1270–1283 Silver WL, Caprio J, Blackwell JF and Tucker D. 1976. The underwater electro-olfactogram: a tool for the study of the sense of smell of marine fishes. Experientia 32: 1216–1217. Skajaa K. 1997. Basic Movement Pattern and Chemooriented Search Toward Baited Gears in Demersal Species: A Field Study of Ling and Edible Crab. MSc thesis, University of Bergen. Skomal GB. 2007. Evaluating the physiological and physical consequences of capture on post-release survivorship in large pelagic fishes. Fish. Manag. Ecol. 14: 81–89. Skud BE. 1978. Factors affecting longline catch and effort: I. General Review. Int. Pac. Halib. Comm. Sci. Rep. 64: 5–14. South Pacific Regional Environment Programme (SPREP). 2001. A review of turtle bycatch in the western and central Pacific Ocean tuna fisheries. A report prepared for the South Pacific Regional Environment Programme (SPREP) by the Oceanic Fisheries Programme, Secretariat of the Pacific Community (SPC). 26 pp. Stiansen S. 2004. Fangstforskjeller Mellom to Typer Kongekrabbeteiner: Adferdstudier i Nærfeltet og Komparative Fiskeforsøk. MSc thesis, University of Bergen (in Norwegian).
139
Stoner AW. 2003. Hunger and light level alter response to bait by Pacific halibut: laboratory analysis of detection, location and attack. J. Fish Biol. 62: 1176–1193. Stoner AW. 2004. Effects of environmental variables on fish feeding ecology: implications for the performance of baited fishing gear and stock assessment. J. Fish Biol. 65: 1445–1471. Stoner AW and Ottmar ML. 2004. Fish density and size alter Pacific halibut feeding: implications for stock assessment. J. Fish Biol. 64: 1712–1724. Stoner AW and Sturm EA. 2004. Temperature and hunger mediate sablefish (Anoplopoma fimbria) feeding motivation: implications for stock assessment. Can. J. Fish. Aquat. Sci. 61: 238–246. Sutterlin AM. 1975. Chemical attraction of some marine fish in their natural habitat. J. Fish. Res. Bd. Can. 32: 729–738. Sutterlin AM and Sutterlin N. 1971: Electrical responses of the olfactory epithelium of Atlantic salmon (Salmo salar). J. Fish. Res. Bd Can. 28: 565–572. Suzuki N and Tucker D. 1971. Amino acids as olfactory stimuli in freshwater catfish, Ictalurus catus (Linn.). Comp. Biochem. Physiol. 40(A): 399–404. Suzuki Z, Warashina Y and Kishida M. 1977. The comparison of catches by regular and deep tuna longline gears in the western and equatorial Pacific. Bull. Far Seas Fish. Res. Lab. 15: 51–73. Swimmer Y, Arauz R, Higgins B, McNaughton L, McCracken M, Ballestero J and Brill R. 2005. Food color and marine turtle feeding behavior: can blue bait reduce turtle bycatch in commercial fisheries? Mar. Ecol. Prog. Ser. 295: 273–278. Swimmer Y, Arauz R, McCracken M, McNaughton L, Ballestero J, Musyl M, Bigelow K and Brill R. 2006. Diving behavior and delayed mortality of olive ridley sea turtles Lepidochelys olivacea after their release from longline fishing gear. Mar. Ecol. Prog. Ser. 323: 253–261. Thorpe JE. 1978. Rhythmic Activity of Fishes. London: Academic Press. 312 pp. Vabø R, Huse G, Fernö A, Jørgensen T, Løkkeborg S and Skaret G. 2004. Simulating search behavior of fish toward bait. ICES J. Mar. Sci. 61: 1224–1232. Walsh WA and Kleiber P. 2001. Generalized additive model and regression tree analyses of blue shark (Prionace glauca) catch rates by the Hawaii-based commercial longline fishery. Fish. Res. 53: 115–131.
140
Fish Behavior near Fishing Gears during Capture Processes
Ward P, Myers RA and Blanchard W. 2004. Fish lost at sea: the effect of soak time on pelagic longline catches. Fish. Bull. 102: 179–195. Ward P, Lawrence E, Darbyshire R and Hindmarsh S. 2008. Large-scale experiment shows that nylon leaders reduce shark bycatch and benefit pelagic longline fishers. Fish. Res. 90: 100–108. Watson JW, Epperly SP, Shah AK and Foster DG. 2005. Fishing methods to reduce sea turtle mortality associated with pelagic longlines. Can. J. Fish. Aquat. Sci. 62: 965–981. Webster DR, Rahman S and Dasi LP. 2001. On the usefulness of bilateral comparison to tracking turbulent chemical odor plumes. Limnol. Oceanogr. 46: 1048–1053. Webster DR and Weissburg MJ. 2001. Chemosensory guidance cues in a turbulent chemical plume. Limnol. Oceanogr. 46: 1034–1047. Weihs D. 1987. Hydromechanics of fish migration in variable environments. Am. Fish. Soc. Symp. 1: 254–261 Weimerskirch H and Jouventin P. 1987. Population dynamics of the Wandering Albatross, Diomedea exulans, of the Crozet Islands: causes and consequences of the population decline. Oikos 49: 315–322. Weimerskirch H, Brothers NP and Jouventin P. 1997. Population dynamics of Wandering Albatross Diomedea exulans and the Amsterdam Albatross D. amsterdamensis in the Indian Ocean and their rela-
tionship with longline fisheries: conservation implications. Biol. Conserv. 79: 257–270. Weimerskirch H, Capdeville D and Duhamel G. 2000. Factors affecting the number and mortality of seabirds attending trawlers and long-liners in the Kerguelen area. Polar Biol. 23: 236–249. Werner EE. 1974. The fish size, prey size, handling time relation in several sunfishes and some implications. J. Fish. Res. Bd. Can. 31: 1531–1536. Wienecke B and Robertson G. 2002. Seabird and seal—fisheries interactions in the Australian Patagonian toothfish Dissostichus eleginoides trawl fishery. Fish. Res. 54: 253–265. Wilson RR and Smith KL Jr. 1984. Effect of near bottom current on detection of bait by the abyssal grenadier fishes Coryphaenoides spp, recorded in situ with a video camera on a free vehicle. Mar. Biol. 84: 83–91. Yamaguchi Y, Nonoda T, Kobayashi H, Izawa K, Jinno T, Ishikura I, Uchida M and Tonogai M. 1983. Effectiveness of artificial bait for obtaining higher hooking rate on bottom set long-line fishing. Bull. Jpn. Soc. Sci. Fish. 42: 1819–1824. Yokota K, Kiyota M and Minamia H. 2006. Shark catch in a pelagic longline fishery: comparison of circle and tuna hooks. Fish. Res. 81: 337–341. Zimmer-Faust RK and Case JF. 1983. A proposed dual role of odor in foraging by the California spiny lobster, Panulirus interruptus (Randall). Biol. Bull. 164: 341–353.
Fish Behavior in Relation to Longlines SPECIES MENTIONED IN THE TEXT albacore, Thunnus alalunga American plaice, Hippoglossoides platessoides Atlantic cod, Gadus morhua bigeye tuna, Thunnus obesus bluefin tuna, Thunnus thynnus carp, Cyprinus carpio coho salmon, Oncorhynchus kisutch conger eel, Conger sp. goldfish, Carassius auratus Greenland halibut, Reinhardtius hippoglossoides haddock, Melanogrammus aeglefinus, hake, Merluccius sp. herring, Clupea harengus ling, Molva molva Pacific cod, Gadus macrocephalus, Pacific halibut, Hippoglossus stenolepis Patagonian toothfish, Dissostichus eleginoides pigfish, Orthopristis chrysopterus
pinfish, Lagodon rhomboides plaice, Pleuronectes platessa rabbitfish, Siganus fuscescens rainbow trout, Oncorhynchus mykiss red hake, Urophycis chuss sablefish, blackcod, Anoplopoma fimbria saithe, Pollachius virens salmon, Salmo salar sea catfish, Arius felis spiny dogfish, Squalus acanthias swordfish, Xiphias gladius turbot, Scophthalmus maximus tusk, Brosme brosme walleye pollock, Theragra chalcogramma white catfish, Ictalurus catus white hake, Urophycis tenuis wolffish, Anarhichas lupus yellowfin tuna, Thunnus albacares
141
Chapter 6 Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues Bjarti Thomsen, Odd-Børre Humborstad, and Dag M. Furevik
sea with low mortality. Pots are thus seen as a benign fishing method and superior to other gear from an environmental point of view (Cole et al. 2003; Dayton et al. 1995; Jennings and Kaiser 1998). This chapter offers a brief overview of the use of fish pots in various parts of the world, describes how they work in relation to fish behavior, and discusses the conservation challenges involved in their operation. Further details on various types of pots from around the world and their construction and mode of operation can be found in work by Furevik (1994), Sainsbury (1996), Slack-Smith (2001), and Gabriel et al. (2005). Discussions on the large-scale fixed fishing gear, the trap, are given in Chapter 7.
6.1 INTRODUCTION A 2000-year-old Greek document on fishing, the Halieutika, described fish traps that “work while their masters sleep” (Bekker-Nielsen 2002). These ancient words underline one of the characteristics inherent in passive fishing gears such as pots and traps in contrast to most modern gears, which are operated actively. According to the FAO fishing gear classification, pots are a subgroup of traps. Fish traps are passive fishing gears that allow fish to enter easily but then make escape difficult. Traps may be constructed as fixed weirs or fences along the shore or in rivers that guide fish into confined spaces. Fish pots, on the other hand, are transportable boxlike or basketlike enclosures designed to capture fish by attracting them to the pot and luring them inside through one or more narrow “one-way” entrances (Fig. 6.1). Fish pots may be deployed individually or in strings of several pots. Pots range in size from a few liters up to a few cubic meters. In some places, small pots are known as “creels.” Pots have several characteristics that are desirable for a modern fishing gear. They are not laborintensive and use less energy in operation than active gears. Pots only affect an area of seabed of the same order of magnitude as the footprint of the gear itself. They deliver the catch alive and with minimal physical damage, which is a prerequisite for a high-quality product. Catching the fish alive also allows unwanted bycatch to be returned to the
6.2 WORLDWIDE USE OF FISH POTS Pots are important fishing gears for crustaceans (crabs and lobsters), but for fish they are less commercially important compared with other fishing gears. On a global scale, all trap fishing (including large-scale trap and small-scale pots for fish and crustaceans) ranks only seventh in catch quantity, preceded by seine, midwater trawl, bottom trawl, gillnet, hook and line, and dredge (Watson et al. 2006). However, fish pots are important in some regions. In tropical waters, coral reefs and outcrops largely prohibit fishing with trawl and many other gears, whereas pots are often the preferred gear for
143
144
Fish Behavior near Fishing Gears during Capture Processes
10
9
1 6
11
4
3
2
5 7 8
current
Figure 6.1. General illustration of fish pots. 1, Rigid frame covered by netting. 2, Entrance, usually cone or wedge shaped and may be equipped with triggers (3) to prevent fish from escape. 4, Bait bag. 5, Escape ring to allow small animals to escape. 6, Break-away panel, which will corrode or rot away to open the pot if lost. 7, Attachment line. 8, Anchor. 9, Buoy line. 10, Surface buoy. 11, Ground line (only used if pots are deployed in a longline). (Drawing: A-B Tysseland, Norway.)
bottom-dwelling fish species. In other parts of the world, pots have found use in specific areas or fisheries due to their special characteristics, mainly with respect to their low impact on the habitat and their ability to capture fish alive. Fish pots are probably best known from the Caribbean, where the Antillean pot has been the principal fishing gear for centuries (Earle 1889; Munro et al. 1971). In many parts of the Caribbean, they still account for more than 50% of landings. The type of pots used throughout the Caribbean is bound by tradition and they are usually constructed from locally available materials. Since around 1920, pots have mainly been constructed from galvanized hexagon-shaped mesh wire on a framework of mangrove or other wooden sticks. The pot fishery in the Caribbean is a typical multispecies fishery and many different species are taken. In Bermuda, pots were also the principal fishing gear until they were banned in 1990 because of overfishing. On the eastern seaboard of the United States, pots are used for a limited number of species, with black
sea bass (Centropristis striata) being the most important. Sea bass typically concentrate in structurally complex habitats, where pots are more effective than mobile fishing gears. Pots account for an average of 45% of the total sea bass landings from this region (Shepherd et al. 2002). On the West Coast of the United States and Canada, as well as offshore Alaska, pots were introduced in the fishery for sablefish (Anoplopoma fimbria), also called black cod, in the early 1970s. For a few years, pots were the most important gear, but their popularity decreased rapidly during the 1980s. However, in the Bering Sea and Aleutian Islands, pots have gained popularity in recent years because longlining has problems due to depredation by whales (Hanselman et al. 2006). Off the coast of Alaska, pots are used for Pacific cod (Gadus macrocephalus). In this fishery, heavy steel frame pots with dimension of 1.8 × 1.8 × 0.9 m are set as single units using purpose-built fishing vessels with pot launchers to handle the heavy gear (Fig. 6.2). Similar pots that are used for Pacific cod in Alaska
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues
145
Figure 6.2. Pacific cod pot used off Alaska. These pots are operated by purpose-built vessels with pot launchers to handle the heavy gear. (Photograph: P. Munro, USA.)
Figure 6.3. Traditional Turkish fish pots on the deck of a local wooden boat. The pots are made of wire and 80 cm in diameter with entrance in top middle. (Photograph: A. Lok, Turkey.)
have been introduced for Atlantic cod (Gadus morhua) in Newfoundland and New England with some success, but wider commercial use remains to be seen (Pol et al. 2005; Walsh et al. 2006). Many Asian countries and island countries in the Indian Ocean have a tradition of using pots. In the Arabian Gulf, wire basket pots (known as gargoor) that are semicircular in cross section with sizes up to 2 m in diameter are used. In Kuwait, pots account for 50% of finfish landings, and in Oman, pot fishing represents up to 19% of all gear used by the artisanal fishery (Al-Masroori et al. 2004). In the Seychelles, pots, mainly constructed of bamboo strips, account for 20% of the reef fish catch (Mahon and Hunte 2001). Since 1990 a live food-fish trade has developed throughout Southeast Asia, and in some places up to 55% of live catches come from pots. Pots are commonly used in various parts of Japan together with other fishing gears. One example is in Hokkaido, where pots were introduced in 1980 in the important fishery for arabesque greenling (Pleurogrammus azonus) (Li et al. 2006). In Australia, pots are an established fishing method in a 700-tonne (metric ton) multispecies fishery, with snappers (Pagrus auratus) accounting for more than half of the value (Stewart and Ferrell 2003). On the northwest shelf of Australia, a pot fishery developed using cylindrical pots to target
species from Nemipteridae, Lethrinidae, Lutjanidae, and Serranidae families. These pots, unique to this area, are 1.5 m in diameter and 0.9 m high (Whitelaw et al. 1991). In New Zealand, a commercial pot fishery for blue cod (Parapercis colias) takes place in depths of 30 m or more (Cole et al. 2003). Pots have a steel frame with dimensions of 1.9 × 1.4 × 0.9 m and conical entrances on the vertical sides. In Europe, the commercial use of fish pots is of minor importance compared with that of other fishing gears. One example is a traditional circular pot (Fig. 6.3) used on a small scale in Turkey and Greece. In Norway there is a limited small-scale fishery for wrasse (Labridae) that are used as cleaner fish in the salmon farming industry (Bjordal 1993). Recently, the development of a new twochamber pot (Figure 6.4) for Atlantic cod, tusk (Brosme brosme) and ling (Molva molva) has enjoyed some success and a few costal vessels have changed from longlining and gillnetting to pot fishing (Furevik and Skeide 2003). Commercial fish pots are thus in use all over the world, and are produced in various sizes and designs, ranging from small pots that can be easily handled by one person to large heavy pots that can only be operated from large purpose-built vessels with special handling equipment.
146
Fish Behavior near Fishing Gears during Capture Processes The first three phases of catching processes of baited fish pots involving chemoreception and food localization are similar to those used with other baited gear, such as baited hooks. Some related discussions on chemoreception and food search are given in Chapter 5.
Figure 6.4. Norwegian two-chamber pot. The pot is made of three frames covered by netting. The bottom frame is made from steel and has some weight. The other two frames are aluminum. Floats are attached to the upper panel to erect the pot when deployed. Two cone-shaped entrances lead fish into the bottom chamber. A wedgeshaped entrance allows fish to enter the upper chamber. Two collapsed pots are seen next to a 12-m-long GRP pot vessel. (Photograph: B. Thomsen, Faroe Islands.)
6.3 FISH BEHAVIOR IN RELATION TO POTS Pots are passive fishing gear and catching success will depend on their ability to attract fish, lure the fish inside, and keep them in captivity until the pot is retrieved (hauled). Fish behavior in relation to pots may be described according to senses of fish influenced by the pot itself as well as additional “long-range” stimulation, such as bait. In tropical coral reef fisheries, pots were traditionally deployed unbaited (Earle 1889), whereas in most other pot fisheries, bait is used to increase the area within which fish may react and be attracted to the pot. The behavior involved in the catching process of baited fish pots can be divided into several phases: • • • • • •
Behavior before stimulation Arousal to the presence of bait Localization of the food odor source Approaching the pot Ingression and entrapment Capture or possible escape
6.3.1 Attraction Attraction to unbaited pots can involve a multitude of interacting factors such as exploratory behavior, adoption of the pot as a shelter or residence, intraspecific social behavior, and predator–prey interactions. In pot fishing on coral reef fish assemblages, several authors stress the importance of the immediate proximity of the pot to the reef (High and Beardsley 1970; Sylvester and Dammann 1972). According to Munro (1983), pots were usually set among corals and should never be more than 2 m from the nearest coral cover. Luckhurst and Ward (1985) studied the composition of pot catches at 3, 20, and 50 m from a patch reef and found that for most species, more fish were caught closest to the reef. High and Beardsley (1970), Munro et al. (1971), and Luckhurst and Ward (1985) observed the clear attraction of a number of fish species to captured conspecifics, resulting in a sharp increase of ingress after the first few individuals had entered the pot. Antillean pots have also been used with bait in coral reefs habitat. Some authors observed no effect of bait (High and Ellis 1973), whereas others saw only short-term effects with an increase in ingress rate as long as the bait lasted (Luckhurst and Ward 1985; Munro 1974). According to Furevik (1994), bait seemed essential in most pot fisheries. Valdemarsen (1977) tested pots for gadoids and found that fish had very low interest in unbaited pots compared with baited pots. Whitelaw et al. (1991) reported that a large amount of suitable bait (approximately 4 kg) was required for effective potting. Wolf and Chislett (1974), however, reduced the amount of bait from 12.5 kg to 2.5 kg per pot and found no noticeable effect on catch rates. The release rate of attractants from the bait is initially high followed by a rapid decline
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues (Løkkeborg 1990), so that the period a baited pot will “fish” efficiently is fairly short, perhaps as little as a few hours. Furevik (1994) reported that most fish arrived at the pot within first 2 hours after deployment and only a few fish arrived after 1 to 2 days. Whitelaw et al. (1991) observed an increasing number of fish in the pot within the first 3 hours with the number leveled and then decreased afterward, indicating that more fish might have escaped from the pot when the bait was depleted. The type of bait can influence catch rates and proportions of different species. For instance, the use of squid instead of herring increased catch rates of Atlantic cod, whereas only a minor difference was observed for tusk (Furevik 1994). Cole et al. (2003) used pilchards (Sardinops neopilchrdus) and paua (Hailiorrtis iris) guts as bait for blue cod and found paua gut–baited pots caught more blue cod but the fish were smaller. In addition, bycatches of other species were greatly reduced using paua guts as bait. Knowledge of the area within which fish are attracted to the pot is important when pots are used as a survey gear for fish abundance studies as well as for the determination of an optimal distance between pots in a commercial fishery. Bjordal and Furevik (1988) tested different pot spacing measures and caught 2.5 fish per pot at 37-m spacing compared with 3.4 fish per pot at 74-m spacing. These differences may be related to fish density and bait competition. Conners et al. (2004) found that there were indications of competition between pots spaced at 556 m apart. 6.3.2 Fish Behavior when Approaching the Pot When fish arrive at the pot, their behavior will be influenced by “short-range” senses such as vision and lateral line stimulation. Luckhurst and Ward (1985) caught a very small number of fish in the large-mesh pots compared with the number caught in the small-mesh pots (even if the fish were large enough to be retained by the large meshes) and speculate that the large-mesh pots have a relative weak visual image and that many fish may simply not respond to their presence in the same way as they do to the small-mesh pots.
147
High and Beardsley (1970) observed interspecies differences in behavior around fish pots. Groupers were consistently found as solitary individuals and approached the pot with caution, while schooling species entered the pot as a group (squirrelfish and goatfish) or independently (parrotfish, bigeyes) and paired fish (butterflyfish and some parrotfish) followed their mates readily into the pots. In the North Atlantic, Furevik (1994) reported qualitative species-dependent behavior patterns vis-à-vis baited pots. Atlantic cod and ling exhibited search behavior and would occasionally butt against the net. Tusk and catfish (Anarhichas lupus) appeared to approach the pot more slowly. Haddock (Melanogrammus aeglefinus) also seemed to be more careful in their approach to the pot than did Atlantic cod and ling. Almost all fish approached baited pots from downstream, where the bait plume had dispersed (Furevik, 1994). Zigzag swimming behavior was often seen (B. Thomsen, personal observations), which seemed to improve the locationing of the bait. Most fish attracted to a baited pot stayed fairly close to the pot in the downstream area independent of the position of the entrance relative to the current (Furevik, 1994). Only a few fish were sufficiently motivated to enter the pot to reach the bait inside. On several occasions, Atlantic cod were seen to take ownership of the pot and to guard the entrance by chasing other fish away from the pot (B. Thomsen, personal observations). As fish arrive at baited pots from downstream, it is obvious that pot entrances should face downstream so that fish can more easily reach the entrance. Several experiments confirmed that the maximum catch efficiency was realized when the pot entrance was oriented downstream (Furevik 1994; Furevik et al. 2008; Valdemarsen et al. 1977). To pursue the idea of an entrance facing downstream, Whitelaw et al. (1991) constructed a rotational pot design as an alternative to the cylindrical pot used in Western Australia, but they did not achieve significantly higher catches. In Norway, better catch rates were obtained in pots anchored at one end and floated off the bottom, which allows the pot to turn in response to changes in current so that the entrance is always facing downstream (Bjordal and Furevik 1988; Furevik et al. 2008).
