The International research Institute for Climate prediction

IRIIRIThe International research Institute

for Climate prediction

l i n k i n g s c i e n c e t o s o c i e t y

The International research Institute

for Climate prediction

l i n k i n g s c i e n c e t o s o c i e t y

Climate and Fisheries

INTERACTING PARADIGMS, SCALES, AND POLICY APPROACHES

The IRI-IPRC Pacific Climate-Fisheries Workshop

Honolulu, 14-17 November, 2001

Co-organized with

The International Pacific Research Center (IPRC)

SUPPORTING CO-SPONSORS OF THE WORKSHOP:The North Pacific Marine Science Organization (PICES)

The Center for Sustainable Fisheries (CSF), University of MiamiThe GLOBEC International Project Office

The IDYLE Project of the French Institut de Recherche pour le Développement (IRD)

OTHER ORGANIZATIONS PROVIDING SUPPORT FOR ATTENDEES:The Intergovernmental Oceanographic Commission (IOC)

The Food and Agriculture Organization of the United Nations (FAO)The NOAA Office of Global Programs

The National Aeronautics and Space Administration (NASA)

C O L U M B I A U N I V E R S I T Y

C O L U MB I A U N I V E R S I T Y

C O L U M -B I A U N I V E R S I T Y

Co-organized with The International Pacific Research Center (IPRC)

Supporting co-sponsors of the workshop:The North Pacific Marine Science Organization (PICES)

The Center for Sustainable Fisheries (CSF), University of MiamiThe GLOBEC International Project Office

The IDYLE Project of the French Institut de Recherche pour le Développement (IRD)

Other organizations providing support for attendees:The Intergovernmental Oceanographic Commission (IOC)

The Food and Agriculture Organization of the United Nations (FAO)The NOAA Office of Global Programs

The National Aeronautics and Space Administration (NASA)

Published by the International Research Institute for Climate Prediction (IRI)Columbia Earth Institute

Palisades, New York, 10964, USAIRI Publication IRI-CW/02/1

ISBN 0-9705907-6-8

©2002 International Research Institute for Climate Prediction. All rights reserved.

Climate and Fisheries

INTERACTING PARADIGMS, SCALES,AND POLICY APPROACHES

The IRI-IPRC Pacific Climate-Fisheries Workshop

Honolulu,14-17 November, 2001

EDITORSAndrew Bakun and Kenneth Broad

2 The IRI-IPRC Pacific Climate-Fisheries Workshop

Contents

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

BACKGROUND TO THE WORKSHOP

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2. The workshop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3. Adopted terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4. General questions to be addressed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

5. Relevance for fisheries policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

WORKSHOP DISCUSSIONS: 1. CURRENT UNDERSTANDING

6. Examples of currently-established climate-fisheries applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

7. Climate variability affecting fish stocks and fisheries [Art Miller, Raghu Murtugudde and Frank Schwing] . . 18

7.1 Biological sensitivities and links to the physic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

7.2 Predictability issues and needed future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

8. Regime shifts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

8.1 Regime-like variations in Pacific salmon stocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

8.2 Indicators of a regime shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

9. Population connectivity [Jeff Shima] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

10. Available integrative concepts in fishery resource ecology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

10.1 The “optimal environmental window” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

10.2 Ocean triads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

10.3 MacCall’s basin model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

10.4 “Member-vagrant” concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

10.5 “Match-mismatch” hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

10.6 “Lasker windows” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

10.7 The “school trap” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

11. Potential implications of rapidly-evolving adaptive response mechanisms . . . . . . . . . . . . . . . . . . . . . . 30

The IRI-IPRC Pacific Climate-Fisheries Workshop 3

WORKSHOP DISCUSSIONS: 2. FUTURE DIRECTIONS

12. Applications to real world needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

13. Role of comparative studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

13.1 Time series analysis: the application of the comparative method in the temporal domain . . . . . 36

13.2 Inter-regional comparative time-series analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

14. Role of modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

15. Sardine regimes and mesoscale flow structure [Alec MacCall] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

15.1 Outline of the Flow-Based Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

15.2 Fish Behavior: Habitat Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

15.3 Physical and Biological Oceanography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

15.4 Synchrony and Teleconnections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

15.5 Concluding Thoughts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

16. Climate and tuna fisheries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

17. OFCCP – A proposal for a GLOBEC Pacific regional project [Patrick Lehodey] . . . . . . . . . . . . . . . . 45

17.1 Project description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

17.2 Monitoring the upper trophic levels of the pelagic ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

17.3 Food web structure in pelagic ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

17.4 Modeling from ocean basin to individual scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

17.5 Socio-economical impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

18. CLIOTOP – A proposed global comparative research effort on large oceanic pelagics [Olivier Maury] . . . 48

19. Potential utility of a multilateral empirical comparative retrospective study . . . . . . . . . . . . . . . . . . . . . 49

20. Proposed formulation of a Pacific regional climate-fisheries project . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

21. Useful application of the current state of scientific understanding . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Appendix 1. List of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

n international workshop on research issuesrelated to interactions between climate varia-tions and fisheries was held at the East-West

Center of the University of Hawaii in Honolulu fromNovember 14th to 17th, 2001. Forty-eight invitedparticipants represented a sampling of top-tier inter-national scientific expertise with respect to climaticeffects on fishery resource populations, fishing oper-ations, and fishery-related socioeconomic issues. Anunusual aspect was the interaction of physical, bio-logical, and social scientists at all levels of the discus-sions. No prepared papers were delivered. Rather, theintended focus was on interdisciplinary and interre-gional “cross-education” and cross-sharing of insightsand ideas among scientists with experience rangingover a variety of species and industry types, intendedto support a collaborative process of:

• identifying alternative conceptualframeworks and ideas that may bettersupport fruitful interdisciplinarycollaborations (particularly betweenclimate scientists and fishery scientistsof both the “ecological/biological”and “social science” types);

• exploring associated implications forinnovative fisheries managementapproaches;

• considering potential applications ofthe comparative method as a means foreffective multilateral research on cli-mate/ecosystems/fisheries issues in thePacific basin;

• exploring in this regard the potentialutility of certain newly available tech-nologies and methodologies.

The discussions both in plenary sessions as wellas in various separate “focus group” sessions were wide-ranging and animated. General consensus emerged ona variety of issues. It was widely agreed, for example,that: (1) as our available records of data and experiencegrow longer, the observations are not adding up to pic-ture that conforms to conventional scenarios. Effects ofenvironmental variability on fish stocks and fisheries

can no longer be ignored, but we remain stuck in a par-adigm that has existed for half a century and that is notsolving the problem in any general way; (2) we need tomove away from focusing so much of our availableeffort on identifying particular specific relationshipsand on producing empirical models fitted to specificsets of data, but rather to undertake efforts at more gen-eral synthesis that can generate testable generalhypotheses (i.e., we need to search for mechanisms andprocesses, not correlations); (3) climate forecasts (e.g.,ENSO forecasts, etc.) do have significant potentialvalue for the fisheries sector, but the information con-tent must be relevant, communicated properly, andcompatible with available decision-support models; (4)downside risks related to reliance on a poor forecastmight in many cases outweigh potential benefits; how-ever, we should not abandon the search for means toproduce good forecasts; (5) inter-decadal-scale “regimeshifts”, along with associated large-scale synchronies inresource population variations and resultant socioeco-nomic issues, probably represent the “hottest” currentset of climate-fisheries research topics. The apparentlarge-scale synchronies would seem to indicate a ratherdirect link of climatic events to resource populationdynamics, which led to optimism among a significantportion of the workshop participants that majorprogress on the “climate to fish” portion of the problemmight be possible on the near term.

On the other hand, there were also areas wherebroad consensus seemed to be lacking. For example,some participants were quite excited about thepotential role of rapidly-evolving adaptive responsemechanisms, but there was a general level of concernover a lack of clear evidence for their actual operationand significance in real ocean ecosystems. Likewise,certain participants advocated the idea of a compre-hensive collaborative global empirical (statistical)study of available time series of relevant data, butthere were questions as to exactly how and by whichgroups such a grandiose multilateral “desk study” ofhistorical data would be conducted.

On the social science side, there was emphasis onidentifying the range of potential decision-makers inthe fishery sector beyond just managers, and matchingthe temporal and spatial scales of their decisions to


A

our current knowledge and predictive capabilitiesconcerning climate and marine ecosystem interac-tions. There was also widespread consensus on theneed to seriously consider communication issues,including interactions with the media, industry, andpolicymakers in order to avoid misinterpretationsand resulting unintended consequences (e.g., exacer-bation of inequity) of the provision of uncertaininformation. Along those lines, there was intense dis-cussion around the question of when is the sciencegood enough to “go public”, specifically in regards tothe topic of regime shifts.

Several interesting ecological hypotheses wereproposed and discussed. One of these that appearedto excite particular interest focused on gyre-widevariations in the content of mesoscale activity in thesystem as being potentially involved in basin-scalesynchronous alterations between sardine and anchovydominance. More generally along this same line, thebasic proposition that the essential linkage of thelarge-scale climate forcing to fish population dynamicsmay act through smaller-scale processes such as occurin mesoscale ocean features excited considerableinterest. Accordingly, the problems of “downscaling”from global-scale and ocean basin-scale models toregional eddy-resolving models were assigned high-priority for focused research efforts.

The subject of marine ecosystem regime shiftsclearly stood out as an issue of major importance toproper scientific management of fisheries and fisheryresource populations. It appears to be not yet entirelyclear whether all, or even most, major marine ecosys-tem regime shifts are causally linked to climaticregime shifts. However, in the case of the mid-1970sshift, the prototype upon which the current conceptof a marine ecosystem regime shift is largely based, aclear climatic regime shift was indeed roughly con-temporaneous to the initiation of the impressivedecadal-scale population increases, as well as someimportant decreases, noted in many of the largestfishery resource stocks of the world, especially in thePacific. It appears that it might be time for someglobally-focused institution (or group of cooperatinginstitutions) to undertake monitoring and rapidlydisseminating available information on evidence for

large-scale ecosystem regime change occurring in theworld’s oceans.

In terms of immediate fruitful application ofstate-of-the-art climatic analysis and prediction tocurrent real-world fisheries problems, the relativelyadvanced available understanding of Pacific ENSO-related phenomena would appear to suggest that themost likely early successes might be attained withrespect to the tuna fishery concerns of Pacific islandnations and fisheries issues faced by the Pacificcoastal nations of Latin America.

The structure of a proposed multi-lateralOceanic Fisheries and Climate Change project(OFCCP GLOBEC) designed to investigate theeffect of climate change on the productivity and dis-tribution of oceanic tuna stocks and fisheries in thePacific Ocean was described and discussed. Its con-ceptual design and its goal of predicting short- tolong-term changes and impacts related to climatevariability and global warming were much-admiredand its implementation was supported. In addition, theidea of initiating a worldwide comparative researchproject on tunas and other oceanic “top predators”, ofwhich the regional OFCCP project in the tropicalPacific could be a cornerstone, received strong sup-port. This global “umbrella” project (code-named“CLIOTOP”) would ideally be developed within theInternational GLOBEC context.

In addition to these tuna-related initiatives,there was consensus that general basic climate-fisheriesresearch issues, common to a variety of species andecosystem types, could be effectively addressed on aregional basis in the Pacific. The participants agreedthat IRI, IPRC, and the PICES Secretariat should beasked to collaborate in identifying an effective con-ceptual framework for such a multilateral Pacific climatefisheries project, as well as a process for developingand implementing it.

There was widespread agreement that theinclusion of a strong group of fisheries-experiencedsocial scientists together with climate scientists,oceanographers, and biologists had been a very pro-ductive and successful aspect of this workshop. It wasrecommended that follow-on research project devel-opment activities adhere to the same model.


Executive Summary

Background to the Workshop

The oceans cover nearly four-fifths of the earth'ssurface and more than a billion people rely on fish astheir main source of animal protein. In some countries,fish are practically the sole source of animal protein.Demand for food fish and various other useful attrib-utes obtainable from the sea has been accelerated bypopulation growth and by the global trend towardpopulation migration to coastal areas.

Fisheries and fish products provide employmentto nearly 200 million people. Globally, the bulk ofthe people employed in fisheries are poor and manyare without acceptable alternative sources of workand sustenance. In addition, fish and fishing areenormously important to the cultural life of manycoastal societies, and may often define a "quality of life"for people having a cultural tradition of harvesting thesea. Hence, maintenance of viable fishery resourcesmay be extremely important to preserving traditionalways of life, associated economic activities, tourism,etc. In addition, fish represent the fastest growingfood commodity entering international trade.Accordingly, fish and fish products represent anextremely valuable source of foreign exchange to manycountries, in some cases providing as much as half oftotal available foreign exchange income.

Fisheries have become a very big business.Modern, highly capitalized fleets now range the oceansof the world, in some cases competing for a limitedcommon property resource with traditional fisher-men and communities. Massive overcapitalization ofthe global fishing industry is, in fact, an enormousproblem. The Food and Agriculture Organization ofthe United Nations a few years ago estimated thatworld fisheries on a global basis operate at an annualdeficit of US$53 billion. Thus present day commer-cial fisheries actually represent a large net burden onother economic sectors. This unfortunate situation isat least to some degree due to lack of a sound scien-tific basis to correctly gauge sustainable productivityof the resources and to effectively manage impacts inthe face of large-amplitude variability in both physicaland biological aspects of ocean ecosystems.

Observations and modeling have led to recentadvances in understanding the influence of climatevariability on ocean processes and to the recognition of

inherent variability of marine biological communitieson various temporal and spatial scales. In parallel,there have been dramatic improvements in our abilityto conduct biological sampling, genetic identification,etc. In addition, there has been a very large surge ofclimate research activity in recent years. But with fewexceptions, there has been little progress in bringingtogether results from these related frontiers in a com-prehensive framework that can consistently rationalizethe accumulating store of information and experienceso as to provide a reliable basis for sorting out the var-ious effects of fishing, natural climatic variability, andchronic alterations of environment and/or habitat. Theresult is that "sustainable fisheries" remains a theoret-ical ideal rather than a realistic operational goal.

One of the major obstacles involves the largecomponent of climate-associated variance that com-monly acts to obscure the results of human actions andto defeat attempts at prediction. Accordingly, waysneed to be found to more effectively involve climatescientists in interdisciplinary research collaborations.Furthermore, the time may have arrived to begin toquestion some of the conventional dogmas, postulatesand working assumptions that have guided, but alsoconstrained, fishery science and managementthrough the twentieth century. (These may include,among others, (i) long-term stationarity of basic fish– environment linkages, (ii) absolute uniqueness ofspecific regional situations, (iii) a predominant effect oftrophic interactions, (iv) an expectation that a climate-driven stock oscillation should have essentially thesame period as the climatic variation driving it, etc.)Also of great importance is the maintenance andexpansion of research interactions with social scien-tists on the management, economics, and politics ofthe increasingly competitive use of marine resourcesand the ocean environment.

Due to major research advances in understand-ing the ENSO phenomenon, as well as significantprogress in predicting it, the prospects for effectiveearly applications of recent climate research progressto marine fisheries issues may be most promising inthe tropical and eastern boundary zones of the PacificOcean. This is the context which led the InternationalResearch Institute for Climate Prediction (IRI) of


1 Introduction

Columbia University (1) to undertake organizationof an international Climate-Fisheries Workshop, whichwhile having global international participation andtreating the climate-fisheries problem in a generalizedcontext, would focus particularly on the Pacific, and(2) to seek the collaboration of the InternationalPacific Research Center (IPRC) of the University ofHawaii in its implementation and in providing aPacific venue for its setting.

One of the driving ideas behind the workshophas been that climate variability (on a variety of scales)may, if we are clever, provide us with "experiments" toprobe the real mechanistic workings of these ecological-biological-social systems for which ordinary experi-mental controls are usually impractical (sustainablefisheries development requiring a more accurateunderstanding of those basic internal mechanisms).Furthermore, a better understanding of the interde-pendent spatial and temporal scales of "openness" or"closedness" of the resource systems on the ecologicalside (i.e., on population exchanges and changes inspatial pattern dynamics) could provide opportunitiesto explore additional options on the fishery manage-ment side. In addition, it was thought that some newideas regarding mechanisms that could conceivablyenable extremely rapid adaptive adjustments bymarine populations might suggest ways to begin tounderstand how non-stationarities may be introducedinto the processes that link fish populations to theenvironments that support them, and might also sug-gest some innovative adaptive management tacticsthat could conceivably develop into important newtools for managing both climatic and anthropogenicperturbations of fishery resource populations.

It was decided that, in general, the workshopfocus would be on:

(1) identifying alternative conceptualframeworks and ideas that may bettersupport fruitful interdisciplinary col-laborations (particularly between cli-mate scientists and fishery scientists ofboth the “ecological/biological” and“social science” types);

(2) exploring associated implications forinnovative fisheries managementapproaches;

(3) considering potential applications ofthe comparative method as a meansfor effective multilateral research on

climate/ecosystems/fisheries issues inthe Pacific basin;

(4) exploring in this regard the potentialutility of certain newly availabletechnologies and methodologies.

The workshop

The Pacific Climate-Fisheries Workshop washeld in Honolulu from November 14th to 17th, 2001,at the East-West Center of the University of Hawaii.As mentioned above, the primary sponsors were theInternational Research Institute for Climate Prediction(IRI), which is located at the Lamont-Doherty EarthObservatory of Columbia University and theInternational Pacific Research Center (IPRC) of theUniversity of Hawaii. IRI was the primary organizer ofthe workshop while IPRC provided the venue andmeeting arrangements. Also joining the effort as sup-porting co-sponsors of the workshop were the NorthPacific Marine Science Organization (PICES), thenew Center for Sustainable Fisheries (CSF) of theUniversity of Miami, the GLOBEC InternationalProject Office, and the IDYLE Project of the FrenchInstitut de Recherche pour le Développement (IRD).Other organizations providing support for attendeesincluded the Intergovernmental OceanographicCommission (IOC), the Food and AgricultureOrganization of the United Nations (FAO), theNOAA Office of Global Programs, and the NationalAeronautics and Space Administration (NASA).Forty-eight invited participants mainly from PacificRim and Pacific Island countries, but also from as faraway as Germany, South Africa and the Seychelles,represented a good sampling of the best availableinternational scientific expertise concerning climaticeffects on fishery resource populations, fishing opera-tions, and fishery-related economic and social issues.(Names and contact information of the invited par-ticipants are listed in Appendix 1.) An unusual aspectwas the interaction of physical, biological, and socialscientists at all levels of the discussions.

It was specifically intended that the emphasisin the workshop deliberations be on interdisciplinary(and also interregional) cross-education and cross-sharing of insights and ideas. Thus, the larger part of


2

the workshop discussions took place in plenary sessionsattended by all of the workshop participants.Discussions in these plenary sessions were chaired byDr. Juergen Alheit, who has been longtime Chairmanof the GLOBEC SPACC (Small Pelagic Fish andClimate Change) Project and the new Chairman ofthe GLOBEC “Focus 1” Working Group onRetrospective Studies. Dr. Alheit was specificallyasked to come to Honolulu to serve as workshopChairman, not only because of his demonstrated skillin running such meetings to the general satisfaction ofall participants, but also to demonstrate the desire ofthe organizers to position this workshop and the sci-entific activities potentially resulting from it firmlywithin the International GLOBEC context. At theworkshop, his skill in enforcing analytical focus withhumor was appreciated by all.

In addition to plenary discussions, the work-shop periodically broke up into separate, slightly lessbroadly multi-disciplinary “focus groups” for detaileddiscussions among specialists on particular topics ofspecial interest to the particular group members. Thefocus groups were arranged as follows:

• Social Science Focus Group (Chair: M. Hamnett)

• Tuna Focus Group (Chair: R. Olson)• Population Connectivity Focus Group

(Chair: R. Cowen)• Boundary Current Frontal Zone Focus

Group (Chair: T. Sugimoto)• Population-Ecosystem Ecology (Eco)

Focus Group (Chair: R. Quinones)• Climate Focus Group (Chair: A. Miller)These groups initially met separately. However

as the workshop progressed, most of the groups electedto combine in various combinations with othergroups for joint discussions.

All sessions of the workshop, whether plenaryor of focus groups, had designated rapporteurs. Thenotes recorded by the rapporteurs, some initial back-ground documents prepared before the workshop byAndrew Bakun and Kenneth Broad, and summaries

prepared by several of the focus group chairs form theprimary basis for the following sections of this report.Several participants produced original contributionsworthy of potential citation. Consequently, where amajor section of the report essentially represents anoriginal input prepared almost entirely (with minimalintervention by the editors) of an individual partici-pant, this is indicated by the contributor’s name inbrackets next to the section title. In other cases, whereone or more individuals made major contributions toa particular section, it is indicated in a footnote ref-erenced on the section title. Respective affiliations ofthese cited individuals may be found in Appendix I.Andrew Bakun and Kenneth Broad, acting on behalfof IRI, with the input and assistance of LorenzMagaard, Executive Associate Director of IPRC, werethe primary planners and organizers of the workshopcontent and participation. Jerry Comcowich, ofIPRC, coordinated operational arrangements inHonolulu, with the aid of Ellen Bahr and other IPRCadministrative staff. Gisela Speidel, IPRC PublicRelations Specialist, arranged interviews on theworkshop with Honolulu newspapers1. Bakun andBroad acted as co-editors of this workshop report andare solely responsible for errors in representing theviews of the participants.

Adopted terminology

In order to promote a truly fully-integrateddiscussion, the following terminology was pro-posed by the workshop organizers and distributedto participants several weeks before the workshop.The term marine resource system (*MRS) was pro-posed to refer to the entire composite systemincorporating:

1. the fishery resource stock itself;2. its regional marine ecosystem, includ-

ing all components of its occupied


3

1 The resulting articles can be read at the newspaper websites: "Climate could reveal secrets of fisheries" http://the.honoluluadvertiser.com/article/2001/Nov/18/ln/ln15a.html; "Experts examine how climate affects fisheries" http://starbulletin.com/2001/11/17/news/index.html. Dr. Speidel has also written a brief summary of the workshop for the IPRC Climate Volume 2, Number 1.

habitat and the entire trophic webwithin which it functions;

3. the characteristic seasonal climatologyand the basic physical-chemical habitatstructure (to which we expect thestock as well as the ecosystem to be“substantially” adaptively tuned);

4. the fishery (or fisheries) on the resourcestock in question;

5. associated economic activities (includ-ing other, interacting uses of theregional marine ecosystem)and socialvalues;

6. the relevant management frameworkand institutions;

7. the political context in which themanagement activities and the eco-nomic context operate.