148
Fish Behavior near Fishing Gears during Capture Processes
6.3.3 Entrance Design and Ingress/Egress Behavior Several examples are cited in the literature of large numbers of fish arriving at a pot but ingress numbers remaining low. Valdemarsen et al. (1977) caught only 16 of 1033 gadoids that entered the potting area and concluded that pots catch only a small proportion of the fish that come into contact with the gear. Hirayama et al. (1999) studied the fishing mechanism of pots by observing the behavior of puffer fish (Lagocephalus wheeleri) in and around pots in the sea and in tanks and concluded that catch efficiency was 2%. Rose et al. (2005) used a high-frequency imaging acoustic camera to observe the behavior of sablefish and Pacific halibut (Hippoglossus stenolepis) around baited fish pots. The pot caught 9 and 10 sablefish in two sets, where sablefish were observed entering the acoustic camera field 2000 to 5000 times. Cole et al. (2004) made continuous videorecordings of entries and exits from blue cod pots and found that less than 8% of approaches to pot entrance during 30-min sets led to pot entries and that 34% of the blue cod that entered were able to escape the pot before it was hauled. Munro (1974) studied coral fish behavior in Antillean pots and found ingress rates to be fairly constant over time, whereas egress was found to be a fixed proportion (around 12% per day) of the number of fish inside the pot. From his observations, Munro derived a model predicting that the catch would approach an asymptote when egress equals ingress. The motivation of a fish to enter a pot may depend on the state of the fish as well on environmental variables. It has been shown that hunger and light level alter the responses of fish to bait (Stoner 2003) and activity and feeding motivation of the fish may be affected by environmental factors, such as temperature (Stoner et al. 2006). Munro et al. (1971) found fish ingress to be dependent on the phase of the moon and the corresponding tidal rhythms. They reported the highest rates of ingress during the new and full moon. Dalzell and Aini (1992), however, found a single peak during the full moon. It is also reported that daytime catches are higher than nighttime catches (High and Beardsley 1970; High and Ellis 1973).
The critical phase in pot fishing is when fish move into the entrance area (Furevik 1994). The design of the entrance is thus crucial to the fishing success of the pot. When species enter, some are more cautious than others. Atlantic cod and catfish may push aside part of a net panel at the entrance to enter the pot, whereas haddock and ling actively search, but they may turn if resistance is met (Furevik 1994). Fuwa et al. (1995) studied the behavior of puffer fish in the entrance funnel of a pot and derived a selection curve for the pot that depended on the ratio between fish length and the width of the entrance. The entrance design may therefore affect which species and what sizes of fish are caught. In Antillean pots, the conical entrance is usually made of chicken wire and terminates in a funnel opening that faces downward like a horse’s neck. If these pots are placed upside down, they catch only few fish (Sylvester and Dammann 1972). Luckhurst and Ward (1985) tested pots with both straight and horse-neck funnels. In pots with horse-neck funnels, there was a steady buildup of fish inside the pot over 7 days, whereas with a straight funnel, the numbers of fish inside the pot fell after 2 to 4 days. They found horse-neck funnels to be very effective in reducing escape among the species that were sampled in their study. Several authors have described experiments aimed at increasing pot efficiency by optimizing the entrance design. Bjordal and Furevik (1988) tested entrances made of net panels and shaped like a wedge pointing into the pot. They tested horizontal and vertical net panels and found no difference in Atlantic cod and tusk catches. Li et al. (2006) tested responses of arabesque greenling to pot entrance design in a tank experiment using different lengths and inclination angles of wedge-shaped entrance panels. They found an inclination angle of 34 degrees to be most efficient, whereas the effect of funnel length was less conclusive. When pots are equipped with a wider entrance, the ingress rate increases but so does the escape rate (Furevik 1994). Munro (1974) concludes that in Antillean pots the rate of escape is the main determinant of catch rates and the development of effective nonreturning devices fitted to the entrance funnels will substantially enhance catch efficiency.
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues However, nonreturn devices often reduce the rate of entry into the pot and thus reduce catch rates to below those obtained without such devices (Munro 1972). One simple nonreturn device is the use of triggers. These are simple fingers of metal or plastic that may be easily pushed inward when fish enters but are not able to bend outwards. High and Ellis (1973) found that fewer fish entered when triggers were fitted but the escape rate was much higher without triggers. Hughes et al. (1970) modified king crab pots for sablefish and found that triggers made the pots more efficient. Carlile et al. (1997) found an approximately 10-fold increase in Pacific cod catches when retention triggers were fitted to crab pots. In baited pots, the importance of triggers may increase after the bait is fully consumed (Salthaug 2002). Another way to improve entrance efficiency is to use two successive entrances, sometimes called a parlor pot. Usually the outer entrance is relatively large while the inner entrance is much smaller. Furevik and Løkkeborg (1994) tested several entrance designs and found that pots with two successive entrances yielded catches of Atlantic cod that were three times as high as those for singleentrance pots but found no difference for tusk. The optimum entrance design depends on the soak time (Furevik 1994). Sheaves (1995) found higher catch rates at short soak times with simple entrances, whereas no difference was found between short and long soak durations when a complex entrance design was used. In baited pots, fish ingress rates are initially high. The number of fish inside the pot may reach a peak around the time when the bait is exhausted (Whitelaw et al. 1991). In baited pots for blue cod, Cole et al. (2004) found the optimal soak time to be 30 to 40 min. Some species are more adept than others at escaping (Munro 1983), and catch composition may therefore change with soak time. 6.3.4 Fish Behavior inside the Pot Behavior of fish inside a pot is species specific, and within the species individual specimens may also behave differently. After they enter the pot, most fish appear to be calm and mill around, although a few immediately display aggressive behavior and try vigorously to find an escape route. Furevik
149
(1994) found that most fish (Atlantic cod, haddock, tusk, and ling) were more active when they first entered the pot and frequently butted the net. They become less active with time inside the pot. In baited pots, some fish show interest in the bait for a short time (Furevik 1994) but soon lose interest. Luckhurst and Ward (1985) also found that most fish ignored the bait after entering the pot. Luckhurst and Ward (1985) found that many fish display butting behavior after entering the pot, causing abrasion of the nape and laceration of pectoral and pelvic fins, leading to secondary infection by fungi and/or bacteria. The most vulnerable were Scaridae, which were dead or dying within 2 to 3 days of entering pots. Munro et al. (1971) found that almost all fish that were trapped for up to 2 weeks displayed signs of physical deterioration or abrasions with secondary fungal infections. Cooke et al. (1998) also reported fungal growth on potted fish. However, Whitelaw et al. (1991) observed no mortality of fish inside pots even after extended periods of detention. Fish inside pots may escape not only through the entrance but, if they are small enough, also through the meshes of the cover whether made of synthetic netting or wire mesh. Mesh size is a determinant of catch rates and fish size at which fish recruit to pot fisheries (Mahon and Hunte 2001). Stewart and Ferrell (2002) found that size selection in rigid welded wire mesh could be accurately predicted by the maximum body depth of the fish and the maximum mesh aperture. However, Luckhurst and Ward (1985) found escape not to be simply a function of body depth. Some fish species turn on their side to escape through meshes, whereas others do not. In Antillean fish pots, Gobert (1998) found a shift in the selection curve, depending on average catch per pot. He concluded that the squeezing behavior of fish was density dependent. Aggressive behavior has been observed with larger individuals chasing smaller individuals and frequently predating on smaller species. In pots containing conger eels (Conger verreauxi), blue cod either attempted to escape or were consumed (Cole et al. 2001). The activity of fish inside the pot may affect fishing efficiency of the pot. High and Beardsley (1970) observed a “saturation effect” as the rate of
150
Fish Behavior near Fishing Gears during Capture Processes
entry dropped sharply when the fish inside the pot reached a certain number. It appeared that the presence of a large number of fish inside the pot scared off other fish in the area. It has been observed that larger pots usually have higher catch rates for most species (Collins 1990). The chance of finding the entrance for escaping from pots is inversely proportional to the area or volume of the pot if the sizes of the entrances are equal (Munro 1974). 6.4 CONSERVATION CHALLENGES AND SOLUTIONS Pots are generally regarded as an environmentally friendly fishing gear, with few undesirable side effects when catching target fish species. That does not mean that there are no negative conservation consequences in pot fisheries. Conservation challenges include issues related to lost gear and ghost fishing, escape and discard mortality, incidental megafauna interactions, and specific issues related to habitat alteration in vulnerable habitats. The prospects of solving these and other potential problems are good, given the inclusion of conservation characteristics in gear development and appropriate management of fisheries in sensitive areas. By using pots, access to fishery resources can probably be retained while benthic species and habitats remain protected. 6.4.1 Undersized and Nontarget Species: Catch Avoidance, Escapees, and Discards Problems related to capture of undersized or nontarget species occur frequently with most fishing gear. Means to solve these challenges in pot gears include improving selectivity by preventing unwanted organisms from entering the pots in the first place, providing effective escape opportunities, or optimizing capture and handling techniques for live release of discards. The most appealing way of solving potential discard and escape mortality is to find methods of preventing unwanted species from locating and entering the pots. Furevik et al. (2008) found that by floating pots off the bottom, catches of Atlantic cod increased, whereas bycatch of red king crab (Paralithodes camtschaticus) was eliminated. Zhou and Kruse (2000) found that horizontal excluders and slick-board ramps both had the potential to
reduce tanner crab (Chionoecetes bairdi) bycatch in Pacific cod pots. Carlile et al (1997) evaluated seven different pot modifications and found that the installation of halibut excluders in standard pots increased Pacific cod catch and reduced Pacific halibut bycatch. Bait selection is another way to reduce unwanted catch, because it can influence the catch rate and the proportions of different species and size classes. The use of squid as bait instead of herring increased the catch rates of Atlantic cod but not tusk (Furevik 1994). Cole et al. (2003) obtained larger catches of blue cod and greatly reduced bycatch species by using paua guts instead of pilchards as bait. Pedersen (2000) found that catch rates of Atlantic cod were higher when pots were baited with squid compared with mackerel bait in June, whereas no differences were found in October and February. He also found that in February, squid-baited pots caught more Atlantic cod larger than 40 cm than did mackerelbaited pots. Optimizing efficiency for target species and sizes may reduce effort, and thereby bycatch, in a quota-controlled fishery. Other ways to avoid unwanted bycatch may include developing speciesspecific deterrents or shorter soak times to avoid slow-moving scavenging invertebrates or simply avoiding problem areas and time periods. By far the most frequently used method to achieve a “cleaner catch” is to mount specific escape rings, panels, or vents to allow unwanted catches to escape after they have entered the pots (e.g., Pedersen 2000; Shepherd et al. 2002; Stewart and Ferrell 2002). Selecting web mesh size (Luckurst and Ward 1985; Robichaud et al. 1999; Sary et al. 1997) in accordance with minimum landing size, for example, is another option, which offers numerous escape routes. Successful reduction of bycatch mortality by improving selectivity relies heavily on the assumption that escapees are unharmed. If they are not, increasing the opportunity for escape may result in a higher level of unaccounted fishing mortality (Chopin and Arimoto 1995). There is currently little information on the survival of escapees from pots. Signs of eagerness to escape may range from subtle pushing against the net to squeezing through meshes (Robichaud et al. 1999), a behavior that may be exaggerated in the presence of predatory fish
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues (Hartsuijker and Nicholson 1981). Laboratory studies have shown that the simple passage of a fish through a netting mesh or other selective device does not necessarily inflict fatal injury (see Chapter 11). If undersized and nontarget species cannot be avoided or do not escape, they may be discarded after being brought on board (e.g., Stewart and Ferrell 2003). Due to the benign nature of the fish capture process, it is expected that the mortality of fish discarded from pots may be low as the catch is usually alive, with low injury rates (Nøstvik and Pedersen 1999) and low capture-related stress (Pilling et al. 2001). Survival rates of discarded bycatch can be inferred indirectly from tagging studies. When cod were caught for tagging, Nøstvik and Pedersen (1999) found that fish pots had one of the lowest injury rates, the highest tagging percentage, and the highest recapture rates, compared with fyke nets, hand line, pelagic trawl, and bottom trawl. Although survival rates of discards in pot fisheries are probably among the highest, passage up through the water column and handling on deck pose several additional threats to survival, including barotrauma, thermal shock, air exposure, physical impact, and predator exposure (Davis 2002). Returning to the depth after discard may also increase vulnerability of predation. For these reasons alone, discards should be avoided. 6.4.2 Lost Gear and Ghost Fishing Ghost fishing is defined as the ability of fishing gear to continue fishing after control of the gear has been lost by the fisherman (Smolowitz 1978). Most studies on ghost fishing by lost pots have been performed in the crab and lobster fisheries. Only a few deal with fish pots, although problems and mitigation measures in the different fisheries may overlap. Moored static gears may be lost for several reasons. Strong currents may force buoys down to depths at which they collapse. End markers may be cut off by propellers. Groundlines and buoy lines may be caught on rough bottoms or otherwise break by wave action during retrieval in bad weather. Whole fleets of pots may be displaced from their original positions due to currents or gear conflicts. Pots are often rigid structures made of strong materials and are likely to maintain their configuration after they
151
are lost. They may therefore continue to catch and potentially kill captured organisms after being lost. Baited pots lose most of their catching efficiency for target species when their bait has been consumed or the release of feeding attractants ends. The potential for continued catches of target species beyond this initial catch is low. Captured species, however, are inadequate evidence to prove ghost fishing mortality. If fish are able to escape and dead bodies cannot be found, the ghost fishing mortality has been defined as zero (Matsuoka et al. 2005). If entrapped fish cannot escape they will eventually die due to starvation or injury resulted from behaviors such as bumping on the webbing inside pots (Bullimore et al. 2001; Matsuoka et al. 2005). Captured fish that die inside the pot may re-bait the pots, enhancing ghost fishing capability. Selfbaiting through the death and decay of trapped fish may attract and capture scavenging species (especially crustaceans), which also may be trapped and die. In the worst case scenario, there may be a continuous cycle of capture, decay, and attraction of scavenging species for as long as the gear remains intact (Carr et al. 1990). In the case of unbaited pots, the capture of species does not rely on olfaction, and they may continue to fish target species and nontarget species for a long period of time. Matsuoka et al. (1997) found that some finfish pots continued ghost fishing for as long as 3 years. No general conclusions can be drawn from published studies regarding ghost fishing mortality in fish pots, simply because published studies are few and differ in many aspects such as area, species, type of pots, circumstance of loss, and study methods. One of the few studies of ghost fishing in fish pots found that, depending on fish density, fish learned to escape from pot openings or bar gaps after 10 days (Ayaz et al. 2006). After 24 days, no fish were observed in the pots. The pots were collapsed after 6 months when the study was terminated. In another field study quantifying catch rates of simulated lost fish pots at five traditional fishing grounds near Muscat and Mutrah, Sultanate of Oman (Al-Masroori et al. 2004), escape rates were low even after long soak times, with an estimated ghost fishing mortality of 95%. However, the adverse effects of fish-pot ghost fishing may easily be solved. Decomposition of the
152
Fish Behavior near Fishing Gears during Capture Processes
gear itself may not be a desirable option in reducing the catching capacity of a lost gear because of the longevity of the materials used. However, escape can be enhanced by mounting degradable panels/ threads and/or galvanic time-release mechanisms (GTRs) into the pots, allowing escape after a predetermined time (Blott 1978; Breen 1990; Scarsbrook et al. 1988; Wyman 1996). Scarsbrook et al. (1988) tested the effectiveness of various escape mechanisms in conical pots used in the sablefish fishery and found that square or triangular panels were more effective than just a “slash” secured with biodegradable twine and that these measures virtually eliminated ghost fishing. Such measures are simple and can be quick and easy to service once installed (Breen 1990). These mitigation measures should be considered if ghost fishing is a problem in both baited and unbaited pots. Retrieval of lost gear might be the ultimate means to reduce ghost fishing (Stevens et al. 2000). Retrieval is important, not only as a means of reducing mortality but also because lost pots constitute a littering problem (Edyvane et al. 2004; Hess et al. 1999; Lee et al. 2006) and the risk of losing more gear will increase as previously lost gear poses new threats to snagging and further losses of gear. Due to a certain reluctance to report losses and to the inherent difficulty of performing quantitative surveys of lost fishing gear, quantitative estimates of lost gears vary considerably (Bullimore et al. 2001). 6.4.3 Megafauna Interaction Incidental interaction between pot gears and megafauna such as whales and turtles seems to be mostly related to groundlines and buoy lines. In some pot fisheries that involve endangered species, such entanglements are of utmost concern and may be a significant part of the mortality of the species. In some other cases, the use of pots instead of other static gears is encouraged to avoid predation on bait or captured target species by birds and marine mammals. In a study by Johnson et al. (2005), entanglements of endangered right whales (Eubalaena glacialis) and humpback whales (Megaptera novaeangliae) were analyzed to determine the
types and parts of gear that were involved. Of all the entanglements, 89% were attributed to pot and gill-net gears, and when the gear part could be identified, 81% involved entanglements in buoy lines and/or groundlines. Consequently, reducing the number of lines in the water column, for example, by using sinking or neutrally buoyant groundline could reduce the risk of entanglement. The risk presented by buoy and surface system lines might be mitigated by a suitable insertion of weak links with sufficiently low breaking strengths. Another solution to reduce the number of buoy lines may be to set fleets of multiple pots instead of setting them singly. The efficacy of such mitigating measures remains to be demonstrated. Pots and their associated ropes may also entangle and drown marine turtles. Dayton et al. (1995) speculated that leatherback turtles (Dermochelys coriace) may mistake marker buoys for jellyfish and become entangled in buoy lines. Sewell and Hiscock (2005) performed a thorough review of marine turtle bycatch in the United Kingdom and Ireland. The leatherback turtle is the only species of turtle that is significantly affected by fisheries in U.K. and Irish waters. Although the global significance of bycatch in this area is not fully known, measures to reduce the impact on the declining global population are encouraged. 6.4.4 Habitat Alteration The potential for habitat disturbance due to potting is low compared with that for mobile fishing gear (Jennings and Kaiser 1998). Pots are static gears, and each pot thus affects an area of seabed of the same order of magnitude as the footprint of the gear itself. Furthermore, the forces on the seabed can be mainly attributed to the weight of the gear, which is usually made as light as possible to minimize production costs and provide ease of handling. At present, the spatial distribution of pots is also limited, and the few reported disturbances of potential significance have been in areas of erect fragile coral epifauna, areas that are probably vulnerable to all bottom-contact gears (Mortensen et al. 2005). In fact, the only long-term negative effect of pot fishing on habitat is based on anecdotal information from British Columbia, where red tree corals were reported to have disappeared in an area where
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues prawn pots were set, because corals became entangled in the mesh of the pots (Risk et al. 1998). Pots are usually regarded as environmentally friendly, and their use is encouraged to retain access to fish resources while protecting benthic species and habitats (Blyth et al. 2004; Kaiser et al. 2000; Mangi and Roberts 2006; Risk et al. 1998). Such habitats indeed sometimes serve as “de facto” reference or control areas that are subjected to minimal disturbance (Kaiser et al. 2000). Stone (2006) found that the greatest potential for disturbance existed when pots are dragged along the bottom during retrieval. He studied the coral habitat in the Aleutian Islands of Alaska with regard to depth distribution, fine-scale species association, and fisheries interactions in areas fished by bottom trawls, longlines, single-set pots, and fleet-set pots by means of video transect analysis. Only one coral site was found where fleet-set crab pots were involved and no sites where single-set pots were used. Fleet-set pots with ropes connecting the adjacent pots cover larger areas and thus come in contact with more epifauna. They may therefore cause more damage than single-set pots. At the one site where disturbance was observed, the seabed was scored to the bare substrate, but no information was offered as to whether any fauna were damaged. Although this finding demonstrates that pots may cause habitat alterations, the study also showed that the frequency of damage by this gear is extremely low (a single incident in 25 video transects, or 0.9% of the total length of video transects). Further, dragging of pots and lines represents an abnormal situation that fishermen strive to avoid due to the risk of losing or damaging their gears. Fishermen normally try to minimize drag forces, because connecting ropes cannot withstand high dragging forces, due to the friction of the pots and moorings being towed along the seabed. Instead, hauling is carried out in such a way that each pot sits in its deployed position on the bottom until it is lifted vertically off the bottom. Precarious situations may arise when strong wind and current conditions force the vessel in the opposite of intended direction or off track, or produce a lifting angle facilitating bottom drag, or when fishing takes place at great depths with difficult positioning or in unpredictable areas with steep irregular bathymetry (Stone 2006). In such areas, it
153
is often necessary to use heavy moorings. This component intuitively poses a threat through direct contact with fauna on deployment (Eno et al. 2001), but the numbers and amount of biomass affected will be relatively low, even with a fishing effort far beyond current and prospective levels. Eno et al. (2001) examined the effects of fishing with crustacean pots on benthic species. They demonstrated few or no immediate effects on several species that are perceived to be sensitive to mechanical disturbance. Notably, sea pens (Penatula phosphorea, Virgularia mirabilis, and Funiculina quadrangularis) bent away due to the pressure wave exerted by the dropping pot, which effectively prevented direct contact with their tips. After smothering and even uprooting, they reestablished themselves when they regained contact with the muddy substrate. Other than damage sustained by large individual Ross corals (Pentapora foliacea), the short-term effects of pot fishing on sensitive benthic species did not appear to be detrimental. The environmental impacts of artisanal fishing gear on coral reef ecosystems were studied in the multigear fishery of southern Kenya (Mangi and Roberts 2006). No direct evidence was offered regarding physical damage by pots to corals. However, coral heads were removed to hold pots on the bottom during fishing and therefore had direct effects on the reef. However, sustainability of coral reef is more likely to be achieved by reducing the use of beach seines and spears, while maintaining the use of pots, gillnets, and handlines, which cause the least damage. Blyth et al. (2004) examined benthic communities at sites within and adjacent to the U.K. Inshore Potting Agreement (IPA) area, which had been subjected to four different commercial fishing regimens (one static gear regimen and three towed gear regimens) since 1978. Significantly greater total species richness and biomass of benthic communities were reported at sites under static gear regimens than those under towed gear regimens. The benthic community biomass under static gear regimens was also significantly greater than that under all other regimens. Similar findings were also reported from the same IPA area by Kaiser et al. (2000), who found that communities within areas closed to towed fishing gears were dominated by higher biomass
154
Fish Behavior near Fishing Gears during Capture Processes
and emergent epifauna that increased habitat complexity, whereas towed areas were dominated by smaller-bodied fauna and scavenging taxa. 6.5 CONCLUDING REMARKS Fish pots have a long history as a fishing gear and are widely used in many parts of the world with a wide variety of designs and modes of operations. However, the volume of fish captured with pots is small and catch efficiency is low in comparison with other gear. Pots possess several superior and appealing characteristics compared with other fishing gear: low labor and energy use, minimal habitat impact, and live fish delivery. Growing concern regarding the sustainability of the marine environment and growing support for responsible utilization of natural resources may be incentives to revive the use of pots as an alternative fishing method. However, fish pots need further development to increase catch efficiency. Most observations on fish behavior in relation to catching mechanisms have identified the entrance as the most important factor in improving the efficiency. Effective bait is also a major contributor to fishing success. The release rate of odor from the bait decreases rapidly, and a system that prolongs the release of bait odor would produce substantial advantages. One potential adverse effect of pot fishing is ghost fishing by lost pots, which could probably be solved by mounting timed-release systems. In comparison with most other gears, fish pots have more favorable characteristics than drawbacks. Fish pots might therefore be developed into a viable alternative to other fishing gear in many fisheries and even become the preferred fishing gear in the future. REFERENCES Al-Masroori H, Al-Oufi H, McIlwain JL and McLean E. 2004. Catches of lost fish traps (ghost fishing) from fishing grounds near Muscat, Sultanate of Oman. Fish. Res. 69: 407–414. Ayaz A, Ozekinci U, Altinagac U and Ozen O. 2006. An Investigation of Ghost Fishing of Circular Fish Traps used in Turkey. Presented at the ICES-FAO WGFTFB meeting, April 2006, Izmir, Turkey. Bekker-Nielsen T. 2002. Nets, boats and fishing in the Roman world. Classica et Mediaevalia. Vol. 53.
Bjordal Å. 1993. Capture techniques for wrasses (Labridae). ICES CM. 1993/B: 22. Bjordal Å and Furevik DM. 1988. Full scale fishing trials for tusk (Brosme brosme) and cod (Gadus morhua) with collapsible fish trap. ICES CM. 1988/B: 33. Blott AJ. 1978. A preliminary study of timed release mechanisms for lobster traps. Mar. Fish. Rev. 40: 40–49. Blyth RE, Kaiser MJ, Edwards-Jones G and Hart PJB. 2004. Implications of a zoned fishery management system for marine benthic communities. J. Appl. Ecol. 41: 951–961. Breen PA. 1990. A review of ghost fishing by traps and gillnets. Proceedings of the Second International Conference on Marine Debris, 2–7 April 1989, Honolulu, Hawaii. pp 571–599. Bullimore BA, Newman PB, Kaiser MJ, Gilbert SE and Lock KM. 2001. A study of catches in a fleet of “ghost-fishing” pots. Fish. Bull. 99: 247–253. Carlile DW, Dinnocenzo TA and Watson LJ. 1997. Evaluation of modified crab pots to increase catch of Pacific cod and decrease bycatch of Pacific halibut. N. Am. J. Fish. Manag. 17: 910–928. Carr HA, Amaral EH, Hulbert AW and Cooper R. 1990. Under-water survey of simulated lost demersal and lost commercial gill nets off New England. In: Coe JM and Rogers DB (eds). Marine Debris: Sources, Impacts and Solutions. pp 171–186. New York: Springer. Chopin FS and Arimoto T. 1995. The condition of fish escaping from fishing gears: a review. Fish. Res. 21: 315–327. Cole RG, Tindale DS and Blackwell RG. 2001. A comparison of diver and pot sampling for blue cod (Parapercis colias: Pinguipedidae). Fish. Res. 52: 191–201. Cole RG, Alcock NK, Tovey A and Handley SJ. 2004. Measuring efficiency and predicting optimal set durations of pots for blue cod Parapercis colias. Fish. Res. 67: 163–170. Cole RG, Alcock NK, Handley SJ, Grange KR, Black S, Cairney D, Day J, Ford S and Jerrett AR. 2003. Selective capture of blue cod Parapercis colias by potting: behavioral observations and effects of capture method on peri-mortem fatigue. Fish. Res. 60: 381–392. Collins MR. 1990. A comparison of three fish trap designs. Fish. Res. 9: 325–332. Conners ME, Munro P and Neidetcher S. 2004. Pacific cod pot studies 2002–2003. AFSC Processed Report 2004.
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues Cooke SJ, Bunt CM and McKinley RS. 1998. Injury and short-term mortality of benthic stream fishes: a comparison of collection techniques. Hydrobiologia. 379: 207–211. Dalzell P and Aini JW. 1992. The Performance of Antillean wire mesh fish traps set on coral reefs in Northern Papua New Guinea. Asian Fish. Sci. 5: 89–102. Davis MW. 2002. Key principles for understanding fish bycatch discard mortality. Can. J. Fish. Aquat. Sci. 59: 1834–1843. Dayton PK, Thrush SF, Agardy MT and Hofman RJ. 1995. Environmental effects of marine fishing. In: Aquatic conservation: Mar. Freshw. Ecosys. 5: 205–232. Earle EM. 1889. The fish pot of the Caribbean Sea. J. Mar. Biol. Assoc. UK. 1: 199–204. Edyvane KS, Dalgetty A, Hone PW, Higham JS and Wace NM. 2004. Long-term marine litter monitoring in the remote Great Australian Bight, South Australia. Mar. Pollut. Bull. 48: 1060–1075. Eno NC, MacDonald D, Kinnear J, Amos SC, Chapman C, Clark R, Bunker F and Munro C. 2001. Effects of crustacean traps on benthic fauna. ICES J. Mar. Sci. 58: 11–20. Furevik DM. 1994. Behavior of fish in relation to pots. In: Fernö A and Olsen S (eds). Marine Fish Behavior in Capture and Abundance Estimation. pp 28–44. Oxford: Fishing News Books. Furevik DM, Humborstad O-B, Jørgensen T and Løkkeborg S. 2008. Floated fish pot eliminates bycatch of red king crab and maintains target catch of cod. Fish. Res. 92: 23–27. Furevik DM and Løkkeborg S. 1994. Fishing trials in Norway for Torsk (Brosme brosme) and Cod (Gadus morhua) using baited commercial pots. Fish. Res. 9: 219–229. Furevik DM and Skeide RL. 2003. Fiske etter torsk (Gadus morhua), Lange (Molva molva) og brosme (Brosme brosme) med tokammerteine langs norskekysten. Bergen: Institute of Marine Research. Fuwa S, Ishizaki M, Sako K and Imai T. 1995. A catching model of fish trap for puffer. Nippon Suisan Gakkaishi. 61: 356–362. Gabriel O, Lange K, Dahm E and Wendt T. 2005. Von Brandt’s Fish Catching Methods of the World. 4th ed. Oxford: Fishing News Books and Blackwell Publishing. 523 pp. Gobert B. 1998. Density-dependent size selectivity in Antillean fish traps. Fish. Res. 38: 159–167. Hanselman DH, Lunsford CR, Fujioka T and Rodgveller CJ. 2006. Alaska Sablefish Assessment
155
for 2007. NPFMC Bering Sea, Aleutian Islands and Gulf of Alaska SAFE, Alaska Sablefish: 341–428. Hartsuijker L and Nicholson WE. 1981. Results of a potfishing survey on Pedro Bank (Jamaica): The relations between catch rates, catch composition, the size of fish and their recruitment to the fishery. FAO/TCO/JAM 8902: Potfishing survey on Pedro Bank. Hess NA, Ribic CA and Vining I. 1999. Benthic marine debris, with an emphasis on fishery-related items, surrounding Kodiak Island, Alaska, 1994– 1996. Mar. Pollut. Bull. 38: 885–890. High WL and Beardsley AJ. 1970. Fish behavior studies from an undersea habitat. Comm. Fish. Rev. 32(10): 31–37. High WL and Ellis IE. 1973. Underwater observations of fish behavior in traps. Helgolander wiss. Meeresunters. 24: 341–347. Hirayama M, Fuwa S, Ishizaki M and Imai T. 1999. Behavior of puffer Lagocephalus and the fishing mechanism of the pot trap. Nippon Suisan Gakkaishi. 65: 419–426. Hughes SE, Worlund DD and Hipkins FW. 1970. Adaption of king crab pots for capturing sablefish (Anoplopoma fimbria). J. Fish. Res. Bd. Can. 27: 1747–1755. Jennings S and Kaiser MJ. 1998. The effects of fishing on marine ecosystems. Adv. Mar. Biol. 34: 201–352. Johnson A, Salvador G, Kenney J, Robbins J, Landry S and Clapham P. 2005. Fishing gear involved in entanglements of right and humpback whales. Mar. Mam. Sci. 21: 635–645. Kaiser MJ, Spence FE and Hart PJB. 2000. Fishinggear restrictions and conservation of benthic habitat complexity. Conserv. Biol. 15: 1512–1525. Lee D-I, Cho H-S and Jeong S-B. 2006. Distribution characteristics of marine litter on the sea bed of the East China Sea and the south of Korea. Estuarine, Costal and Shelf Science 70: 187–194. Li Y, Yamamoto K, Hiraishi T, Nashimoto K and Yoshino H. 2006. Behavioral responses of arabesque greenling to trap entrance design. Fish. Sci. 72: 821–828. Løkkeborg S. 1990. Rate of release of potential feeding attractants from natural and artificial bait. Fish. Res. 8: 253–261. Luckhurst B and Ward J. 1985. Behavioral dynamics of coral reef fishes in Antillian fish traps at Bermuda. Proc. Gulf. Carib. Fish. Inst. 38: 528–546. Mahon R and Hunte W. 2001. Trap mesh selectivity and the management of reef fishes. Fish Fish. 2: 356–375.
156
Fish Behavior near Fishing Gears during Capture Processes
Mangi SC and Roberts CM. 2006. Quantifying the environmental impacts of artisanal fishing gear on Kenya’s coral reef ecosystem. Mar. Pollut. Bull. 52: 1646–1660. Matsuoka T, Nakashima T and Nagasawa N. 2005. A review of ghost fishing: scientific approaches to evaluation and solutions. Fish. Sci. 71: 691–702. Matsuoka T, Osako T and Miyagi M. 1997. Underwater observation and assessment on ghost fishing by lost fish-traps. Proc. Asian Fish. Forum 4: 179–183. Mortensen PB, Mortensen LB and Gordon DC Jr. 2005. Effects of fisheries on deepwater gorgonian corals in the Northeast Channel, Nova Scotia. Am. Fish. Soc. Symp. 41: 369–382. Munro JL. 1972. Large volume stackable fish trap for offshore fishing. Proc. Gulf Carib. Fish. Inst. 25: 212–228. Munro JL. 1974. The mode of operation of Antillean fish traps and the relationships between ingress, escapement, catch and soak. ICES J. Mar. Sci. 35: 337–350. Munro JL. 1983. The composition and magnitude of trap catches in Jamaican waters. In: Munro JL. (ed). Caribbean Coral Reef Fishery Resources. ICLARM Studies and Reviews. No. 7. Manila, the Philippines: International Center for Living Aquatic Resources Management. Munro JL, Reeson PH and Gaut VC. 1971. Dynamic factors affecting the performance of the Antillean Fish Trap. Proc. Gulf Carib. Fish. Inst. 23: 184–194. Nøstvik F and Pedersen T. 1999. Catching cod for tagging experiments. Fish. Res. 42: 57–66. Pedersen K-A. 2000. Effekter av agntype, maskevidde og settetidspunkt på fangsteffektivitet og størrelsessammensetning av torsk i fiske med teiner. University of Tromsø, Norway (in Norwegian). Pilling GM, Purves MG, Daw TM, Agnew DA and Xavier JC. 2001. The stomach contents of Patagonian toothfish around South Georgia (South Atlantic). J. Fish Biol. 59: 1370–1384. Pol M, Walsh P and Marcella R. 2005. Cod potting in Massachusetts. A report submitted to Northeast consortium. Available online: http://northeastconsortium. org/ProjectFileDownload.pm?report_id=661 &table=project_report. Risk JR, McAllister DE and Behnken L. 1998. Conservation of cold and warm water seafans: Threatened ancient gorgonian groves. Sea Wind 12(1). Ocean Voice International. Robichaud D, Hunte W and Oxenford HA. 1999. Effects of increased mesh size on catch and fishing
power of coral reef fish traps. Fish. Res. 39: 275–294. Rose CS, Stoner AW and Matteson K. 2005. Use of high-frequency imaging sonar to observe fish behavior near baited fishing gears. Fish. Res. 76: 291–304. Sainsbury J. 1996. Static gear. In: Commercial Fishing Methods: An Introduction to Vessels and Gears, 3rd ed. Oxford: Fishing News Books and Blackwell Science. Salthaug A. 2002. Do triggers in crab traps affect the probability of entry? Fish. Res. 58: 403–405. Sary Z, Oxenford HA and Woodley JD. 1997. Effects of an increase in trap mesh size on an overexploited coral reef fishery at Discovery Bay, Jamaica. Mar. Ecol. Prog. Ser. 154: 107–120. Scarsbrook JR, McFarlane A and Shaw W. 1988. Effectiveness of experimental escape mechanisms in sablefish traps. N. Am. J. Fish. Manag. 8: 158– 161 Sewell J and Hiscock K. 2005. Effects of fishing within UK European marine sites: guidance for nature conservation agencies. Report to the Countryside Council for Wales, English Nature and Scottish Natural Heritage from the Marine Biological Association, Plymouth: CCW Contract FC 73-03214A. 195 pp. Sheaves MJ. 1995. Effect of design modifications and soak time variations on Antillean Z fish trap performance in a tropical estuary. Bull. Mar. Sci. 56: 475–489. Shepherd GR, Moore CW and Seagraves RJ. 2002. The effect of escape vents on the capture of black sea bass, Centropristis striata, in fish traps. Fish. Res. 54: 195–207. Slack-Smith RJ. 2001. Fishing with traps and pots. FAO Training Series 26. Rome, FAO. Smolowitz RJ. 1978. Trap design and ghost fishing: Discussion. Mar. Fish. Rev. 40: 59–67. Stevens BG, Vining I, Byersdorfer S and Donaldson W. 2000. Ghost fishing crab (Chionoecetes bairdi) pots off Kodiak, Alaska: pot density and catch determined from sidescan sonar and pot recovery data. Fish. Bull. 98: 389–399. Stewart J and Ferrell DJ. 2002. Escape panels to reduce by-catch in the New South Wales demersal trap fishery. Mar. Freshw. Res. 53: 1179–1188. Stewart J and Ferrell DJ. 2003. Mesh selectivity in the New South Wales demersal trap fishery. Fish. Res. 59: 379–392. Stone RP. 2006. Coral habitat in the Aleutian Islands of Alaska: depth distribution, fine-scale species
Fish Pots: Fish Behavior, Capture Processes, and Conservation Issues associations, and fisheries interactions. Coral Reefs. 25: 229–238. Stoner AW. 2003. Hunger and light level alter response to bait by Pacific halibut: laboratory analysis of detection, location and attack. J. Fish Biol. 62: 1176–1193. Stoner AW, Ottmar ML and Hurst TP. 2006. Temperature affects activity and feeding motivation in Pacific halibut: Implications for bait-dependent fishing. Fish. Res. 81: 202–209. Sylvester JR and Dammann AE. 1972. Pot fishing in the Virgin Islands. Mar. Fish. Rev. 34: 33– 35. Valdemarsen JW. 1977. Analysis of pot as a bottom gear and studies of some factors influencing the catch efficiency. Dept. of Fishery Biology, University of Bergen, Bergen, Norway. Valdemarsen JW, Fernö A and Johannessen A. 1977. Studies on the behavior of some gadoid species in relation to traps. ICES CM. 1977/B: 42.
157
Walsh PJ, Hiscock W and Sullivan R. 2006. Fishing trial for cod (Gadus morhua) using experimental pots. St. John’s: Fisheries and Marine Institute of Memorial University of Newfoundland. Watson R, Revenga C and Kura Y. 2006. Fishing gear associated with global marine catches I. Database development. Fish. Res. 79: 97–102. Whitelaw AW, Sainsbury KJ, Dews GJ and Campbell RA. 1991. Catching characteristics of four fish-trap types on the North-West Shelf of Australia. Aust. J. Mar. Freshw. Res. 42: 369–382. Wolf S and Chislett G. 1974. Trap fishing. Explorations for snapper and related species in the Caribbean and adjacent waters. Mar. Fish. Rev. 37: 49–61. Wyman E. 1996. Selective groundfish pots offer solution to bycatch problems. University of Alaska Sea Grant Report 96-03. Zhou S and Kruse GH. 2000. Modifications of cod pots to reduce Tanner crab bycatch. N. Am. J. Fish. Manag. 20: 897–907.
158
Fish Behavior near Fishing Gears during Capture Processes
SPECIES MENTIONED IN THE TEXT Atlantic cod, Gadus morhua arabesque greenling, Pleurogrammus azonus black sea bass, Centropristis striata blue cod, Parapercis colias catfish, Anarhichas lupus conger eel, Conger verreauxi haddock, Melanogrammus aeglefinus humpback whales, Megaptera novaeangliae ling, Molva molva leatherback turtle, Dermochelys coriace Pacific cod, Gadus macrocephalus Pacific halibut, Hippoglossus stenolepis
paua, Hailiorrtis iris pilchard, Sardinops neopilchrdus puffer fish, Lagocephalus wheeleri red king crab, Paralithodes camtschaticus right whale, Eubalaena glacialis Ross coral, Pentapora foliacea sablefish, black cod, Anoplopoma fimbria sea pens, Penatula phosphorea, Virgularia mirabilis, and Funiculina quadrangularis snapper, Pagrus auratus tanner crab, Chionoecetes bairdi tusk, Brosme brosme wrasse, Labridae
Chapter 7 Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges Pingguo He and Yoshihiro Inoue
servation challenges including interactions of the gear with megafauna species and mitigation measures.
7.1 INTRODUCTION In a broad sense, traps can be defined as stationary fishing gears to which fish or shellfish are misled, drift into, or are attracted to an enclosure and cannot found their way out, therefore being “trapped.” In that sense, traps include large-scale stationary netting structures such as cod traps in Newfoundland, set nets in Japan, and small pots such as lobster pots. These small fish or shellfish pots are also called traps in many regions of the world. To avoid confusion, traps and pots are distinguished in this chapter. Traps here refer to large stationary fixed structures (including trap nets, set nets, pound nets, and weirs) to lead and trap fish, whereas pots refer to small-scale baited enclosures to attract and retain fish. This chapter is concerned with traps. Fish capture by baited pots is discussed in Chapter 6. Traps are stationary fishing gears set in coastal waters for capturing various fish and shellfish species. They have been called set nets (or setnets), bag nets, pound nets, stake nets, weirs, traps, and other names. In this chapter, we refer to Japanese traps as “set nets” and other traps as “traps” or other names to coincide with literature. Here we discuss worldwide trap fisheries and fish behavior patterns and exemplify specific fish behavior near Newfoundland cod traps, Japanese set nets, and Baltic traps for salmon and whitefish. The chapter also describes species and size selectivity and con-
7.2 TRAP FISHERIES AND TRAP DESIGNS Traps are used all over the world’s marine and freshwater environment. Traps are fuel-efficient compared with many other fishing gears (Nomura 1980). Fish trapped in the net are usually alive before harvested and allow for selective harvesting and live release. Japan may be the leading nation in the use of the gear and research activities on the subject. In 1986, there were more than 1600 large-scale set nets and 15,000 medium-size and small-scale set nets in use in Japan (Inoue and Arimoto 1989). Annual landing by set nets in later 1980s was about 650,000 metric tons, about 25% of the total landing from coastal waters in Japan. Another area where traps are widely used is eastern Canada, particularly Newfoundland and Labrador. Traps in Newfoundland are primarily used for Atlantic cod (Gadus morhua) but also Atlantic mackerel (Scomber scombrus), Atlantic herring (Clupea harengus), and capelin (Mallotus villosus). In 1920s, there were reported 7365 cod traps in use in the area. In 1989, just before the cod
159
160
Fish Behavior near Fishing Gears during Capture Processes
moratorium, there were approximately 4000 cod traps around the province of Newfoundland and Labrador (He and Nemoto 1999) and accounted for 57% of the cod landings at that time. Cod traps virtually disappeared in 1992 when the Northern cod was closed to commercial fisheries. 7.2.1 Early Traps Traps are one of the oldest commercial fishing gears in the world. Stonewalls constructed to trap fish in tidal waters, estuaries, and rivers are reported to date back to Neolithic times or earlier (Nishimura 1964). Graffiti found in a grotto on an Italian island depicted trapping systems made of palm tree branches for bluefin tuna (Thunnus thynnus) some 4000 years ago (Sara 1980). Sara (1980) further described more modern tuna traps from various sources as early as Aristotle’s time. Trap designs from 2 a.d. by Oppiano very much resemble the present-day traps with “entrances, gateways and access roads.” Reef nets were used by native peoples in Vancouver Island on the Canadian west coast many centuries and probably thousands years ago. Earlier reef nets use kelps spliced into cordage made of natural fiber to form walls to guide salmon into the net between two canoes (Claxton and Elliott 1994).
The natives believed that salmon are more comfortable to enter the reef nets surrounded by kelps. Remains of wooden stakes with formation of chevron- and heart-shaped traps were recently found in Vancouver Island (Greene 2005). Isotope aging indicated that these remains are more than 1000 years old. At the beginning of the 1900s, Europeans introduced large-scale salmon traps to the same fishing grounds using the same fish migration and other behavior patterns. The large traps consist of wooden poles with the netting hung onto them. These large traps were used until1958, when more restrictive fishing regulations, increased cost of material and operation, and development of the modern seine fishery made the trap less feasible. In mid-1990s, a new floating salmon trap (Fig. 7.1) was tested on the same fishing ground where the traditional reef net and staked salmon traps were once operated (He and Walsh 1997, 1998). Fishing for salmon by traps has a long history in Alaska. The first reported use of a salmon trap was in 1885 (Hipkins 1968). The first salmon traps were pile traps and could only be used in shallow water with a good soft bottom so that the piles could be drilled. The first floating salmon trap was built in 1890 and was reported to have been imported from Norway. In 1930, there were 423 floating traps in
Figure 7.1. A new floating salmon trap tested in Vancouver Island, British Columbia. (Courtesy Gordon Curry.)
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges use in Alaska. This trap was abolished from Alaskan waters in 1959 due to its high fishing efficiency. By 1967, only three traps were operated by Indian Reserves. In 1994, only one salmon trap was operated by an Indian Reserve on Annette Island in Alaska. Weirs made of sticks were used to catch Atlantic salmon (Salmo salar) as early as 1606 in the Canadian Maritime provinces (Anderson and Brimer 1976). There were records of partition to erect fish weirs in New Hampshire in 1729 according to the same source. 7.2.2 Japanese Set Nets A variety of stationary fishing gears resembling the present-day set nets were used in Japan in the early years of the second millennium. A prototype of set net known as oshiki-ami (Fig. 7.2) was used during the Edo period around 1600 a.d. (Inoue et al. 2002). It had a triangle-shaped bag net with a leader net. A similar set net but with a box-shaped bag net
161
called daibo-ami (Fig. 7.2) appeared around 1900. The net was, however, liable to deformation by currents and allowed easy escape of fish. Therefore, the net and schools of fish had to be monitored by watchers constantly and the net had to be hauled soon after the entry of a fish school. Many workers and many boats were required to haul the entire net. Otoshi-ami, which has additional chambers and a bag net, was developed from the daibo-ami around 1910. Otoshi-ami is the most popular set net in Japan at the present. In comparison to the daibo-ami, the escape of the fish is effectively prevented, and the net can be hauled at a certain time of the day rather than immediately after the entry of a school of fish. Only the bag net is hauled up, so it requires less labor and less time to haul. Large set nets are used to catch large pelagic species such as tuna, yellowtail, and salmon as well as demersal species such as sea bream. These set nets are normally owned by fishery cooperatives and involve as many as 75 fishermen in the setting
Figure 7.2. Evolution of Japanese set nets.
162
Fish Behavior near Fishing Gears during Capture Processes
and hauling (Inoue 1988). These large-scale set nets can have variations in the number of entrances and bag nets. The bag net can also be located on the surface, in mid-water, or on the seabed. Mediumscale pelagic set nets are used for small schooling pelagic fish. The largest trap is reported to be 1000 m long and 500 m wide with a leader of as long as 5 km. One of the most important species groups caught by set nets is chum salmon (Oncorhychus keta). About 100,000 tons of salmon are harvested annually in Japan, of which 60% are from coastal waters. More than 95% of coastal salmon landings are from set nets. Occasionally, 10,000 chum salmon are caught in one haul (Inoue and Arimoto 1989). Figure 7.3 illustrates a larger-scale Japanese salmon set net of otoshi-ami style showing the names of the parts of the net. Traditionally set nets are set on the surface, but recently mid-water set nets have been developed and used in northern Japan. Mid-water set nets are more resistant to bad weather and sea conditions and have been reported to catch more salmon than surface set nets.
Figure 7.3.