Thus, the *MRS was considered to be anintrinsically interlinked entity, with each elementfeeding back on other elements and no element beingimmune to such feedbacks from other elements. Apriori, the *MRS was not expected to be necessarilystationary in any way (in underlying mean state, inlong-term trend, etc.). Further it was proposed that,for the workshop discussions, low-frequency (nonsea-sonal) climatic variability would be considered anexternal driving force acting on the *MRS.

The more limited view that is often taken ofa marine resource system was assigned the term fish-habitat system (*fhs). This term was intended to referto the subsystem of the *MRS consisting of elements(1), (2), and (3).

General questions to be addressed

It was envisioned, and communicated to theparticipants, that the Honolulu Workshop wouldseek to identify potential research approaches to thefollowing general questions:

(a)What are the effects of climate variabilityon the *MRS (and can we potentiallyforecast these effects based on climatemonitoring or forecasting)?

(b)What can climate variability tell usabout the internal dynamics of the*MRS (i.e., can we use climate vari-ability as “experiments” to probe themechanisms controlling the internaldynamics of the *MRS – our lack ofunderstanding of which upsets our abil-ity to confidently answer question a)?

(c)What productive management actionscould be enabled by significant scien-tific progress on specific aspects ofquestions a and b?

Moreover, it was emphasized that priority fordiscussion of particular aspects of questions a and bwould be strongly determined by their potential rel-evance to question c (i.e., there would be less priorityplaced on discussing ways to develop the ability tosay “the “steam roller is coming and you can’t get outof way” than the ability to say “the steam roller iscoming and here is what you can do (or try) in orderto get out of the way”).

Relevance for fisheries policy

It was recognized a priori that there are paralleltrends in the discourse and development of fisheriesregulatory mechanisms that focus on very differentscales and embrace differing views on the centralizationof management. These trajectories have implicationsfor the potential uses of climate information for fish-eries management (and fisheries related decision-making in general):

• On the one hand, there is an emphasison local level involvement by multiplestakeholders in decisions regardingestablishment of regulatory mecha-nisms. This emphasis is motivated byan increasing acceptance that withoutco-management, or community man-agement, we are less likely to achieveregulatory compliance, anticipatesocioeconomic feedbacks from vari-ous policies, or incorporate qualitative

1 0 The IRI-IPRC Pacific Climate-Fisheries Workshop

4

5

knowledge about the local ecosystemfrom long-time inhabitants (e.g.,McCay and Acheson 1987, McCayand Svein 1996; Jentoft et al. 1998;Jentoft 2000, Hughey et al. 2000;Berkes et al. 2001).

• On the other hand, we are witnessingthe development of an increasingnumber of national and supranationalagreements – complex bilateral andmultilateral treaties on the managementof open ocean resources, of shared(straddling) stocks, and regardingecosystem linkages between widelyseparated metapopulations (Ward etal. 2000). Underlying the establishmentof such agreements is the acceptancethat distant bodies of water, lands, andstocks of living resources are linkedthrough large-scale oceanographic andbiological processes, and activities inone locale may impact another.

Attention to the traditional approach of limitingthe number of fish caught as determined by stockassessment models (i.e., Total Allowable Catch) hasresulted in a large body of literature on the con-straints and incentives for the adoption of differentfishing restrictions. Further, there is extensive docu-mentation of failures using these approaches due tosocial and ecological factors (e.g., McGoodwin 1990)Intractable limitations in data availability, assump-tions of stationarity of living resources, assumptionsof human behavior in open access situations, insuffi-cient understanding of multi-species interaction, andclimatic influences have been cited as factors thatseverely limit the efficacy of the many numericalapproaches and the subsequent regulations based onmodel outputs.2

Alternative approaches that embrace the com-plexity, spatial heterogeneity, climatic influences, anddata realities are gaining credence in the scientificcommunity. These approaches emphasize an under-standing of ecosystem interactions that can result innon-stationarity, and what some characterize as the

chaotic nature of the abundance variations of many fishstocks (Bakun 2001). Such interactions are recognizedto be important at multiple spatial (e.g., mesoscaleevents such as wind influenced larval survivability,and macroscale processes – basin wide asynchrony inspecies regime shifts) and temporal scales (i.e., days,seasonal to interannual, decadal). Of relevance arepotential increases in our capability to monitor climaticand biological processes that may improve our abilityto predict short term fluctuations in stock abundance.Meanwhile, awareness of fluctuations on longertimescales suggests the importance of understandinghow the short-term fluctuations in abundance relate tolonger term, more dramatic changes.

In summary, it appears that biological andhuman activities on multiple scales must beaddressed. If we consider the implications of thechange in approach to management on an ecosystembasis, we are led to consider alternative approaches tomanaging resources that force us to think aboutactivities at a range of scales. From the policy per-spective, certain key questions loosely guided muchof the discussion during this workshop that is sum-marized further in the report:

• What is “the state of the art” in thecurrent ability to predict fluctuationsand movement of the major pelagicspecies in the Pacific on the varioustime scales (i.e., temporal and spatialresolution, forecast skill)?

• In what forms are this informationavailable (i.e., probabilistic forecasts,hindcasting only) and to whom?

• How do we know when this informa-tion is “reliable enough” to be used inpolicymaking, or should we considerthese only as hypothetical scenarios?

• How, if it all, is this information beingused by policymakers, fishing groups,national governments, and other stake-holders in their decision-making?

The IRI-IPRC Pacific Climate-Fisheries Workshop 1 1

2 For (a controversial) overview of limits of numerical modeling in relation to fisheries management, see Wilson et al. (1994).

• Given current understanding and pre-dictive capabilities, what are thepotential uses of this information bythe different stakeholders?

• How, if at all, is this knowledge beingincorporated into agreements amongfishing groups (from local to regionallevels). How could it be better utilized?

• What are the constraints and incentivesfor the use of this information, and dothey differ from the traditional onesidentified in the literature?

• What are the potential unintendedconsequences of the introduction ofor reliance on this information?

• What new tools and techniques areavailable that may influence currentmanagement of fisheries (e.g., eggpump, genetic marking)?

• How might this information be used toalter current management approaches?

• What are the relative trade-offs ininvesting in increasing observationalcapability of key indicators versusdevelopment of predictive models?


Workshop Discussions

1.CURRENT UNDERSTANDING

Examples of currently-establishedclimate-fisheries applications

During the first plenary session, there was dis-cussion of current uses of climate information forfisheries management. This provoked several partici-pants to raise points that remained salient throughoutthe meeting, and also highlighted the varied (andoften unanticipated) ways that climate informationhas been used to date. Several of the fisheries scien-tists noted the use of climate information (e.g., SST)in stock assessment (primarily to analyze catchabilityand estimate recruitment) and ecosystem models,emphasizing the point that climate information isnot just of use to managers, but plays a crucial role inresearch as well. Other target audiences, or “endusers” of climate information were mentioned,including members of the financial sector linked tofisheries, brokers of fish products, fleet and plantmanagers, and those who set fisheries regulations.Most examples were in the context of seasonal-tointerannual variability (i.e., ENSO) versus longertimescale events. Many of the more detailed exampleswere drawn from the cases of small pelagics off thecoast of S. America, tuna fisheries in Pacific smallisland states, and salmon in the NE Pacific. Below wehighlight some of the main points from the discussion.This section is NOT intended to be a review of the use ofclimate information in the fisheries sector, but to provideexamples of the application of climate information thatemerged in the workshop.

Western S. America pelagics: Peru and Chile have long been aware of the

impact of climate variability, notably ENSO, on theirmarine ecosystem. During warm events there is bothhorizontal (southward) and vertical migration ofsmall pelagics (the resource supporting one of theworld’s largest fishmeal industries), and duringextreme warm events reproduction, and even sur-vival, are seriously compromised (Barber and Chavez1983, Arntz et al. 1985, Sharp and McLain 1993,

Serra 1987, Serra 1991, Tarazona and Castillo 1999).3There are numerous examples of the use of climateforecasts by members of their fisheries sector (see Table1 for examples of decisions that could be affected byclimate information), some of which were anticipatedby researchers prior to the dramatic advances in fore-casting ENSO (Glantz 1979, 1986, Glantz andThompson 1981, Walters 1989). For example, theoceanographic agency in Peru anticipates the need toincrease sampling prior to an ENSO event, andbudgets cruises accordingly; preemptive vedas(restricted fishing periods) to protect stressed stockshave been implemented during strong ENSO events,but are often cut short due to political pressurearound election time, highlighting the inextricablerole of national politics in affecting regulations.Similarly, restricting fishing of the shared stock ofanchoveta on the border between Chile and Perucontinues to remain politically hazardous, evokingnationalistic protests in both countries, with Peruviansarguing “why should we protect our fish if they arejust going to be caught by the Chileans when theymigrate southward” (for details concerning the use ofvarious sorts of climate information by the fisheriessector in Peru during the 1997-98 El Niño, see Broadet al., in press, Carr and Broad 2000).

Given the scientific uncertainty of both theclimate forecasts and the associated impacts on species(not all ENSO events have similar impacts on theecosystem), fisheries regulators are faced with difficultoptions. One workshop participant involved with the1997-98 ENSO gave the example of the dilemmafaced by Peruvian regulators: when it was knownthat it was going to be an event of great magnitude,there were two seemingly equal logical managementapproaches: 1) to impose extreme conservation, withthe societal result of unemployment, defaulted loans,and social unrest during an already depressed eco-nomic period, or 2) to allow heavy fishing since thefish may likely disappear (and not reproduce) due tothe severity of the event, even with no fishing pressure.As might be expected, Peru took middle ground,alternating closed and “exploratory” fishing periodsas pressure from the industry mounted.


6

3 For an overview of the effects of the 1997-98 El Niño on physical mechanisms and biogeochemical cycles, see, e.g., McPhaden (1999) and Chavez et al. (1998).

Many banking decisions hinge on futureexpectations of ENSO events. Loan officers monitorforecasts on the internet as an aid in decidingwhether to make a loan or not, assuming negativeimpacts of warm events on the pelagic fisheries.During the 1997-98 event, some Peruvian fishingfirms were waiting for large loans and were alleged tohave put pressure on the media and government sci-entific agencies to downplay the effects of El Niño,while another subset of the industry wanted to use ElNiño as an excuse to have a state of emergencydeclared to reduce interest rates on existing loans,and to leverage refinancing packages. The fact thatmany financial investments (e.g., vessel and plantconstruction) are played out over multi-year timehorizons reveals some of the difficulty in changingbehavior, patterns of overinvestment, etc. with ashort lead (3-6 month) ENSO forecast, and con-versely, the potential for better understanding oflonger term fluctuations to potentially influence suchinvestment decisions.

Much of the debate over the intensity and tim-ing of the 1997-98 El Niño played out in the media,who exhibited the tendency to sensationalize infor-mation to increase sales, often turning probabilistic

statements by the scientific community into deter-ministic headlines. Conflicting messages reached thepublic, leading people to ignore the “noise”. Again,this highlights the issue of communication of infor-mation in affecting societal decision-making.

The impact of El Niño on small pelagicsreceives the most attention. However both positiveand negative impacts on species important to arti-sanal fisheries and to aquaculture (e.g., abundance ofscallops) also occur (for details, see Tarazona andCastillo 1999). For example, the arrival of potentiallyvaluable tropical species such as mahi mahi (dolphin-fish) near shore provide an opportunity for small-scalefishers; however due to lack of capital to buy appro-priate gear, or lack of export markets for the products,artisanal groups can not always take advantage of thisopportunity. Further, during strong events such as1997-98, tropical species may be so abundant, fromEcuador to Central Chile, that prices drop to suchlow levels that it is not worth the cost of fuel to catchthe fish. And finally, to some regions, El Niño bringsdevastating floods, wave damage, and other infra-structure damage, reducing the ability to get productsto market.


Tropical Pacific tuna: The use of climate information as a gauge of

tuna availability appears to be important for a numberof actors in the small island states of the tropicalPacific region. For instance, the tendency for a generaleastward shift of distributions of skipjack and yellowfintuna during El Niño events (Lehodey et al. 1997) hasobvious effects on the level of catches in the exclusiveeconomic zones of Pacific island nations and results in

important impacts on their economies. In addition,there are changes in schooling patterns, size, fat con-tent, ratio of mixed species, and other characteristics,posing a challenge to those catching and processingthese fish. Tuna are caught for commercial purposesby foreign vessels and for subsistence by locals inmost small island states in the tropical Pacific region.There has been little study of the use of climate infor-mation by small-scale (subsistence) fishers, and only


GROUP GOAL DECISIONS

Industrial purse seine fishers Large industrial catch, Build / Repair vessels, Change / Alter nets, and processors Conglomerate profits Relocate fleet, Hire personnel, Layoff

personnel, Install refrigeration system, Upgrade plant technology, Stop fishing, Stockpile products, Change product ratios (fishmeal vs. canning), Diversify into other industries

Artisanal fishers Large artisanal catch, Change fishing gear, Reject non-traditional(net fishermen, purse seine Fishing tradition gear, target new species, Change household(<30 gross registered tons), production options (e.g., spouse works more),longline fisherman, trawlers, Negotiate five-mile limit with industrydivers (shellfish and spearfishermen)Labor High and full employment Change household production options (e.g.,

Adequate and stable income children sent to work), MigrationBanks Returns on investments Accept or reject loan request, Refinance debt,

ForecloseRegulatory administrators Sustainable fishery, Agency Establish closed seasons (vedas), Establishand scientists funding, job security, quotas, Gear restrictions / allowances, Increase

Prestige and consulting jobs enforcement, Increase sampling and observation, Permit / reject new licenses, Allow experimental fishing, Misrepresent skills / information

Conservation groups Sustainable fishery Support bans on fishing, Lobby for fleet size reduction

Politicians Re-election, Overall welfare Support regulatory measures, Support variousconstituencies (firm owners, labor)

Media Sales Exaggerate impacts of El Niño, Attribute(subscriptions, advertising) impacts to El Niño, Inject false certainty

informationForeign interests Low-cost resource products Substitute protein source (soy), Structural

(e.g., fishmeal), Debt adjustment policiesrepayment

TABLE 1. Goals of actors in the Peruvian fishing sector and climate-related decisions (Broad et al. in press, Climatic Change)

recently has there been study of the use of climateinformation by groups linked to the commercialindustry, including national governments. We focus onthe commercial industry below, drawing heavily onthe recent work by Hamnett and Anderson (1998)who, through the Pacific ENSO Applications Center,work on the use of climate information with decision-makers in small island states of the Pacific.

One group that demonstrated a capability toquickly adapt to climate variability was the purseseine fleet. Accustomed to using advanced technology(i.e., satellite imagery, aircraft, FADs), some of themore experienced captains anticipated the spatialshift in resources due both to their own monitoringof conditions and to the climate forecasts issued byNOAA and other agencies. During the 1997-98 ElNiño episode, however, this group encountered diffi-culty with the change in characteristics of the fishcaught (size, flesh texture, fat content, etc.). This ledto initial problems with “mushy tuna syndrome”until consultants were flown in from Hawaii to provideadvice on modifying handling techniques. Otherproblems encountered by vessels included difficultysetting nets in “abnormal” current conditions, inabil-ity of equipment to handle the larger school sizes,and problems in catching the more deeply schoolingfish. Processing plants also encountered initial diffi-culty due to the unusual characteristics of the fish,and had to modify equipment in response.

On a more macro scale, the individual islandstates are affected in diverse ways by the shift in spatialdistribution of tuna, with some states being favoredand others negatively affected by reductions in revenuefor foreign fishing licenses and losses of labor in theprocessing plants (both of which account for a highpercentage of the GDP and the basis for employmentin some independent island states). Advanced warningof the 1997-98 El Niño allowed some governmentsto plan ahead for the changes in average revenue.

Taking a more regional perspective, participantspointed out the potential importance of the ways inwhich climate information may be utilized in multi-lateral fisheries management – namely by the newly

established tuna commission for the western andcentral Pacific.4 Under this regime, new institutionalmechanisms will be developed to strengthen the sci-entific basis of tuna management (for example, in orderto apply widely the precautionary approach). (SeeTarte 2001 for political analysis of tuna managementin this region). In this context it will be important toenhance knowledge of climate variability and itsimpact on the marine environment. Such informationmay also be useful for addressing economic andpolitical issues within the Commission, such as theallocation of catch and effort quotas between memberstates. Tarte (2002) related the potential use of climateinformation to the particular situation in Fiji, wherethere has been controversy over the alleged illegalissue of fishing licenses by corrupt government offi-cials. It has been suggested that the number of tunalicenses issued in the past year was four times thatrecommended as sustainable for the fishery in Fiji’sexclusive economic zone (EEZ). Moreover, accordingto the Oceanic Fisheries Program of the Secretariat forthe Pacific Community (SPC), there was seriousunder-reporting and under-estimating of catch andlandings in Fiji, underscoring claims that the tunafishery was in imminent danger of collapse.5

Tarte goes on to pose the question of what is asustainable level of fishing activity in Fiji’s exclusiveeconomic zone? Determining this depends on arange of factors (not least the accuracy of catch andeffort data). It also requires consideration of the pos-sible impact of climate variation on stock levels.Policymakers in Fiji have suggested that Fiji has ben-efited from climatic changes that have caused tunamigrations through Fiji’s EEZ to increase in recentyears. Thus there was “immense potential” to exploitthe tuna industry further by issuing more tuna licensesthan were currently recommended.6

The above example points to the value andpossible danger of the use of climate information infisheries management decisions. It suggests that suchinformation may be used by governments to issuemore licenses, and by industry to catch more fish.As was noted in a 2000 workshop in Noumea,


4 Information on multilateral treaties draws extensively on Tarte (2002).5 ‘Fiji Fisheries in Bad Shape’, Pacific Magazine, November 2001, pp.19-20.6 Statement by Fiji’s Minister for Fisheries and Forests, Fiji Times, 28 November 2001, p.5.

“such responses are not necessarily helping with sus-tainable management”7. Therefore, in addition to moreunderstanding of climate-fish linkages, there is a needto explore ways of integrating such information into asustainable management strategy and to consider“institutional, socioeconomic and political questions”.

North Pacific salmon: The case of salmon in the North Pacific was

also raised during the workshop as an example of theimportance of shifts in the mean state of the oceanfor understanding changes in biological productivity.This case has been studied from the biophysical per-spective (e.g., Beamish et al. 1997, 1999) as well asfrom a policy perspective, bringing into relief severalimportant points that remain to be adopted by policy-makers. Pacific salmon catches in Alaska have variedinversely with catches from the U.S. West Coast duringthe past 70 years and appear to be related to climateforcing associated with the Pacific Decadal Oscillation(Hare et al. 1999, Mantua et al. 1997). Miller (1996,2000), applying a game theory model, has relatedthese shifts in productivity and catch to ongoingtreaty negotiations, and the breakdown in cooperationbetween the US and Canada in setting harvest alloca-tions under the Pacific Salmon Treaty. Miller empha-sizes that institutional factors will determine theextent to which the management of such resourcescan adapt effectively to climate variability or long-termclimate change.

Salmon also provide an example of the potentialuse of seasonal-to-interannual climate informationfor managing complex ecosystems. Pulwarty andRedmond (1997) analyze the use of climate informa-tion in the management of the Columbia River system,which attempts to balance hydropower productionwith salmon restoration and multiple stakeholdergroups. Despite some clear potential uses for seasonalforecasts in reducing uncertainty, their study foundthat the “complexity of the management environment,the lack of well-defined linkages among potentialusers and forecasters, and the lack of supplementarybackground information relating to the forecastspose substantial barriers to future use of forecasts”.

The above three cases are not intended toaddress all uses of climate information in the fisheriessector, but merely to highlight the many actors,overlapping scales, conflicting goals, and potentialunintended consequences that may be impacted byintroducing climate information into society. Thesecases cover a wide geographical, cultural, and ecosystemrange, and represent some of the most detailed analysesof the actual and potential use of various sorts ofclimate information available to date. There remainspromise, by most researchers’ accounts, for pursuingindividual applications of climate information in theseand similar cases.

Climate variability affecting fishstocks and fisheries [Art Miller,Raghu Murtugudde and FrankSchwing]

Climate changes in physical oceanographicvariables have been clearly linked with oceanicecosystem changes on many temporal and spatialscales. This physical forcing is especially obvious onseasonal (and shorter) timescales where variations insea-surface temperature (SST) and upwelling, forexample, strongly control productivity, growth andmigration. This forcing is also prominently evidenton interannual timescales associated with El Niñoand La Niña episodes, when tropical ecosystems aredirectly affected by the severely altered oceanic con-ditions or when eastern Pacific boundary ecosystemsare remotely forced by the concomitant oceanic andatmospheric teleconnections. On longer timescales,such as those associated with decadal regime shiftsthere is still a high degree of uncertainty concerningthe mechanisms that are involved when changes inphysical oceanographic conditions influence oceanbiology (e.g., Alheit and Bernal 1993, Miller andSchneider 2000, Hare et al. 2000).

The mechanisms by which physical forcingaffects oceanic ecosystems range from those associatedwith the smallest scales of dissipation, turbulent mixing


7 Report on Workshop on Inter-annual Climate Variability and Pelagic Fisheries, International Research Institute for Climate Prediction, New Caledonia, 6-24 November, 2000, p.20.

7

and diffusion, to those acting on the mesoscale asso-ciated with fronts, eddies, and upwelling, to thoseoperating on the basin scale associated with gyres, ElNiño, and the thermohaline convective circulation(Denman et al. 1996). The regional and local mani-festations of this physical forcing on the biology canoccur instantaneously or with a time delay. Climateforcing apparently acts to modulate the complicatedecosystem in both linear and non-linear ways.

The most important physical oceanographicvariables that influence biology are thought to be seasurface temperature (SST), mixed-layer depth (MLD),thermocline depth, upwelling strength, upper-oceancurrent fields and sea ice. SST, which is strongly corre-lated to atmospheric sea level pressure, is the bestobserved oceanic physical variable over climatetimescales. Partly for this reason, studies often attemptto link SST changes to ecosystem changes. But thedirect influence of SST on ecosystems is obscured bythe fact that many physical processes cause SST tochange (e.g., direct surface heating, horizontal currentadvection, upwelling, changes in mixing) so that SSTanomalies can be symptomatic rather than causal.

The oceanic MLD, in concert with the nutri-cline and photic zone depths, can influence primaryproduction by affecting new nutrient availability andthe average light intensity to which individualautotrophs are exposed. Primary production in thesubtropics tends to be limited by nutrients, in contrastto the subpolar regions where light tends to be thelimiting factor. Long-term changes in thermoclinedepth along boundaries can directly influence thepreferred habitats of benthic fauna or change thecharacteristics of mesoscale eddy and filament forma-tion to fundamentally affect upwelling processes andnear-surface nutrient enrichment.