Japanese set nets are also used for demersal fish species such as flatfishes and Gadidae. The Japanese-style cod trap in use in Newfoundland (described later) was introduced from northern Japan, where it was used to catch Pacific cod (He and Nemoto 1999). Set nets are also used to catch shellfish species such as squid in Japan. These shellfish traps are normally small in scale and operated in coastal waters of shallow depths. The set net technology has been extensively developed in Japan; however, it still possesses the fundamental characters of passive fishing gear. For example, the catches are not very stable because fishing depends on the migration of fish, which may vary with environmental conditions. In recent years, various improvements have been rendered to make this fishery more stable and profitable. 7.2.3 Newfoundland Cod Traps The cod trap is one of the most cost-efficient fishing gears and can be spectacularly successful when used under the appropriate conditions. Since its “invention” in the 1860s, the cod trap has been
An otoshi-ami–style Japanese set net with names of different parts of the net.
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges modified to catch squid, mackerel, herring, and capelin. In more than 120 years of development, three distinct cod trap styles have emerged—the traditional Newfoundland cod trap, the modified Newfoundland cod trap, and the Japanese-style cod trap. In 1989, there were about 4000 active cod traps in Newfoundland, of which 48.6% were the traditional traps, 37.6% were modified traps, and 13.7% were Japanese traps. The traditional Newfoundland cod trap consists of a square box and a leader. A typical cod trap is 60 fathoms on the round and 10 fathoms deep. The
163
trap is box-shaped and without winker panels (Fig. 7.4). The box has a bottom but does not have a roof. The trap is normally set on the seabed with its floatlines 1 to 2 fathoms below the water surface. The footropes of the trap are weighted with lead ropes and sand bags are added on the corners. The length of the leader is 200 to 300 m depending on the location and bottom topography. In many cases, the leader is fastened at the shore. The Japanese-style cod trap was introduced into the Newfoundland cod fishery in the early 1960s from northern Japan, where the traps were used to
Figure 7.4. Three popular styles of Newfoundland cod traps. (A) Traditional Newfoundland cod trap. (B) The modified Newfoundland cod trap. (C) The Japanese-style cod trap.
164
Fish Behavior near Fishing Gears during Capture Processes
catch Pacific cod (Fig. 7.4). The major difference between the Japanese trap and the traditional Newfoundland traps is that the former has a porch and a funnel leading to the box-type net. There is a roof over the porch and in the box. The Japanese traps are often submerged 5 to 7 m below the surface. The modified Newfoundland trap was styled from the traditional Newfoundland trap by adding two winker panels at the entrance. The modified traps are typically set in waters 3 to 5 m deeper than the depth of the net with its headlines submerged. There is no roof on the trap. 7.2.4 Baltic Fish Traps In the Baltic Sea, traps are used for Atlantic herring, Atlantic salmon, and European whitefish (Coregonus lavaretus). Whitefish traps were evolved from the hoop net in the middle of nineteenth century in the west coast of Finland and their use spread across the bay to the Swedish coast (Toivonen et al. 1992). A typical whitefish/salmon trap consists of a leader, a wing net, one or more sections of middle chambers, and a bag net where fish are raised and harvested (Fig. 7.5) (Lehtonen and Suuronen 2004; Toivonen et al. 1992). There were 200 to 300 whitefish traps along the Finnish coast in the early 1990s, targeting the spring migration of whitefish as they enter Bothnia Bay. A trap can catch as much as
10,000 kg of whitefish in one fishing season, with salmon and brown trout as bycatch. The annual landing of whitefish from traps grew from 10 tons to about 110 tons in the 1980s (Toivonen et al. 1992). Each whitefish trap is different in design. A survey by Toivonen et al. (1992) of fishermen from the Finnish coast indicated that the total trap lengths ranged from 206 to 468 m, of which 150 to 400 m were leaders. The width of the trap (between wings) was 32 to 55 m. Leaders are made of large-mesh polyethylene materials, and other parts of the net are made of nylon. Traditionally, salmon traps use mesh sizes that result in meshing. Whitefish traps use thick twines and mainly use a guiding mechanism to lead the fish into the bag net. Salmon and whitefish traps in the Baltic Sea have undergone tremendous changes in recent years due to increasing interaction between traps and seals. Seals eat fish that are captured or concentrated in the traps and can be accidentally caught by the trap. A discussion on seal-proof traps is provided later. Herring traps in use in the Baltic Sea are also called herring pound net and were introduced from North America at the end of nineteenth century (Toivonen et al. 1992). Most herring traps operate in spring during the spawning migration. The general shapes of herring traps are very similar to
Figure 7.5. A typical salmon/whitefish trap in use in the Baltic Sea. (Redrawn after Lehtonen and Suuronen 2004.)
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges that of whitefish trap, except that herring traps have much smaller mesh size (Tschernij et al. 1993). A typical trap can land 15 to 20 tons of herring for a season of 1 to 2 months. 7.2.5 Other Traps Scottish bag nets are commonly used for catching salmon in waters beyond defined estuaries. However, their numbers are decreasing as the abundance of wild stock of salmon decreases. These bag nets take advantage of salmon swimming along the coast to reach their spawning river. Bag nets are floated off the seabed. Juvenile bluefin tunas are captured by traps in the Mediterranean Sea and transferred to cages to grow up and fatten for the Japanese sushi market. Bluefin tuna traps are also used on the Canadian east coast. These traps are large-scale fishing gears. Bluefin tuna are large and they need to swim constantly to stay alive; therefore, the box of the trap needs to be large to hold the fish. The main part of trap measures about 44 × 41 m. Whitefish traps are also used in inland waters of Canada and the United States. Traps, or pound nets, are used in the mid-Atlantic coast of the United States, targeting a variety of pelagic species primarily used for bait in the blue crab fishery in Chesapeake Bay. Mackerel and herring traps in use in Newfoundland have very similar designs; in many cases, fishermen use the same trap for both species. A heart-shaped herring trap, or weir, is used in the coastal waters of Quebec and Nova Scotia. Capelin is a small pelagic species found in the North Atlantic, notably Norway and Newfoundland. Capelins spawn on sand beaches along Newfoundland, and some stocks spawn on offshore banks. Capelin is caught with traps and purse seines in Newfoundland during June and July. Capelin traps are very similar to cod traps in shape, except that they have much smaller meshes (around 48 mm) and they are set near the surface. Traps are used widely in other parts of the world. Various types of traps, some with designs similar to the Japanese set nets, are used in various locations along the Chinese coast (Feng et al. 1987). The use and designs of other traps in different parts of the world are described by Gabriel et al. (2005).
165
7.3 FISH BEHAVIOR IN AND AROUND TRAPS As mentioned earlier, a good understanding of fish behavior is very important for successful trap fishing. As a large-scale fishing gear, these traps are often set in a fixed location for the entire season. Failure to understand fish behavior, especially the migration and movement patterns, can result in catch failure for the season. Here we discuss fish behavior in three major trap fisheries—Newfoundland cod trap, Japanese salmon set net and Baltic salmon/whitefish trap. 7.3.1 Cod Behavior near a Newfoundland Cod Trap He (1993) reported a fish behavior study near a modified Newfoundland cod trap using a seabedmounted sonar and an underwater video camera. The trap under observation was a typical cod trap of the “modified” design. The trap was 120 m on the round and 22 m deep. The trap was set at 26-m water depth at the trap entrance, with its floatline 4 m below the surface. The following summarizes findings reported by He (1993). Cod at a distance from the trap. Whether fish at a distance from the trap swim into the trap depends on their route of migration, presence of prey or predator near the trap, or water temperature. Capelin are reported to play a significant role in the inshore migration of cod in Newfoundland. Cod near the leader. The leader of a trap is designed to intercept and guide the fish toward the trap entrance. The basic behavior of fish when blocked by a wall of netting is to swim along the leader toward the deep end (Fig. 7.6). However, if fish approached the trap from the trap box area, the fish may be guided toward the shore (Fig. 7.6). Fish were very reluctant to swim through meshes during voluntary swimming. During the experiment with the cod traps, very few cod were observed to swim through the leader with a mesh size of 203 mm during fishing conditions. The leader is visible, especially when grown with algae, after it is in water for some time. Cod at the entrances of the trap. There are two entrances separated by the leader in cod traps in Newfoundland. Cod were seen to swim into the trap from both entrances. Likewise, they swam out from
166
Fish Behavior near Fishing Gears during Capture Processes
Figure 7.6. Movements of schools of cod (Gadus morhua) near a modified Newfoundland cod trap. (A) Fish guided away from the trap entrance. (B) Fish guided toward the trap entrance. (Redrawn from He 1993.)
both entrances. There are a number of examples indicating that schools of fish swam into the trap from one entrance and a portion of it exited the other entrance. On one occasion, a large school of more than 1100 cod was observed by camera to swim into the trap in 3 min, representing an entering rate of 367 fish/min. However, not all the schools of fish swam into the entrance in a smooth manner. In many instances, a school of fish paused at the entrance while only a portion of it swam into the trap. In one occasion, a large school of cod was spotted with the use of sonar to swim into and subsequently exit the trap in a period of 3.5 min (He 1993). The swimming speed of cod entering and exiting the entrances of the trap was less than one body length per second (BL/s), well below the maximum sustained swimming speed for this species of similar size (approximately 50 cm). Startled schools of fish showed a speed of 4 BL/s. Cod of 50 cm can
swim at 2 BL/s for an indefinite duration and up to 10 BL/s during a brief burst (see Chapter 1). The depth of fish near the trap seemed to relate to the ground swell, which in turn was related to both the wind direction and strength. When there were heavy ground swells, fish stayed 10 to 15 m off the bottom (26-m water depth). On a calm day with no ground swells, fish stayed very close to bottom, often as close as 0.5 m. Cod inside the trap. Cod inside the trap were observed to swim in circles similar to that observed in a large tank. They swam in either a clockwise or a counterclockwise direction without preference. Cod remained about 1 m from the netting of the trap wall. Large schools of cod were observed to swim into and exit out of the trap at ease, indicating that the design of the entrance can be modified to reduce a large-scale exodus. Fish were also observed to swim through 92-mm mesh netting “drying twine” during hauling. This escape behavior may be related
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges
167
Figure 7.7. Distribution of cod (Gadus morhua) in relation to wind-induced water temperature variations and vulnerability to cod traps set at water depth of 22 m (12 fathoms) in northeastern Newfoundland. (Adapted after Lear et al. 1986.)
to stress associated with reduced space and crowding. Effect of temperature. The success or failure of cod trapping in Newfoundland has largely been related to fish abundance and environmental factors. Temperature is a very important factor affecting the route and timing of migrations, as well as relative distances from the shore and depth of water they reside (Rose 1993). Cod in the Newfoundland area prefer waters with temperature between 0.7° and 4.2°C (Lear 1984). In most of the Newfoundland coast, the peak cod trap season was June and July. Strength and direction of prevailing winds can affect the availability of cod to cod traps set near the shore, as illustrated in Figure 7.7. 7.3.2 Fish Behavior near Japanese Set Nets Fish behavior around the set net The behavior of large schools of sardine (Sardinops melanosticta) in and around the set net was analyzed from sonar image recordings made with a scanning sonar (Inoue 1988; Kim et al. 1995). When large schools of sardine moved along the outside of the set net, the shape of the school gradually changed—the front portion extended forward in the direction of movement and the rear portion concentrated in the same direction such that the school retained its original shape. When large
schools of sardine entered the main net of the set net, the school formation was loosened and pointed in various directions before forming a dense school pattern again and then moved directly to the slope net. When the size of the front portion of the fish school enlarged, the maximum recorded moving speeds were 1.76 and 2.77 m/s for schools inside and outside the set net, respectively. Fish behavior near the leader. Many schools of fish were caught due to the effect of blocking and leading by the set net leader. The behavior of the schools of fish of several species was investigated in the set net fishing grounds around the coast of Japan using a scanning sonar to determine the function of the leader (Inoue 1988). The leader was effective in blocking off the course of fish school, although its mesh size was large enough for the fish to pass through. Only 8% of the schools of fish passed through the leader (Fig. 7.8), whereas 76% of the schools of fish encountered the leader moved along it (Inoue 1988). Of those swimming along the leader, three times as many schools of fish moved offshore toward the set net as they did toward the shore. The leader influenced schools of fish as far as 60 m from the leader. The distance between the leader and fish school depends on species. Barracuda (Sphyraena pinguis) remained 6 to 7 m from the leader, while jack mackerel remained 5 to 20 m and
168
Fish Behavior near Fishing Gears during Capture Processes
Figure 7.8. Proportion of fish schools guided by leaders or penetration through leaders of Japanese set nets. (Redrawn from Inoue and Arimoto 1989.)
yellowtail (Seriola quinqueradiata) remained 10 to 15 m away (Nomura 1980). Different mesh sizes are used for the leader according to different species targeted. The ratio of the mesh size of trap leader to that of gillnet targeting the same species can range from 0.4 to as great as 18 (Nomura 1980). For species such as sardine (Sardinia melanosticta), the ratio can be 18 times because they usually stay far from the leader. The ratio for flounder (Paralitchthys oblivaceus) was only 0.4 because smaller mesh sizes used in the leader can avoid meshing of fish in the leader. Behavior of fish in the playground. The behavior of yellowtail schools in the playground of a
large-scale set net was investigated in relation to the catching function of the funnel net by use of scanning sonar (Inoue 1988; Kim and Inoue 1998). More schools of fish were observed in the playground in the morning but more schools were observed in the bag net in the afternoon. The fish remained in the playground for a long time. Yellowtail schools changed the shape when passing the funnel net. The rate of entering the bag net was 24% among the schools of fish heading toward the funnel net (Kim and Inoue 1998). The rate of exit to the playground from the bag net was 27% among those heading toward the funnel (Fig. 7.9). It seems that the funnel was not very effective in leading in
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges
169
Figure 7.9. Fish behavior at the slope net. (A) The proportion of schools entering into the slope net. (B) The proportion of schools exiting from the slope net. (Redrawn from Kim and Inoue 1998.)
the fish but was effective in preventing fish from escaping. Capture efficiency of set nets. By counting the number of salmon schools near the trap and the number of schools swimming into the trap, Inoue (1988) found that almost half of the salmon schools near the trap were captured by the trap. Some 46% of the schools observed with sonar within the 500-m range were captured by the trap (Fig. 7.10). 7.3.3 Salmon and Whitefish Behavior near Baltic Traps Research indicates that Atlantic salmon in the Finnish coast of Baltic Sea follow isothermal surface water between 13° and 14°C during spawning migration (Westerberg 1982). Whitefish, however, swim near the seabed, feeding on gastropods during that period (Toivonen and Hudd 1993a). Whitefish change swimming behavior later in the season and prefer to swim near the surface. These behavioral differences were used to reduce salmon bycatch in whitefish traps (see details later in this chapter). Lunneryd et al. (2002) used tags and an acoustic positioning system and tracked whitefish in relation to an experimental leader net with a net enclosure in Sweden. Three leader mesh sizes were used: 100,
300, and 800 mm. They found that none of the 28 individuals (mean weight 0.85 kg) in 478 h of observation swam through the meshes, even though the size of large meshes (800 mm) was large enough for the largest fish (1.396 kg) to do so. Using a random walking model, they projected that the individuals would have 120 net encounters if they were not avoiding the leader net (Lunneryd et al. 2002). They also found that turning distance is larger with the small-mesh leader when compared with the medium-size leader; presumably the small-mesh leader is easier to detect. Lunneryd et al. (2002) recorded the distance at which whitefish react to a netting panel. They found that there were two peak frequencies: one at 3 to 5 m and the other at 15 to 20 m. They argued that the reaction at 3 to 5 m may be visual, but the reaction at the 15- to 20-m range could not be explained. Acoustic pressure from the netting was not strong enough to elicit a response, as measured sound pressure level produced by the leader structure were not of sufficient intensity to be detectable by whitefish at greater than 4 m (Wahlberg et al. 2000). Both salmon and whitefish showed diurnal activity. They were most active between 09:00 and 11:00 h (Toivonen and Hudd 1993a). As a result, they speculated that “fishermen may have disturbed
170
Fish Behavior near Fishing Gears during Capture Processes
Figure 7.10. Proportion of fish schools guided by the leader in Japanese set nets as observed by scanning sonars. (Redrawn from Inoue 1988.)
the entering of fish into their trap by hauling between 06:00 and 09:00.” However, there was little difference of turning distance between night and day. 7.4 FISH BEHAVIOR AND TRAP DESIGNS Traps are passive fishing gears as they are strategically placed on the fishing grounds for fish to swim into them and be trapped. Therefore, knowledge of fish behavior, particularly schooling and migration patterns, is very important. The behavior of fish differs among species, but some aspects can be generalized. Ancient traps such as reef nets use kelp (dune grass) existing on the fishing ground. It was believed that “salmon entering the reef net felt safely surrounded by the dune grass in swift running tides” (Claxton and Elliott 1994). Traps are set at a fixed location awaiting fish and intercept them on their migration route. Prior
knowledge and predictable migration patterns in terms of route and timing are extremely important to successful trapping operations. Salmon are known to return to native rivers after spending some time at sea. Various early salmon traps such as reef nets were thus developed to catch returning salmon by native residents of the North American west coast (Claxton and Elliott 1994). Atlantic cod migrate to inshore waters of Newfoundland and Labrador during summer months to feed on capelin that spawn on beaches (DFO 1988). Cod traps set during these months near the shore trapped a large amount of cod. On the other hand, a cod trap introduced to northern Norway to catch spawning migrating cod that remained in relative deeper waters was not very successful. The areas that schools of fish frequently pass are referred to as the “fish route.” A successful trap operation depends on the position and angle of net in relation to the fish route. Important factors that
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges
171
Figure 7.11. Illustration of the leader and the trap in relation to bottom contour, headland, and typical fish swimming route.
determine the formation of a fish route include (1) characteristics of the coastal topography and isobaths, (2) presence and location of natural reefs in the area, (3) sea states, and (4) consistency of the seabed. With knowledge of these factors, it is possible to predict types of fish species, size of school, and seasonal changes that occur in a given fish route. Trap locations are called “berth.” Good trap berths are determined through generations of fishing in the area. Most fish seek deeper waters when they encounter danger. Therefore, fish normally swim toward the deeper end of a net set across depths. Traps should therefore be located in the deeper end of the leader to accept fish intercepted by the leader, as illustrated in Figure 7.11. Fish are guided into the playground and, with time, they may swim into the bag net. The history of improving trap and set net designs has therefore focused on two conflicting
goals: (1) allow fish to enter easily and (2) prevent fish from escaping once they have been trapped. Most fish swim against the current; this is called rheotropism. In a laboratory flume tank, fish can be induced to swim against current until they cannot sustain the current and duration. In the field, fish were observed to swim against the current when searching for food and migrating upstream, the latter especially in salmonid. Fish take minimum risk during their routine activities. Experiments in tanks show that fish avoid large-mesh netting panels, even though the mesh sizes are several times larger than their body (Glass et al. 1993). In the field, fish avoid netting, such as the leader of a cod trap with mesh size that is considerably larger than what would effectively gill them (He 1993). Many fish spend their lives in schools; this is particularly true of small pelagics. Fish in a school
172
Fish Behavior near Fishing Gears during Capture Processes
are more affected by their neighbors. In traps and in other fish gears, fish were observed to swim into the nets en masse due to schooling behavior. Presence of the same species of fish on the other side of the leader can cause the fish to swim through the leader, which otherwise does not usually occur. Visibility of the trap and the leader is important in fish avoidance or penetration through the leader. Whitefish (Coregonus lavaretus) show net avoidance both during the day and during the night (Lunneryd et al. 2002). The fish were more likely to detect the net and turn at 15 to 20 m and at 3 to 5 m. The 3- to 5-m detection may be due to visual response, whereas the 15- to 20-m detection may be due to detection mechanisms other than visual, as it does not relate to turbidity and light conditions. Generally speaking, the more visible the leader, the more effective it is in leading the fish along the leader. Visibility of netting underwater is determined by the relative contrast of the twine against background. Wardle (1986) demonstrated how the same twine could have different visibility and appearance depending on the direction at which the twine is being viewed. Yellow twines were reported to have the greatest visibility in clear water, but visibility reduced in turbid water (Nomura 1980). Scientists from Aberdeen (SOAFD 1992) found that glow net was best in blocking Atlantic mackerel (Scomber scombrus) against a blue background in a large tank, whereas monofilament net was the poorest (see Chapter 8). Jarvik (1985) found that the white net leader is better in guiding herring than is dark net in the Estonia coastal herring pound net fishery. 7.5 SIZE AND SPECIES SELECTIVITY AND MORTALITY OF ESCAPEES AND DISCARDS 7.5.1 Size Selectivity Mesh size is the most important factor affecting the size of fish captured in traps. In Newfoundland cod traps, the minimum “dry twine” mesh size is 89 mm. Studies indicated that if the mesh size was increased, the proportion of the small cod (less than 43 cm) would be substantially reduced. As undersized cod can be as much as 66% of the total cod trap landings when using the 89-mm mesh size netting (Brothers
and Hollett 1991). Increasing mesh size would reduce landing in the short term but would contribute to stock recovery. In addition to large mesh sizes in the “dry twine” area, grids were also tested (Brothers 2002). Square mesh panels were tested in the west coast of Newfoundland in 1997 (Brothers 2000). A 6 × 6 m square mesh panel of 117-mm mesh size (85-mm bar length) installed in the drying twine seemed to improve size selectivity and reduce undersized cod (less than 43 cm). Fujimori et al. (2000) reported that diamond mesh hung at 70% has a better passing ratio than at other hanging ratios and than square mesh of the same mesh size when used in the bag net of Japanese salmon set net. In his laboratory tank experiment, the passing ratio, which can be considered as the escape ratio in real fishing situations, increased with time passed and stabilized after approximately 60 s. 7.5.2 Species Selection and Bycatch Reduction Traps generally have good species selection characteristics because traps target schooling fish that are usually of similar species and sizes. However, on some occasions, different species mix and are caught together by the traps. On the Canadian east coast, Atlantic salmon (Salmo salar) has been closed to commercial fishing since early 1990s. However, they were caught in various fish traps, such as cod and capelin traps. In the early 1990s, various experiments were carried out to reduce salmon bycatch in these traps, and some gear modifications were quite successful. The modifications were made primarily to the leader of the cod traps: the sunken leader, the large mesh top panel, and the deflector panels (Fig. 7.12). The sunken leader design (Fig. 7.12A) was the best design and reduced salmon bycatch by 88% (Brothers 1996). Trap leaders made of large mesh sizes in the top portion were also tested in capelin traps with varying degrees of success (Brothers 1997). In the Baltic Sea off the coast of Finland, a large number of Atlantic salmon were caught in whitefish traps during spring shoreward migration. This has resulted in the ban of all stationary gears, including traps, in the state waters during spring and early summer. Research indicates that during spawning
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges
Figure 7.12.
173
Cod trap modification to reduce salmon bycatch. (Redrawn from Brothers 1996.)
migration in the spring, salmon follows isothermal surface water between 13° and 14° while whitefish are feeding on gastropods near the seabed (Toivonen and Hudd 1993a). Installing a 3-m-deep prohibiting net (Fig. 7.13) near the surface made of large-mesh polyethylene material in front of the trap entrance reduced salmon bycatch by 62% without reduction in whitefish catch during May and June (Toivonen and Hudd 1993a). However, the same design resulted in a large reduction (17%) of whitefish as they come near the surface in the fall. 7.5.3 Survival of Fish Discarded from Traps Very few studies were conducted on trap-caught and discarded fish. Judging from the high survival
rate of cod transferred from trap to farming cages for grow-out, the mortality rate of discards from traps may be quite low if handled properly. In fact, several studies have used trap-caught fish for tagging studies (Jokikokko 2002). Mortality of fish discarded from traps is affected by physical injury caused by contact with netting or other gear parts. This is especially serious when there is strong water currents that deform the shape of the netting. Stress due to confinement and hypoxia, especially when there is a large catch, can cause mortality, as seen in purse seines (Lockwood et al. 1983). Undersized salmon less than 60 cm in length caught in traps are required to be released in the
174
Fish Behavior near Fishing Gears during Capture Processes
Figure 7.13. Whitefish trap modifications to reduce salmon bycatch in the Baltic Sea. (Adapted from Toivonen and Hudd 1993.)
Finnish Baltic salmon trap fishery. It is thus important to know whether those salmon released from traps die due to capture and handling processes. An experiment conducted in 1993 indicated that undersized Atlantic salmon caged for 10 days had a mortality rate of 35% (Toivonen and Hudd 1993b). This mortality rate was considered high and might have been caused by caging stress, high surface water temperature, strong currents, and handling. Indeed, a subsequent experiment using tag and recapture method indicated a mortality rate of about 11% (4% to 21%). Furthermore, migration behavior of the trap-caught and released salmon was not altered due to the capture and handling process (Siira et al. 2006). 7.6 CONSERVATION ISSUES AND MITIGATION MEASURES IN TRAP FISHERIES As a type of large-scale fishing gear, fish traps are particularly susceptible to interactions with mega-
fauna species in the sea, including mammals, turtles, and birds. Most traps target schooling pelagic species, such as herring and capelin, which are often the food of the megafauna species. Some mammals take advantage of traps that concentrate fish to catch prey (Fjalling 2005; Suuronen et al. 2006). Conflict arose when humans and animals target the same species, resulting in mortality of animals and loss of or damage to fishing gears (Lien et al. 1989; Wickens 1995). 7.6.1 Interaction of Marine Mammals and Sea Birds with Newfoundland Traps Collisions and other interactions of whale or other large marine animals with cod traps have been reported in Newfoundland, especially before the cod moratoria in 1992 (Lien et al. 1989, 1992). Annual reports from Lien and his Whale Research Group of Memorial University of Newfoundland between 1979 and 1993, as analyzed by He and Howse (1994), indicated that an annual average of 24 large
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges whales, mostly humpbacks (Magaptera noveangliae), were killed as a result of interaction with traps, primarily cod traps in Newfoundland and Labrador. Another 54 were released live as a result of the entrapment and release effort by the Whale Research Group. Whale collisions damage or take away traps, as well as kill or wound the whale. Research to prevent this conflict include the use of whale alarms, acoustic signaling devices that warn the whale of the presence of the net (Fig. 7.14). Other megafauna and nontarget species that interact with cod traps in Newfoundland include dolphins, seals, seabirds, and sharks (Lien et al. 1993). Many marine mammals possess ability to echolocate objects, such as prey items in the water, but their capability varies with species and acoustic characteristics of the habitat (Evans 1973). These animals may also be able to perceive the existence of the net in water through echo-location techniques. Todd (1991) found that traps that use smaller mesh size netting (e.g., capelin traps) have
Figure 7.14. A prototype whale alarm developed by the Whale Research Group of Memorial University of Newfoundland.
175
a stronger acoustic signal than do cod traps that use large meshes. Consequently, more humpback whales were entrapped in cod traps than in capelin traps. These factors, as well as behavior and sensory capability of cetaceans and mechanisms of their detection and interaction with stationary fishing gears, were reviewed by Nelson and Lien (1992). One device tested by Lien et al. (1992) was an acoustic alarm that produces a 3- or 6-s sound at 4-kHz peak frequency with intensity of 135 dB (re 1 μPa at 1 m). Testing of the alarm in cod traps in Newfoundland indicated a significant decrease in the collision and entrapment rate when the alarms were used (Lien et al. 1992). The use of the alarm did not reduce target species (cod) catch during the test period. 7.6.2 Seals and Baltic Fish Traps Interaction of traps and seals in the Baltic are twofold: seals may be caught and killed by traps and traps and captured fish may be damaged by seals. Westerberg et al. (2006) reported that 462 grey seals (Halichoerus grypus) were caught by commercial fishing gears in Sweden, some of which were by traps targeting whitefish and salmon. A recent survey in the northern Baltic Sea by Kauppinen et al. (2005) found that seals, particularly grey seals, damaged at least 37% of salmon (S. salar) in Bothnian Sea and 3% to 9% in other areas in older traps. Damage to whitefish ranged from 5% to 7%. Gear damages due to seal ranged from 2% to 15% per trap hauls. More interactions between seals and traps may have resulted from increased seal population and more offshore setting of traps (Kauppinen et al. 2005). Fjalling (2005) demonstrated that there are hidden effects of seals to the trap fisheries in addition to the observed remains of the damaged fish. Some fish may be consumed whole as seals seem to prefer smaller fish to larger fish, or small parts of remaining fish may have fallen out of the net. The presence of seals in the trap area may discourage fish from swimming into the trap. Damage to the net due to seals allows more fish to escape through the holes made by the seals. Traditional estimates of seal damages based on the number of fish remains in the net could underestimate the seal effect by as much as 46% (Fjalling 2005).
176
Fish Behavior near Fishing Gears during Capture Processes
Various measures were tested or being tested to protect catch from damage from seals. Lehtonen and Suuronen (2004) tested a wire grid at the entrance of the fish bag located at the last section of the trap and a strong and larger fish bag. The grid was made of 2-mm steel cables with 175-mm spacing and was fitted to the entrance vertically (Fig. 7.15). The experimental fish bag was made of Dyneema material of 1-mm diameter and 80-mm mesh size. The commercial fish bag was made of 210/30 twisted nylon and 130- to 140-mm mesh size. The experimental fish bag was almost twice as large in volume. The result indicated that the experiment trap with a grid and a strong, large bag net was able to reduce seal-induced salmon damage in the fish bag by 70%. Catch was higher in the experimental trap, indicating that salmon were not
prevented from entering into the fish bag by the wire grid but seals were prevented. There were no damages to the strong fish bag, indicating that the strong material prevented seals from causing damage. Seals were observed to turn away from the wire grid during underwater observations. Lunneryd et al. (2003) used large meshes (400 mm) in side panels of the first and second sections of a modified trap to allow escape of salmon and trout chased by seals. While the amount of catch of salmon and trout was similar between the two traps, the standard trap with 200-mm mesh size sustained considerably more damage due to seal feeding on fish enmeshed in the side panel. Seal activity around the standard trap was 16 times greater than that around the modified trap. It was estimated that 65% of potential catch might have
Figure 7.15. Grid designs to reduce interaction of seals with salmon and whitefish traps in the Baltic Sea. (Lehtonen and Suuronen 2004.)
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges
177
Figure 7.16. A rope leader design to reduce turtle bycatch in the mid-Atlantic pound net fishery. (Adapted from DeAlteris and Silva 2007.)
been lost in the standard trap due to seal damages, while 52% of potential catch might have been lost due to escape through the large meshes. Using large mesh in the trap makes it less rewarding to seals to feed inside the trap and may lead to fewer seals on the trap grounds in the long term (Lunneryd et al. 2003). Suuronen et al. (2006) tested two traditional and five modified salmon traps during 2003 and 2004 on the Finnish coast of Baltic and found that a combination of large meshes in the wings together with a protected fish bag and a wire grid at the entrance to the fish bag can reduce seal damage and improve catch. More recent work includes develop-
ment of a pontoon trap to catch live seals and release them far away from the traps (Lehtonen and Suuronen 2010). 7.6.3 Turtles and the Pound Net Fishery in the Mid-Atlantic Coast of the United States Several species of sea turtles, including hawksbill (Eretmochelys imnricates), green (Chelonia mydas), leatherback (Dermochelys coriacea), Kemp’s Ridley (Lepidochelys kempii), and loggerhead (Caretta caretta), seasonally visit Chesapeake Bay in the eastern United States. They are occasionally caught in the leader and the bag net of pound nets
178
Fish Behavior near Fishing Gears during Capture Processes
targeting various fish species such as weakfish (Cynoscion regalis), croaker (Micropogonias undulates), harvest fish (Peprilus alepidotus), butterfish (Peprilus triacanthus), and threadfin shad (Dorosoma petenense). To reduce interaction with turtles and subsequent mortality, DeAlteris and Silva (2007) tested a modified leader made of a combination of vertical ropes (top two-thirds) and netting (lower one-third) (Fig. 7.16). The theory was that pelagic species such as harvestfish, butterfish, and threadfin herring would be guided by vertical lines toward the bag net located at the deep end of the leader, whereas turtles, which are usually also near the surface, would pass through the gaps between the ropes (61-cm spacing) without being caught. Comparative fishing between the modified leader and commercial leaders made of 280-mm mesh size indicated a substantial reduction in turtle interaction. Twenty-three turtles interacted with the leaders of four pound nets with a control leader and none with the same nets with an experimental leader. The use of the modified leader effectively eliminated turtle encounters during the experimental time period. As a result, pound net leader specifications similar to that tested were put into rule by the National Marine Fisheries Service (Federal Register 2006). The rule requires that all offshore pound net (greater than 3 m in depth at the land end of the leader) must use hard lay ropes of greater than 8-mm diameter in the top part of the leader (twothirds) with spacing not less than 61 cm. The mesh size of the bottom one-third netting must not be greater than 203 mm. 7.7 CONCLUDING REMARKS Trap is one of the oldest fishing gears used in commercial fisheries. Traps are used all over the world, especially Japan and Newfoundland, catching a variety of fish and shellfish species. Successful operation of the trap requires good understanding of fish behavior, especially migration pattern and reaction to fishing gears. The design of a trap is a compromise between two conflicting objectives: easy access for the fish and prevention of fish escape. Challenges to trap fisheries include interactions with mammal mammals and other charismatic species. These interactions cause morality to the animal and damage to the target species and fishing
gears. Mitigation measures are promising for some fisheries, with enhanced research efforts required in other fisheries. REFERENCES Anderson GJ and Brimer AE. 1976. Salar: The Story of the Atlantic Salmon. New York: The International Atlantic Salmon Foundation. 74 pp. Brothers G. 1996. Salmon bycatch in cod traps. CASEC Project Summary. Department of Fisheries and Oceans, St. John’s, Newfoundland, Canada. 6 pp. Brothers G. 1997. Salmon bycatch in capelin traps. CASEC Project Summary. Department of Fisheries and Oceans, St. John’s, Newfoundland, Canada. 5 pp. Brothers G. 2000. Testing square mesh panels in trap nets to reduce the catch of juvenile Atlantic cod. ICES CM/J: 15: 8 pp. Brothers G. 2002. Cod trap selectivity: an experiment to reduce the catch of small fish with the use of rigid grates. DFO Proj. Sum. EACT-25.2002.DFO (FDP 374). 5 pp. Brothers G and Hollett J. 1991. Effect of mesh size and shape on the selectivity of cod traps. Can. Tech. Rep. Fish. Aquat. Sci. 1782: 73 pp. Claxton E Jr and Elliott J Jr. 1994. Reef net technology of the saltwater people. Brentwood Bay, BC, Canada: The Saanich Indian School Board. 55 pp. DeAlteris J and Silva R. 2007. Performance in 2004 and 2005 of an alternative leader design on the bycatch of sea turtles and the catch of finfish in Chesapeake Bay pound nets, offshore Kiptopeake, VA. Final summary report submitted to National Marine Fisheries Service. DFO. 1988. The science of cod. Fo’c’sle. 8(2): 29 pp. St. John’s, Newfoundland: Canadian Department of Fisheries and Oceans. Evans WE. 1973. Ecolocation by marine dolphinids and one species of freshwater dolphin. J. Acoust. Soc. Am. 54: 191–199. Federal Register. 2006. Sea turtle conservation; modification to fishing activities. 50 CFR Parts 222 and 223. Vol. 71, No. 121. June 23, 2006. Feng S, Huang X and Ma S. 1987. China Atlas of Marine Fishing Gears. Hangzhou, Zhejiang, China: Zhejiang Sci. Technol. Pub. 386 pp. Fjalling A. 2005. The estimation of hidden sealinflicted losses in the Baltic Sea set-trap salmon fisheries. ICES J. Mar. Sci. 62: 1630–1635. Fujimori Y, Abe K, Zhimizu S and Miura T. 2000. Analysis of the escape behavior of juvenile salmon
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges Oncorhynchus keta from the bag-net for bycatch prevention in a set-net fishery. Fish. Sci. 66: 424–431. Gabriel O, Lange K, Dahm E and Wendt T. 2005. Von Brandt’s Fish Catching Methods of the World. 4th ed. Oxford: Blackwell. 523 pp. Glass CW, Wardle CS and Gosden S. 1993. Behavioral studies of the principles underlaying mesh penetration by fish. ICES Mar. Sci. Symp. 196: 92–97. Greene N. 2005. A new angle on northwest coast fish trap technologies: GIS total station mapping of intertidal wood-stake features at Comox Harbor, BC. Presented at the Canadian Archaeology Association 2005 Annual Conference, Nanaimo, BC. He P. 1993. The behavior of cod around a cod trap as observed by an underwater camera and a scanning sonar. ICES Mar. Sci. Symp. 196: 21–25. He P and Howse K. 1994. Groundfish Harvesting Technologies: An Annotated Bibliography. Fishing Technology Unit Report 11/94. St. John’s, Newfoundland: Fisheries and Marine Institute. 115 pp. He P and Nemoto M. 1999. The Newfoundland cod trap: its origin, development and fishery. Trap Net Fish. 95: 34–43. (in Japanese). He P and Walsh P. 1997. Behavior of salmon near the new Sooke trap. Fishing Technology Unit Report 02/97. St. John’s, Newfoundland: Fisheries and Marine Institute. He P and Walsh P. 1998. Behavior of salmon near the new Sooke trap. Fishing Technology Unit Report 02/98. St. John’s, Newfoundland: Fisheries and Marine Institute. 18 pp. Hipkins FW. 1968. Construction and operation of a floating Alaska salmon trap. US Fish Wildl. Bur. Comm. Fish. Leaflet 611: 12 pp. Inoue Y. 1988. Fish behavior in the set net fishing grounds using a sonar. Bull. Nat. Res. Inst. Fish. Eng. Japan. 9: 227–287 (in Japanese with English abstract). Inoue Y and Arimoto T. 1989. Scanning sonar survey on the capturing process of trapnet. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 417– 421. St. John’s, Newfoundland: Marine Institute. Inoue Y, Matsuoka T and Chopin F. 2002. Technical guide for set-net fishing. International Set Net Fishing Summit in Himi, Kita-Nihon Kaiyo Center, Himi, Japan. 42 pp. Jarvik A. 1985. Possibility for rationalization of a spring spawning herring poundnet fishery. Finnish Fish. Res. 6: 118–126.
179
Jokikokko E. 2002. Migration of wild and reared Atlantic salmon (Salmo salar L.) in the river Simojoki, northern Finland. Fish. Res. 58: 15–23. Kauppinen T, Siira A and Suuronen P. 2005. Temporal and regional patterns in seal-induced catch and gear damage in the coastal trap-net fishery in the northern Baltic Sea: effect of netting material on damage. Fish. Res. 73: 99–109. Kim MK and Inoue Y. 1998. Studies on the behavior of fish schools in the main-net of a large scale setnet using scanning sonar V. The behavior of yellowtail Seriola quinqueradiata school entrapped in a large set-net and the catching function of the funnel-net. Bull. Kor. Soc. Fish. Tech. 34(1): 13–20 (in Korean with English abstract). Kim MK, Inoue Y and Park JS. 1995. Studies on the behavior of fish schools in the main-net of a large scale set-net using scanning sonar II. The behavior of large school of sardine Sadinops melanosticta in and around the Set-net. Bull. Kor. Soc. Fish. Tech. 31(1): 8–14 (in Korean with English abstract). Lear WH. 1984. The winter distribution of cod in NAFO Division 2J, 3K and 3L, based on research vessel catches during 1978–1983. NAFO SCR Doc. 84/VI/24. 799: 9 pp. Lear WH, Baird JW, Rice JC, Caescadden JE, Lilly GR and Akenhead SA. 1986. An examination of factors affecting catch in the inshore cod fishery of Labrador and Eastern Newfoundland. Can. Tech. Rep. Fish. Aquat. Sci. 1469: 71 pp. Lehtonen E and Suuronen P. 2004. Mitigation of sealinduced damage in salmon and whitefish trapnet fisheries by modification of the fish bag. ICES J. Mar. Sci. 61: 1195–1200. Lehtonen E and Suuronen P. 2010. Live-capture of grey seals in a modified salmon trap. Fish. Res. 102: 214–216. Lien J, Barney W, Todd S. et al. 1992. Effect of adding sounds to cod traps on the probability of collisions by humpback whales. In: Marine Mammal Sensory Systems. pp 701–707. Thomas JA, Kastelein RA and Supin AY (eds). New York: Plenum. Lien J, Barney W, Ledwell W. et al. 1993. Incidental entrapments of marine mammals by inshore fishing gears reported in 1992, some results of bycatch monitoring and tests of acoustic deterrents to prevent whale collisions in fishing gear. Whale Research Group, Memorial University of Newfoundland, St. John’s, Newfoundland. 28 pp. Lien J, Stenson GB and Ni IH. 1989. A review of incidental entrapment of seabirds, seals and whales
180
Fish Behavior near Fishing Gears during Capture Processes
in inshore fishing gear in Newfoundland and Labrador: a problem for fishermen and fishing gear designers. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 67–71. St. John’s, Newfoundland: Marine Institute. Lockwood SJ, Pawson MG and Eaton DR. 1983. The effect of crowding on mackerel (Scomber scombrus L.)—Physical condition and mortality. Fish. Res. 2: 129–147. Lunneryd S-G, Fjalling A and Westerberg H. 2003. A large-mesh salmon trap: a way of mitigating seal impact on a coastal fishery. ICES J. Mar. Sci. 60: 1194–1199. Lunneryd S-G, Westerberg H and Wahlberg M. 2002. Detection of leader net by whitefish Coregonus lavaretus during varying environmental conditions. Fish. Res. 54: 355–362. Nelson D and Lien J. 1992. A review of gear and animal characteristics responsible for incidental catches of cetaceans in fishing gear, along with proposed solutions. Proceedings of the World Fisheries Congress, Athens, Greece, May 3–8, 1992. Nishimura A. 1964. Primitive fishing methods. Ryukyuan Culture and Society. pp. 67–77. Nomura M. 1980. Influence of fish behavior on use and design of setnets. In: Fish Behavior and its Use in the Capture and Culture of Fishes. ICLARM Conf. Proc. 5: 446–472. Bardack JE, Magnuson JJ, May RC and Reinhart JM (eds). Manila, Philippines: International Center for Living Aquatic Resource Management. Rose GA. 1993. Cod spawning on a migration highway in the northwest Atlantic. Nature. 366: 458–461. Sara R. 1980. Bluefin tuna trap fishing in the Mediterranean. ICCAT Col. Vol. Sci. 11: 129–144. Scottish Office Agriculture and Fisheries Department (SOAFD). 1992. Marine Laboratory Annual Review 1990–1991. Scottish Office Agriculture and Fisheries Department, Aberdeen, UK. 84 pp. Siira A, Suuronen P, Ikonen E, and Erkinaro J. 2006. Survival of Atlantic salmon captured in and released from a commercial trap-net: Potential for selective harvesting of stocked salmon. Fish. Res. 80: 280–294.
Suuronen P, Siira A, Kauppinen T, Riikonen R, Lehtonen E and Harjunpaa H. 2006. Reduction of seal-induced catch and gear damage by modification of trap-net design: Design principles for a seal-safe trap-net. Fish. Res. 79: 129–138. Todd SK. 1991. Acoustic Properties of Fishing Gear: Possible Relationships to Baleen Whale Entrapment. MSc Thesis, Memorial University of Newfoundland. 213 pp. Toivonen A and Hudd R. 1993a. Behavioral differences of Atlantic salmon (Salmo salar) and whitefish (Coregonus lavaretus) as the basis for improving the species selectivity of whitefish trapnets. ICE Mar. Sci. Symp. 196: 51–58. Toivonen A and Hudd R. 1993b. Survival of undersized salmon after release from the trap net. ICES CM/B:10: 6 pp. Toivonen A, Hudd R and Heikkilä P. 1992. European whitefish trap net fishing gears in the southern part of the Bothnian Bay (Baltic). Pol. Arch. Hydrobiol. 39: 879–884. Tschernij V, Lehtonen E and Suuronen P. 1993. Behavior of Baltic herring in relation to a poundnet and the possibility of extending the poundnet season. ICE Mar. Sci. Symp. 196: 36–40. Wahlberg M, Lunneryd SG, Bégout-Anras ML and Westerberg H. 2000. Whitefish leader net avoidance: Possible role of auditory cues. Adv. Fish Telemetry. Fish. Aquaculture Sci. 137–147. Wardle CS. 1986. Fish behavior and fishing gear. In: Pitcher TJ (ed). The Behavior of Teleost Fishes. pp 463–495. London: Croom Helm. Westerberg H. 1982. Ultrasonic tracking of Atlantic Salmon (Salmo salar L.) – I. Movements in coastal regions. Rep. Inst. Freshw. Res. Drottoinghoim. 60: 81–101. Westerberg H, Lunneryd A-G and Fjalling A. 2006. Reconciling fisheries activities with the conservation of seals throughout the development of new fishing gear: a case study from the Baltic fishery-grey seal conflict. Am. Fish. Soc. Symp. 49: 587–597. Wickens PA. 1995. A review of operational interactions between pinnipeds and fisheries. FAO Fish. Tech. Pap. 346: 86 p.