7.1 Biological sensitivities and links to the physics

These physical environmental changes occur-ring in the ocean affect viruses, bacterioplankton,phytoplankton, and so on up to whales. Correlationsbetween these physical variables and long-term changesin ecosystems have routinely been identified, but thespecific mechanisms involved are usually unclear.There are a number of reasons for this: the ecosystemcan be contemporaneously influenced by manyphysical variables; the ecosystem can be very sensitive

to the seasonal timing of the anomalous physicalforcing; and the ecosystem can generate intrinsicvariability on climate time scales. Determining base-line levels of "biological noise" in the absence ofaltered climatic conditions would be an importantachievement. But limited observations of both thephysics and biology also confound these variousinterpretations. Consequently, only the generalizedinfluences of large-scale physical forcing rather thanprecise linkages are normally invoked in explainingdecadal ecosystem variations.

Long-term decreases in macrozooplankton inthe Southern California Bight and CaliforniaCurrent System have drawn considerable attention asbeing fundamental to the health of the entire ecosystemand have been hypothesized as being indicative ofglobal warming influences (Roemmich and McGowan1995). The long time series of the CaliforniaCooperative Fisheries Investigations (CalCOFI) sur-veys gives a unique long-term perspective to physicalforcing of biological systems on decadal timescales.Long-term warming of the CalCOFI region is oftencited as a determinant of decreased productivitythrough weaker upwelling; however, data-based esti-mates of upwelling changes are found to be oppositeto that expectation.

The grand challenge is to understand themechanisms of the responses to these instantaneousand delayed physical environmental changes and thevarious modes of significant feedback through thetrophic web. The basin-scale nature of climatechange appears to organize patterns of response infishery resource populations. For example, salmonstocks in Alaska tend to vary in phase with each otherbut out of phase with salmon stocks of the U.S.Pacific Northwest (Mantua et al. 1997). Similarly,basin-wide correlations and anti-correlations occurbetween geographically disparate sardine andanchovy populations (Schwartzlose et al. 1999).

Many ideas have been advanced to explainthese linkages, including direct influences of temper-ature anomalies on fish migration routes and con-comitant predation factors, the influence of regionalzooplankton productivity differences on feeding con-ditions and early life history effects on populations.

The problems inherent in relying on catch dataare rather obvious. Catchability, migration, andrecruitment can all be aliased and misinterpreted onefor another. For example, the warming in the tropical


Atlantic during 1983 was initially assumed to havecaused loss of fisheries whereas later analysis indicatedthat the catchability was simply low due to a deeperthermocline and resulting lower position of the tunain the water column.

Another challenge is to separate effects of lifecycle changes due to climate impact from changesdue to migration. The former could be related toenvironmental changes that affect such things asocean triad configurations (see Section 10.2) whilethe latter may simply be the response of the adultpopulations to changes in location of preferred con-ditions (temperatures, prey concentrations, etc.).Seasonal to interannual variability in the fisheryresource populations of different regions or of differentspecies in the same region may appear uncorrelateddue to their adaptations to respond to the expectedlocal large amplitude environmental changes such asEl Niño or the migration of the fronts or the variabilityin the upwelling. However, regime shifts or relativelysmall amplitude changes that persist for extendedperiods appear to produce synchronous responses inregions far removed from each other and amongunrelated and non-competitive species.

7.2 Predictability issues and needed future work

At this stage, appropriate diagnosis is generallymore important than forecasting. However, in somecases, predictable components of the physical oceansystem can be identified. For example, a portion of thevariance of SST in the region off the coast of Japan(Figure 2) can be predicted with significant skill at one-to-three year lead times (Schneider and Miller, 2001).Determining the influence of these subtle and compli-cated physical changes on ecosystem dynamics holdspromise for predicting ecosystem changes as well.

State-of-the-art tools such as general circulationmodels (GCMs), individual based models (IBMs),and trophodynamic (nitrate-phytoplankton-zoo-plankton-detritus, NPZD) models can be employedtogether with data mining and retrospective analysesof all available data to better understand the variabili-ty seen in fisheries. It is paramount to be able to

properly downscale from the large-scale climaticindices to their impacts on triads, fronts, eastern andwestern boundary currents, and coastal upwellings.Coupled operational general circulation (OGCM)primary production models exist that can be forcedwith fairly reliable “reanalyses” winds and fluxes andcombined with ocean reanalysis products to quantifythe variability of the subtropical/subarctic gyres,transports in the eastern and western boundarycurrents, thermocline depths, etc.

Regime shifts 8

Over the past decade and a half, earlier beliefsin an essential stability of marine ecosystems havebeen largely displaced by a growing appreciation ofthe importance of large-amplitude low-frequencyvariability occurring in many regions of the world’soceans. For example, Venrick et al. (1989) reported adoubling of depth-integrated chlorophyll in the sub-tropical North Pacific starting in the mid-1970s.Brodeur and Ware (1992, 1995) indicated a dou-bling of biomass of zooplankton and of pelagic fishand squid in the subarctic North Pacific in the 1980scompared to the late 1950s to early 1960s period.Roemmich and McGowan (1995) reported a 70%decrease in zooplankton biomass in the CaliforniaCurrent since the 1950s, along with correspondingdrastic declines in certain seabird species. A verysharp 60%-70% decline in zooplankton biomass offPeru in the mid-1970s, following the collapse of theanchoveta was also reported (Carrasco and Lozano1989, Loeb and Rojas 1988, Alheit and Bernal1993). Remarkably, a significant component of thismode of variability appears, at least in the period ofthe 1970s and 1980s, to have been synchronized oververy large spatial scales (Kawasaki 1983, Lluch-Beldaet al. 1989, 1992). Accordingly, some manner oflinkage via large-scale climatic (atmospheric) telecon-nections would appear to be a logical necessity.

This mode of low frequency variability seemsoften to take on the appearance of periods of relative


8 Alec MacCall, Jeff Polovina, and Skip McKinnell made important contributions to this section.

8

stability over time scales of one to several decadeswhich are interspersed by comparatively sudden“regime shifts”. These regime shifts tend to be charac-terized by radical expansions or contractions of occu-pied habitat of important populations and/or byreplacement of one dominant species of fish byanother. The question of how to properly account forregime shifts in fisheries management and endan-gered species protection has become a major issue.

A shift that occurred around 1976 has receivedthe most attention because it is strongly evident in avery wide variety of biological and physical measure-ments (Ebbesmeyer et al. 1991, Graham 1994,Trenberth and Hurrell 1994, Miller et al. 1994,Polovina et al. 1994, Parrish et al. 2000, Hare et al.2000) Previous regimes and transitions showed a vari-ety of phase relationships, whereas the 1976 event wasnearly synchronous. Also, the evidence for most ofthe earlier transitions is less quantitative, and may besubject to other plausible explanations, for instance,development of canning technology in the rise of ear-lier sardine fisheries.

Some other proposed regime shifts include:

• the early 1940s off the West Coast of North America,

• the late 1960s in the Humbolt system,• the late 1980s in the Humbolt system, off

Japan and in parts of the North Pacific,• the late 1990s off the West Coast of

North America.9

We may be using the term “regime shift”rather loosely. For example, many biological changesassociated with the late 1980s event are evident, butthe underlying physical changes are not as clear.

The properties of regime shifts may provideinsight into their underlying mechanisms. In contrastto ENSO events (i.e., both El Niño and La Niña),which are relatively transient, regimes are long-termphenomena. The transitions or shifts betweenregimes include changes on a variety of time scales,nonlinear effects and phase relationships. Howeverthe general character is a bimodal tendency10 inwhich the mean is a relatively rare condition. Much

of the evidence for changes in regimes is biological,and the differential responses of various species mayallow a natural classification system. Even if somebiological changes result from a redistribution inspace, rather than a change in abundance, that redis-tribution may be a valid indicator of a change inunderlying physical and biological conditions. Smallchanges in the timing of physical events and condi-tions relative to the seasonality of fish spawning couldbe a sensitive mechanism of physical-biological inter-action. A gradual change in physics can result in asudden shift in the biological system.

In describing the patterns of regime shifts, itis hard to judge when an event actually starts. It iswell known that matching time series can be arbi-trary and easily produces spurious relationships, butthat approach continues to be used in describing pat-terns of shifts. The nature of biological control,whether top-down or bottom-up, may also influencethe order of events. The actual physical or biologicaltrigger may be hard to see. This may be further com-plicated by the influence of pre-existing conditions,for example, it has been hypothesized that the recentgrowth of the sardine population off California start-ed during a period of “biological opportunity” andwas released by a physical change (of course the keyto understanding is to identify the nature of the“opportunity” and the “change”). In some cases, clas-sical predator-prey phase relationships are apparent,but are subject to alternative explanations and couldbe misleading.

Atmospheric interactions tend toward wide-scale synchronicity (e.g., north and south of theEquator), but deep ocean mechanisms allow long lagtimes (e.g. on east and west sides of the Pacific).ENSO events contribute a large portion of the vari-ance in physical and biological measurements, andwe might be able to interpret regime-scale eventsmore clearly if the ENSO-related phenomena couldbe filtered out. Shifts in winds can cause largechanges in coastal circulation and patterns of primaryand secondary productivity. Although we are notable to model food chain effects satisfactorily, boththe Humbolt Current and California Current systemshave demonstrated severe decreases in primary and


9 Evidence is not yet complete.10 Actually, it remains to be determined whether the character of regimes is bimodal or polymodal, as our detailed observational base

is short, and long-term qualitative historical indicators such as sardine presence and absence tend to be binary.

secondary productivity since the early 1970s. Becauseof the complexity of these physical and biologicalsystems, it would appear important to use multiplevariables to define regimes and regime shifts.

The emphasis we have placed on a few popularand highly visible indicator species such as sardinesand anchovies may be misleading. Other species,such as horse mackerel off South America, may beexperiencing large regime-related changes in abun-dance, but our monitoring is inadequate. Even thesardine-anchovy system does not necessarily behavesimilarly in different systems. In the Humbolt system,both anchoveta and sardine show large and inversevariability. In the Japan and California system, sar-dines fluctuate strongly, but anchovies are much lessvariable. Fisheries can further complicate the inter-pretation of biological shifts. Intense harvesting canchange the abundances of both target species andspecies that function as predators, prey and competi-tors. However, fishing will not influence the physicalsystem. The relative effects of fishing and climatehave been difficult to separate in the relatively shorthistory of quantitative data, but resolution shouldbecome easier as the time series of observations isextended, and with increased opportunities for inter-system comparisons.

The recent apparent regime shift in 1998/99has had dramatic effects on coastal zooplankton com-munities (Mackas et al. 2001). Coastwide increasesin juvenile salmon abundance were first identified in1999 at the Coastal Ocean/Salmon Ecosystem eventin Nanaimo (Peterson and McKinnell 2000a) andhave persisted through 2000 (McKinnell 2001) andin 2001, newspaper headlines such as “Salmonreturning to area rivers in record numbers” were notuncommon.

8.1 Regime-like variations in Pacific salmon stocks

Attention was focused on climate and Pacificsalmon issues by Beamish and Bouillon (1993) whenthey reported the coherence of low frequency, largeamplitude variation in catches of sockeye, pink, andchum salmon around the Pacific Rim. The feedingmigration of these three species, along with steelheadtrout, occur in the oceanic waters of the NorthPacific Ocean whereas coho and particularly chinooksalmon tend to occupy coastal areas to a greater

extent. Beamish & Boullion reported that dramaticincreases in the abundance of sockeye, pink, andchum salmon, particularly in Alaska and Japan, werecoincident with the 1976/77 climate regime shift(Ebbesmeyer et al. 1991). Remarkably, just as dra-matic increases in sockeye, pink, and chum salmonwere first reported on basin-wide scales in the early1990s, the catches of chinook and coho salmon, andsteelhead trout declined dramatically in the southernparts of their range in North America such that manyindividual populations were listed as EndangeredSpecies in the United States. In Canada, somesalmon fisheries that had continued for over 100years were closed as a conservation measure for thefirst time (McKinnell et al. 2001b).

The fisheries management response to theseabrupt declines was not immediate, in part becausetheir view (and the view of many scientists) was thatmarine survival of salmon was a stochastic process,rather than an autocorrelated regime-like process.Fisheries on coho salmon, for example, continued athigh exploitation rates even when salmon genera-tions would not have replaced themselves even in theabsence of fishing (Bradford and Irvine 2000).

8.2 Indicators of a regime shift

In current practice, the population trends ofquickly responding pelagic fish species such asanchovy tend to be watched for indications of regimeshift. For example in the Humboldt Current system,the anchovy is considered to be a primary indicator,particularly when the several anchovy stocks distrib-uted from northern Peru to southern Chile exhibitsimultaneous changes in abundance trends. When thishappens, other species such as the common sardine(Strangomera bentincki), that tend like the anchovy tobe favored in cool periods in this system, seem to varysimultaneously and synchronously. Where two pri-mary small pelagic species seem to fluctuate out ofphase (e.g., sardines and anchovies), the earliest indica-tion may be gleaned from the population dynamics ofthe subdominant (suppressed) species, as the subdom-inant species sometimes has begun to increase towardeventual dominance while the currently dominantspecies was still abundant (Schwartzlose et al. 1999).

Changes in the geography of habitat utilized bysmall pelagic fish stocks, particularly reproductivehabitat, tend to accompany major abundance trends.


Consequently, such changes might be a valuable earlyindicator for a regime shift. Egg pump surveys(Checkley et al. 2000) would provide a good way tomonitor the distribution of current reproductive habi-tat and therefore to perhaps spot incipient regime shifts.

Another potential regime shift indicator pro-posed in the workshop discussions was surfacechlorophyll measured from satellite, which has thedistinct advantage of being data that is readily avail-able. However, there was considerable controversyconcerning the utility of surface chlorophyll for iden-tifying regime shifts. It has been difficult to find clearrelationships between variations in chlorophyll and infisheries. For example, the four major eastern bound-ary current systems around the world have very simi-lar chlorophyll biomass levels, but the fish productionthat they typically support differs by as much as 20fold or more. Certainly, the use of satellite remotelysensed chlorophyll has proved useful to identify con-vergent fronts, to monitor the interannual dynamicsof these fronts, and to explain aspects of the move-ment of albacore in the North Pacific (Polovina et al.2001). However this application doesn’t link chloro-phyl biomass changes to fish dynamics, but ratheruses the fact that sharp chlorophyll gradients may bea proxy indicator for a convergent surface front.

Zooplankton species composition appears to bea sensitive indicator of water mass changes that may bekey elements in marine ecosystem regime shifts.Examples of changes in zooplankton volume andspecies composition have been associated with low fre-quency changes in the California Current and betweenthe density of certain Atlantic zooplankton species andthe North Atlantic Oscillation. Attention may need tobe paid to the zooplankton size structure since effectson higher trophic levels (i.e. fish) may be size-specific.

There was considerable debate at the work-shop about what may constitute a regime shift. Forexample, it was mentioned that in recent years therehas been a significant change in zooplankton speciescomposition off Oregon, which is likely a response toan increase in the southward transport of TransitionZone waters. Some participants felt that this southwardmovement of the Transition Zone pelagic ecosystemrepresented a regime shift while others felt that to

qualify as a regime shift, the changes must go beyondmere spatial movements of water masses and theirentrained planktonic communities.

In any case, it was generally agreed that to qual-ify as a regime shift, there must be organized changesin a variety of biological and environmental character-istics. Thus, an appropriate single index of regimeprobably must be substantially multi-dimensionaland multivariate. An example of a multivariateapproach was the use of 100 physical and biologicaltime series to identify 1977 and 1989 regime shifts inthe Northeast Pacific (Hare and Mantua 2000).

Population connectivity 11

Processes driving variability in fish populations(and fisheries) operate on a variety of spatial and tem-poral scales. It appears that current conceptual modelsregarding “fisheries” and “climate” may tend toemphasize the time domain while under-representingthe spatial domain. Such models might benefit frombroadening the focus to include spatial dynamics, ofwhich population “connectivity” is a key element.

There are at least four levels of connectivitythat might be important.

(1) Connectivity within populations. Many marine speciesuse separate locationsfor spawning, larvaldevelopment, juvenilefeeding, and adultfeeding. Such complexlife histories requiresuccessful connectionsbetween spatially seg-regated locations to“close” the life-cycleand successfully pro-duce subsequent generations (Fig. 1). Fishing pressureand/or climate could potentially affect any of these


9

11 This section reports “Connectivity” Focus Group deliberations involving Robert Cowen (chairman), Suam Kim, Simon Thorrold, Ian Perry, Mary-Elena Carr and was prepared by Jeff Shima (Focus Group rapporteur)

Adult Sites

Spawning Sites

Nursery Sites

FIGURE 1.Connectivity within populations.

critical linkages. Climatic shifts, for example, coulddisrupt connectivity within populations throughalterations of ocean circulation patterns. In addition,more localized impacts resulting from climatic shifts(e.g., changes in upwelling regimes at adult sites;changes in terrestrial runoff at estuarine nursery sites)and/or fishing practices (e.g., overfishing at adult sites;by-catch of undersized individuals at nursery sites)may break the “flow” of individuals between sites.

Pertinent in this respect is the potential role of“fish culture” (i.e., learned behavior) in maintainingpatterns of connectivity or isolation. Cod, groupers,and herring (Hay et al. 2001) were cited as possibleexamples, whereby older individuals might indicate toyounger generations the routes to particularly fruitfulfeeding/spawning grounds. We discussed how intensivesize-selective fisheries could remove these individualsleading to a breakdown in historical migration routes,and by implication, use of historical feeding/spawning

grounds. For species where “culture” might play animportant role, we also discussed how an increase in therelative abundance of young (i.e., naïve) individualsmight result in a greater “sampling” of the environ-ment, leading to populations more equipped to trackoptima in changing climatic regimes.

(2) Connectivity among populations. Marine species with complex life histories may haveseparate populations (or stocks) that share a commonlocation during one portion of their life-history (e.g.,shared larval-, juvenile-, or adult feeding habitat),with continued reproductive isolation maintained bysome mechanism (e.g., natal homing; Thorrold et al.2001). Consequently, genetically distinct popula-tions may be directly connected with one another atsome stage of their life-cycle, facilitating competitiveinteractions and/or related responses to spatially dis-crete fisheries/climate disturbances. Examples of such


Adult Population 1

Spawning Sites

Nursery Sites

Adult Population 2

Spawning Sites

Nursery Sites

Pop 1

Pop 4

Pop 3

Pop 2Upstream “source”

Pop 1

Pop 4

Pop 3

Pop 2

Shifted “source”

Regime Shift B. Regime 2A. Regime 1

FIGURE 2. Connectivity among populations.

FIGURE 3. Connectivity between populations. Panel A indicates “normal” regime with southward currentand an upstream source of propagules. Panel B depicts potential alteration in connectivity resulting from ashifted hydrodynamic regime.

connectivity among populations include stocks ofAtlantic herring, which mix as adults in the North-West Atlantic, but retain discrete spawning locations.

(3) Connectivity between populations.Many marine populations may be composed of discretesubunits (i.e., local populations) connected by a disper-sal phase (typically the larval phase but can includemigratory juvenile or adult stages). Such a grouping oflocal populations may be best thought of as a“metapopulation” (sensu Hanski 1991). A metapopu-lation may exhibit dynamics that diverge substantiallyfrom a well-mixed stock, highlighting the importanceof considering spatially explicit dynamics for specieswhere a metapopulation framework may be representa-tive. The nature of the connections between local pop-ulations constituting the metapopulation is likely to beinfluenced by hydrographic conditions, which mayaffect transport and/or survival and successful recruit-ment (Fig.3). Changes in hydrodynamic conditions(potentially stemming from climatic regime shifts) mayalter patterns of connectivity between local populations(e.g., Cowen et al. 2000). Similarly, changes in thereproductive output of any source stock (through local-ized fishing of adults, or climatic shifts that alter localprimary productivity, etc.), may influence the “flow” ofindividuals between local populations.

(4) Ecosystem Connectivity. Although local marine populations may be reproduc-tively isolated from one another (i.e., genetically dis-tinct), they may not fluctuate independently fromone another because of a shared resource base (e.g.,

primary production depleted by one stock may notreach another stock) or a shared predator/fishery(e.g., predators/fisheries may switch to another stockdepending upon the dynamics of a focal stock). Thenature of these shared ecosystem connections may actto synchronize dynamics of demographically discon-nected populations. Climate/hydrodynamic shiftscould therefore influence separate populationsthrough actions upon predator/fishery behavior ordynamics, and/or via effects on a shared resourcebase, independent of any potential effects on direct(i.e., demographic) population connectivity.

Available integrative concepts infishery resource ecology

There are several available schematic construc-tions that serve to integrate a number of processes tocapture an aspect of the perceived reality of marineecosystem operation within an easily recognized con-ceptual “package”. These serve as a useful commu-nicative “shorthand” among colleagues working infisheries oceanography, facilitating efficient discussionby encapsulating a relatively complex set of processesand interactions in a single jargon-type terminology.It may be useful here to identify and describe severalof the most prominent of these.

10.1 The “optimal environmental window”

Within the last fifteenyears, effective nonlinear meth-ods of empirical analysis havebeen introduced to marine ecol-ogy and fisheries science(Mendelssohn and Cury 1987,Mendelssohn and Mendo 1987).These methods were applied byCury and Roy (1989) to thePeruvian anchoveta, the Cali-fornia sardine, the Moroccansardine, and the Senegalese andIvoirian sardinellas. The result wasa consistently domed-shaped


10

Population 1 Population 2

Predation/Fishing

Resources

Population 1 Population 2

Predation/Fishing

Resources

Clim

ate Variability

FIGURE 4. Ecosystem Connectivity. Reproductively isolated local populations may be indi-rectly connected via a shared resource (i.e., “bottom-up connectivity”) or a shared predator orfishery (i.e.,“top-down connectivity”). Climatic shifts may affect any of these ecosystem levels.

relationship (Fig. 5)where reproductivesuccess appearedhighest at an inter-mediate wind inten-sity and decreased atboth higher and lowerintensities.

Over the pastseveral decades, empir-ical analyses of therelationships betweenlocal wind effects andrecruitment variabilityin different easternocean pelagic fishpopulations oftenyielded differing and,therefore, rather un-satisfying results. In view of the results of Cury andRoy, it is no mystery that the previous results mayhave been inconsistent. In cases in which most of thedata may have been on the "left flank" (low windspeed side) of the optimal environmental window, apositive relationship with wind speed would befound. Conversely, if most of the data were on the"right flank" (high wind speed side), an inverse rela-tionship with wind speed would be found. If the datawere rather evenly distributed across both flanks ofthe window, linear methods could pick up no rela-tionship at all.

Expansion of Cury's and Roy's analysis toadditional stocks or data sets has consistently yieldedthe same characteristic "window" feature such thatmaximum recruitment corresponded to an interme-diate wind intensity level during the larval period.These include California anchovy (Cury et al. 1995),a new series of recruitment estimates for the Moroccansardine (Kifani 1991, Roy et al. 1991), an analysis ofpre-1945 data for the California sardine (Ware andThomson 1991), and the Chilean sardine (Serra etal.1998). Various studies associated with the elabora-tion and interpretation are reported in the “CEOSvolume” (Durand et al. 1998).

Constructing an interpretation of the optimalenvironmental window result is quite straightfor-ward. It readily conforms to, and incorporates, severalprominent current hypotheses concerning variabilityin larval survival. The control on the "right flank",

i.e., the "high wind"side, could comeabout either through(1) excessive offshoretransport leading tooffshore loss of pelagiclarvae from the favor-able coastal habitat(e.g. Parrish et al.1981, Sinclair 1988)or through (2) overlyintense turbulent mix-ing which could dis-perse fine-scale con-centrations of appro-priately sized foodparticles needed forsuccessful first feeding(Lasker 1975, 1978)

as well as inhibit basic photosynthetic production bymixing phytoplankton cells beyond their "criticaldepth" (Sverdrup 1953, Steele 1974). Strong turbu-lence might also impair a larva's ability to physicallycapture prey (MacKenzie et al. 1994). An obviousexplanation for the "left flank", or "low wind" side, isa lack of nutrient-enrichment by wind-inducedupwelling or mixing, leading to inadequate produc-tion of appropriate larval food (Cushing 1969). Inaddition, it is possible that under conditions wherethe interaction of feeding behavior with stable fine-scale food particle structure may be less importantthan the energy savings produced by turbulent diffu-sion of food particles toward feeding larvae, themechanism of Rothschild and Osborn (1988) mayexert some control on the "left flank" by increasinglarval survival toward the slightly higher wind speedswithin the "window".

10.2 Ocean triads

Comparative studies of fish habitat climatologyhave tended to identify three major classes of physicalprocesses that combine to yield favorable reproductivehabitat for coastal pelagic fishes and also many othertypes of fishes:

(1) enrichment processes (upwelling, mixing, etc.)


WEAK MODERATE STRONG

~ 6 m s-1

W I N D I N T E N S I T Y

OPTIMALENVIRONMENTAL

WINDOW

RE

CR

UIT

ME

NT

FIGURE 5. The optimal environmental window (Cury and Roy 1989)where reproductive success (recruitment) is highest at an intermediatewind intensity level and declines at both higher and lower intensity levels.

(2) concentration processes (convergence,frontal formation, water column sta-bility) and

(3) processes favoring retention within(or drift toward) appropriate habitat.

This set of types of processes has been calledthe "fundamental triad" underlying reproductivehabitat suitability by Bakun (1996). Ocean zoneswhere the triad elements coexist in configurationsthat are evidently favorable for population growth oflocally reproducing species are sometimes called“ocean triads” (Bakun 1998).

The importance of enrichment processes iswidely recognized and appreciated. Perhaps less widelyappreciated is the importance of concentrationprocesses. For small organisms, such as fish larvae,sea water represents quite a viscous fluid; major energyexpenditures may be necessary just to move fromfood particle to food particle. Thus large amounts ofenergy, needed for the rapid growth that is requiredfor quick passage through the various size-related levelsof intense predation that pervade the ocean environ-ment, may be expended in feeding activity.Consequently, availability of processes whereby foodparticles are concentrated (e.g., Fig. 6) tends to be vital.

This is probably a major reason why varioustypes of interfaces, or ergoclines (Legendre and Demers1985), tend to be sites of enhanced biological activity in

the ocean. Such interfaces tend either to be maintainedby, and/or to maintain, mechanisms of concentration(Bakun 1996). Ocean fronts (Fig. 6) are obvious exam-ples. The importance of processes occurring in or nearocean fronts is suggested by the widespread attractionof fish and other marine animals to drifting objects. Theactual convergent water motions associated with afront may be too subtle to be directly sensed by pelagicorganisms operating in an environment devoid of fixedreference points. However, drifting objects tend to becarried into and to accumulate within frontal structuresAn innate attraction to drifting objects serves to positionthe fish within the zones of enhanced biological activ-ity and correspondingly improved feeding conditions.

Conversely, turbulence is a dispersive processand so tends to act counter to concentration processes.Thus, intense turbulent mixing events have appearedto be detrimental to larval survival (e.g., Lasker,1978, 1981a, 1981b). On the other hand, extremelysmall-scale turbulence might actually act like a con-centration mechanism by increasing the encounterrate of small organisms with food particles(Rothschild and Osborn, 1988).

The third element in the triad is retention. Lifecycles of marine organisms tend to include at leastone stage of passive larval drift. Thus, in a dispersivefluid medium, loss of early life stages from the popu-lation habitat may represent serious wastage ofresources. Consequently, fish populations tend tospawn in locations and seasons that minimize suchlosses (Parrish et al. 1981, Sinclair 1988).

A set of examples of favorable ocean triad con-figurations for eastern ocean coastal upwelling systemscan be found in Bakun (1998), and for a semi-enclosed sea in Agostini and Bakun (2002). Otherexamples are described in Bakun (1996).

10.3 MacCall’s basin model

A particularly fully-elaborated theoreticalconstruction is MacCall’s (1990) “basin model” con-ceptualization (Fig. 7), where the favorability of thehabitat for population growth and maintenance ateach location is plotted on a diagram on a scale thatincreases downwards, visually tracing out a “basin”.Density-related unfavorable factors (crowding, com-petition for food, etc.) oppose the accumulation of allthe fish in the most favorable location (the “deepest”point in the basin”). In fact, the model postulates that


Lower Density Water(fresher and/or warmer)

Higher Density Water(more saline and/or cooler)

Surface Front

FIGURE 6. Schematic diagram of a front between waters of differing density.Arrows indicate density-driven flows associated with the front. "Particle"symbols indicate planktonic organisms capable of resisting vertical displacement.(Scales are distorted: vertical scale greatly expanded relative to horizontal;particles greatly magnified; surface waves not to scale, etc.) After Bakun(1996).

the fish will spread themselves through the habitat insuch a way that reduced density counteracts the effectsof poorer habitat, such that all the fish have equal netbenefit regardless of their position, a condition calledan “ideal free distribution” in ecological theory.MacCall does not provide a mechanism as to how thefish might accomplish such an equitable allocation ofhabitat space in any particular instance but proposesthat such a distribution can be approached if fish havea tendency to follow gradients of perceived improve-ment in habitat favorability. The characterization isprobably apt in a long-term mean sense and might inany particular instance be considered as a most prob-able distribution of biomass concentration. In thelong-term mean case, one could imagine individualsof a population as being analogous to molecules of aliquid, which would fill the “basin” up to a given level,with the number of molecules (individuals) at anyhorizontal location being proportional to the depth(habitat favorability) of the basin at that location.

Thus in Fig.7a, a high biomass phase might bevisualized as filling the basin up to one of the upperdashed lines, where the best habitat and therefore the

greatest density of individuals would be located inthe deeper southern sub-basin, but the populationlevel is high enough so that some of the populationoverflows the intermediate “shallows” in habitat suit-ability, to fill up the secondary northern sub-basin.This signifies the habitat expansion that tends toaccompany higher biomass phases of sardines andother small coastal pelagic species. Then if the popu-lation were to decline sufficiently (e.g., to a levelbelow the level of the shallow area between the twosub-basins) the population may be entirely confinedto the deeper sub-basin. As long as the basic meanenvironmental configuration (i.e., the basin topogra-phy) remains unchanged, the population distributionwill continue to cycle through essentially the samepattern of geographical expansion and contraction.

However, if a climate shift altered the underlyingbasin topography (e.g., if the northern sub-basinbecame deeper than the southern one as indicated inFig. 7b) the population could actually shift its locationin a low biomass phase to become concentrated inarea of the northern sub-basin.


0

LOCATION

0

a.

b.

Northern sub-basin

Southern sub-basin

Northern sub-basin

Southern sub-basin

FIGURE 7. Schematic diagrams based on MacCall's "basin model" concept of density-dependenthabitat suitability (MacCall, 1990). The area circumscribed by the intersection of each dashedhorizontal line with the basin floor is functionally related to total population size. The line cor-responding to zero population growth implies the "carrying capacity" of the habitat. The habitatsuitability and corresponding steady-state population density are implied by the depth of thebasin at any location. (a) Before a climatic shift. (b) Following a climatic shift.

10.4 “Member-vagrant” concept

Sinclair (1988) presented an elegant treatmentof the role of local retention of early life stages byphysical ocean structural features in maintainingmarine populations. He referred to the individualsretained as “members” and those lost to the localpopulation as “vagrants”. Sinclair’s essay has donemuch to call attention to the essential linkages ofocean hydrodynamics to marine population dynamics,particularly to the fact that it is only the “member”portion of the surviving offspring that are importantto year-to-year population variation, and to theessential question of degree of “openness” or “closed-ness” of a local population segment. Thus the term“member-vagrant” has come to connote a specificclass of extremely important issues in the climate-fisheries research area.

10.5 “Match-mismatch” hypothesis

Cushing (1971, 1975, 1990) has popularizedthe idea that year class success in relatively high-lati-tude systems, because of the short growing seasonsfor phytoplankton production, may be cruciallydependent on the timing of spawning with respect tothe timing of availability of sufficient concentrationsof larval food particles (e.g., copepod nauplii), withyears with a good “match” yielding good reproductivesuccess and strong year classes, and years of substantialmismatch resulting in poor reproductive success andweak year classes. Because of the very important effectof temperature on rates of biological development,interannual temperature variations are importantelements of this way of thinking about populationvariation. Likewise, because the establishment ofwater column stability tends to be what determinesthe onset of the spring phytoplankton bloom in sub-polar systems, incidence of late winter or early springstorms may also be extremely important to annualrecruitment strength.

Cushing mainly focused his attention onNorth Atlantic herring, but the idea obviously maybe applicable to a number of different types ofspecies of higher latitudes, including cod, haddock,flounder, etc. (e.g., Brander and Hurley 1992, Fortieret al. 1995, Gotceitas et al. 1996, Brander et al.2001) and also animals other than fishes. For exam-ple, Bertram et al. (2001) showed that early spring

warming results in early arousal from diapause inNeocalanus cristatus. When the plankton grow andreturn to depth in diapause early, they are not availableto the important planktivorous bird, Cassin’s auklet,and as a consequence there is reproductive failure inthe auklet.

10.6 “Lasker windows”

Although the idea has been in a state of relativeeclipse in recent years following the decline of theIOC-FAO International Recruitment Program andthe ascendancy of the GLOBEC Program, Lasker’s(1975, 1978) stable ocean hypothesis was a major fac-tor in structuring fisheries-oceanographic research inthe 1980s and early 1990s. In a remarkable series oflaboratory experiments and field investigations,Reuben Lasker and his colleagues concluded thatfine-scale patches of highly concentrated food parti-cles were necessary for successful first feeding ofanchovy larvae off Southern California, and thatstorm events could destroy this fine scale structure,leading to recruitment failure. Bakun and Parrish(1980) noted that it is probably not the average mixingintensity that is crucial, but the existence of adequatetemporal "windows" during which the production ofturbulent mixing energy remains low for a longenough period for concentrated food particle stratato accumulate and for larvae to accomplish successful“first-feeding”. For example, if five full days may berequired following a storm-related mixing event inorder to form sufficiently concentrated food strata toprevent larval starvation, any calm period five days orless in duration would have no influence in increas-ing survival of a year class. It is only on the sixth dayand the succeeding days of continued calm that larvaemay pass through this critical “first feeding” survivalwindow. And as soon as another mixing event tookplace, the process would have to begin again withthe necessity of another unbroken five-day stringbefore any more larvae could be successful in avoid-ing starvation.

Following this idea, Peterman and Bradford(1987) succeeded in empirically relating interyear vari-ability in larval anchovy mortality rate off California tofrequency of calm periods of sufficient duration forfine scale strata of food organisms to form. Such calmperiods have been called variously "Lasker windows","Lasker events", “Lasker gates”, etc.


Following Pauly (1989), one can define a“Lasker {x,y} window” as representing a full day inwhich the wind speed did not exceed "x" m s-1,which directly follows an unbroken sequence of atleast "y" preceding days with wind speeds not exceed-ing the same "x" m s-1 limit. This allows a rationalparameterization of crucial variability occurring onshorter scales than the time increment (e.g., a fullspawning season) used in analysis. Such a parameter-ization, which refers specifically to a defined set ofmechanisms, represents an advance over using mereaverages as proxies for aggregate effects of shorter-scale variability. Peterman and Bradford (1987) chose10 m s-1 for the wind speed limit in applying theLasker window approach in their study of Californiaanchovy. Mendelssohn and Mendo (1987) chose alesser wind speed limit of 5 m s-1 for their study ofPeruvian anchoveta.

One should note that "Lasker windows" repre-sent a physical effect that passively allows concentra-tions of food particle distributions to build up andpersist, as opposed, for example, to the process ofactive concentration by the physical process of con-vergence in frontal formations. In this case the activeconcentrating agent is biological, produced either byactive swarming behavior or by in situ growth.Existence of the physical event (i.e., a sufficient Laskerwindow) permits the biological process to be effective.

10.7 The “school trap”

During the collapse of the California sardinepopulation in the 1940s and 1950s, pure schoolsbecame less frequent and sardines were found mixedwith anchovy or chub mackerel (Radovich 1979).During the period of very low abundance in the late1960s, the only California sardines ever reportedwere observed in very small numbers entrained with-in schools of much more numerous jack mackerel ofsimilar size. Likewise, off Peru, large sardines havebeen found in mixed schools with horse mackereland chub mackerel of similar size, whereas smallersardines have been observed nearer the coast mixedwith anchoveta (Jordán et al. 1978).The existence ofmixed-species fish schools appears to underscore theextreme strength of the schooling imperative (Bakun

and Cury 1999). Associated larger than optimalschool size distributions and extreme reluctance ofindividuals to voluntarily leave a school may lead tolowered levels of resources available to individuals.Related mechanisms may account for observed featuresof density-dependent somatic growth, density-dependent habitat selection and species alternationsin small pelagic fishes12.

An interesting point brought out by Bakun andCury is the possibility of adverse effects on a minorityspecies operating within schools of a more abundantspecies. For example, because sardines are known to bemore migratory than anchovies, and presumablytherefore better adapted for sustained strong efficientswimming, anchovies entrained in sardine schools mayfind themselves needing to swim harder than theirlevel of optimal efficiency just to keep up. Because sar-dines are able to effectively filter much smaller foodparticles, these entrained anchovies may experiencedifficulties in capturing sufficient food in many of theareas through which a sardine-dominated schoolmight elect to traverse. Thus the school trap could bea factor in the fact that population oscillations ofanchovies and sardines tend to be in opposite phase,almost never in modern fishery experience reachinghigh levels of biomass at the same time.

Potential implications of rapidly- evolving adaptive response mechanisms

Bakun (2001, 2002a) has elaborated a premisethat (1) durable adaptive responses may be evolvingwithin fish populations on much more rapid timescales (~10 – 50 years) than possible via classicalgenetics-based Darwinian evolution and (2) if so,there are major implications for the way we may viewfish population dynamics. For example, any degree ofunderstanding of the adaptive mechanisms involved,or even a well-founded belief that such mechanismsare operating, could conceivably open up a broadrange of new considerations for adaptive manage-ment responses, design of mitigation actions, andforesight as to the nature and course of climatic or


11

12 A. Bakun, P. Cury, P. Fréon, C. Roy (under revision) The “school trap”–The price to pay for remaining together.

anthropogenic impacts. It was, in fact, suggested thatsuch rapidly-evolving response processes operatingwithin exploited resource populations could causesome conventional fishery management responses tobe absolutely the wrong ones (e.g., serving to trapand maintain the populations within low-productivityphases) and that moreover, unless the mechanismsinvolved are understood, it may be very difficult to eversort out, understand, and adequately allow for, theinteracting effects of low frequency climate variability.One particular aspect of the school-mix feedbackmechanism is that it could conceivably be an importantelement of regime shifts occurring in marine ecosys-tems. Another is that the mechanism could provide aprocess by which mobile fish populations might altertheir customary zones of operation so as to withdrawaway from areas of major fisheries.

The author of these ideas, Andrew Bakun,being one of the organizers of the workshop, distrib-uted an extract from Bakun (2002a) to the invitedparticipants prior to the workshop in order to try tomotivate discussion on this class of issues. Workshopparticipants generally agreed that there is clear evi-dence that fisheries pressure can act to alter populationcharacteristics such as age and size of first reproduc-tion, sex ratio, etc. It was also agreed that it is ratherclear that what may have seemed to be populationcollapses, have been so in reality. Participants weregenerally unaware of situations where stock abundancehad been essentially maintained, while only the dis-tribution had changed. While many felt a need for aparadigm shift in fisheries thinking, most tended toagree that more evidence is needed before the premisethat adaptive responses may be evolving within fishpopulations under strong selective pressures (espe-cially fisheries as a selective pressure) and that theseresponses are passed on to succeeding generations.

Most were more comfortable with more con-ventional ideas of direct adaptive responses tochanges in environmental conditions that, ratherthan representing the type of rapidly-evolvingresponse envisioned by Bakun (2001, 2002a), wouldseem to be innate elements of the animals basic phys-iology, or behavioral responses that are ingrained inthe species by very long time-scale biological(Darwinian) evolution. For example, Alec MacCallreported an observed adaptive response of sardines inthat they increase egg production when food avail-ability is high, which is likely a direct physiological

response to the individual’s excess available energy forelaborating reproductive products.

As another example, it was noted that duringcalm winters previous to the climatic regime changethat occurred about 1976, the Japanese sardine,which was in a low population phase at the time, hadthe habit of spawning in bays during late winter.After 1976, intense monsoon forcing associated withharsh winters may have caused deep mixing of thebays, making them non-suitable for adult feeding orfor survival of offspring, thereby causing the popula-tions to move out of the bays. It was in this period,when the earlier adapted strategy may have becomenon-viable, that the sardine population underwentits dramatic expansion. Thus vertical stability hasbeen hypothesized as the agent for switching the sar-dines between the two phases of population abun-dance, yielding the apparent paradox of a change topoor environmental conditions in an adaptively-selected reproductive habitat leading to a massivepopulation increase (some participants were dubi-ous). The process by which the population wasmoved out of the bays in a durable manner was notspecified. One possibility is that each generation ofthe fish went back and tested the suitability of thebays and each time found them lacking. Otherwise,some sort of rapidly-evolving adaptive responsemechanism would seem to have been required tokeep new year classes of the population, which tendto be segregated in separate schools because of sizedifferences, from re-entering the bays.

In general, the feeling among workshop partic-ipants was that there may be inadequate data todirectly resolve the school-mix feedback issue.However, some felt that there might be ways that the“school mix-feedback” hypothesis could be testedempirically using otolith microchemistry to identifyrelative abundance of individuals with differing“affinities” for a particular region. This approachmight be most applicable to species that exhibit spe-cific behavioral attributes (e.g., natal homing). Someparticipants were of the opinion that the hypothesisseems too focused on a simple predator-prey interac-tion; for example, the existence of replacement mech-anisms among species at the ecosystem level, includingchanges from vertebrate to invertebrate dominanceor changes from a “chitinous” to a “gelatinous” foodweb, emphasize the importance of considering morecomplex multispecies interactions than a simple


predator-prey interaction. In any case, in order to testthe school-mix feedback” hypothesis, very explicitdefinition of its assumptions and critical mechanismsand processes to be falsified are necessary. However,under time constraints of this workshop, deeperexamination of this issue was not possible.

However, the discussion produced a clearrecognition of the fact that we are always limited by

data and that the practitioners of fisheries oceanogra-phy need to allocate more energy to synthesis effortsto generate testable hypothesis about the mechanismsand processes involved in the relation between climateand fisheries variability. We are in need of new ideasthat can break the conceptual deadlock and producereal progress.


Workshop Discussions

2. FUTURE DIRECTIONS

Applications to real world needs 13

Although the application of knowledge of cli-mate and societal interactions is a relatively youngfield, there have been enough lessons gleaned fromrecent use of climate information in parallel cases (e.g.,technology transfer) that can be brought to the tablewhen considering how current and future informationcould be used in the fisheries sector (see examples ofconstraints and opportunities related to the use of cli-mate information in other sectors in Agrawala andBroad 2002, Barrett 1998, Broad and Agrawala 2000,Glantz 1986, 1995, 1996, Hansen 2001, Letson et al.2001, Mjelde et al. 1988, Mjelde and Keplinger 1998,Orlove and Tosteson 1998, Pulwarty and Redmond1997, Pfaff et al. 1999, Rayner et al. 1999, Roncoli2000, Sarewitz et al. 2000, Stern and Easterling 1999).The attention of the “social science” focus group(which included anthropologists, political scientists,biological and physical scientists with experience in cli-mate forecast applications, a FAO fisheries specialist,and a US government economist involved in fisheriesmanagement) was on addressing the issue of whatdecisions could potentially be enabled, or enhanced,by significant scientific progress on understanding theeffects of climate variability on the marine resourcessystem. Several themes emerged early in the discussionand remained central throughout the meeting. Theseincluded:

(a) definitional issues regarding relevant decision-makers who could potentially use climate information;

(b) distinguishing timescales of importance and identifying key decisions made on those timescales;

(c) recognition of the importance of communicationof probabilistic scientific information;

(d) the risk of unintended consequences of the use ofclimate information.

Current and potential understanding of thelinkage between climate and fisheries on multiple

timescales is theoretically relevant to many of thegroups that make up the “fisheries sector”. Forexample, in addition to the scientists who make rec-ommendations on regulations and the managersthemselves, we must consider the financial decision-makers (e.g., bankers who make loans), plant ownersand labor, industrial and artisanal fishers, and eventheir families who make household-level decisionsbased on expectations of the coming season(s) catch.Central to the discussion was the point that it is notonly fisheries managers who can, and will, use thelatest available scientific information for regulatorypurposes. Thus, these other groups should be consid-ered as integral to the communication process fromthe start.

In the current dialogue concerning use of cli-mate information, there remains division over whatare the critical timescales for various fisheries man-agement decisions. Again, if we define this issuebroadly, as discussed above, based on current knowl-edge of reactions to information by the differentactors, different timescales have implications for dif-ferent decisions:

Synoptic: Improvements in real-time observational capabilitiestend to favor the industry versus the regulators.Industry is usually on the cutting edge of develop-ing/adapting new technologies and are already adeptat using satellite information to choose optimal fish-ing grounds. General regulations, on the other hand,are difficult to change in response to synoptic scaleinformation because of insufficient lead time to makechanges in the regulatory process. However, improve-ments in sampling devices (i.e., egg pump counter)and satellite tracking devices could be used for finergrain management and enforcement decisions in nearreal-time, especially once an event is underway.