Large-Scale Fish Traps: Gear Design, Fish Behavior, and Conservation Challenges SPECIES MENTIONED IN THE TEXT albacore, Thunnus alalunga alewife, Alosa pseudoharengus American shad, Alosa sapidissima Atlantic cod, cod, Gadus morhua Atlantic herring, herring, Clupea harengus Atlantic mackerel, mackerel, Scomber scombrus Atlantic salmon, Salmo salar barracuda, Sphyraena pinguis blueback herring, Alosa aestivalis bluefin tuna, Thunnus thynnus bluefish, Pomatomus saltatrix butterfish, Peprilus triacanthus capelin, Mallotus villosus chum salmon, Oncorhynchus keta coho salmon, Oncorhynchus kisutch croaker, Micropogonias undulates European whitefish, Coregonus lavaretus flathead, Platycephalus bassensis green turtle, Chelonia mydas grey seal, Halichoerus grypus haddock, Melanogrammus aeglefinus harvest fish, Peprilus alepidotus hawksbill turtle, Eretmochelys imnricates humback whale, Magaptera noveangliae jack mackerel, Trachurus japonicus Japanese mackerel, chub mackerel, Scomber japonicus kawakawa, Euthunnus affinis
181
Kemp’s ridley turtle, Lepidochelys kempii lake whitefish, Coregonus lavaretus leatherback turtle, Dermochelys coriacea loggerhead turtle, Caretta caretta lemon shark, Negaprion brevirostris leopard shark, Triakis semifasciata Pacific bonito, Sarda chiliensis plaice, Pleuronectes platessa redfish, Sebastes marinus sailfish, Istiophorus platypterus saithe, Pollachlus virens sardine, Sardinops melanosticta seabass, Dicentrarchus labrax skipjack tuna, Katsuwonus pelamis sockeye salmon, Oncorhynchus nerka sole, Solea solea sprat, Sprattus sprattus striped bass, Morone saxatilis striped mullet, Mugil cephalus threadfin shad, Dorosoma petenense wahoo, Acanthocybium solandrei weakfish, Cynoscion regalis white marlin, Tetrapturus albidus whiting, Gadus merlangus winter flounder, Pseudopleuronectes americanus yellowfin tuna, Thunnus albacares yellowperch, Perca flavescens yellowtail, Seriola quinqueradiata yellowtail flounders, Pleuronectes ferruginea
Chapter 8 Fish Behavior near Gillnets: Capture Processes and Influencing Factors Pingguo He and Michael Pol
8.1 INTRODUCTION Gillnets are simple and versatile gears that catch a variety of fish and shellfish. Unlike mobile gears such as trawls, gillnets do not need to be towed or moved to catch fish; unlike baited gears such as hooks and pots, gillnets do not require the addition of bait; and unlike fixed gears such as traps and weirs, gillnets do not require fixed structures and are much more easily portable. Gillnets may be one of the simplest fishing gears in design with a plain sheet of webbing salvaged to frame ropes. They are used in every region of the world and operated from small boats of a few meters in length to highly mechanized offshore vessels. On closer examination, however, greater complexity is revealed. Even small details of gillnet construction appear to affect species and size selectivity. Although gillnets are simple in design and operation, the behavior of fish during the gillnet capture process is largely undocumented and not well understood. The history of gillnetting may be as old as that of net making, and references can be traced back to 3000 years ago in Egyptian tombs. The modern commercial gillnet fishery in the Northwest Atlantic dates to the mid-1800s when natural fibers such as cotton and hemp were used to knit the netting. Gillnetting expanded in this region after migration of hauling technology from the Laurentian Great Lakes to Massachusetts in the 1930s. Synthetic materials were tested in fishing nets in the 1950s
and became very popular on both sides of the North Atlantic due to the large catch increase observed and the almost maintenance-free nature of the material (He 2006a; Pol and Carr 2000; Potter and Pawson 1991). Modern gillnet webbings are made as invisible as possible to mesh fish before they can avoid them. Fish meshing into the net are often caught behind their gills, or “gilled,” and thus the term “gillnet” is used, although other methods of capture are also common in gillnets. Gillnets and entangling nets are one of the nine basic fishing gear categories in the U.N. World Food and Agricultural Organization (FAO) classification of fishing gears (Nedelec and Prado 1990). Set gillnets, driftnets, trammel nets, fixed gillnets, and encircling gillnets are five major subtypes in this category. Some key features of these nets are listed in Table 8.1. These different types of nets or the same type of nets of different mesh sizes and rigging may be combined to form a “combination gillnet.” A typical gillnet consists of webbing and frame ropes (headrope and footrope) (Fig. 8.1). The webbing is manufactured as diamond mesh in a single piece, including a selvedge of usually double monofilament along the top and bottom. The webbing is cut to length and lashed at intervals to the headrope and the footrope with hanging line. The headrope may have internal floatation or a series of floats attached to it. Footropes are often
183
184
Fish Behavior near Fishing Gears during Capture Processes
Table 8.1. Types of Gillnets and Their Key Features Gear Type
Important Features
Set gillnets
Anchored/weighted to the bottom; relatively stationary; can be set on the bottom, in midwater or near the surface Not fixed to the bottom; drift with the current; usually near surface; either tied or not tied to the vessel. Three layers of nets; a middle net with a smaller mesh size and two outer nets with larger mesh sizes Hung onto stakes to form a wall or “fence”; usually in tidal and shallow waters or in rivers
Drift gillnets
Trammel nets
Fixed gillnets
constructed from braided line with built-in lead (“leadline”), although a series of single weights may also be used. Typically, the gear is also anchored at both ends with solid weights or Danforth anchors using bridle lines. Buoy lines with buoys and/or highflyers are used to mark both ends of the gear at the surface. Depending on the fishing area, jurisdiction, and the type of fishery, a gillnet may require attachment of tags, pingers, weak links, breakaway swivels, radar reflectors, or acoustic gear-finding transmitters. There may be specific requirements on the size, strength, and density of ropes (e.g., NOAA 2008), as well as the maximum number of nets allowed. 8.2 CAPTURE MECHANISMS, GEAR DESIGNS, AND FISHING EFFICIENCY Four basic mechanisms of fish capture by gillnets can be identified: gilling, wedging, snagging, and entangling (Hovgård and Lassen 2000), as shown in Figure 8.2. • Gilling—caught with the mesh behind the gill cover • Wedging—caught by the largest part of the body
• Snagging—caught by the mouth or teeth or other part of the head region • Entangling—caught by spine, fins, or other parts of the body as a result of struggling Fish may be caught by more than one of these mechanisms in the same gillnet. Key design features of a gillnet include netting material and color, twine diameter and number of filaments, mesh size or opening, vertical and horizontal hanging ratios, and net dimension (length and height). The mesh size of a gillnet determines to a great extent the size of fish caught in the net as proposed by Baranov (1948) in his geometric similarity theory. He predicted that the majority of fish retained by a gillnet would have their length within 20% of the optimal length (modal length). In practice, this relationship may not be that simple, considering a wide range of species, gear design features, and operational conditions (Hamley 1975). While surveying young Atlantic cod (Gadus morhua) in Greenland waters, Hovgård (1996) found that more cod were caught by gilling, and less by other mechanisms, as mesh size was increased. Comparative fishing trials using gillnets of 127and 140-mm mesh size on the south coast of Newfoundland targeting redfish resulted in 3.6 times more fish caught in the smaller mesh size nets (Brothers and Yetman 1982). However, catch rates of Greenland halibut Reinhardtius hippoglossoides increased with larger meshes (Melindy and Flight 1992). Gillnets made of 203-mm mesh size caught 38% more fish in weight than those with 140-mm mesh size (Melindy and Flight 1992). The increase in catch rates was due to the increase in the size of fish caught for the larger-mesh nets as the average weight of fish caught in the large-mesh nets was 3.5 times heavier than those in the smallmesh nets. Comparing gillnets of 180- and 220-mm mesh sizes in the Barents Sea for Greenland halibut, Nedreaas et al. (1993) found that the modal length was 55 cm for the smaller mesh size compared with 66 cm for the large-mesh size. While regulating mesh size to reduce undersized fish has been a common management measure in many fisheries, for some species, such as paddlefish (Polyodon spathula), size selectivity cannot be established,
Fish Behavior near Gillnets: Capture Processes and Influencing Factors
185
Figure 8.1. Schematic illustration of a string of gillnet while fishing. Inset: an anatomy of a gillnet with names of gear components. (He 2006a.)
perhaps due to unusual morphology (Scholten and Bettoli 2006). Hanging ratio is another factor for design consideration. Horizontal hanging ratio is the ratio of rope length (headrope or footrope) to the stretched length of the attached webbing. Vertical hanging ratio is similarly the ratio of the skirt line to the stretched length of the webbing. Both are expressed as a decimal or percentage, with lower values indicating more “slackness” and affecting the shape of the mesh (Fig. 8.3) and resistance to penetration. Takagi et al. (2007) modeled forces acting on a bottom sink gillnet and determined that discontinuities and strong localized forces developed on the net surface. These results suggest that hanging ratios create different relative forces on the meshes, which may increase or decrease the amount of force required
for mesh penetration by fish. Comparative fishing trials between gillnets of different headrope hanging ratios (0.5–0.7) indicated that the best hanging ratio for catching Atlantic cod (Gadus morhua) was 0.6 instead of the traditional 0.5 used by Norwegian fishermen (Angelsen et al. 1979). For European dab, nets with a hanging ratio of 0.2 caught twice as many fish as nets with a hanging ratio of 0.6 (Hovgård and Lassen 2000). Samarayanka et al. (1997) reported 40% more catch of tuna (mostly skipjack tuna, Katsuwonus pelamis, and yellowfin tuna, Thunnus albacares) and sharks with a hanging ratio of 0.5 versus 0.6. Slackly hung gillnets have been found to result in more fish becoming entangled than gilled, which results in poorer size selectivity (Angelsen et al. 1979; Hamley 1975; Samaryanka et al. 1997; Stewart 1987). When
186
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.2. Fish capture by gillnets, illustrating four modes of capture: gilling, wedging, snagging, and entangling.
studying Tilapia, Hamley (1975) obtained a size range (90% of catch) of 18 to 23 cm in a tightly hung net but 8 to 22 cm in a slackly hung net. Sulaeman et al. (2000) attributed the improved
catch efficiency of slackly-hung gillnets to an increase in the mesh per unit area of the net panel. However, a 25% increase in catch was observed by Samarayanka et al. (1997) even when adjusted for the increased area. Webbing material, numbers of filaments in the twine, and twine size affect visibility of the netting in water and the “softness” of the netting, which in turn affects the mechanism of fish capture. Monofilament nets are less visible and generally produce larger catches than multifilament nets (Collins 1979; Larkins 1963, 1964; Pristas and Trent 1977). Larkins (1964) reported that the monofilament nets in a monofilament and multifilament string caught 1.9 to 4.1 times more Pacific salmon than did the multifilament nets in the same string. A larger percentage of fish are gilled in monofilament gillnets than are in nets made of multifilament and multimonofilament, which tend to result in tangling. Thinner twines generally catch more fish (Holst et al. 2002; Hovgård 1996; Hovgård and Lassen 2000), as they are less visible and softer, but they may have poorer size selection (larger selection range) due to elongation when a fish pushes into the mesh (Hansen 1974) and ease of entanglements (Yokota et al. 2001). Turunen (1996) reported no change in size frequency but a 190% increase in catch of pikeperch (Stizostedion lucioperca) when comparing 0.15-mm twine and 0.20-mm twine. Hovgård and Lassen (2000) reported that monofilament nets with 0.16-mm twine caught 2 to 3 times more European dab (Limanda limanda) than a net with 0.28-mm twine. Holst et al. (2002) found that gillnets made of four-strand No. 1.5 multimonofilament twine (0.28-mm diameter) caught about 1.5 times that of six-strand No. 1.5 twine (0.36-mm diameter) for Baltic cod (Gadus morhua). Hovgård (1996) found that fishing efficiency was inversely related to the ratio of twine thickness to mesh size for a number of species in Greenland waters. However, nets made of thin twines are more easily damaged, which may result in increased costs and lost fishing time. Further, they may produce increased catch of undesired species such as crustaceans. Net dimension (height and length) may also affect catch, although net length generally does not alter the species or size composition of catches—it
Fish Behavior near Gillnets: Capture Processes and Influencing Factors
187
Figure 8.3. Explanation of hanging ratio of a gillnet. (He 2006a.)
merely increases effort. Fish may lead along the nets, and length may affect the likelihood of striking the mesh. Net height, however, appears to alter species selectivity and affect fishing efficiency of some species. Height in gillnets is affected by a number of factors—the amount of buoyancy in the headrope, the presence of “tie-down” lines (lines connecting footrope and headrope that restrict the height of a gillnet), and the strength of currents and resistance of the webbing. In the Northwest Atlantic, Atlantic cod are targeted with nets with a great deal of buoyancy, so that they reach their maximum height. Flatfish are targeted using headropes with less floatation and with tie-down lines. 8.3 SIZE SELECTIVITY OF GILLNETS The selection of fish by a fishing gear is the process that causes the catch of the gear to have a different composition, either of sizes or of species, than that of the population on the fishing grounds (Wileman et al. 1996). Selectivity is the quantitative assessment of this selection process. Gillnet selectivity processes, mechanisms, and analysis methods are reviewed by Hamley (1975), Millar and Fryer 1999, Hovgård and Lassen (2000), and Fujimori and Tokai (2001). Gillnet size selectivity curves are approximated as Gaussian or bell-shaped and may have two or more peaks reflecting different mecha-
nisms of capture (discussed earlier) or multiyear class population. Bimodal curves were found to provide the best fit in several studies (Fonseca et al. 2005; Madsen et al. 1999; Moth-Poulsen 2003); in others, the normal scale curve (Revill et al. 2007) or the lognormal (M. Pol, unpublished data) provided the best fit. Efficiency of gillnets is affected by mesh size, webbing material, hanging ratio, twine size, and fish behavior as discussed earlier. However, mesh size is likely the most important factor affecting gillnet size selectivity. Experiments confirm that larger meshes result in catches of more large fish, shifting the selectivity curve to the right as shown in Figure 8.4A; these results conform Baranov’s (1948) geometry similarity rule on fish size and mesh size. If the x-axis is expressed as length divided by the mesh size, selectivity can be expressed as one master curve as seen in Figure 8.4B. Gillnets generally catch larger fish compared with other gears, if the proper mesh size and netting materials are used. Comparative fishing trials have demonstrated that gillnets caught more large fish than other fishing gears. When used simultaneously on the west coast of Greenland, gillnets caught more large Greenland halibut than did longlines (Boje 1991). Although both gears caught fish of the
188
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.4. Selectivity curve of gillnets of different mesh sizes for European hake (Merluccius merluccius), and master curve using transformed length. (Redrawn from data in Fonseca et al. 2005.)
same peak length of about 70 cm and had a similar length range from 45 to 115 cm, longlines caught a larger proportion of fish between 50 and 65 cm, while gillnets caught a larger proportion of fish between 65 and 85 cm. Cod on the Flemish Cap in the Northwest Atlantic are fished by several fleet sectors that can be identified by gear (trawl/mesh size, gillnet, longline) and by country (Boje 1991). Portuguese gillnetters caught the largest cod with an average weight of 2.5 kg, whereas Spanish and Portuguese freezer trawlers caught the smallest cod
with an average weight of 0.4 kg and 0.9 kg, respectively (Boje 1991). Lowry et al. (1994) compared gillnets and trawls with the same mesh sizes ranging from 105 to 130 mm targeting Baltic cod and found that gillnets caught fish with peak lengths 7 to 16 cm longer than fish caught with trawls using the same mesh size. Nedreaas et al. (1993) and Huse et al. (1999) compared 220-mm mesh size gillnets with a 135-mm codend mesh size trawl and No. 12/0 EZbaiter hooks in a longline targeting Greenland halibut in the Barents Sea off northern Norway.
Fish Behavior near Gillnets: Capture Processes and Influencing Factors They found gillnet catches were composed of mostly mature females of large size, whereas the trawl and longline had a much lower percentage of large mature females. The average length of gillnet fish was 65.9 cm, the longline caught fish that averaged 59.6 cm, and the trawl caught fish that averaged 50.1 cm. Comparison of three gear types targeting cod and haddock showed similar results (Huse et al. 2000). The selection range of gillnets are also narrower than that for other gears (Erzini et al. 2003). Santos et al. (2002) compared gillnets to longlines in a hake (Merluccius merluccius) fishery and found gillnets had a narrower size selectivity but found the longlines yielded better-quality fish, attributed to long soak times (more than 8 h) in the gillnets. 8.4 FISH BEHAVIOR AND GILLNET FISHING Fish availability, vulnerability, and mobility are the most important factors influencing fishing efficiency of stationary gears. Horizontal and vertical migrations are well known in many fish species. Diurnal vertical migration related to light levels and semidiurnal vertical excursions related to tide can affect gillnets set on the seabed. Increases in the rate of horizontal movement increases the probability of fish encountering gillnets. The amount of horizontal movement is especially important for set gillnets that await encounter on predicted fishing routes or foraging grounds. Fishing operators therefore need local knowledge of fish availability to set nets in the right place at the right time to be successful. Temperature may be the most important factor affecting distribution, movement, and swimming capacity. Vertical and horizontal temperature distribution patterns can cause localized concentrations and dispersal of fish and make them more or less vulnerable to gillnets (Perry and Neilson 1988; Rose and Leggett 1989; Woodhead 1964). The fishing range (the size of fishing area) of a gillnet and encounter rate of a fish (Engås and Løkkeborg 1994; He 2003; McQuinn et al. 1988) may be reduced at lower temperatures, influencing fishing efficiency of gillnets (Stoner 2004). Swimming speed of fish in relation to temperature has been discussed by Wardle (1975, 1980) and He (1991,
189
1993, 2003) and in Chapter 1. In general, swimming speeds are lower at lower water temperatures (Figure 8.5). He (2003) measured the rate of movement of winter flounder (Pseudopleuronectes americanus) on fishing grounds using a video camera and found the rate of movement was reduced by 70% when bottom water temperature was reduced from 4.4° to −1.2°C. He (2003) further discussed how fishing areas or the active fishing space of a gillnet may be altered due to a change in water temperature and soaking durations (Fig. 8.6). Swimming speed is also related to fish body length, as discussed in Chapter 1. Consequently, larger fish can swim faster and have larger geographical ranges than smaller fish. Gillnets (and other stationary gear) may have an intrinsic size selection property as larger fish are likely to reach the net and become available to the nets set some distance away (Fig. 8.6). This length-related difference in encounter probability in gillnets was discussed by Rudstam et al. (1984), who also applied size-related differences in encounter probability to correct abundance estimates for some freshwater species in the Laurentian Great Lakes. Other factors affecting swimming and local movement may include satiation or hunger (Robinson and Pitcher 1989) and prey density (Asaeda et al. 2001), light level (Gjelland et al. 2004), oxygen level (Beamish 1978), and spawning condition. Robinson and Pitcher (1989) found that the swimming speed of herring (Clupea harengus) was highest when they were hungry, presumably related to active food searching behavior. When prey species were plentiful (high prey density), swimming speed may be slowed down (Asaeda et al. 2001). Angelsen (1981) reported that more male spawning cod and halibut are caught by gillnets than are females because they are more active on spawning grounds. Reports of underwater observations of fish behavior near gillnets in the field are scarce and limited to freshwater or coral reef environments where shallow water and good lighting conditions provided better opportunities for observations either by the naked eyes or by underwater video cameras. Laboratory tank observations of fish capture by gillnets (Potter and Pawson 1991) revealed that Atlantic
190
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.5. Reduction of activity and swimming capacity due to lower water temperature, assuming 1 at the higher end of temperature. (From He 2003.)
salmon (Salmo salar) initially struggled powerfully for less than 30 s. This powerful struggle was followed by a long period of weak activities, similar to the behavior observed in hooked fish (Bjordal and Løkkeborg 1996). Gilled fish tended to swim forward, pulling the net with them. Smaller fish may escape by squeezing through the meshes. Eleven salmon that escaped by squeezing through meshes did so in less than 25 s. Tangled fish were more likely to wrench their head or tail and to swim backward or alongside the net. The fate of capture or escape was also largely determined during the first 25 s (Potter and Pawson 1991). Fujimori et al. (1994) classified rainbow trout (Oncorhynchus mykiss) behavior in laboratory experiments in two ways: swimming straight into the net head-on or contacting the net with abdomen or tail. Visibility (or invisibility) of the net is the most important aspect of gillnet design and operation. Fish vision and underwater visual characteristics of fishing gear components have also been discussed in Chapter 2. The visibility of the net is determined by the fish’s visual characteristics, material of the net, light level and composition, water clarity, contrast of the net, and relative position of the fish to
the net. Angelsen and Huse (1979) tested seven nets made from different materials and colors and found that monofilament nylon was least visible and multifilament nylon was the most visible at various water depths. Wardle (1989) illustrated how different shaded twines hung vertically in water had different visibilities when viewed at different angles (Fig. 8.7). White twines disappeared toward the surface, black twines disappeared near the bottom, and grey twines were least visible when viewed horizontally. Transparent monofilament lines hung vertically are almost invisible when viewed horizontally (Gabriel et al. 2005). Comparison of mesh penetration or avoidance of four types of twines of different colors by Atlantic mackerel (Scomber scombrus) indicated that the fish more readily penetrate the naturally colored (transparent) monofilament netting, whereas nets made from glow twine were more likely to be avoided (Fig. 8.8) (SOAFD, 1992). Faulkner (1994) discussed a “window-pane” gillnet with a highly visible large mesh in the top section and regular gillnet webbing in the bottom section. The highly visible netting on the top was believed to drive surface-swimming fish into the deep water, where
Fish Behavior near Gillnets: Capture Processes and Influencing Factors
191
Figure 8.6. Predicted fishing range of gillnets at different water temperatures and at different soaking durations. (From He 2003.)
they were subsequently caught in the bottom part of the net. Once the fish are in the vicinity of the net, those individuals unaware of the presence of the net may swim into it and become caught. Visibility of the net is reduced when there is low contrast between the net and its background. Smaller-diameter materials are also less visible. Nighttime hours, periods with no moon, high-latitude winter days, deep water, and turbid water (near estuaries, tidal area with muddy bottom) all contribute to lower visibility of the net. Fish are not able to detect the netting easily at lower light levels, increasing the chance of
being caught by gillnets. However, lower light level conditions also cause fish to slow down (Gjelland et al. 2004), which reduces encountering probability with gillnets. The lower visibility of synthetic nets made of thin twines may be largely responsible for the increase in catch rates compared with natural fibers (Potter and Pawson 1991). Fish may also be deterred by strong smells from preservatives used in natural fibers. Vibrations from water current passing through meshes of the netting may also make fish aware of the net (Gabriel et al. 2005). Fish seeing or otherwise detecting the net may turn and swim
192
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.7. Contrast of white, grey, and black twines hung vertically in water in relation to viewing angle. (Redrawn from Wardle 1989.)
parallel to the net, similar to the behavior observed in the leader of a trap (see Chapter 7). Acosta and Appeldoorn (1995) described observations of gillnet catch in relation to soak time and found that catch rate initially decreased after setting for 6 to 10 h but increased between 10 and 20 h. It was argued that the initial decline in catch efficiency may be related to reduction of fish density soon after the net was set as well as increased visibility of the net with meshed fish. Acosta and Appeldorn (1994) observed that several fish turned away after seeing a struggling fish caught in the net. The geometry of a gillnet may also be affected by the fish caught in the net, especially if it is entangled and twisted into several meshes. Grant (2002) observed walleye (Sander vitreus) using a stereo camera system in a shallow lake. He examined three mesh sizes and counted the number of fish blocked (turned away), passing through meshes, and caught in the meshes. Many small fish passed through meshes of large-mesh nets while many large fish were blocked by the small-mesh net, as expected. Of 147 fish observed, 35 were caught, while 46 swam through the meshes, 29 escaped after becoming temporarily wedged or entangled, and 39 were
blocked (turned away) and never contacted the net. These specific numbers are important in determining catchability of specific gillnets during stock assessment surveys. The possibility of increasing catches in gillnets through the use of additional stimuli has been investigated. Properties of fish attraction using bait are discussed in Chapter 5. Baited gillnets were tested in Norway (Engås et al. 2000; Kallayil et al. 2003). The idea of using bait in gillnets came from unexpectedly high catch rates when gillnets were set for a longer duration. It was postulated that fish caught at the beginning of the soak period might have acted as bait to attract more fish to the area (Engås et al. 2000). In some fisheries, such as the deepwater Greenland halibut fishery off Labrador, fishermen may set nets for longer than 2 weeks (Melindy and Flight 1992). In that research, however, catch rates of Greenland halibut were reduced and the amount of spoiled fish increased when soaked for a longer period of time. Baited gillnets caught more Atlantic cod, saithe (Pollachius virens), ling (Molva molva), and Greenland halibut in a study by Engås et al. (2000) but not in one by Kallayil et al. (2003). Analysis of tagged and acoustically tracked Atlantic cod near baited and nonbaited gillnets indicated that fish spent more time near baited gillnets and have more encounters with the nets (Kallayil et al. 2003). Fish were observed to turn and make directional movements toward the bait as far away as 800 m. Fish swam more slowly near baited gillnets than near nonbaited gillnets. Kallayil et al. (2003) attributed a decrease in catch rates to this slower swimming behavior despite more observed encounters. Slower swimming may give fish more time to avoid gillnets. The presence of bait may have interrupted swimming and made fish more aware of the net. In their experiment, Kallayil et al. (2003) noticed that relatively more cod were tangled than gilled or wedged in baited gillnets compared with nonbaited gillnets, indicating that most of fish were not directly swimming into the baited gillnets. 8.5 MEASURES TO REDUCE BYCATCH AND DISCARDS IN GILLNETS Gillnets can catch a variety of species, with many of them being considered as bycatch and subsequently
Fish Behavior near Gillnets: Capture Processes and Influencing Factors
193
Figure 8.8. Avoidance of Atlantic mackerel (Scomber scombrus) to nets of different colors. (Redrawn from data in SOAFD 1992.)
discarded. While the rate of discard from gillnet fisheries was considered low on a global scale (0.5% by weight), discard in some specific gillnet fisheries is quite high (Kelleher 2005). Morizur et al. (1996, cited in Kelleher 2005) reported that up to 100% of fish caught in offshore gillnets soaked for more than
6 days may be discarded due to poor quality. Murawski (1993) reported that 44% of fish by weight were discarded from Gulf of Maine groundfish gillnets in 1991. In this case, the majority of discards from gillnets were due to a lack of market for the bycatch species (Alverson et al. 1994).