Seasonal-to-interannual: This has been the timescale of greatest focus to datein terms of application of forecast, in large partbecause of our understanding of the physical mecha-nisms that enable the ability to predict of one of themain drivers of seasonal climate variability, the ElNiño-Southern Oscillation (ENSO). ENSO has direct


12

13 Guillermo Podesta made important contributions to the preparation of this section.

impacts on some commercially important species(e.g., Peruvian anchoveta, tropical Pacific tuna popu-lations). While forecasts of ENSO events have someskill with up to 6 months lead time, there remainproblems with implementing major changes in fish-eries within this time frame for a number of reasons:the industrial sector (including vessels, fishmealplants, canneries, etc.) is usually heavily indebted,with investments that are calculated to be paid offover multiple years. Again, only when the timing ofthe investment happens to coincide with relevantinformation about upcoming conditions that mayaffect the coming seasons could a decision-makermake good use of information (e.g., give a loan/don’tgive a loan; buy a new engine or not, etc.).Otherwise, lenders (i.e., banks) and borrowers (i.e.,firms) are locked into a pattern with a different tem-poral scale than ENSO-like events; artisanal (smallscale) fishing groups tend to lack the capital to switchapparatus or the skills to temporarily switch profes-sions given the short lead time; current forecasts havedifficulty predicting the magnitude and duration ofevents, and each event is associated with differentimpacts on the living marine resources.

Decadal-interdecadal (regime shifts): Much of the meeting focused son the issue of largescale fluctuations in major fisheries populations.Such occurrences have obvious dramatic implicationsfor the human populations that depend on theseresources. However, given our level of understandingof such shifts, the realistic potential for this informa-tion to affect fisheries-related decisions are not clear.We are still unable to answer basic questions neces-sary for societal decision-making, for example, “arewe in the rise, peak, or decline phase of a fluctua-tion?” or “how can we tell a regime shift from othersorts of variability?”, or “what is the relationshipbetween interdecadal variability and events thatoccur on other timescales?”. This does not imply,however, that we “throw the baby out with the bath-water” – from a societal perspective; understandingthese shifts is arguably a key factor to reducing whatis increasingly agreed to be the central problem inglobal fisheries – overcapitalization of the industry.As mentioned above, many of the decisions leadingto overcapitalization (e.g., lending practices) are notgoing to be solved/reversed with short-term climateinformation, but in theory, more rational decisions

could be made by groups ranging from governmentplanners to plant laborers if there were a sense ofwhat the next few years could hold.

There was general consensus in the group thatthere needed to be a more systematic study of therange of actors and decisions made on the varioustimescales. Drawing on experiences from the applica-tion of climate information in the agricultural andwater management sectors, one product that wasidentified as likely to be useful for our understandingof the potential of different sorts of climate informa-tion was a “decision calendar” (see Pulwarty and Melis,2001, for an example of a case study that bridgestimescales and demands of multiple stakeholders).This product would include key decisions by the dif-ferent actors, key periods when these decisions aremade, the lead times, and the type of informationnecessary for influencing the decision. Such a prod-uct is a first step for (a) identifying points of insertionof climate information of all types/timescales, (b)identifying unintended consequences of poor fore-casts, and (c) identifying which groups might befavored or hurt by the introduction of climate infor-mation into the decision-making process.

Considering the issue of application of scien-tific knowledge brings up the question of when is thescience good enough to “go public”? Of course, thereis not a clear benchmark, and ultimately, applyingscientific knowledge becomes an iterative learningprocess based on successes and failures. Past lessonshave led to the realization that effort must be put onthe communication process, as decision-makersinterpret probabilistic information in a variety ofways based on the societal and individual characteris-tics from which they operate. Thus a bank manageraccustomed to dealing with the uncertainty associatedwith financial forecasts may be better suited to useprobabilistic climate information than a vessel cap-tain who must maintain a high degree of certainty inactions that affect the safety of the crew would be.This may necessitate different sorts of approaches/products/education for different decision-makerswithin the same sector – a time consuming and costlyendeavor that the scientific community may not beable to undertake.

As with all sorts of applications of new infor-mation and technology transfer, there is a risk ofpotential unintended consequences of introducingnew information into a social system. Unintended


outcomes may result from erroneous information,misunderstanding of information, malicious use ofinformation, or a range of exogenous factors outsidethe control of the decision-maker. Further, given ourrelative lack of understanding of many aspects ofregime shifts, the scientific community may risk losinglegitimacy by promoting the use of what informationwe have to fisheries managers and other decision-makers, thereby reducing the chances of being takenseriously in the future. Finally, rational decisions byindividuals may not result in an outcome that is goodfor a collective group. For example, detailed knowl-edge of migratory patterns of some species inresponse to climatic trends may allow more advancedplanning of fleet location, gear needs, etc., thus leadingto increased, even more efficient extraction and anensuing drop in prices (good for the consumer, badfor the fishing industry). Fisheries that lack enforce-ment (e.g., tuna outside of the EEZ) may be especiallyvulnerable. Differential understanding of informa-tion during the 1997-98 El Niño in parts of S. Americademonstrated how benefits may accrue unequally,and how probabilistic information can be subject toalternative, and often self-serving, interpretations(see for example, Broad et al. in press).

Role of comparative studies

The great biologist, Ernst Mayr has called theexperimental method and the comparative method“the two great methods of science” (Mayr 1982).Drawing valid scientific inference requires multiplerealizations of the process of interest, preferably overa range of differing conditions, in order to separatecausality from happenstance with some reasonabledegree of confidence. The experimental method,wherein experimental controls are imposed thatallow the scientist to systematically vary conditions ofinterest while holding other factors constant, is per-haps the most direct approach to assembling theneeded suite of realizations. But climatic linkages tomarine ecosystems and to associated human economicand cultural activities are hardly amenable to experi-mental controls.

Fortunately, the comparative method presentsan available alternative. For example, Mayr (ibid.)

credits the comparative method for nearly all of therevolutionary advances in evolutionary biology,which likewise involves scales (mostly temporal inthat case) that can not be encompassed by the usualexperimental approaches. The comparative methodassembles the separate realizations needed for sci-entific inference by a process of recognition ofinformative patterns of naturally-occurring temporaland spatial variations in naturally existing conditionsand phenomena. That is, different sets of geographicalsettings, encompassing a range of natural variabilityin the conditions and mechanisms, replace controlledexperimental "treatments".

Of the available integrative concepts enumer-ated in Section 10, the optimal environment window,ocean triads, school trap and the member-vagrantresults are clear examples of products of applicationof the comparative method. In addition, the basinmodel, while perhaps fundamentally a theoretical(intellectual) concept, was obviously framed on thebasis of MacCall’s early knowledge of the similar pat-terns of correspondence between population abun-dance and extent of occupied habitat in theCalifornia and Peru-Humboldt regional systems.

13.1 Time series analysis: the application of the comparative method in the tem-poral domain

Of course, one may look for independent real-izations of a process in time as well as in space. Thustime series analysis is essentially another mode ofapplication of the comparative method where theseparate realizations required for drawing valid infer-ences are separated by time rather than space.Essential to the application of most commonmethodologies for drawing scientific inference viatime series analysis is the requirement for stationarityof the basic mechanism being investigated. Thisworks well in physics and chemistry, and in biologyfor mechanisms acting at the molecular level. Butorganisms and groups of organisms have abilities toadapt habits and behaviors, which may introduceimportant nonstationarities into the operation ofenvironmental – biological linkage mechanisms (asdiscussed in Section 11). Since fitting of models andempirical functions to data series based on anassumption of stationarity has become such a basicelement of fisheries science, a major challenge


13

appears to be to find ways to ensure that stationarityassumptions are appropriate when we use them.Another may be to find ways to validly utilize thecomparative method in the time domain to investigatethe nonstationary processes and mechanisms that areintrinsic to biological systems.

Another problem is the fact that in fishery-environmental science, reliance on linear statistics andempirical methods has been very much the fashion.This is in spite of the fact that one would intuitivelyexpect dome-shaped relationships rather than linearones. Fish stocks would have a natural tendency toadapt their spawning habits to represent choices ofseasonality and geography that would most oftenyield the most favorable combinations of the princi-ple factors controlling recent reproductive success.That is how natural selection works. Accordingly,fish populations would tend to be adapted to, andtherefore fare best under, conditions which are rathertypical of their habitual spawning habitats. Thereforeit would seem that highest success should be associ-ated more with typical conditions than with atypicalconditions on the spawning grounds (unless, forexample, the atypical conditions represented favor-able circumstances which were not normally availableelsewhere within the range of the population).Consequently, one would generally expect dome-shaped relationships, with highest success at interme-diate values of a crucial factor and lower success atmore extreme values on either the high or low sides.For example, temperature can either be too high ortoo low, with the optimum for a given species atsome intermediate value.

Thus it may not be surprising that empiricalstudies of environment-recruitment linkages haveoften yielded inconsistent, and therefore intellectuallynon-satisfactory, results. For example, if an empiricalstudy addressed a situation where in most instancesconditions were on one flank of such a dome-shapedrelationship (e.g., near an extreme end of speciesrange, etc.), than a linear analysis might pick up a sig-nificant relationship. Likewise, in another situationwhere most of the data were on the other flank, anequally significant result, but having opposite “sign”,could be found. In such a case, comparison of the

two situations would yield directly opposing results,even though the underlying dome-shaped relation-ship held consistently in all cases. And of course, ifdata were distributed on both flanks of such a dome-shaped relationship, linear methods would probablyfail completely to pick up any significant empiricalrelationship at all.

A problem with introducing the possibility ofnonlinear relationships is that it is much easier to fitdata when one has an indefinite choice of functionalforms. Consequently, without the discipline of a singlea priori choice, such as linearity, the problem of spu-rious fits becomes even worse than usual.

13.2 Inter-regional comparative time-seriesanalysis

In such circumstances, requiring comparativeinterregional consistency in functional form mayoffer a useful alternative. In order to help relieve theproblems introduced by the very short time series ofannual data points normally available for fisheries-environmental research, Bakun (1985) suggestedarraying empirical models, derived on similar basesfrom different similarly-structured regional ecosys-tems, and attempting to recognize informative patternsamong model parameters. Cury and Roy (1989) tookup this idea and added the additional aspect of non-linear analysis methodologies. The result was thefamous, domed-shaped, "optimal environmentalwindow" relationship (Fig. 5, Section 10.1). Thisfinding, and its follow-on extensions (see Durand etal. 1998) represents the most consistent empirical“environment-population dynamics” result that wehave in fishery science. It has provided tangibleempirical support for concepts (e.g., the ocean triadsidea) that had been emerging inferentially over theprevious decade. Most importantly, the fact that thelocal wind emerges consistently as the factor yieldingthe a priori-expected dome-shaped response makes itrather clear that the global climatic signals that seem tobe imposing large-scale marine population synchronyare, in the case of eastern-ocean coastal upwellingecosystems, most probably transmitted primarily bythe action of the local wind on the sea surface.


Role of modeling 14

Modeling has become an essential componentof integrated ocean research efforts. Models providethe basic “bookkeeping” for testing how well what webelieve we understand about the controlling mecha-nisms operating within a system of interest indeedadds up, based on the observed “inputs” to the system,to what is actually observed in terms of “outputs”.Thus they are an essential discipline enforced uponour scientific credulity in situations where the systemmay be sufficiently complex that one cannot easilyencompass mentally all the significant details of itsoperation. This is particularly true where the processesin the system have feedback linkages or where non-linearities may be involved. And we know thatmarine ecosystem processes are rife with interlinkingfeedbacks and nonlinearities.

However, one must always keep in mind thefamous “garbage in, garbage out” warning. A model’sresults do not represent any new factual knowledgegenerated internally to the model itself, but are onlythe direct reflection of the knowledge, insights, andassumptions that went into the model’s formulation.This is a fact that must be emphasized in fisheries sci-ence, where model outputs representing the results offairly arbitrary assumptions, have come to be viewed,in a situation where real objective reality has been hardto come by, as acceptable substitutes for actual reality.

But, with this admonition in mind, it is unde-niable that process-oriented models, which are nor-mally executed on computers, are extremely usefulcomponents of a scientific investigation. Model studiestend to be very inexpensive compared, for example,to at-sea operations. They are an extremely econom-ical way to identify data requirements, and to devel-op hypotheses that may be amenable to testing. Theycan be run in simulation mode to help clarify andcommunicate ideas and to identify key junctures inthe system. For example, one may use different modelswith the same forcing to understand how the out-comes may differ based solely on the different modelformulations. A computer laboratory, based on a setof models, may provide a virtual environment that

can be used to follow an experimental approach andto run controlled experiments (i.e., altering the windforcing, the currents at the boundary, the sea floortopography, the nutrient composition of the water,mortality of larvae, etc.). In this way, computer modelscan be used to complement field studies.

Currently, biological models tend to not be veryuseful in a predictive sense because they are, as yet, notvery good. And the degree of uncertainty increases asthe results are compounded through the various levelsof an ecosystem. But lack of skill in specific predictiondoes not necessarily connote lack of utility (consider,for example, the current situation in El Niño predic-tion, where a number of different models are used toassemble a current prognosis; the fact that a particularmodel may perform well in one situation but “miss”widely in another does not in any way mean that whatwe have learned about the evolution of and precursorsto El Niño episodes is not extremely valuable in a verypractical sense). Arraying together the outcomes ofruns of similarly formulated models applied to differ-ent regional situations and attempting to rationalizeinformative patterns of correspondence and non-correspondence in these results may be a productivemethodology for pursuing inferences and insights viathe comparative method (thus seeking to make anymodels developed “transportable” to different regionalsituations is a very good idea).

The problem of “down-scaling” from globalatmospheric models or basin-scale ocean models, toregional (e.g., “coastal band” models, etc.) capable ofresolving mesoscale features is an important one. Theresponse of the open ocean to major climatic signals orevents such as ENSO, PDO (Pacific DecadalOscillation, Mantua et al. 1997) and NAO (NorthAtlantic Oscillation, Hurrell and van Loon 1997) isextensively studied and documented15. While thesestudies have considerable implications for understand-ing large-scale climate dynamics of the earth, theirregional manifestation and especially their conse-quences at smaller scales such as the scale of the ocean’scontinental shelves remain poorly documented andunderstood. Climate and oceanographic basin scaleanalyses are often based on spatially smoothed datasets


14

14 Claude Roy and Frank Schwing made important contributions to the preparation of this section.15 It is notable that some global climate models (e.g. Timmermann et al. 1999) predict an increase in the amplitude and frequency of

ENSO evacents in the equatorial Pacific due to atmospheric accumulation of greenhouse gases.

or on large-scale models, with a resolution too coarseto resolve the mesoscale structures such as upwellingthat occurs over the continental shelf. As a conse-quence, these analyses tend to submerge indications ofthe coastal response within those of the adjacentlarger-scale oceanic environment. For example, ananalysis of SST and wind along the coastal shelf offWest Africa indicated that the link between ENSOand this part of the Atlantic ocean is far stronger thanpreviously thought (Roy and Reason, 2001).

Surprisingly, the impact of large-scale climatevariability on the ecology of coastal ecosystems isquite often better known than the impact on thecoastal marine physical components. For example,there is a abundant literature on the impact of the1982-1983 and 1997-1998 El Niño episodes on themarine biota and resources of California and Perubut much less is known on the El Niño impact on thelocal physical processes (see for example, the volume“Pacific Climate Variability and Marine EcosystemsImpacts” (McKinnell et al. 2001a)).

A comparative research focus on the impact ofthe dominant large-scale climatic signals on coastalecosystems might be particularly relevant from asocio-economic perspective. A good candidate for apilot study would be the eastern boundary regions ofthe Pacific and Atlantic oceans where atmosphericforcing plays a dominant role in controlling keyecosystem processes such as coastal upwelling. Itcould start by investigating the local signature of themajor climatic signals in atmospheric forcing fields aswell as at the oceanic boundaries of the coastalregions. This can be done using historical time seriesas well as output from basin-scale atmospheric andoceanic models. A next step would be to implementhigher-resolution regional-scale coastal models overthe continental shelf. These models, forced at theirboundaries by atmospheric and oceanic signal mayprovide particularly efficient tools for exploring theresponse of the dynamics and structure of coastalecosystems to major basin-scale (or global) climaticsignals (Penven et al. 2001, Marchesiello et al. 2002).

Sardine regimes and mesoscale structure (an integrative hypothesis)[Alec MacCall]

An Hypothesis Explaining Biological Regimes inSardine-producing Pacific Boundary Current Systems(South America, North America and Japan):Implications of Alternating Modes of Slow, MeanderingFlow and Fast Linear Flow in the Offshore Region

Introduction and acknowledgementBecause this hypothesis is the product of a

workshop and draws on ideas contributed by a longlist of excellent scientists, it seems appropriate tobegin rather than end with some acknowledgments.The following hypothesis was developed while par-ticipating in the Boundary Current - Frontal SystemsWorking Group at the Pacific Climate and FisheriesWorkshop, held November 14-17, 2001 at the East-West Center, University of Hawaii. Especially notableis the contribution of Takashige Sugimoto, who orig-inally described most aspects of this hypothesis forthe Japanese ecosystem. The contribution I made wasto generalize and extend his hypothesis to the otherPacific boundary current systems. Don Olson pro-vided an oceanographer’s insight as to how flowconditions in these three systems might be linked.Other scientists who were not at the workshop (e.g.,Richard Parrish of the Pacific Fisheries EnvironmentalLaboratory, Pacific Grove, Paul Smith of theSouthwest Fisheries Science Center, La Jolla, andDaniel Lluch-Belda of the Centro Interdisciplinariode Ciencias Marinas, La Paz) have nonetheless con-tributed significantly to development of the hypothesis.Finally, a major acknowledgement goes to AndrewBakun, who not only convened this workshop, butalso provided the foundation for these ideas throughhis extensive work on the oceanography of pelagicfish reproduction.

15.1 Outline of the flow-based hypothesis

The major sardine-producing systems havetwo distinct pelagic habitats, a nearshore coastalhabitat and an offshore boundary current habitat.The characteristic long-term patterns of biologicalvariability in these systems are associated with inter-decadally alternating strong and weak modes of


15

boundary current flow and related reproduction-related physics of the nearshore and offshore habitats,and from habitat switching by the larger species (i.e.,sardines, but not anchovies) to utilize the offshorehabitat when conditions are favorable. Under condi-tions of slow, meandering boundary current flow,retention of eggs and larvae spawned in the offshorehabitat is greater than under conditions of fast,straight flow.

15.2 Fish behavior: habitat switching

Some species such as anchovies (Engraulis spp.)and juvenile coho salmon (smolts of Oncorhynchuskisutch) are restricted to the coastal habitat, probablybecause of their small size and limited swimmingability. Other generally larger species such as sardines(Sardinops spp.), mackerel (Scomber japonicus) andhorse mackerel (Trachurus spp.) are able to utilizeboth nearshore and offshore habitats, and switchbetween nearshore and offshore habitats according totheir relative favorability for feeding and reproduction.The evidence for habitat switching by populations ofthe larger pelagic species varies among geographicregions. During their recent increase of sardine abun-dance in the Japanese system, the population clearlyexpanded into far offshore regions associated withthe Kuroshio Current (Wada and Kashiwai 1989). Inthe California Current, the far offshore region hasnot been well sampled even by the CalCOFI ichthy-oplankton surveys, especially after the regime shift of1976. Importantly, sardine eggs were collected 200 to300 miles off California and Oregon in early ichthy-oplankton surveys conducted in 1931 and 1939 duringthe earlier period of high sardine productivity (Smith1990). In the early 1990s, the Russian trawlerNovodrutsk conducted exploratory fishing forTrachurus 200 miles off the California coast andfound unexpected abundances of Sardinops andScomber in those offshore waters (D. Abramenkoff,NMFS La Jolla, personal communication). Recentsardine egg surveys using the Continuous UnderwayFish Egg Sampler provide improved offshore moni-toring (Checkley et al. 2000), but frequently fail toreach the offshore edge of the distribution.

Bakun’s (1996) “fundamental triad” of enrich-ment, concentration and retention provides a frameworkfor evaluating the suitability of inshore and offshorehabitats. The role of enrichment in this hypothesis is

unresolved, and may vary among systems (see below),but the meandering patterns associated with weakerboundary current flow have clear implications forconcentration and retention. Of the elements of thetriad, variability in larval retention is by far the mostimportant aspect of this hypothesis.

15.3 Physical and biological oceanography

The offshore boundary current tends to exhibittwo alternative modes:

1. A fast transport mode, in which thecurrent is relatively straight (reducedmotion perpendicular to the mainflow).

2. A slow transport mode, in which thecurrent meanders (increased motionperpendicular to the main flow) andhas complicated structure, a relativelylarge frontal area, and greater tendencyto form persistent mesoscale eddies.

Note that temperature anomalies associatedwith flow strength are governed by the source watertemperature: The slow, meandering mode is associatedwith warming in the California and Peru Currentswhich have high latitude sources, but a slowing of theKuroshio system produces a cooling because of itstropical source. In the eastern Pacific systems, the slowmeandering mode characteristic of warm regimesmay also occur during El Niño events embeddedwithin cool regimes (e.g., the El Niño of 1957-58off California, during which sardines experiencedtwo years of good reproduction in a decade that other-wise showed consistent recruitment failures). It ishypothesized that with the exception of El Niño per-turbations, these boundary currents tend to stay inthe same flow mode for periods of many years, andthen switch suddenly to the opposite mode in a so-called “regime shift”.

Logerwell and Smith (2001) have shown thatoffshore mesoscale eddies in the California Currentare associated with significantly elevated densities ofsardine larvae. Logerwell et al. (2001) modeled thespatial bio-energetics of such an offshore eddy, andconcluded that they are likely to be a significant sourceof sardine recruitment. An important unanswered


question is whether the hypothesized decreased currentvelocity and increased meandering associated withthe slow-flow current is sufficient to improve reten-tion of sardine larvae, or whether the flow must formpersistent eddies with closed circulation and well-developed structure for sardine populations toincrease. In either case, the extent of meanderingand/or eddy formation is presumed to be enhancedduring periods of weaker boundary current flow.

Modes of boundary current flow exhibitregional variations that may explain regional differ-ences in nutrient patterns. Richard Parrish (NMFSPacific Fisheries Environmental Laboratory, personalcommunication) has shown that the post-1976 slowmode of the California Current drew nutrient-poorwater from the equatorward side of the North Pacifictransition zone, whereas the pre-1976 fast modedrew richer water from a higher latitude. Parrish’smechanism helps explain the post-1976 decline inplankton volumes seen in the California Current(Roemmich and McGowan 1995); a similar mecha-nism in the southern hemisphere could explain theparallel decline in plankton abundance off Perureported by Carrasco and Lozano (1989). Planktonvolumes should be interpreted with caution, as theyinclude gelatinous forms and do not necessarilyreflect trends in abundance of zooplankton thatwould serve as forage (Smith 1985). In contrast tothe eastern Pacific systems, the slow-flow mode of theKuroshio system may experience an enrichment dueto increased intrusion of nutrient-rich OyashioCurrent water, as happened during the mid-1980s,ending suddenly in1988 (Sugimoto et al. 2001).