194
Fish Behavior near Fishing Gears during Capture Processes
Bycatch of nontarget species may be reduced by understanding and using differences in the vertical and horizontal distribution of fish. Different species may occupy different levels of the water column and some spend a considerable time on the substrate. Video camera observations in the natural environment indicated that winter flounder spent 33% to 68% of time on the substrate with a larger proportion of time at lower temperatures (He 2003). The fish were never observed to rise to more than 0.6 m from the seabed. Although some flounder species take prolonged excursions to higher levels in the water column (Cadrin and Westwood 2004; Walsh and Morgan 2004), they usually reside very close to the seabed. In commercial practice, gillnets targeting flounders often have a reduced vertical height created through reduced floatation or tiedown lines, thus avoiding or reducing catch of other species that live higher off the bottom. Pol (2006) tested the effect of reduction of gillnet height through the addition of spaced weights on the headrope and using nets with double footrope and no headrope. Flatfish catch was maintained while bycatch of Atlantic cod was reduced by 49% and 58%, respectively, compared with standard flatfish gillnets in the Gulf of Maine. He (2006b) tested two low vertical height nets in the Gulf of Maine.
The 8 meshes deep (MD) experimental gillnets caught significantly less cod than the regular 25 MD net, whereas the catch efficiency for flounders (mainly American plaice [Hippoglossoides platessoides]) was similar (He 2006b). The extended gillnets with an extra 10 meshes of webbing (35 MD) caught significantly more Atlantic cod than the standard 25 MD nets in tests in Newfoundland (Yetman 1989). Norwegians use much higher gillnets (60 MD) when targeting cod (Engas et al. 2000). Similarly, nets with tie-down lines caught more flounder and other bottom-dwelling animals (e.g., lobsters) but less cod than the standard cod net due to a reduced vertical profile and a large amount of slack netting near the seabed (He 2006b). Crabs and lobsters are strongly substrateassociated and thus are often caught in groundfish gillnets (Godøy et al. 2003; He 2005, 2006b). In some jurisdictions, retention of crustacean species is prohibited in gillnet fisheries or they are so abundant as to become a nuisance. To avoid the catch of these species that live on the seabed, the footrope of a gillnet may be raised. Norwegian researchers tested a gillnet rigged with “Norsel lines” of 0.5 m long (Fig. 8.9). Although there was some reduction in the targeted Atlantic cod, the Norsel nets significantly reduced the catch of king crabs (Paralithodes
Figure 8.9. Design of the Norsel net to avoid bycatch of king crabs (Paralithodes camtschaticus) in northern Norway. (Redrawn based on Godøy et al. 2003.)
Fish Behavior near Gillnets: Capture Processes and Influencing Factors camtschaticus) (Godøy et al. 2003). Similar experiments were carried out in Newfoundland to reduce snow crab bycatch in Greenland halibut fishery (Brothers 2002). Preliminary experiments in the Gulf of Maine to reduce cod catch when targeting haddock (Melanogrammus aeglefinus) and pollock (Pollachius virens) with Norsel nets provided encouraging but limited results (Eayrs and Salerno 2008). Bycatch species escaped, released, or discarded from gillnets may experience physiological and physical impact, and eventual mortality, related to capture and discard processes (Suuronen 2005; also see Chapter 11). Buchanan et al. (2002) estimated that traditional gillnets targeting other species caused 35% to 70% of total mortality on coho salmon (Oncorhynchus kisutch) in the Pacific coast. Shorter soak durations together with an onboard recovery procedure reduced mortality rates significantly. The mortality rate from 40-min set was 6.7% and those from 140-min set was 52% to 72% after being held in net pens for 48 h (Buchanan et al. 2002). Vander Haegen et al. (2004) found an immediate survival rate of greater than 95% for spring chinook salmon (Oncorhynchus tshawytscha) caught by several tangle nets and gillnets using tagging and recapture methods, but fish released from tangle nets recovered better than those from gillnets. The recovery rate from 114-mm mesh size tangle nets was 1.9 times that of 203-mm gillnets. They found that fish in small-mesh tangle nets were often snagged by the snout rather than gilled and argued that snagging would have reduced injury compared with gilling or wedging and also allowed the fish to continue to respire. 8.6 INTERACTION OF MARINE MAMMALS, SEABIRDS, AND SEA TURTLES WITH GILLNETS Bycatch and related mortality of charismatic animals, including marine mammals, seabirds, sea turtles, and others, has created negative images for gillnets. As a result, gillnets and driftnets are banned in some areas. While the interaction of megafauna species with fishing gears and mitigation measures are discussed in detail in Chapter 13, a brief account of issues related to gillnet fishing is provided here.
195
Seabird bycatch occurs in almost all gillnet types, especially those nets set near the surface, adjacent to bird colonies, and in shallow waters (DeGange and Day 1991; Forney et al. 2001; Lewison et al. 2004; Lien et al., 1989; Melvin et al. 1999). During the capelin spawning season in Newfoundland, intense inshore feeding by birds and peak commercial fishing activities with both bottom and surface gillnets coincided, resulting in significant seabird mortality (Lien et al. 1989). The greatest bycatch of birds by gillnets was near bird breeding colonies, with diminishing bycatch as distance from the colony was increased (Lien et al., 1989). Common murres (Uria aalge) were most often caught in monofilament groundfish gillnets, whereas Atlantic puffins (Fratercula arctica) were more often caught in surface gillnets for salmon. Descriptions of mitigation measures to reduce seabird mortality are limited (Manville 2005). Hayase and Yatsu (1993) submerged high-sea driftnets 2 m below the surface and significantly reduced seabird entanglement. However, there was a substantial reduction in targeted species. Faulkner (1994) theorized that a “window-pane” gillnet containing a thicker twine and larger mesh top panel and regular gillnets underneath it might reduce seabird bycatch. Melvin et al. (1999) used a similar principle and tested a modified salmon gillnet with the top 20 meshes of the webbing made of highly visible white multifilament twine, and they were able to significantly reduce bycatch of common murre and rhinoceros auklet (Cerorhinca monocerata) (Fig. 8.10). Melvin et al. (1999) also found that acoustic pingers (1.5-kHz frequency in 4-s bursts at 120 dB re 1 μPa) attached to the headrope of a gillnet were able to reduce seabird bycatch. Because most seabird bycatch occurred at dawn and dusk and during certain times of the year, a combined measure of gear modification, time restriction, and area closure may reduce seabird bycatch by 70% to 75% without significant reduction in target species (Melvin et al. 1999). Interactions between marine mammals and gillnets can result in animal mortality and severe damage to the fish and the fishing gear (Lewison et al. 2004; Lien 1995; Lien et al. 1989; Northridge 1991). Globally, more than 80,000 small cetaceans have been reported killed annually in coastal waters,
196
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.10. Design of a gillnet with high contrast white netting on the top to reduce seabird bycatch in Puget Sound sockeye salmon driftnet fishery. C, control net; E1, experimental net with 20 mesh white netting on the top; E2, experimental net with 1.5-kHz pingers. (Redrawn from data in Melvin et al. 1999.)
with many of them killed as a result of fishing activities (Jefferson and Curry 1994). According to Lien et al. (1989), an annual average of 24 humpback whales (Megaptera novaeangliae) were entrapped in Newfoundland groundfish gillnets between 1978 and 1987. Belden et al. (2006) reported that 2292 marine mammals were caught in the U.S. Northeast groundfish gillnets in 2004, with another 231 animals in the Mid-Atlantic coastal gillnets. Carretta et al. (2005) reported marine mammal, sea turtle, and bird mortality in the California driftnet fishery for swordfishes and sharks between 1996 and 2002 with a variety of mortality rates for different species. Bycatch of marine mammals by large-scale drift nets resulted in a ban of driftnet fishing in the high seas. Harbor porpoises (Phocoena phocoena) are incidentally caught in gillnets throughout their distribution range in northern waters (Perrin et al. 1994). Acoustic pingers tested in groundfish gillnets in the Gulf of Maine and Bay of Fundy reduced mortality of harbor porpoises, and the use of the devices has become mandatory in the fishery (Kraus et al. 1997; Trippel et al. 1999). Kraus et al. (1997) demonstrated that 10-kHz pingers were able to reduce porpoise catch while maintaining target species
catch of cod and pollock. However, the mechanism by which acoustic pingers were able to reduce porpoise bycatch was not clear (Dawson et al. 1998; Kraus et al. 1997). Harbor porpoise feed on herring, and it was argued that herring can hear high-frequency sound and might have avoided gillnets with pingers. Less herring near gillnets with pingers may have resulted in less porpoise bycatch (Kraus et al. 1997; Trippel et al. 1999). Additionally, Cox et al. (2004) monitored bottlenose dolphin (Tursiop truncatus) in an inshore gillnet site and found that dolphins stayed farther away from gillnets with active pingers. Borodino et al. (2002) reduced Franciscana dolphin (Pontoporia blainvillei) bycatch in gillnets using pingers but found increased pinniped depredation on target species. Mixing of barium sulfate (BaSO4) within monofilament nylon increases reflectivity of the net (Cox and Read 2004; Mooney et al. 2004) and has been reported to reduce harbor porpoise bycatch without a reduction in target species (cod, haddock, and pollock) (Trippel et al. 2003). Trippel et al. (1996) found that the majority (96%) of porpoise bycatch was on the upper twothirds of the gillnet that had a standup height of approximately 4 to 5 m. A reduced height gillnet
Fish Behavior near Gillnets: Capture Processes and Influencing Factors may have a positive effect on reducing porpoise bycatch in the gillnet fishery. For some bottomdwelling species such as flounder, the height of the net may be reduced without affecting the catch of the target species (He, 2006b). Other measures to reduce bycatch of harbor porpoises focus on prevention of entanglement through the use of stiff, neutrally buoyant or sinking ropes or on escape once entanglement occurs, through weak ropes or weak links. These measures are required in New England gillnet and pot fisheries (NOAA 2008). 8.7 DERELICT GILLNETS: GHOST FISHING PROBLEMS AND SOLUTIONS Gillnets and other fishing gears can become lost due to adverse weather or sea conditions or by conflict with other fishing gears or vessel traffic (Matsuoka et al. 2005). Gillnets, whether lost unintentionally, abandoned, or otherwise discarded at sea, have a similar effect on animals and the environment. The derelict gillnets may continue to fish for an extended period of time, causing additional mortality to fish and other organisms. This phenomenon is called “ghost fishing.” Gear designs and modification to eliminate ghost fishing or to reduce the fishing capacity of ghost gears are called “de-ghosting” technologies. With the introduction of synthetic materials in gear construction, these derelict gillnets may continue to fish for several years before they become inactive. In the Northeast Atlantic, including the Mediterranean, as many as 25,000 gillnets were reported lost each year, with loss rates as great as 3.2% (Macfadyen et al. 2009). Canadian Fishery Consultant Ltd (CFCL 1994) estimated that around 5000 gillnets were lost annually in Atlantic Canadian waters. Gear loss in this region may be aggravated by increased use in deep waters for species and in more hostile sea conditions such as in the Greenland halibut fishery. Cooper et al. (1988) conducted video camera surveys by a remotely controlled underwater vehicle in the Gulf of Maine and estimated that there might be 2497 nets (91 m each) on a 64-nm2 area of traditional gillnet grounds on Stellwagen Bank and Jeffries Ledge, equivalent to 39 derelict nets per square nautical mile.
197
Direct observations of derelict gillnets or simulated lost gillnets confirm that these nets continue to fish (Carr et al. 1985; Cooper et al. 1988). Gillnets deliberately set over wrecks in U.K. coastal waters continued to catch and kill fish for at least 2 years (Revill and Dunlin 2003). It was estimated that lost salmon nets might fish for 2 years for fish and 6 years for crabs (High 1985). In shallower waters, derelict gillnets may become overgrown by algae, or “biofouled.” Because these algae-laden nets are more visible, their fishing capacity is correspondingly reduced. Takagi et al. (2007) estimated that the height of bottom gillnets declined rapidly after 15 days and to zero after 25 days of deployment. Fishing capacity decreased to about 15% to 20% of a typical net in the first few weeks (Carr and Cooper 1988; Revill and Dunlin 2003). However, shallow water is usually rich in marine life, so catch rates can still be considerable. Erzini et al. (1997) set “damaged” gillnets in 15 and 18 m off southern Portugal. They estimated that the lifetime of a ghost gillnet was between 15 and 20 weeks. Observations after 8 to 11 months indicated complete destruction or heavy colonization by algae resulting in incorporation into a reef. They estimated that a lost 100-m length of gillnet will catch 314 fish over a 17-week period (Erzini et al. 1997) but thought that this number was likely an underestimate due to predation and scavenging. Prevention of ghost fishing may include prevention of gear loss, derelict gear retrieval, and deghosting technologies. Inshore gillnets use floats on the headrope and leadline on the footrope to spread the net vertically. Therefore, use of degradable material that causes the lost gillnet to lose floatation could reduce the vertical profile and hence fishing capacity. Carr et al. (1992) tested degradable plastic plates for attaching floats to the headrope of gillnets (Fig. 8.11). The gear was set to simulate ghost fishing for a period of 220 days. Two types of degradable plastic plates were used. Divers made underwater observations to check net profiles and catch. Only 2 of the 20 degradable attachment panels partially degraded after 220 days—apparently the panel can be modified to increase degrading process. No significant differences in catch were observed between sections of the net rigged with degradable float attachments and those with
198
Fish Behavior near Fishing Gears during Capture Processes
Figure 8.11. Degradable plastic panel for attaching floats to the headrope of a gillnet. (He 2006a.)
regular rigging. Although there was a report of degradable fishing nets being developed in Japan (Anon. 1993), the commercial application of degradable nets has not been implemented. The final attempt to reduce ghost fishing is to retrieve the derelict gillnets and “clean up” the fishing grounds. Several countries conduct gear retrieval operations regularly (see Macfadyen et al. 2009; Matsuoka et al. 2005). Mandatory reporting of gear loss and the use of acoustic transponders can facilitate retrieval of nets if they become lost. 8.8 CONCLUDING REMARKS Gillnets are very size selective, landing only a narrow range of fish size. Size selectivity of gillnets is closely related to mesh size and changes in type of webbing material, twine size, and hanging ratio. Although gillnets are simple in design and operation, the behavior of fish near gillnets and their capture processes are not well understood. Research is needed to better understand the influence of tide and other factors on net height and on vertical distribution of target species. There are two major conservation challenges facing gillnet fisheries. Gillnets have poor species selectivity, resulting in bycatch and discards and causing mortalities to marine mammals, sea birds, and turtles. Research on mitigation measures has shown progress, includ-
ing successful implementation of pinger use in some fisheries, but more work is needed. Ghost fishing of derelict gillnets remains a challenge, except in the countries that conduct active retrieval of derelict net. REFERENCES Acosta AR and Appeldoorn RS. 1995. Catching efficiency and selectivity of gillnets and trammel nets in coral reefs from southwestern Puerto Rico. Fish. Res. 22: 175–196. Alverson D, Freeberg M, Murawski S and Pope J. 1994. A global assessment of fisheries bycatch and discards. FAO Fish. Tech. Pap. 339: 233 pp. Angelsen KK and Huse I. 1979. Qualitative contrast comparison of different colored gillnets. ICES CM. 1979/B: 20. 3 pp. Angelsen KK. 1981. Engineering and fish reaction aspects of gillnetting—a review. ICES CM. 1981/B: 34. Angelsen KK, Haugen K and Floen S. 1979. The catching efficiency of cod gillnets with different hanging ratio (E) and different floatline buoyancy. ICES CM. 1979/B: 19. Anon. 1993. Biodegradable fishing nets. World Fishing. Oct. 1993. Asaeda T, Priyadarshann T and Manatunge J. 2001. Effects of satiation on feeding and swimming behavior of planktonvores. Hydrobiologia. 443: 147–157.
Fish Behavior near Gillnets: Capture Processes and Influencing Factors Baranov FI. 1948. Theory and assessment of fishing gear. Pishchepromizdat, Moscow. (Translated from Russian by the Ontario Department of Lands and Forests). Beamish FWH. 1978. Swimming capacity. In: Fish Physiology. Vol. 7. Locomotion. pp 161–187. Hoar WS and Randall DJ (eds.). New York: Academic Press. Belden DL, Orphanides CD, Rossman MC and Palka DL. 2006. Estimates of cetacean and seal bycatch in the 2004 northeast sink gillnet and mid-Atlantic coastal gillnet fisheries. NOAA Northeast Fisheries Science Center. Ref. Doc. 06-13: 24 pp. Boje J. 1991. A comparison of selectivity in longlines and gillnets in the fishery for Greenland halibut in West of Greenland. NAFO SCR Doc. 91/39. Borodino P, Kraus S, Albareda D, Fazio A, Palmeiro A, Mendez M and Botta S. 2002. Reducing incidental mortality of Franciscana dolphin Pontoporeia blainvillei with acoustic warning devices attached to fishing nets. Mar. Mam. Sci. 18: 833–842. Bjordal Å and Løkkeborg S. 1996. Longlining. Oxford: Fishing News Books. 156 pp. Brothers G. 2002. Reducing snow crab bycatch in turbot gillnets. Project Summary. EACT-10.2002. DFO (FDP 299). Brothers G and Yetman L. 1982. A study to determine the impact of mesh size on the redfish catches. Department of Fisheries and Ocean Newfoundland Region. Industrial Development Branch. FDN1981/82 -06. 5 pp. Buchanan S, Farrell AP, Fraser J, Gallaugher P, Joy R and Routledge R. 2002. Reducing gillnet mortality of incidentally caught coho salmon. N. Am. J. Fish. Manag. 22: 1270–1275. Cadrin SX and Westwood AD. 2004. The use of electronic tags to study fish movement: a case study with yellowtail flounder off New England. ICES CM. 2004/K:81. Carr HA and Cooper RA. 1988. Manned submersible and ROV assessment of ghost gillnets in the Gulf of Maine. IEEE Conf. Proc. 1987: 622–625. Carr HA, Amaral EA, Hulbert AW and Cooper R. 1985. Underwater survey of simulated lost demersal and lost commercial gill nets off New England. NOAA Tech. Memo. NMFS. 54. Carr HA, Blott AJ and Caruso PG. 1992. A study of ghost gillnets in the inshore waters of southern New England. Mar. Technol. Soc. Conf. Proc. 1992: 361–367. Carretta JV, Price T, Peterson D and Read R. 2005. Estimates of marine mammal, sea turtle, and seabird
199
mortality in the California drift gillnet fishery for swordfish and thresher shark, 1996–2002. Mar. Fish. Rev. 66(2): 21–25. CFCL. 1994. Review of fishing gear and harvesting technology in Atlantic Canada. Ottawa, Canada: Fisheries and Oceans Canada, Fishing Industry Services Branch, Fishing Operations. Collins JJ. 1979. Relative efficiency of multifilament and monofilament nylon gillnet towards lake whitefish (Coregonus clupeaformis) in Lake Huron. J. Fish. Res. Bd Can. 36: 1180–1185. Cooper RA, Carr HA and Hulbert AH. 1988. Manned submersible and ROV assessment of ghost gillnets on Jefferies and Stellwagen Banks, Gulf of Maine. NOAA Undersea Res. Prog. Res. Rep. 88–4. Cox TM and Read AJ. 2004. Echolocation behavior of harbor porpoises Phocoena phocoena around chemically enhanced gill nets. Mar. Ecol. Prog. Ser. 279: 275–282. Cox TM, Read AJ, Swanner D, Urian K and Waples D. 2004. Behavioral responses of bottlenose dolphin, Tursiops truncatus, to gillnets and acoustic alarms. Biol. Conserv. 115: 203–212. Dawson SM, Read A and Slooten E. 1998. Pingers, porpoises and power: uncertainties with using pingers to reduce bycatch of small cetaceans. Biol. Conserv. 84: 141–146. DeGange AR and Day RH. 1991. Mortality of seabirds in the Japanese land-based gillnet fishery for salmon. The Condor. 93: 251–258. Eayrs S. and D.J. Salerno. 2008. Testing raised—webbing gillnets to reduce bycatch of cod while targeting pollock. Online: http://northeastconsortium.org. Accessed: 05/31/2009. Durham, NH: Northeast Consortium. Engås A and Løkkeborg S. 1994. Abundance estimation using bottom gillnet and longline—the role of fish behavior. In: Marine Fish Behavior in Capture and Abundance Estimation. pp 134–165. Fernö A. and Olsen S (eds). Oxford: Fishing News Books. Engås A, Jørgensen T and Angelsen KK. 2000. Effect on catch rate of baiting gillnets. Fish. Res. 45: 265–270. Erzini K, Monteiro CC, Ribeiro J, Santos MN, Gaspar M, Monteiro P and Borges TC. 1997. An experimental study of gill net and trammel net “ghost fishing” off the Algarve (southern Portugal). Mar. Ecol. Prog. Ser. 158: 257–265. Erzini K, Gonçalves JMS, Bentes L, Lino PG, Ribeiro J and Stergiou KI. 2003. Quantifying the roles of competing static gears: comparative selectivity of longlines and monofilament gill nets in a multi-
200
Fish Behavior near Fishing Gears during Capture Processes
species fishery of the Algarve (southern Portugal). Sci. Mar. 67: 341–352. Faulkner G. 1994. Getting the most out of your gillnet. National Fisherman, Sept. 1994: 32–33. Fonseca P, Martins R, Campos A and Sobral P. 2005. Gill-net selectivity off the Portuguese western coast. Fish. Res. 73: 323–339. Forney KA, Benson SR and Cameron GA. 2001. Central California gillnet effort and bycatch of sensitive species 1990–1998. In: Proc. Seabird Bycatch: Trend, Roadblocks, and Solutions. Univ. Alaska Sea Grant, AK-SG-01-01. pp. 141–160. NOAA. 2008. Guide to the Atlantic large whale take reduction plan. Online: http://www.nero.noaa. gov/whaletrp/plan/ALWTRPGuide.pdf. Accessed: 05/31/2009. US National Oceanic and Atmospheric Administration. Fujimori Y, Tokai T and Matuda K. 1994. Effect of diurnal activity of rainbow trout and light intensity on gillnet catching in water tank experiments. Nippon Suisan Gakkaishi. 60: 577–583. Fujimori Y and Tokai T. 2001. Estimation of gillnet selectivity curve by maximum likelihood method. Fish. Sci. 67: 644–654. Gabriel O, Lange K, Dahm E and Wendt T. 2005. Von Brandt’s Fish Catching Methods of the World. 4th ed. Oxford: Blackwell. 523 pp. Gjelland KØ, Bøhn T, Knudsen FR and Amundsen P-A. 2004. Influence of light on the swimming speed of coregonids in subarctic lakes. Ann. Zool. Fennici. 41: 137–146. Godøy H, Furevik D and Løkkeborg S. 2003. Reduced bycatch of red king crab (Paralithodes camtschaticus) in the gillnet fishery for cod (Gadus morhua) in northern Norway. Fish. Res. 62: 377–384. Grant GC, Radomski P and Anderson CS. 2002. Using underwater video to directly estimate gear selectivity: the retention probability for walleye (Sander vitreus) in gill nets. Can. J. Fish. Aquat. Sci. 61: 168–174. Hamley JM. 1975. Review of gillnet selectivity. J. Fish. Res. Bd Can. 32: 1943–1964. Hansen RG. 1974. Effect of different filament diameters on the selective action of monofilament gillnets. Trans. Am. Fish. Soc. 103: 386–387. Hayase S and Yatsu A. 1993. Preliminary report of a squid subsurface driftnet experiment in the North Pacific during 1991. Int. North Pac. Fish. Comm. Bull. 53: 557–576. He P. 1991. Swimming endurance of cod, Gadus morhua L. at low temperatures. Fish. Res. 12: 65–73.
He P. 1993. Swimming speeds of marine fish in relation to fishing gears. ICES Mar. Sci. Symp. 196: 183–189. He P. 2003. Swimming behavior of winter flounder (Pleuronectes americanus) on natural fishing grounds as observed by an underwater video camera. Fish. Res. 60: 507–514. He P. 2005. Characteristics of bycatch of porcupine crabs, Neolithodes grimaldii (Milne-Edwards and Bouvier, 1894) from deepwater turbot gillnets in the northwest Atlantic: Fish. Res. 74: 35–43. He P. 2006a. Gillnets: gear design, fishing performance and conservation challengers. Mar. Technol. Soc. J. 40(3): 11–18. He P. 2006b. Effect of the headline height of gillnets on species selectivity in Gulf of Maine. Fish. Res. 78: 252–256. High WL. 1985. Some consequences of lost fishing gear. Proceedings of the Workshop on the Fate and Impact of Marine Debris, 26-29 November 1984, Honolulu, Hawaii. NOAA—TM—NMFS— SWFC—54: 430–437. Holst R, Wileman D and Madsen N. 2002. The effect of twine thickness on the size selectivity and fishing power of Baltic cod gill nets. Fish. Res. 56: 303–312. Hovgård H 1996. A two-step approach to estimating selectivity and fishing power of research gill nets used in Greenland waters. Can. J. Fish. Aquat. Sci. 53: 1007–1013. Hovgård H and Lassen H. 2000. Manual on estimation of selectivity for gillnet and longline gears in abundance surveys. FAO Fish. Tech. Pap. 397: 84 pp. Huse I, Gundersen AC and Nedreaas KH. 1999. Relative selectivity of Greenland halibut (Reinhardtius hippoglossoides, Walbaum) by trawls, longlines and gillnets. Fish. Res. 44: 75–93. Huse I, Løkkeborg S and Soldal AV. 2000. Relative selectivity in trawl, longline and gillnet fisheries for cod and haddock. ICES J. Mar. Sci. 57: 1271–1282. Jefferson TA and Curry BE. 1994. A global review of porpoise (Cetacea: Phocoenidae) mortality in gillnets. Biol. Conserv. 67: 167–183. Kallayil JK, Jørgensen T, Engås A and Fernö A. 2003. Baiting gill nets—how is fish behavior affected? Fish. Res. 61: 125–133. Kelleher K. 2005. Discards in the world’s marine fisheries. An update. FAO Fish. Tech. Pap. 470: 131 pp. Kraus SD, Read AJ, Solow A, Baldwin K, Spradlin T, Anderson E and Williamson J. 1997. Acoustic alarms reduce porpoise mortality. Nature 388: 525.
Fish Behavior near Gillnets: Capture Processes and Influencing Factors Larkins HA. 1963. Comparison of salmon catches in monofilament and multifilament gill nets. Commer. Fish. Rev. 25(5): 1–11. Larkins HA. 1964. Comparison of salmon catches in monofilament and multifilament gill nets—Part II. Commer. Fish. Rev. 26(10): 1–7. Lewison RL, Crowder LB, Read AJ and Freeman SA. 2004. Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol. Evolut. 19: 598–604. Lien J. 1995. Conservation aspects of fishing gear: cetaceans and gillnets. In: Solving bycatch: Considerations for today and tomorrow. pp 219– 224. Alaska Sea Grant College Program Report 9603. University of Alaska Fairbanks. Lien J, Stenson GB and Ni IH. 1989. A review of incidental entrapment of seabirds, seals and whales in inshore fishing gear in Newfoundland and Labrador: a problem for fishermen and fishing gear designers. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 67–71. St. John’s, Newfoundland: Marine Institute. Lowry N, Knudsen LH and Wileman DA. 1994. Mesh size experiments in the Baltic cod fishery. ICES CM. 1994/B: 13. Macfadyen G, Huntington T and Cappell R. 2009. Abandoned, lost or otherwise discarded fishing gear. FAO Fish. Aquacult. Tech. Pap. 523: 115 pp. Madsen N, Holst R, Wileman D and Moth-Poulsen T. 1999. Size selectivity of sole gill nets fished in the North Sea. Fish. Res. 44: 59–73. Manville AM, II. 2005. Seabird and waterbird bycatch in fishing gear: next steps in dealing with a problem. USDA Forest Service. Gen. Tech. Rep. PSW-GTR191: 1071–1082. Matsuoka T, Nakashima T and Nagasawa N. 2005. A review of ghost fishing: scientific approaches to evaluation and solutions. Fish. Sci. 71: 691–702. McQuinn IH, Gendron L and Himmelman JH. 1988. Area of attraction and effective area fished by a whelk (Buccinum undatum) trap under variable conditions. Can. J. Fish. Aquat. Sci. 45: 2054–2060. Millar RB and Fryer RJ. 1999. Estimating the sizeselection curves of trawls, traps, gillnets, and hooks. Rev. Fish. Biol. 9: 89–116. Melindy S and Flight J. 1992. Development of a deep water turbot fishery by inshore gillnetters. CAFID Report. 30 pp. Melvin EF, Parrish JK and Conquest LL. 1999. Novel tools to reduce seabird bycatch in coastal gillnet fisheries. Conserv. Biol. 13: 1386–1397.
201
Mooney TA, Nachtigall PE and Au WWL. 2004. Target strength of a nylon monofilament and an acoustically enhanced gillnet: predictions of biosonar detection ranges. Aquat. Mam. 30: 220–226. Morizur Y, Pouvreau N and Guénolé A. 1996. Les rejets dans la pêche artisanale française de Manche occidentale. Plozané, France: IFREMER. 123 pp. Moth-Poulsen T. 2003. Seasonal variations in selectivity of plaice trammel nets. Fish. Res. 61: 87–94. Murawski SA. 1993. Factors influencing bycatch and discard rates: analysis from multispecies/multifishery sea sampling. NAFO SCR Doc. 93/115. 17 pp. Nedelec C and Prado J. 1990. Definition and classification of fishing gear categories. FAO Fish. Tech. Pap. 222 (Rev 1): 92 pp. Nedreaas K, Soldal AV and Bjordal Å. 1993. Performance and biological implication of a multigear fishery for Greenland halibut (Reinhardtius hippoglossoides). NAFO SCR Doc. 93/118. 15 pp. Northridge SP. 1991. An updated world review of interactions between marine mammals and fisheries. FAO Fish. Tech. Pap. 251 (Suppl. 1): 219 pp. Perrin WF, Donovan G P and Barlow J. (eds.). 1994. Gillnets and Cetaceans: Report of the International Whaling Commission, Special Issue 15. Cambridge: International Whaling Commission. Perry RI and Neilson JD. 1988. Vertical distributions and trophic interactions of age-0 Atlantic cod and haddock in mixed and stratified waters of Georges Bank. Mar. Ecol. Prog. Ser. 49: 199–214. Pol M. 2006. Testing of Low-Profile, Low CodBycatch Gillnets: Phases I and II. Report to the Northeast Consortium, Durham, NH, USA. Online: http://www.northeastconsortium.org. Accessed 05/31/2009. Durham, NH: Northeast Consortium. Pol M and Carr HA. 2000. Overview of gear developments and trends in the New England commercial fishing industry. Northeast. Naturalist. 7(4): 329–336. Potter ECE and Pawson MG. 1991. Gill netting. Ministry of Agriculture, Fisheries and Food, Directorate of Fisheries Research, Lowestoft, Laboratory Leaflet, No. 69. 34 pp. Pristas PJ and Trent L. 1977. Comparisons of catches of fishes in gillnets in relation to webbing material, time of day, and water depth in St. Andrew Bay, Florida. Fish. Bull. 75(1): 103–108. Revill AS and Dunlin G. 2003. The fishing capacity of gillnets lost on wrecks and on open ground in UK coastal waters. Fish. Res. 64: 107–113. Revill A, Cotter J, Armstrong M, Ashworth J, Forster R, Caslake G and Holst R. 2007. The selectivity of
202
Fish Behavior near Fishing Gears during Capture Processes
the gill-nets used to target hake (Merluccius merluccius) in the Cornish and Irish offshore fisheries. Fish. Res. 85: 142–147. Robinson CJ and Pitcher TJ. 1989. The influence of hunger and ration level on shoal density, polarization and swimming speed of herring, Clupea harengus L. J. Fish Biol. 34: 631–633. Rose GA and Leggett WC. 1989. Interactive effects of geophysically-forced sea temperatures and prey abundance on mesoscale coastal distributions of a marine predator, Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 46: 1904–1913. Rudstam LG, Magnuson JJ and Tonn WM. 1984. Size selectivity of passive fishing gear: a correction for encounter probability applied to gillnets. Can. J. Fish. Aquat. Sci. 41: 1252–1255. Samaranayaka A, Engås A and Jørgensen T. 1997. Effects of hanging ratio and fishing depth on the catch rates of drifting tuna gillnets in Sri Lankan waters. Fish. Res. 29: 1–12. Santos MN, Gaspar MB, Monteiro CC and Vasconcelos P. 2002. Gill net and long-line comparisons in a hake fishery: the case of southern Portugal. Sci. Mar. 66: 433–441. Scholten GD and Bettoli PW. 2006. Lack of size selectivity for paddlefish captured in hobbled gillnets. Fish. Res. 83: 355–359. SOAFD. 1992. Marine Laboratory Annual Review 1990–1991. Scottish Office Agriculture and Fisheries Department, Aberdeen, UK. 84 pp. Stewart PAM. 1987. The selectivity of slackly hung cod gillnets constructed from three types of twine. J. Cons. Int. Explor. Mer. 43: 189–193. Stoner AW. 2004. Effect of environmental variables on fish feeding ecology: implications for the performance of baited fishing gear and stock assessment. J. Fish Biol. 65: 1445–1470. Sulaeman M, Matsuoka T and Kawamura G. 2000. Effect of hanging ratio on size selectivity of gillnet. Nippon Suisan Gakkaishi. 66: 439-445. (In Japanese with English abstract). Suuronen P. 2005. Mortality of fish escaping trawl gears. FAO Fish. Tech. Pap. 478: 72 pp. Takagi T, Shimizu T, and Korte H. 2007. Evaluating the impact of gillnet host fishing using a computational analysis of the geometry of fishing gear. ICES J. Mar. Sci. 64: 1517–1524. Trippel EA, Wang JY, Strong MB, Carter LS and Conway JD. 1996. Incidental mortality of harbor
porpoise (Phocoena phocoena) by the gillnet fishery in the lower Bay of Fundy. Can. J. Fish. Aquat. Sci. 53: 1294–1300. Tripple EA, Strong MB, Terhune JM and Conway JD. 1999. Mitigation of harbor porpoise (Phoceona phocoena) bycatch in the gillnet fishery in the lower Bay of Fundy. Can. J. Fish. Aquat. Sci. 56: 113–123. Trippel E, Holy N, Palka D, Shepherd T, Melvin G and Terhune J. 2003. Acoustic reflective net reduces porpoise mortality. Mar. Mam. Sci. 19: 40–43. Turunen T. 1996. The effects of twine thickness on the catchability of gillnets for pikeperch (Stizostedion lucioperca (L.)). Ann. Zool. Fennici. 33: 621–625. Vander Haegen GE, Ashbrook CE, Yi KW and Dixon JF. 2004. Survival of spring Chinook salmon captured and released in a selective commercial fishery using gill nets and tangle nets. Fish. Res. 68: 123–133. Walsh SJ and Morgan MJ. 2004. Observations of natural behavior of yellowtail flounder derived from data storage tags. ICES J. Mar. Sci. 61: 1151–1156. Wardle CS. 1975. Limit of fish swimming speed. Nature 225: 725–727. Wardle CS. 1980. Effect of temperature on the maximum swimming speed of fishes. In: Environmental Physiology of Fish. pp 519–531. Ali MA (ed). New York: Plenum. Wardle CS. 1989. Understanding fish behavior can lead to more selective fishing gears. Proc. World Symp. Fish. Gear and Fish. Vessel Design. pp 12– 18. St. John’s, Newfoundland: Marine Institute. Wileman D, Ferro RST, Fonteyne R and Millar R. 1996. Manual of methods of measuring the selectivity of towed fishing gears. ICES Coop. Res. Rep. 215: 126 pp. Woodhead PMJ. 1964. Changes in the behaviour of the sole, Solea vulgaris, during cold winters, and the relation between the winter catch and sea temperatures. Helgol. Wiss. Meeresunters. 10: 328–342. Yetman L. 1989. Comparison of extended and standard gillnets for harvesting Atlantic cod. DFO NFLD Atl. Fish. Dev. Proj. Sum. 16: 3 pp. Yokota K, Fujimori Y, Shiode D and Tokai T. 2001. Effect of thin twine size on gill net size-selectivity analyzed with the direct estimation method. Fish. Sci. 67: 851–856.
Fish Behavior near Gillnets: Capture Processes and Influencing Factors SPECIES MENTIONED IN THE TEXT American plaice, Hippoglossoides platessoides Atlantic cod, Gadus morhua Atlantic mackerel, Scomber scombrus Atlantic puffin, Fratercula arctica Atlantic salmon, Salmo salar Baltic cod, Gadus morhua bottlenose dolphin, Tursiop truncatus chinook salmon, Oncorhynchus tshawytscha coho salmon, Oncorhynchus kisutch common murre, Uria aalge European dab, Limanda limanda Franciscana dolphin, Pontoporia blainvillei Greenland halibut, Reinhardtius hippoglossoides haddock, Melanogrammus aeglefinus hake, Merluccius merluccius
203
harbor porpoise, Phocoena phocoena herring, Clupea harengus humpback whale, Megaptera novaeangliae king crab, Paralithodes camtschaticus ling, Molva molva paddlefish, Polyodon spathula pikeperch, Stizostedion lucioperca pollock, Pollachius virens rainbow trout, Oncorhynchus mykiss rhinoceros auklet, Cerorhinca monocerata saithe, Pollachius virens shearwater, Puffinus spp. skipjack tuna, Katsuwonus pelamis winter flounder, Pseudopleuronectes americanus walleye, Sander vitreus yellowfin tuna, Thunnus albacares
Chapter 9 Electric Senses of Fish and Their Application in Marine Fisheries Hans Polet 9.1 INTRODUCTION Discussion of the use of electricity in fisheries can be traced to the middle of eighteenth century. In 1765, Job Baster, in his work “Natuurkundige Uitspanning—Part 2,” speculated that electricity would have an effect on shrimp as electric shocks had similarities to the shocks produced by the electric eel, and he proposed further investigation (De Groot and Boonstra 1974). However, it was not until the middle of the nineteenth century that the first experiment with electric fields on animals in water was carried out. In 1863, a British patent was granted to Isham Baggs for electric fishing (Snyder 2003). At that time, the four basic reactions of fish to electric fields had been distinguished: fright, anodic electrotaxis, tetanus, and electrocution. Research and application of electricity in fisheries were started in the freshwater environment in the 1920s and 1930s (e.g., Scheminzky 1924). A successful and widely used application of electric fishing (electrofishing) in freshwater followed as a research and management technique. Nowadays it is mainly used in shallow rivers, brooks, and lakes for stock assessment; fish health surveys; tagging; catching broodstocks; and eliminating undesirable species. The technique has been used around the world, but there has been growing concern among fishery biologists and managers regarding its potential for harming fish and other marine organisms. Snyder (2003) pointed out that electrofishing involves a very dynamic, complex, and often misunderstood mix of physics, physiology, and behavior. Although often not externally obvious or fatal,
spinal injuries and associated hemorrhages sometimes have been documented in over 50% of fish examined internally. Other harmful effects, such as bleeding at the gills or vent and excessive physiological stress, also are of concern and electrofishing over spawning grounds can harm embryos. The research of B. M. Bary (1956) on the behavior of marine fish in electric fields was the start of a boom in the research into applications of electricity in capturing fish at sea. The possibility of increasing shrimp catches with electricity was started in the United States in the early 1950s. During the next four decades, this work was also taken up in Europe, the former USSR, India, and China (van Marlen 1997). Studies of fish behavior in electric fields in the laboratory and in the sea resulted in electrofishing techniques for marine fisheries. Commercial application was, however, rarely reported. By the end of the 1980s, all work in this field was stopped in Europe due to a ban on the method by the European Union. In contrast, the fishing method flourished in China, where, by the end of the twentieth century, more than 3000 vessels used electric fields to capture shrimps. However, mismanagement and abuse of the method led to a national ban in 2001 (Yu et al. 2007). The only countries where the literature indicates that marine electrofishing is still considered as a realistic option for future commercial application are the Netherlands (van Marlen et al. 2006) and Belgium (Polet 2003). Fifty years of research worldwide to find ways of using electricity in marine fishing and great
205
206
Fish Behavior near Fishing Gears during Capture Processes
spending of financial resources have led to a very poor outcome, not in terms of knowledge but in terms of commercial products. In going through the literature, it is stunning to see so many positive reports of field trials that never led to any longerterm commercial application. If any future new design is to be successful, it will have to do better in both prospects for economic performance and minimization of ecosystem effects. It will have to take into account the complexity of electrofishing, the vulnerability of the technology at sea, and the often-unnoticed side effects on biota.
9.2 PROPERTIES OF AN ELECTRIC FIELD IN WATER Figure 9.1 shows a variety of types of electric currents that can be used for electric fishing (Snyder, 2003). Alternating current (AC) and direct current (DC) are the two main types that are not interrupted. Those that can be interrupted are referred to as pulsed current. Interrupted AC current is rarely used for electrofishing. Interrupted DC currents will be referred to as pulsed DC or PDC. Pulsed waveforms generated by electrofishing equipment occur in a wide variety, most often square, half-sine, quarter-sine, or capacitor discharge. According to Snyder (2003), the frequencies applied in the field (mainly rivers and lakes) lie between 15 and 120 Hz but experimentally they can range from 1 to about 500 Hz. The pulse pattern can be simple or complex. In the simple pattern, the same single pulse is generated at a certain frequency, but the complex pattern consists of a high primary frequency interrupted secondarily at a much slower frequency, thus producing bursts or trains of higher frequency pulses. The main variable determining the reaction of fish is the electric field itself (i.e., the strength of the electric field). Field strength can be described by the voltage gradient (E in V/cm) and/or the current density (J in μA/cm2) and/or power density (D in W). In the case of an interrupted electric field, the pulse type, the frequency (Hz) and pulse duration or length (ms) are also determining factors. Duty cycle is sometimes used as a characteristic for the electric field, such as in studies on the effect of electrofishing on the health of fish. It depends on the frequency and the pulse duration and the time
during which the electric field is active expressed as a percentage. In natural waters, electric fields are usually heterogeneous—that is, the current and thus the field lines do not run parallel to each other, depending on the shape of sources (electrodes) and the medium between them. In such fields, the field lines radiate and spread widely around and between the electrodes (Fig. 9.2). In the laboratory, an almost homogeneous field can easily be created by using two platelike electrodes in a tank with sides having the same dimensions as the electrodes. In this case, the current flows parallel to the sides of the tank directly from one electrode to the other, providing a constant voltage gradient, current density, and power density. A homogeneous field simplifies experimental conditions and is ideal for lab experiments, but it may be difficult to extrapolate to commercial electrofishing operations, during which the electric fields will usually be heterogeneous. Depending on the conductive properties and porosity of the substrate, the electric field can extend into the bottom. Haskell (1954) and Zalewski and Cowx (1990) found that a substrate with fine particles and organic material has a better conductivity (the electric field will penetrate deeper) compared with coarse gravel and rubble. Snyder (2003) also noted that interactions between the water and an object or substrate with different conductivity distort the electric field (Fig. 9.3). At the boundary between the water and the object, the current in the water near its surface progressively concentrates (current density increases, voltage gradient declines) if the adjacent medium is more conductive than the water and vice versa. This is the case for fish and boats made of metals. It is therefore advised against using metal boats in the vicinity of an electric field. B. M. Bary (1956) states that to obtain electric control of fish in freshwater lakes and streams, high field strength is required because the conductivity of the fish is higher than that of the water. This is, however, more than offset by the reduced current needed to maintain the field strength. Thus, the overall power requirement for obtaining a particular response is comparatively low. Seawater, by contrast, has a high conductivity compared with that of the fish so that a high current is required to maintain a given field strength and therefore the overall
Sine wave, alternating current (AC)
A
Volts
Volts
+ 0
Direct current (DC), smooth, generated by a battery or DC generator
B
+
– 0 Time C
Time D
DC, rippled, generated by partially filtered, full-wave, rectified AC
Half-sine, full-wave pulsed direct current (PDC), generated by unfiltered, half-wave, rectified DC +
Volts
Volts
+
0
0 Time E
Time F
Half-sine, full-wave PDC, generated by unfiltered, full wave, rectified AC
Square PDC, generated by interrupting smooth or rippled DC +
Volts
Volts
+
0
0 Time G
Time H
Quarter-sine, half-wave PDC, generated by controlled, half-wave, rectified AC
Exponential PDC, generated by capacitor discharge +
Volts
Volts
+
0
0 Time I
Time J
Square pulse train (PDC) (Coffelt's CPS©)
Hybrid pulsed direct current (PDC/DC), PDC on top of a DC base +
Volts
Volts
+
0
0 Time
Figure 9.1.
Time
Different waveforms used in electrofishing. (Snyder 2003.)
207
+100
–100 50
–30 20 10
–20 5
0
5
–10
Closely spaced spherical electrodes Current lines
Constant voltage lines
+100 50 20 10
Isolated Spherical Electrode Note: Total current between any two current lines is the same; zone illustrations assume both polerites are equally effective AC accelerator fields. Voltage levels are for illustration only. Flectrade
Danger zone
Differential zone
Figure 9.2. Hypothetical three-dimensional diagrams of heterogeneous electric fields around and between electrodes. Contrary to the diagrams, current flows from negative to positive electrodes. (Novotny 1990.)
208
Electric Senses of Fish and Their Application in Marine Fisheries
γ1