Differences in nearshore oceanography associ-ated with these flow modes is less clear. In theEastern Pacific systems, coastal upwelling and nutri-ent enrichment may be somewhat stronger duringfast flow modes, but substantial coastal upwellingalso occurs during slow flow modes. However, thereduction in nutrients and the deeper thermoclinedepths associated with the warm, slow flow mode inthe Eastern Pacific systems may reduce nutrientenrichment associated with coastal upwelling underthose conditions. The convoluted island and shore-line topography of the Japanese coastal ecosystemcontrasts strongly with the linear, exposed coastlineof the Eastern Pacific systems, and may provide resi-dent fishes with good nearshore egg and larval reten-tion during periods of fast Kuroshio Current flow.

15.4 Synchrony and teleconnections

At the decadal scale, a general synchrony ofregime shifts in the various Pacific coastal boundarycurrents has been observed (Kawasaki 1983, 1989),but at finer time scales this synchrony is only approx-imate. Interdecadal variability in intensity of themajor North and South Pacific gyres appears to belinked through hemispheric and global atmosphericcirculation so that changes of intensity of both gyresare approximately in phase. In the recent record, themajor unexplained departure from synchrony is theapparent time lag of about one decade in theCalifornia Current relative to the Japanese andPeruvian systems (MacCall 1996).

Trachurus benefits from attaining a larger sizebefore individuals migrate to the cold and energeti-cally demanding offshore and high latitude waters. Adetailed demographic comparison of life-historiesand habitat preferences for these characteristicboundary current species would be useful. There maybe an opportunity for more detailed predictability ofabundance fluctuations in the full list of pelagicspecies, and not only sardines and anchovies.

15.5 Concluding thoughts

In previous efforts to understand “the regimeproblem” (e.g., Schwartzlose et al. 1999) we havebeen distracted by temperature relationships evidentin the eastern Pacific systems (i.e., warm conditionsfavor sardine productivity), and have tended to thinkof the Japanese system (in which cold conditions areassociated with sardine productivity) as being differ-ent, and requiring a different explanation. Byacknowledging that temperature anomalies are pri-marily the result of flow patterns, and that reproduc-tively important aspects of species ecology and lifehistory are more closely associated with properties ofthe flow itself, this hypothesis unifies our under-standing of the three major boundary current systemsin the Pacific.

We have also been distracted by habitat selec-tion theories such as the basin model (MacCall 1990)and the ideal free distribution selection (Wada andKashiwai 1989). These models suggested that theoffshore expansion of the sardine spawning grounds(considered as “effect”) is in response to the increasein abundance (considered as “cause”). This new


flow-based hypothesis reverses the interpretation ofcause and effect, so that the improvement of offshorespawning habitat (now the “cause”) leads to improvedreproduction and increased abundance (now the“effect”). However, the flow-based hypothesis is notexclusive of the previous habitat selection theories,and there may well be habitat selection effects withinthe nearshore and offshore systems. This is consistent

with the geographic expansion of sardine spawningwithin the southern California Bight during theincreasing abundances of the late 1980s (Barnes et al.1992, Smith 1990). MacCall’s basin model may bebest suited to the anchovy (for which it was originallydeveloped), the species that is least able to switchbetween nearshore and offshore habitats.


Mode of current flow Weak, slow flow Strong, fast flow

Coastal sea level Higher Lower

Water motion Enhanced meandering Reduced meandering

Frontal area Increased Decreased

Offshore larval retention Favorable Unfavorable

Temperature anomaly

Eastern Pacific Warm Cool

Japan Cool Warm

Nutrient supply

Eastern Pacific Reduced (lower lat. source) Enhanced (higher lat. source)

Japan Enhanced (Oyashio intrusion) Reduced

Sardine abundance Increased Decreased

Anchovy abundance

California Slight decrease Slight increase

Peru/Chile Strong decrease Strong increase

Japan Slight decrease Slight increase

Table 2. Comparison of Pacific Ocean coastal ecosystem properties under weak and strong modes of boundary current flow.

Climate and tuna fisheries 16

Tunas are particularly valuable fishery resources,supporting highly capitalized, technologically advancedfishing operations and important international trade(a single large bluefin tuna in prime condition forsashimi preparation may sell for US$40,000 or morein the Tokyo fish market). They are a diverse group offishes; for example, comparing the biology of frigateand bullet tunas (Auxis spp.) with that of Atlanticbluefin tuna (Thunnus thynnus) encompasses a broadr-k continuum. In recent years, the annual world catchof the four main tropical tuna species (skipjack, yel-lowfin, bigeye, albacore) approached 4 million tones,with two thirds of the catch coming from the Pacific(Lehodey 2002).

ENSO fluctuations appear to have majoreffects on Pacific tuna populations, both on distribu-tions and migrations (Lehodey et al. 1997) and onrecruitment strength and population abundance(Fournier et al. 1998). The effect of El Niño episodesappears to be positive on skipjack recruitment andnegative on albacore recruitment, resulting in a neg-ative correlation in recruitment between these twospecies. This is reflected in an interdecadal (regime-scale) out-of-phase relationship between the twospecies, with albacore recruitment appearing todecline to a lower average level after the 1976 climaticshift (see Section 9) to a more El Niño-dominatedphase of the Pacific Decadal Oscillation, while skip-jack recruitment generally increased (Lehodey 2002).

Regime-type fluctuations also seem to occur intuna populations. For example, it appears that theyellowfin population has experienced two differentrecruitment regimes (1975-1984 and 1985-1999),the second being higher than the first. The tworecruitment regimes correspond to two regimes inbiomass, the higher recruitment regime producinggreater biomasses.

A summary of current needs in tuna researchis presented in the document “Research Implicationsof Adopting the Precautionary Approach toManagement of Tuna Fisheries” (FAO 2001). Quotingfrom the executive summary of that document:“There is a growing acknowledgement worldwide that

abiotic and biotic environmental changes significantlyaffect the distributions, and perhaps also the productivity,of various tunas. Thus it is important to determine thenature and extent of the impact of climate variabilityupon the pelagic ecosystems and the tuna stocks. Thisnatural variability should be taken into account as anadditional source of uncertainty in stock assessmentand management.”

The discussions in the tuna focus group at theHonolulu workshop revealed a number of salientpoints concerning climate-fisheries research issues.Clearly, there is a need to learn more about the habitatpreferences of tunas, (i.e., more research is needed onthe behavior and physiology of individual fish). Arehabitat preferences stationary or plastic? The effectsof climate change on tunas depend on the fact thattunas are mobile. Climate effects are, therefore, morecomplex than for sedentary fishes. The environmentand the status of the population mitigate rates, timing,and routes of migration or movement. For example,albacore migrate from the west to the east followingfronts. Usage of the fronts may depend on the pro-ductivity of the fronts. Different age-classes may havedifferent migration behaviors.

Temperature and forage are considered to beimportant determinants of tuna movements. Notmuch is known about the relationships betweentunas and their forage as a whole, and the effect ofclimate on trophic processes such as ecological transferfrom primary production to the middle trophic levelsis an important topic of research. The lengths of foodchains leading to tunas probably vary over largeregions (e.g., western vs. eastern Pacific). Data onvariability trends in the forage base are not available.Physics plays a role in concentrating prey and makingit available to the predators (for example, the concen-tration of prey in the upper mixed layer by a shoalingthermocline).

It is important that a clear distinction be madebetween catchability and availability of tunas, and theeffects of physical processes on catchability. In thewestern Pacific, changes in the vertical thermal struc-ture associated with ENSO fluctuations (i.e., shallowerthermocline during El Niño events) seem to have onlya minor influence on the skipjack catchability but


16

16 Bob Olson and Patrick Lehodey made important contributions to the preparation of this section.

increase the catchability of yellowfin by surface gear(pole and line and purse seine). However, for tunalonglining in the tropical Pacific the MLD is knownto change the catchability of fishing gear (Lu et al.1998, Lu et al. 2001).

Situations exist which may lend themselves totesting hypotheses regarding rapidly evolving mecha-nisms. For example, sometimes when the fish aresmall schools of tunas are composed of two or morespecies. Since different species grow at different rates,these fish will be of slightly different ages whenschooling together. Yet at that particular ontogeneticstage they are sharing a common behavior by schoolingtogether. This aspect of tuna schooling behaviormight be used to design experiments to test hypotheses.

Although spawning by yellowfin tuna in theeastern Pacific appears to be quite widespread(Schaefer 1998), rather specific spawning sites arethought to exist for this species in the Atlantic (anexample is the Gulf of Guinea in the first quarter ofeach year). This site fidelity may provide a source tostudy school-mix feedback mechanisms reminiscentof the albacore example provided by Bakun (2001).Tagging studies and size composition analysis of thecatch might be useful tools for this purpose.

Floating objects and seamounts are known toaggregate tunas. In the eastern Pacific, prior to 1982the only floating objects utilized by purse seine vesselswere natural logs. After 1982, artificial fish aggregatingdevices (FADs) began to be used by the purse-seinefishery to attract tunas, markedly increasing the catch-ability of smaller yellowfin and bigeye. Thus, a largerproportion of the total catch originated from floating-object sets after 1982 than before 1982. There is apossibility that this aspect of the fishery might havealtered the yellowfin and bigeye populations byremoving fish that may have been predisposed toaggregating at floating objects versus fish that mightbe more likely to school in unassociated or dolphin-associated schools. It was suggested at the workshopthat fisheries data stratified before and after 1982might provide a tool to elucidate rapidly-evolvingadaptations in yellowfin and bigeye populations relatedto changes in fisheries selectivity.

There is a need to use fisheries-independentsampling platforms to study tunas and to testhypotheses generated by spatial models (e.g.,Lehodey 2001). Research is constrained by data gen-erated by the commercial fisheries alone. Researchers

generally have access to data about the fish only inthe locations where the fisheries are operating; nothingis known about the growth rates (for example) oftunas in other areas. Some regions are predicted bymodels to contain good habitat for certain tunas, butthis cannot be validated because the fisheries may notoperate in those areas during those seasons. Fisheries-independent sampling platforms might includeexperimental longline vessels and dedicated purse-seine vessels.

An important problem is the fact that ecosys-tem models adapted to an ecosystem-managementapproach are still at an early stage of development.Such an approach implies the integration of spatio-temporal and multi-population dynamics and theconsideration of interactions among populations ofdifferent species and between populations and theirphysical and biological environment. The modelingof such a complex system is certainly a challenge,requiring combining two viewpoints which are mostoften pursued separately by separate groups of scien-tists (i.e., population dynamicists focused on tempo-ral variations in population abundance and fisheriesoceanographers/ecologists who have generally, in thecase of tunas, centered on species distribution).

Due to the extended distributions and migra-tional habits and capabilities of large oceanic pelagicfishes, there have in the past been problems in con-fidently knowing even if the populations are varyingat all, much less knowing how much they are vary-ing, and even less knowing why they may be varying.Incorporation of statistical functions, i.e. likelihoodfunctions, in population dynamics models withmultiple types and sources of data (catch-effort,length frequencies, tagging) now produce much bet-ter estimates of population abundance indices(recruitment, biomass). The reliability of these seriesis believed to be increasing, while at the same timesuggesting larger fluctuations that what is predictedfrom classical virtual population (VPA) analysis, andblind tests show that these models can predict realtrends within a range of error less than 10-20%.Although there is no way to produce direct confir-mation of these estimates, the correlation betweenfluctuation of abundance indices with climateindices constitutes an indirect validation. Anotherindirect validation is the correspondence of biomassand recruitment estimates based on statistical modelsto those derived from models (such as SEPODYM,


Lehodey 2001) that are constrained only by environ-mental variables.

The design, rationale, and preliminary findingsassociated with the GLOBEC-OFCCP project (seeSection 17, below), were presented at the workshop inplenary session by Patrick Lehodey. This presentationexcited much interest and favorable comment. Earlyresults of an integrated physical-ecosystem model inwhich forage dynamics are simulated in an underlyingadvective-diffusive ocean model linked to phytoplank-ton, zooplankton, forage, and tuna components arepromising. The model has been successful in explaining~50% of the variance in spatially-disaggregated catchdistribution (monthly one-degree squares) of Pacificskipjack catch between 1972 and 1992 (Lehodey2001b). Specifically, it successfully models interannualvariability associated with El Niño events, as well as alower frequency shift at about 1976 (see Section 9,above) toward lower albacore and higher skipjackcatches. The sources of the modeled variability areENSO-related effects on physical factors (advectionand temperature) and biological productivity (primaryand secondary production) in two spawning regions.

This work stimulated discussion about thedesirability of including major modeling efforts toaddress the primary issues not only concerning tunabut in terms of Pacific climate and fisheries in general.It also brought out the fact that there are some inter-esting unexplained issues, such as why yellowfin andskipjack tuna populations seem to have fluctuated inopposite phase with respect to El Niño episodes, thatmay offer potentially productive entry points into thegeneral tuna/climate research problem. It also madeit clear that the comparative method, applied interre-gionally on a global basis, might be a particularly eco-nomical and efficient way to acquire real insights intokey aspects of the operation of the climate/ecosystem/tuna systems.

This led to the suggestion for a global“umbrella”-type project directed at large oceanicpelagic fish resources, which would operate somewhatanalogously to the GLOBEC SPACC project focus-ing on small pelagic fish resources of ocean boundaryregions and peripheral seas and which likewise reliesheavily on the comparative approach. This project,code-named “CLIOTOP” (see Section 18, below), ofwhich the OFCCP project in the tropical Pacificwould be a cornerstone, would ideally be developedwithin the International GLOBEC context.

OFCCP – A proposal for a GLOBEC Pacific regional project [Patrick Lehodey]

The Oceanic Fisheries and Climate Changeproject (OFCCP GLOBEC) has been designed tosystematically investigate the potential effects of cli-mate change on the productivity and distribution ofoceanic tuna stocks and fisheries in the Pacific Oceanwith the goal of predicting short- to long-termchanges and impacts related to climate variability andglobal warming.

During the last decade, considerable progresshas been achieved in the development of physical andbiogeochemical 3D numerical ocean models.Biologists have also developed a diverse range ofmodeling approaches that reflects both the complex-ity of marine ecosystems and interests in the studyand exploitation of the ocean. Studies of physiologyand behaviour of higher pelagic predators like tunaor salmon led to elaboration of energetics and indi-vidual-based (IBM) models (Dagorn and Freon1999, Kirby et al. 2000). The science of fisheriesstock assessment has shown significant progress inthe recent years; for example, the integration of sta-tistical functions allowing extraction from multiplesources of data the best estimates for populationdynamics (Fournier et al. 1998). At larger scale, thecommunity and its food web appear increasinglycomplex. Their mathematical descriptions requireseveral levels of simplification, for example usingfunctional groups of species or using a size spectrumin which larger organisms feed on smaller ones.Combinations of these different approaches may alsobe possible. For instance, the model SEPODYM(Bertignac et al. 1998, Lehodey et al. 1998, Lehodey2001b) combines a production model for the foragefunctional group with an age-structured populationmodel of tuna predator species and their multi-fisheries. The dynamics of forage and predators areconstrained only by environmental variables (tem-perature, oxygen, oceanic currents, and primaryproduction) that can be predicted from coupledphysical-biogeochemical models. Therefore, it nowseems possible, using these different modelingapproaches, to explore the underlying mechanismsby which climate–induced environmental variabilityaffects the pelagic ecosystem and tuna populations.


17

17.1 Project description

The ultimate goal of the project is to conductsimulations with ecosystem models that include themain tuna species, using an input data set predictedunder a scenario of climate change induced by green-house warming as defined by the IntergovernmentalPanel on Climate Change (IPCC). This should leadto the first tentative understanding of how green-house warming will affect, at the ocean and globalscales, the abundance and productivity of marinepopulations in the pelagic ecosystem, focusing on themajor exploited species and fisheries, by a real cou-pling between atmospheric, oceanic, chemical andbiological processes. Potential feedbacks from thechanges in the pelagic ecosystem, and socio-econom-ical consequences will be investigated to proposeadaptation measures for the future. However, analysesof simulations based on retrospective series of oceanicand fishing data sets are necessary intermediate stepsto increase the reliability in the predictive capacity ofthe models. In particular, realistic prediction by themodels of changes and fluctuations observed at short(e.g., ENSO) and decadal (e.g., Pacific DecadalOscillation, PDO) time scales in the ocean ecosystemand the tuna populations are necessary before predic-tion based on the global warming projection areincorporated. In addition, diverse studies are neededto improve the parameterization (e.g., energy transferfrom primary to secondary production), the modelingof key processes (e.g., recruitment, movements, andfeeding), and to validate the results of the simula-tions. Four major components have been identifiedto achieve these objectives.

17.2 Monitoring the upper trophic levels of the pelagic ecosystem

The past decade has generated significantprogress in understanding ocean processes and thecoupling of ocean and atmosphere in regulatingEarth’s climate. These accomplishments were madepossible by the concurrent development of new tech-nology and instrumentation, as well as substantialprogress in numerical modeling (ocean general circu-lation models and new conceptual models of lowertrophic level food webs). Despite the increase com-plexity when considering all the pelagic ecosystem, asimilar approach, closely associating observation and

modeling, seems the most appropriate to investigatethe dynamics of upper trophic levels (frommacroplankton to higher predators). However, whilethere has been substantial progress in acoustical tech-nology or individual electronic tracking devices, theinstrumentation allowing large-scale and long-termrecording of key upper trophic components is stillmissing. One of these key components is themicronekton for which there is relatively little infor-mation. Comparatively, there is much more infor-mation on large pelagics (the predators of themicronekton) that are usually valuable exploitedspecies. The fisheries for these species provide keyinformation (catch, size) allowing population dynam-ics models to predict the population biomass, andeventually their spatial distribution, especially wherelarge-scale tagging programs have been carried out forthese valuable species. Therefore, while climatechange concerns as well as recent ecosystem-basedmanagement requirements necessitate rapid develop-ment of numerical ecosystem models, we have only avery preliminary idea of the biology, ecology anddynamics of the intermediate key components of thepelagic food web. Indeed, there are not even enoughobservations to produce a mean spatial distribution ofthe micronekton biomass at the scale of ocean basins.

It is proposed in the present project to useexisting technologies, and also to develop new instru-mentation for monitoring the upper trophic levels ofthe pelagic ecosystem. Observation will combine bothextensive studies at the ocean basin-scale and intensivestudies in some sub-areas and key sites. Extensivestudies aim at building ocean data sets for micronektonbiomass and large pelagics biomass or individualrecords, using acoustic (micronekton biomass), sonar(tuna biomass), and electronic tracking (individuals)devices. Intensive studies will focus on importantprocesses and behavior (e.g., prey-predator interac-tion, habitat, schooling and aggregation of tunas,reproduction, composition and dynamics ofmicronekton, etc). In addition, at each scale of obser-vation there will be a corresponding modeling devel-opment, e.g., large-scale ecosystem, populationmodels or individual-based models. Observationswill be used to parameterize and improve the models.Eventually, models could provide real-time predic-tion for operational activities at sea, and help in thevalidation of dynamic processes or hypotheses arisingfrom model simulations.


17.3 Food web structure in pelagic ecosystems

Production at higher trophic levels (usuallyexploited species) depends on the production at lowerlevels (bottom-up control) and may be modulated bythe physical forcing and the structure of the marinefood webs. Ecological concepts suggest for instancethat the structure of the food web can be controlledby the biodiversity within the system and/or by high-er predators (top-down control). However, concern-ing pelagic ecosystems, there is very little observationto illustrate such controls. In association with the datacollected by the monitoring component of this proj-ect, it is essential for modeling the pelagic ecosystemto identify the functional groups, how energy andmatter flow through these groups and how they areaffected by physical and biological changes as well asby human activities (fisheries).

Two kinds of analyses will be helpful in thistask: a classical approach based on the study of stomachcontents to establish the prey-predator interactions,and the more recent isotope-ratio approach, thatappears a promising way for describing the energytransfer through the food web (Rau et al. 1983, Fry1988). The success of these approaches also relies onthe multiplicity of studies in different regions of theocean(s) and in different periods of time. The com-parative study necessitates developing standardizedprotocols, reference databases and controlled labora-tory experiments. Retrospective analyses based on thenumerous diet studies published or still in archives ofmany institutes should be also encouraged.Information obtained from these studies and fromthe monitoring will be used in individual energeticsmodels (IBM), mass-balance models (ECOPATH-ECOSIM) and spatial ecosystem models(SEPODYM).

17.4 Modeling from ocean basin to individual scale

Close association between observation andmodeling has been a permanent guide in conceptual-ization of this project. Recognizing the diversity ofspace-time scales processes overlapping in pelagicecosystem dynamics, a second key idea is that a generalframework is needed to integrate studies at differenttime and space scales with potential connections

between them. Therefore, at each scale the modelsand sources of data collected in the studies of the firsttwo components are indicated. There is a large rangeof models represented in the project covering global toindividual scales. At global or basin scales, predictionsfrom three different coupled physical-biogeochemicalmodels will be used over the period 1950-present.The global model will also provide predictions forthe next century using a scenario of greenhousewarming. These predictions will be used to run theecosystem models of upper trophic levels on whichthe economical and social analyses rely. At least oneof the physical-biogeochemical models should pro-vide prediction at high resolution in one or a fewidentified sub-regions (first step for a nested modelapproach) where intensive process studies are con-ducted. A similar approach will be investigated forthe spatial ecosystem models. This would allow con-nections between large and small scales (low and highfrequencies) processes and testing the mechanismsthat control the system when moving from one scale(frequency) to the other.

17.5 Socio-economical impacts

The interannual climate variability due toENSO events has important socio-economic impactson the tuna fishery and the industry at the globalscale that in turn may affect the tuna populations(e.g., higher/ lower catch) and the pelagic ecosystem(by-catch, interaction between species, top-downeffects). Several causes drive the fluctuations of tunastocks and catches. While economic rather than bio-logical reasons limit (today) the catch increase of themost productive tuna species (skipjack) in the Pacific,the intense fishing effort on the highly valuablebluefin tuna, perhaps combined with environmentalforcing, has led to a decline in this population fromthe 1960s to the 1980s. Interactions amongst speciesand between the multiple and diverse fisheries, aswell as potential cascade effects in the ecosystem raiseimportant questions for management with potentialstrong socio-economic repercussions.

Based on an existing spatial model (Chakravorty1997, Chakravorty and Sibert 1999, Chakravortyand Nemoto 2001) developed for investigating opti-mal spatial allocation of fishing effort, new studieswill include climate variability to obtain informationon alternative scenarios regarding the interannual as


well as spatial variability of fish stocks. A key issuewill be in classifying alternative scenarios of climatechange that could be translated into the spatial andtemporal distribution of fish species, as well as theirmovement. Other issues that will need to be modeledinclude multiple fleets with different efficiency char-acteristics, and the presence of fishing vessels frommultiple political jurisdictions, that will imply differ-ent costs of fishing through differences in materialand labor costs. The presence of multiple countrieswill allow the use of strategic behavioral models. Forinstance, Nash Equilibria could be calculated forregions that are subject to fishing by multiple coun-tries. These results could be compared to optimaloutcomes under single ownership of the fishery andthe impacts of feasible management measures couldbe simulated.

Investigations of these interactions and effectsoccurring with ENSO would help to assess the vulner-ability and impacts in a scenario of global warming,and to eventually propose adaptations and/or miti-gation measures for the future.

CLIOTOP — A proposed global comparative research effort on large oceanic pelagic fisheries [Olivier Maury]

In the current context of very strong fishingpressure, climatic variability and potential climatechange, understanding, modeling and eventually fore-casting the dynamics of open ocean ecosystems associ-ated with apex predators is becoming crucially impor-tant. Adequately responding to this major challengewill require major improvements of our knowledgeconcerning the processes involved in offshore ecosys-tem dynamics at various scales and complexity levels.

The Honolulu Climate-Fisheries Workshopprovided very stimulating prospective discussionson these issues. One important outcome is the ideathat time has come to organize on a global scale acomparative study of open sea pelagic ecosystem/climate coupling to identify interaction processes.An international cooperative research effort, code-named the CLIOTOP (CLimate Impacts on openOcean TOp Predators) project, was proposed.

Indeed, there is now a remarkable correspon-dence between the scales at which oceanographers,climatologists and biogeochemists are developingstate-of-the-art physical (OGCM, ect.) and biological/ecosystem (e.g., “nutrient-phytoplankton-zooplankton(NPZ)) models and the basin scales at which openocean apex predator populations may operate. Butdue to the large extent and complexity of the consid-ered ecosystems, adequate observational and experi-mental approaches are very difficult to implementand most often must be restrained to local regionalphenomena and special cases. As a consequence, ourability to analyze and understand observed phenom-enon in terms of causal processes and mechanisms isseriously compromised.

The process of generalizing those availablelocal observations and findings may be facilitated bythe large-scale comparative approach (see Section 14,above) that would be the core of the CLIOTOPproject framework. Such an international projectwould facilitate the sharing of ideas and informationamong a large number of independent research proj-ects pursuing GLOBEC-type approaches involvingobservation and modeling in a variety of open oceanregions and over a range of species. This could pro-vide the basis for informative applications of thecomparative method, which would benefit all thecollaborating parties (i.e., the whole, in such a case,being much more than the sum of the parts).

It is clear that such sharing of information andexperience among regional programs has yieldedgreat dividends at relatively minor cost in the pro-gression of international comparative research projectsaddressing small pelagic fish systems that hasoccurred over the past 15 years (i.e., first the IOC-FAO Sardine-Anchovy Recruitment Project (SARP),then the NOAA-IRD-ICLARM Climate andEastern Ocean Systems (CEOS) Project, and finallythe current GLOBEC Small Pelagics and ClimateChange (SPACC) Project). The experience andinsights gained by these projects could probablyallow us even to shorten the process of setting up anefficient “open sea” project.

In analogy to SPACC, the proposedCLIOTOP program could be an "umbrella”-typeprogram covering various independent and potentiallycollaborating regional projects17 in the three oceans18.There is indeed a very large, multi-national commu-nity of marine ecologists, fishery scientists, climate


18

scientists and oceanographers that is highly interestedin such a large-scale “open sea” project. TheCLIOTOP project would serve to promote and facil-itate information flow and exchange of ideas amongvarious projects and institutions throughout theworld. It would organize meetings for comparativeanalysis and informative pattern recognition andwould facilitate publication of reports focused on thisproblem area. It would be organized around scientificquestions focusing on processes and mechanisms, theultimate goal being to improve our predictive capa-bilities concerning the operation of open oceanecosystems in a global change context.

Studying the dynamics of large pelagic ecosys-tems in such a climatic change perspective involvesseveral fields ranging from climate modeling andoceanography to population biology and ecology,fishery science and socio-economics. Given thecomplex nature of its foci, an “open sea” GLOBECprogram should strongly encourage the co-operationand exchanges with other IGBP programs such asSOLAS, GAIM and JGOFS as well as WorldClimate Research Project (WCRP) programs such asCLIVAR. Being able to make use of the tools andexpertise provided by those international programs isa crucial need for such a future “open sea” project.

A collective intent letter has been written tothe GLOBEC Scientific Steering Committee andsigned by a number of individual scientists involvedin the discussions which have followed the meeting(Dr. Juergen Alheit, Baltic Sea Research Institute,Germany; Dr. Andrew Bakun, IRI, USA; Dr. RobertCowen, RSMAS/MBF, USA; Dr. Jean-MarcFromentin, IFREMER, France; Dr. Patrick Lehodey,SPC, New Caledonia; Dr. Olivier Maury, IRD,Seychelles; Dr. Arthur J. Miller, Scripps Institution of

Oceanography, USA; Dr. Raghu Murtugudde,ESSIC, USA; Dr. Ian Perry, Fisheries and Oceans,Canada; Dr. Jeffrey Polovina, NMFS Hawaii, USA;Dr. Claude Roy, IRD, South Africa). Its purpose is topropose to the GLOBEC-SSC the adoption ofCLIOTOP as new International GLOBEC project.

Potential utility of a multilateral comparative retrospective data analysis effort 19

Climate is often implicated as the driving forcein many of the large-scale shifts in marine ecosystems.Evidence for climate-induced ecosystem changes ondecadal scales has been accumulating. This evidencehas been gained mostly from local retrospective studiesand, where a comparative approach among a numberof similar systems has been possible, the results haveproven quite significant (e.g. the optimal environmentalwindow result discussed in Sections 10.1 and 13.2).

However, there always has been a serious prob-lem of spurious correlations in fisheries-environmentalscience. There are several reasons for this20:

1. The lack of abundant degrees of free-dom available in time series (availabledata series tend to be too short tocontain sufficient realizations of theimportant variations of interest);

2. The pertinent data series also tend tocontain substantial autocorrelation


19

17 Proposed initial CLIOTOP pilot projects:• the OFCCP project (Oceanic Fisheries and Climate Change in the Pacific; contact: P. Lehodey, SPC, Nouméa)• the RSMAS billfish project on billfish early life history (contact: R.K. Cowen, CSF, Miami)• the STROMBOLI/MERCATOR Atlantic bluefin tuna project (contact J.M. Fromentin, Ifremer, Sète).• the THETIS project (Thons tropicaux: Environnement, sTratégies d’exploitation et Interactions biotiques dans les écoSystèmazes

hauturiers; contact: F. Marsac, IRD la Réunion; O. Maury, IRD Seychelles),• the TUNABAL project (Bluefin Tuna Eggs and Larval Study in the Azores ; contact ; A. Garcia, COM, Malaga).

18 As in the case of SPACC, the CLIOTOP project would rely mostly on interregional applications of the comparative method, whereas its member projects might be involved with actual field operations, data acquisition or modelling exercises. At this very preliminary stage, five pilot research projects have already been identified and could potentially be involved in CLIOTOP.

19 Ian Perry made important contributions to the preparation of this section.20 For further elaboration and examples, see Section 11.1 in Bakun 1996.

(which further reduces degrees offreedom);

3. An inability to realistically estimatesignificance levels of an empiricalfinding in a situation where we almostnever are made aware of the “failures”,only of the “successes”.

Item 3 may be the most problematic. Forexample, (case a) if a single researcher were to searchthrough twenty different combinations of data series,and found one that met a significance criterion ofp<0.05, this is clearly no more than what he couldexpect to find merely by chance in twenty such trieswith different series of random numbers. If he thenpublished this relationship as a significant finding,we might consider that individual to be either naiveor cynically dishonest. However, if (case b) each oftwenty different researchers (each entirely independ-ently of one another, each basing the test on a logicala priori hypothesis, etc.) tested a different one of thesame group of data sets, and one of them found thesame significant relationship, which he then pub-lished, we would certainly consider this to be proper.However, since the others who did not find relation-ships would not publish that fact (no journal welcomespapers reporting studies that did not yield significantresults), the real result of case (b) is identical to thatof case (a). Thus, unless we know the potential distri-bution from which an empirical result is drawn, it isimpossible to judge its true significance.

Thus it may be time to take the process out ofthe hands of individuals and, as a community, simplyassemble the largest available number of potentiallypertinent data series from a variety of regional envi-ronmental settings, species types, etc., assemble theminto a large data base, and simply “grind out” theinter-series data structures with shear “raw” moderncomputer power. Then we would have the distributionsof successes and failures so that we can realisticallyjudge the true significance of our findings.

The important point is that one might hope touse variations in climate, in particular as manifestedin different regions of the world’s oceans, as naturalexperiments to understand how these systems respondto climate variability. Moreover, in addition to con-ventional tests for linear predictors, etc., one mightbe able to perform informative meta analyses (tests on

distributions of tests). An example (Bakun 2002b) ofa potential test for the existence of life-cycle lengthdependent adaptive response mechanisms (such asschool-mix feedback, see Section 11), was distributedto workshop participants prior to the meeting.

Discussion at the workshop, and ultimatelyconsensus, centered about the utility of conductingthese types of comparisons solely on a global basis.Much useful information and understanding can begained by conducting these analyses on regionalscales, as well as on global scales. Much of the imme-diate forcing to fish populations occurs through localprocesses, for example upwelling, and the local man-ifestations of these forcing processes may differamong locations. Such local instances of forcing maybe connected over very large scales by “teleconnec-tions”, through which the characteristics of the localprocess may be altered or subsumed by the telecon-nection process. There is, however, a great need to doboth types of analyses: comparisons at the local/regional level, and on a global basis.

In addition to purely statistical comparisons offish populations and climate variables, there is a needto develop a model-based approach. Using models,such as coupled physical-biological models, to definea framework for the analyses would help increaseunderstanding of the processes or mechanismsunderlying the statistical relationships. Models permittools from different disciplines to be combined, andhelp to define and to improve the questions beingasked of the data. Models also enable more sophisti-cated questions to be asked, and tested, about thelinkages between climate and fisheries than is possi-ble with purely statistically-based analyses. Modelspermit the incorporation of additional data (e.g., theNCEP reanalyses of atmospheric data, into the com-parisons). Model output might also be considered asdata for input into the fisheries comparisons, e.g. asfor difficult-to-observe variables such as sub-surfacewater properties, or primary production. Models canhelp to alleviate problems with statistical significanceby narrowing the range of explanatory variables usedin the climate – fisheries comparisons.

Another approach, which can be thought of asintermediate between the statistical and modelingapproaches and between the local and global compar-ison scales, is to look for coherent regional-scaleclimate structures or perturbations, and look to see ifthere are relationships among fish populations to


these regional physical systems. An example is theapparent coherent physical manifestation of the cli-mate system in the North Pacific, that is reflected inthe Pacific Decadal Oscillation Index (PDO)(Mantua et al. 1997). Finding relationships andbuilding understanding of the mechanisms linkingsuch regional-scale climate features to responses byfish populations (and their marine ecosystems) maybe easier because connections at these scales may havethe least amount of “filtering” compared to widely-dispersed systems which are connected through indi-rect teleconnections.

Teleconnections, however, are potentially stillimportant, and they are important to understand inthe context of responses of the *MRS to globalchanges. Most progress has perhaps been made onthe study of global teleconnections of fluctuations ofsmall pelagic fishes (e.g. Schwarztlose et al. 1999).Whether small pelagic fishes are more susceptible toglobal climate changes because they live pelagicallyand have short life-spans, or whether they are theonly group that has been investigated in a coherentmanner, is unclear. Other species groups need to beinvestigated for global synchrony, and comparedwith the patterns of small pelagic fishes to under-stand the potentially different responses by specieswith different life-history strategies.

In summary, there is an urgent need for under-standing climate–fish–fisheries connections by mak-ing comparisons between populations within species,and between different species groups, at spatial scaleslarger than local. These comparisons ultimately needto be done on a global scale, but valuable under-standing can be gained by working from regional toglobal scales. Regional analyses should strive toidentify coherent structures or processes in the phys-ical climate system to which fish populations mayrespond, and then look for differences in reactionsamong these populations or species to similar forcingevents. At global scales, analyses need to look for large-scale coherent responses that may be synchronized byas-yet poorly understood teleconnection processes.For both scales of analyses, a combination of statisti-cal and model-based approaches are recommended toboth provide new data sets and test mechanisticunderstanding. The analysis may require somedegree of coordination and/or central planning sothat numbers of significant and non-significant testscan be recorded, and relationships that might not be

significant individually but which may be significantin aggregate can be recognized. “Knotty” issues suchas the proprietary nature of data and the mechanicsof (and incentives for) making data accessible to thebroad community would need to be addressed.

Proposed formulation of a Pacificregional climate-fisheries project

The Pacific Ocean appears to offer severaldistinct advantages for studying climate-scale effectson fisheries and fishery resource populations:

1. Interyear variability dominates in the Pacificregional ecosystems to a much greater degree than inthe systems of other oceans and seas. Moreover, thefirst-order mechanisms behind the largest compo-nent of this variability, ENSO, are relatively wellunderstood;

2. There appears also to exist in the Pacific astrong component of rather coherent interdecadalvariability (the Pacific Decadal Oscillation), at leastin the northern part of the ocean basin;

3. The fish communities on the two sides ofthe Pacific, and in both the northern and southernhemispheres, seem to be composed of similar speciescomplexes (for example, in strong contrast to the sit-uation in the much smaller Atlantic Ocean, a singlespecies of sardine, Sardinops sagax, appears to be adominant component in all four widely-distantquadrants of the Pacific: Californian, Peru-Humboldt, Japanese and Australian.) This wouldappear to facilitate effective application of the com-parative method.

4. The fact that many of the most importantfishery resource stocks appear to be varying in adegree of synchrony on interdecadal time scales indi-cates that much of the potential feedback complexityin trophic and other ecosystem linkages must beoperating somewhat in the background and that thedominant mechanisms must be acting rather simplyand directly on the resource populations themselvesand be driven by basin-scale climatic teleconnections;


20

5. There are available some prominent enig-mas that appear to defy understanding according tothe conventional conceptual framework. Thesemight turn out to be extremely useful entry points tothe process of identifying and addressing the missingpieces in our current understanding. These enig-matic issues include:

(a)Why is the Peru-Humboldt Currentecosystem able to produce so muchgreater tonnage of fish in comparisonto other similar eastern ocean upwellingsystems?

(b)Why do sardines, which are a speciesobviously adapted to highly produc-tive ocean conditions (upwelling areas,etc.) appear to do better, at least in theeastern Pacific, during El Niño episodessince these are characterized by abruptlylowered primary productivity?

(c)How is it that fisheries can apparentlyfluctuate in phase even though theyexist in very different types of ecosys-tems that appear to respond very dif-ferently to common large-scale forcing?

6. The Pacific Ocean’s great distances and associ-ate separations of regions, markets, cultural traditions,states of economic and technological development,etc., may render the Pacific regions particularlyamenable to applications of the comparative methodto socio-economic aspects of climate-fisheries researchand research applications.

Accordingly, in addition to the proposed newOFCCP and CLIOTOP efforts to be focused partic-ularly on tunas and other large oceanic pelagic fishes,it was proposed at the workshop in Honolulu that aregional project focusing more broadly on variousresource stocks, important ecosystem components,and fisheries of the Pacific region might offer anavenue to some substantial progress within the generalclimate-fisheries research area. This idea attractedconsiderable interest and discussion.

It appears that a first step might be to under-take a full assessment of the degree of climate/ oceansynchrony in the Pacific Ocean, including whether

anomalies of a nature that might affect fisheries inthe southern and northern hemispheres are in or outof phase. One of the challenges may be to developeffective proxies for climatic data in the SouthPacific, which has traditionally been poorly observed.

Among hypotheses that have been developed are:

1. Bakun’s (2001, 2002a) rapid adaptive/behavioral responses (e.g. school - mixfeedback) model (See Section 11)

2. Cushing's (1971, 1975, 1990) match/mismatch hypothesis, e.g., peaks ofabundance of fish larvae and theirprey (zooplankton) in phase or not(See Section 10.5.)

3. Lehodey’s advection and larval food/predator ratio model which combinesfood web forcing with spatial andtemporal dynamics (See Section 17)

4. MacCall’s advection / retention e.g.mesoscale variability model (SeeSection 15)

5. Gargett’s (1997) Optimal StabilityWindow hypothesis that relatesPacific salmon survival to regime-likevariations in water column stability.

6. The possibility that shifts in food par-ticle size spectra, perhaps in responseto changes in relative importance ofthe “microbial loop”, might shiftfavorability between filter-feedingspecies with different gillraker meshsize (sardines having much finer filter-ing capabilities than do anchovies(Louw et al. 1998)).

Recognizing that there were too many partici-pants at the workshop to effectively propose andadopt, in the limited time available, an appropriateframework for such a major Pacific-scale project, theworkshop participants agreed to request that PICES,IRI and IPRC consult among one another on develop-ment of such a framework. It was suggested that thePICES North Pacific Transitional Areas Symposium,to be held in La Paz, Mexico, on 23-25 April, 2002,might be a possible venue for such a consultation.


Useful application of current state of scientific understanding

Many workshop participants emphasized theneed to achieve a balance in approaches to acquiringand applying understanding of climate-fisheries inter-actions, with an important first step being improvedcommunication between the fisheries and climateresearch communities. Over the last few years, therehas been much interest from the climate communityin demonstrating the predictive capability gleanedfrom the last few decades of work. Simultaneously thebiological community is moving away from forecast-ing of single species dynamics toward understandingmultiple species interaction through ecosystemmodels. Such fisheries research will necessitate closeinteraction with the climate community on projectsinvolving hindcasting, modeling studies, and reanalysisstudies within the biological community.

Participants were also keenly aware that multi-ple economic, political, cultural, and environmentalfactors interact to affect the sustainability of fisheries.Better understanding of climate’s role on livingmarine resources will not be the “silver bullet” thatreverses the many years of overcapitalization and lackof enforcement that characterize global fisheries.However the group was enthusiastic about the poten-tial for the use of various sorts of climate informationto better inform decisions, ranging from the negotia-tion of multilateral treaties to the private sector’slong-term investment strategies to optimal design ofobservational programs, that may help to alleviatesome of these chronic problems. As with all sorts ofscientific information and technological break-throughs, care must be taken to pay proper attentionto issues of communication of this knowledge,involving stakeholders in the process from the start.

The science, of course, will never be perfectand one challenge addressed at the workshop was toconsider how strongly, and in what circumstances, topromote the use of climate information in addressingreal-world fisheries issues given our current state ofscientific knowledge and understanding. While therewas much optimism that we may be close to significantconceptual and scientific breakthroughs, the abovediscussion has focused in large part on the currentlimitations of our knowledge. A concept that wasraised was that of “foreseeability”, referring to the fact

that even if we cannot predict exactly what is goingto happen, based on experience and existing data,decision makers should be able to anticipate likelyoccurrences (and thus may be accountable for nottaking action). Bearing these issues in mind, therewas agreement that there are activities and productsthat would be useful to pursue given even our currentstate of somewhat limited scientific understanding.These include:

“Picking low-lying fruit”In certain areas of the world (e.g., many Pacific

Island states, coastal Peru), fisheries systems, includ-ing the human dimensions, react relatively consis-tently to certain predictable climate events such as ElNiño. Thus, we are at a stage to focus on these areasfor direct application of climate information. Suchactivities should be pursued (however, keeping inmind the potential unintended consequences dis-cussed above).

Decision calendars/mapsFor areas where we believe there is a discernible

climate impact or trend, we need to study in detail thedecisions and their required lead times that may beaffected by introducing this knowledge. Such socioe-conomic research takes time and should proceed inparallel with the ecologically-oriented research. Localknowledge should be incorporated, as there are clearexamples where the biological system may give indi-cations of coming environmental events prior to thoseensuing from physical/dynamical forecasts.

Knowledge products Although we may not have definitive answers

to give managers and other key actors in fisheries,particularly with respect to longer timescales, thegroup was of the opinion that we should strive for aparticipatory approach, beginning with basic educationabout the current state of the knowledge for relevantdecision-makers, including those in the financial sector.There was emphasis that we must be transparent inrepresenting the limitations of our knowledge inorder to avoid a backlash and loss of legitimacy.While we may not be able to predict regime shifts,knowledge by actors within the fisheries sector thatdramatic fluctuations can occur may help proactiveplanning and serve as additional impetus to an adap-tive management approach.


21

Impacts mapsDrawing on experience in other spheres

involving climate patterns and their impacts, it wassuggested that spatial maps and corresponding timeseries of the climatic/oceanographic patterns in thePacific Ocean associated with the known large-scalepatterns of biological variations and fisheries impacts

would be a useful product for the research and publiceducation communities. The idea is to provide a con-ceptual framework for organizing the data so thatone might conveniently juxtapose environmental fac-tors and expected impacts by species, fishery, andother variables.



Agrawala, S. and K. Broad. 2002. Technology transferperspective on climate forecast applications.Research in Science and Technology Studies 13,45-69.

Alheit, J. and P. Bernal. 1993. Effects of physical andbiological changes on the biomass yield of theHumboldt Current ecosystem. p. 53-68. In: K.Sherman, L. M. Alexander, and B. Gold (eds.)Large Marine Ecosystems—Stress, Mitigation,and Sustainability. American Association for theAdvancement of Science.

Arntz, W. E., A. Landa and J. Tarazona. 1985. ElNiño: Su impacto en la fauna marina. Bol. Inst.Mar., vol. ext., Perú-Callao, 222 pp.

Bakun, A. 1996. Patterns in the Ocean: Ocean Processesand Marine Population Dynamics. University ofCalifornia Sea Grant, San Diego, California, USA,in cooperation with Centro de InvestigacionesBiológicas de Noroeste, La Paz, Baja CaliforniaSur, Mexico. 323 pp.

Bakun, A. 1998. Ocean Triads and Radical InterdecadalStock Variability: Bane and Boon for FisheryManagement Science. pp. 331-358. In: T.J.Pitcher, P.J.B. Hart and D. Pauly (eds.)Reinventing Fisheries Management. KluwerAcademic Publishers, Dordrecht, Netherlands.

Bakun, A. 2001. ‘School-mix feedback’: a differentway to think about low frequency variability inlarge mobile fish populations. Prog. Oceanogr.49, 485-511.

Bakun A. 2002a. Seeking an expanded suite of man-agement tools: implications of rapidly-evolvingadaptive response mechanisms (such as “school-mix feedback”). In: N. Erhardt (ed.) Proceedingsof the World Conference on SustainableFisheries. Center for Sustainable Fisheries, Univ.Miami (in review).

Bakun A. 2002b. An empirical test for the existence ofa generation-time-dependent mechanism for rapidadaptive response of fish populations to variationsin ocean climate, predation or fishery exploita-tion. pp. 2-4. In: C.D. Van der Lingen, C. Roy, P.Fréon, M. Barange, L. Castro, L. Gutierrez, L.Nykjaer and F. Shillington (eds.) Report of aGLOBEC-SPACC/IDYLE/ENVIFISH work-shop on spatial approaches to the dynamics ofcoastal pelagic resources and their environmentin upwelling areas. GLOBEC Rep. 16.

Bakun, A. and P. Cury. 1999. The "school trap": amechanism promoting large-amplitude out-of-phase population oscillations of small pelagicfish species. Ecol. Letters 2, 349-351.

Bakun, A. and R. Parrish. 1980. Environmentalinputs to fishery population models for easternboundary currents. pp. 67-104. In: G.D. Sharp(ed.) Workshop on the effects of environmentalvariation on the survival of larval pelagic fishes.IOC Workshop Rep. 28., IntergovernmentalOceanographic Commission, UNESCO, Paris.323p.

Barber, R. T. and F. P. Chavez. 1983. Biological con-sequences of El Niño. Science 222, 1203-1210.

Barnes, T., L. Jacobson, A. MacCall, and P. Wolf.1992. Recent population trends and abundanceestimates for the Pacific sardine (Sardinopssagax). Calif. Coop. Oceanic Fish. Invest. Rep.33, 60-75.

Barrett, C. B. 1998. The Value of Imperfect ENSOForecast Information: Discussion. Amer. J. Agricul.Econ. 80 (5), 1109-1112.

Beamish, R. J., and D. R. Bouillon. 1993. Pacificsalmon production trends in relation to climate.Can. J. Fish. Aquat. Sci. 50, 1002-1016.

References


Beamish, R. J., C. E. M. Neville, et al. 1997.Production of Fraser River sockeye salmon(Oncorhynchus nerka) in relation to decadal-scale changes in the climate and the ocean. Can.J. of Fish. Aquat. Sci. 54, 543-554.

Beamish, R.J., D.J. Noakes, G.A. MacFarlane, L.Klyashtorin, V.V. Ivanov and V. Kurashov. 1999.The regime concept and natural trends in theproduction of Pacific salmon. Can. J. of Fish.Aquat. Sci. 56, 516-526.

Berkes, F., R. Mahon, P. McConney, R. Pollnac andR. Pomeroy. 2001. Managing Small-ScaleFisheries. Alternative Directions and Methods.International Development Research Centre,Canada. 320p.

Bertram, D.F., D.L. Mackas and S.M. McKinnell.2001. The seasonal cycle revisited: interannualvariation and ecosystem consequences, pp. 283-307 In: S.M. McKinnell, R.D. Brodeur, K.Hanawa, A.B. Hollowed, J.J. Polovina and C.-I.Zhang (eds.) Pacific Climate Variability andMarine Ecosystem Impacts. Prog. Oceanogr. 49.

Bradford, M. J., and J.R. Irvine. 2000. Land use,fishing, climate change, and the decline ofThompson River, British Columbia, cohosalmon. Can. J. Fish. Aquat. Sci. 57, 13-16.

Brander, K.M. and P.C.F. Hurley. 1992. Distributionof early-stage Atlantic cod, (Gadus morhua), had-dock (Melanogrammus aegelfinus) and witchflounder (Glyptocephalus cynoglossus) eggs on theScotian Shelf: A reappraisal of evidence on thecoupling of cod spawning and plankton produc-tion. Can. J. Fish. Aquat. Sci. 49, 238-251.

Brander, K.M., R.R. Dickson and J.G. Shepherd.2001. Modelling the timing of plankton produc-tion and its effect on recruitment of cod (Gadusmorhua). ICES J. Mar. Sci. 58, 962-966.

Broad, K., and S. Agrawala. 2000. The Ethiopiafamine: Uses and limits of seasonal climate fore-casts. Science 289, 1693-1694.

Broad, Kenneth, Alex Pfaff, and Michael H. Glantz(in press) Effective & Equitable Disseminationof Seasonal-to-Interannual Climate Forecasts:policy implications from the Peruvian fisheryduring El Niño 1997-98. Climatic Change.

Brodeur, R.D., and D.M. Ware. 1992. Long-termvariability in zooplankton biomass in the subarcticPacific Ocean. Fish. Oceanogr. 1, 32-38.

Brodeur, R.D., and D.M. Ware. 1995. Interdecadalvariability in distribution and catch rates ofepipelagic necton in the North Pacific Ocean.pp. 329-356. In Beamish, R.J., ed. ClimateChange and Northern Fish Populations. Can.Spec. Publ. Fish. Aquat. Sci. 121.

Carr, M.-E. and K. Broad. 2000. Satellites, Society,and the Peruvian fisheries during the 1997-98 ElNiño. pp. 171-191. In: D. Halpern (ed.)Satellites, Oceanography and Society. ElsevierScience, Amsterdam.

Carrasco, S. and O. Lozano. 1989. Seasonal andlong-term variations of zooplankton volumes inthe Peruvian Sea, 1964-1987. pp. 82-85. In:Pauly, D., P. Muck, J. Mendo, and I. Tsukayama,Eds. The Peruvian upwelling ecosystem: dynamicsand interactions. ICLARM ConferenceProceedings 18, 438p.

Chakravorty, U. 1997. “Optimal regulation of theHawaii pelagic fishery: Proposal,” Pelagic FisheriesResearch Program, 1997-2002.

Chakravorty, U. and K. Nemoto. 2001. Modelingthe effects of area closure and tax policies: A spatial-temporal model of the Hawaii longline fishery.Marine Resource Economics 15, 179-204.

Chakravorty, U. and J. Sibert, eds. 1999. Ocean-Scale Management of Pelagic Fisheries: Economicand Regulatory Issues, Joint Institute of Marineand Atmospheric Research Contribution 99-321, 102 pp.


Checkley, D., R. Dotson and D. Griffith. 2000.Continuous, underway sampling of eggs ofPacific sardine (Sardinops sagax) and northernanchovy (Engraulis mordax) in spring 1996 and1997 off southern and central California. Deep-Sea Res. II 47, 1139-1155.

Cowen R., K.M.M. Lwiza, S. Sponaugle C.B. Parisand D.B. Olson. 2000. Connectivity of marinepopulations: open or closed? Science 287, 857-859.

Cury, P., and C. Roy. 1989. Optimal environmentalwindow and pelagic fish recruitment success inupwelling areas. Can. J. Fish. Aquat. Sci. 46,670-680.

Cury, P., C. Roy, R. Mendelssohn, A. Bakun, D.M.Husby, and R.H. Parrish. 1995. Moderate is bet-ter: nonlinear climatic effect on Californiaanchovy (Engraulis mordax). pp. 417-424. In: R.J.Beamish (ed.). Climate Change and Northern FishPopulations. Can. Spec. Publ. Fish. Aquat. Sci.121. 739 pp.

Cushing, D.H. 1969. Upwelling and fish produc-tion. FAO Fish. Tech. Pap. 84. 40pp.

Cushing, D.H. 1971. The dependence of recruitmenton parent stock in different groups of fishes. J.Cons. int. Explor. Mer. 33, 340-362.

Cushing, D.H. 1975. Marine Ecology and Fisheries.Cambridge University Press, Cambridge. 278p.

Cushing, D.H. 1990. Plankton production and year-class strength in fish populations: an update ofthe match/mismatch hypothesis. Adv. Mar. Biol.26, 249-293.

Dagorn, L. and P. Freon. 1999. Tropical tuna associ-ated with floating objects: a simulation study ofthe meeting point hypothesis. Can. J. Fish.Aquat. Sci. 56, 984-993.

Durand, M.-H., P. Cury, R. Mendelssohn, C. Roy, A.Bakun and D. Pauly (eds.) 1998. Global VersusLocal Changes in Upwelling Systems. ORSTOMEditions, Paris. 594 pp.

Ebbesmeyer, C.C., D.R. Cayan, D.R. McLain, F.H.Nichols, D.H. Peterson, and K.T. Redmond,1991. 1976 step in the Pacific climate: fortyenvironmental changes between 1968-1975 and1977-1984. In: J. Betancourt and V.L. Tharp(eds.) Proc. 7th Annual Pacific ClimateWorkshop. Calif. Dept Water Resources,Interagency Ecol. Stud. Prog. Tech. Rep. 26,129-141.

FAO. 2001. Research implications of adopting theprecautionary approach to management of tunafisheries. FAO Fish. Circ. 963. 88 pp.

Fortier, L., D. Ponton and M. Gilbert. 1995. Thematch/mismatch hypothesis and the feeding suc-cess of fish larvae in ice-covered southeasternHudson Bay. Mar. Ecol. Prog. Ser. 120, 11-27.

Fournier,D.A., J. Hampton and J.R. Sibert. 1998.MULTIFAN-CL: a length-based, age-structuredmodel for fisheries stock assessment, with appli-acation to South Pacific albacore, Thunnusalalunga. Can. J. Fish. Aquat. Sci. 55, 2105-2116.

Fry, B. 1988. Food web structure on Georges Bankfrom stable C, N, and S isotopic compositions.Limnol. Oceanogr. 33, 1182-1190.

Gargett, A.E. 1997. The optimal stability ‘window’:a mechanism underlying decadal fluctuations inNorth Pacific salmon stocks? Fish. Oceangr. 6,109-117.

Glantz, M.H. 1979. Science, politics, and economicsof the Peruvian anchovy fishery. Marine Policy,July, 1979, 201-210.

Glantz, M.H. 1986. Man, state, and fisheries: Aninquiry into some societal constraints that affectfisheries management. Ocean Development andInternational Law 17, 191-270.

Glantz, M.H. (ed.) 1995. Usable Science II: ThePotential Use and Misuse of El NiñoInformation in North America. NCAR, 260 pp.

Glantz, M.H. 1996 (2001). Currents of Change: ElNiño's impact on climate and society. CambridgeUniv. Press, 194 pp.


Glantz, M.H. and J.D. Thompson (eds.) 1981.Resource Management and EnvironmentalUncertainty: Lessons from Coastal UpwellingFisheries. John Wiley and Sons, New York.

Gotceitas, V., V. Puvanendran, L.L. Leader and J.A.Brown. 1996. An experimental investigation ofthe 'match/mismatch' hypothesis using larvalAtlantic cod. Mar. Ecol. Prog. Ser. 130, 29-37.

Graham, N.E. 1994. Decadal-scale climate variabilityin the 1970s and 1980s: observations and modelresults. Climate Dyn. 10, 135 162.

Hamnett, M.P. and C.L. Anderson C.L. 1998.Impact of Climate Variability on PelagicFisheries in the Pacific Islands. NOAA Office ofGlobal Programs Progress Report, April 1998.Social Science Research Institute, Univ. Hawaii.

Hansen, J. W. (ed.) 2001. Forthcoming: Special Issueon Applying Seasonal Climate Prediction toAgriculture. Agricultural Systems.

Hanski I. 1991. Metapopulation dynamics: a briefhistory and conceptual domain. Biological J.Linn. Soc. 42, 3-16.

Hare, S. R. and N. J. Mantua. 2000. Empirical evi-dence for North Pacific regime shifts in 1977and 1989. Prog. Oceanogr. 7, 103-145.

Hare, S. R., N. J. Mantua and R.C. Francis. 1999.Inverse production regimes: Alaska and WestCoast Pacific salmon. Fisheries 24, 6-14.

Hare, S.R., S. Minobe, W.S. Wooster, and S.McKinnell. 2000. An introduction to the PICESsymposium on the nature and impacts of NorthPacific climate regime shifts. Prog. Oceanogr. 47,99-102.

Hay, D.E., P.B. McCarter, and K.S. Daniel. 2001.Tagging of Pacific herring Clupea pallasi from1936-1992: a review with comments on hom-ing, geographic fidelity, and straying. Can. J.Fish. Aquat. Sci. 58, 1356-1370.

Hughey, KFD, R. Cullen and G.N. Kerr. 2000.Stakeholder groups in fisheries management.Marine Policy 24, 119-127.

Hurrell, J. W. and H. van Loon. 1997. Decadal varia-tions in climate associated with the North Atlanticoscillation. Climatic Change 36, 301-326.

Jordán, R., J. Csirke and I. Tsukayama. 1978.Situacion de los recursos anchoveta, sardina,jurel y caballa al Junio 1978. Instituto del Mardel Peru, Informe No. 56. 32 p.

Jentoft, S. 2000. The community: A missing link offisheries management. Marine Policy 24, 53-59.

Jentoft, S., B. J. McCay and D. C. Wilson. 1998.Social theory and fisheries management. MarinePolicy 22, 423-436.

Kawasaki, T. 1983. Why do some pelagic fishes havewide fluctuations in their numbers? Biologicalbasis of fluctuation from the viewpoint of evolu-tionary ecology. pp. 1065-1080 In: Sharp, G. D.and J. Csirke (eds.). Proceedings of the ExpertConsultation to Examine Changes inAbundance and Species of Neritic FishResources. San Jose, Costa Rica, 18-29 April,1983. FAO Fish. Rep. 291(2,3).

Kawasaki, T. 1989. Long-term variability in thepelagic fish populations. pp.47-60 In: T.Kawasaki, S. Tanaka, Y. Toba and A. Taniguchi(eds.), Long-term variability of pelagic fish pop-ulations and their environment. 402p.

Kifani, S. 1991. Approaches spatio-temporelle desrelations hydroclimat-dynamique des espècespélagiques en région d'upwelling: cas de la sardinedu stock central marocain. Thèse de Doctorat,Spécialité Océanographie halieutique. Univ.Bret. Occid. 290 pp.

Kirby, D.S., O. Fiksen and P.J.B. Hart. 2000. Adynamic optimisation model for the behaviour oftunas at ocean fronts. Fish. Oceanogr. 9, 328-342.


Lasker, R. 1975. Field criteria for survival of anchovylarvae: the relation between inshore chlorophyllmaximum layers and successful first feeding.Fish. Bull. U.S. 73, 453-462.

Lasker, R. 1978. The relation between oceanograph-ic conditions and larval anchovy food in theCalifornia Current: Identification of the factorsleading to recruitment failure. Rapp. P.-v. Réun.Cons. int. Explor. Mer. 173, 212-230.

Lasker, R. 1981a. Factors contributing to variablerecruitment of the northern anchovy (Engraulismordax) in the California Current: Contrastingyears 1975 through 1978. Rapp. P.-v. Réun.Cons. int. Explor. Mer. 178, 375-388.

Lasker, R. 1981b. The role of a stable ocean in larvalfish survival and subsequent recruitment. p. 80-87. In: R. Lasker (ed.) Marine Fish Larvae: mor-phology, Ecology and Relation to Fisheries.Univ. Washington Press, Seattle. 131pp.

Legendre, L., and S. Demers. 1985. Auxiliary energy,ergoclines and aquatic biological production.Naturaliste can. Rev. Écol. Syst. 112, 5-14.

Lehodey, P. 2001. Sepodym skipjack analysis. 14thStanding Committee on Tuna and Billfish.Secretariat of the Pacific Community,Noumea,New Caledonia. Working Paper SKJ-2. 24 pp.(http://www.spc.int/OceanFish/Html/TEB/Env/SKJ2.pdf)

Lehodey, P. 2002. Impacts of the El Niño SouthernOscillation on tuna populations and fisheries.FAO, Rome. In press. (http://www.spc.int/OceanFish/Html/TEB/Env/FAO_Paper.pdf)

Lehodey, P., M. Bertignac, J. Hampton, A. Lewis, A.and J. Picaut. 1997. El Nino Southern Oscillationand tuna in the western Pacific. Nature 389,715-718.

Lehodey, P., J-M. Andre, M. Bertignac, J. Hampton,A. Stoens, C. Menkes, L. Memery and N. Grima.1998. Predicting skipjack tuna forage distribu-tions in the equatorial Pacific using a coupleddynamical bio-geochemical model. Fish.Oceanogr. 7, 317-325.

Letson, D., I. Llovet, et al. 2001. User perspectives ofclimate forecasts: Crop producers in Pergamino,Argentina. Climate Res. 19, 57-67.

Lluch-Belda, D., R.J.M. Crawford, T. Kawasaki,A.D. MacCall, R.H. Parrish, R.A. Schwartzloseand P.E. Smith. 1989. World-wide fluctuationsof sardine and anchovy stocks: the regime prob-lem. S. Afr. J. mar. Sci. 8, 195-205.

Lluch-Belda, D., R.A. Schwartzlose, R. Serra, R.H.Parrish, T. Kawasaki, D. Hedgecock, and R.J.M.Crawford. 1992. Sardine and anchovy regimefluctuations of abundance in four regions of theworld oceans: a workshop report. Fish.Oceanogr. 1, 339-347.

Loeb, V.J. and O. Rojas. 1988. Interannual variationof ichthyoplankton composition and abundancerelations off northern Chile, 1964-83. Fish. Bull.U.S. 86 1-14.

Logerwell, E. and P. Smith. 2001. Mesoscale eddiesand survival of late stage Pacific sardine (Sardinopssagax) larvae. Fish. Oceanogr. 10, 13-25.

Logerwell, E., B. Lavaniegos, and P. Smith. 2001.Spatially-explicit bioenergetics of Pacific sardinein the Southern California Bight: are mesoscaleeddies areas of exceptional prerecruit produc-tion? Prog. Oceanog. 49, 391-406.

Louw, G.G., C.D.van der Lingen and M.J. Gibbons.1998. Differential feeding by sardine Sardinopssagax and anchovy Engraulis capensis recruits inmixed shoals. S. Afr. J. mar. Sci. 19, 227-232.

Lu, H.-J., K.-T. Lee, and C.-H. Liao. 1998. On therelationship between El Niño/Southern Oscillationand South Pacific albacore. Fish. Res. 39, 1-7.


Lu, H.-J., K.-T. Lee, H.-L. Lin, and C.-H. Liao.2001. Spatio-temporal distribution of yellowfintuna, Thunnus albacares, and bigeye tuna,Thunnus obesus, in the tropical Pacific Ocean inrelation to large temperature fluctuation duringENSO episodes. Fish. Sci. 67, 1046-1052.

MacCall, A.D. 1990. Dynamic Geography of MarineFish Populations. University Washington SeaGrant Program, Seattle.158 pp.

MacCall, Alec D. 1996. Patterns of low frequencyvariability in fish populations of the CaliforniaCurrent. Calif. Coop. Oceanic Fish. Invest. Rep.37, 100-110.

Mackas, D.L., R.E Thomson and M. Galbraith.2001. Changes in the zooplankton communityof the British Columbia continental margin,1985-1999, and their covariation with oceano-graphic conditions. Can. J. Fish. Aquat. Sci. 58,685-702.

MacKenzie, B.R., T.J. Miller, S. Cyr and W.C. Leggett.1994. Evidence for a dome-shaped relationshipbetween turbulence and larval fish ingestionrates. Limnol. Oceanogr. 39, 1790-1799.

Mantua, N., S. Hare, Y. Zhang, J. Wallace, R. Francis.1997. A Pacific interdecadal climate oscillationwith impacts on salmon production. Bull. Am.Meteor. Soc. 78, 1069-1079.

Marchesiello P., J.C. McWilliams and A. Shchepetkin.2002. Equilibrium structure and dynamics ofthe California Current System. J. Phys. Oceanogr.In press.

Mayr, E. 1982. The Growth of Biological Thought.Harvard Univ. Press, Cambridge, Mass. 974 pp.

McCay, B.J. and J. Acheson, 1987. The Question ofthe Commons: The Culture and Ecology ofCommunal Resources. Univ. Arizona Press.Tucson.

McCay, B.J. and Svein, J. 1996. From the bottom up:participatory issues in fisheries management.Society and Natural Resources 9, 237-250.

McGoodwin, J.R., 1990. Crisis in the World’sFisheries. Stanford Univ. Press.

McKinnell, S.M. 2001. 2001 coastal ocean/salmonecosystem event. PICES Press Vol. 9 (2), 26-27.

McKinnell, S.M., R.D Brodeur, K. Hanawa, , A.B.Hollowed, J.J. Polovina and C.-I. Zhang (eds.)2001a. Pacific Climate Variability and MarineEcosystem Impacts. Prog. Oceanogr. 49. 639 pp.

McKinnell, S.M., C.C. Wood, D.T. Rutherford,K.D. Hyatt and D.W Welch. 2001b. The demiseof Owikeno Lake sockeye salmon. N. Amer. J.Fish. Manage. 21, 774-791.

Mendelssohn, R., and P. Cury. 1987. Forecasting afortnightly abundance index of the Ivoiriancoastal pelagic species and associated environ-mental conditions. Can. J. Fish. Aquat. Sci. 44,408-421.

Mendelssohn, R., and J. Mendo. 1987. Exploratoryanalysis of anchoveta recruitment off Peru andrelated environmental series. pp. 294-306. In: D.Pauly and I. Tsukayama (eds.) The PeruvianAnchoveta and Its Upwelling Ecosystem: ThreeDecades of Change. ICLARM Studies andReviews 15. Instituto del Mar del Peru(IMARPE), Callao, Peru; Deutsche Gesellschaftfür Technische Zusammenarbeit (GTZ),GmbH, Eschborn, Federal Republic of Germany;and International Center for Living AquaticResources Management (ICLARM), Manila,Philippines. 351pp.

Miller, A. J., and N. Schneider. 2000. Interdecadalclimate regime dynamics in the North PacificOcean: Theories, observations and ecosystemimpacts. Progr. Oceanogr. 47, 355-379.

Miller, A.J., D.C. Cayan, T.P. Barnett, N.E. Graham,and J.M. Oberhuber. 1994. Interdecadal vari-ability of the Pacific Ocean: model response toobserved heat flux and wind stress anomalies.Clim. Dyn. 9, 287 302.


Miller, K.A. 1996. Salmon stock variability and thepolitical economy of the Pacific Salmon Treaty.Contemporary Economic Policy 14(3), 112-129.

Miller, K.A. 2000. Pacific salmon fisheries: Climate,information and adaptation in a conflict-riddencontext. Climatic Change 45, 37-61.

Mjelde, James W. and K. Keplinger 1998. Using theSouthern Oscillation to Forecast Texas WinterWheat and Sorghum Crop Yields. J. Climate 11,54-60.

Mjelde, J.W, H.S.J. Hill, and J. F. Griffiths. 1988. Areview of current evidence on climate forecastsand their economic effects in agriculture. Amer.J. Agri. Econ. 80, 1089-1095.

Orlove, B.S. and J. Tosteson. 1998. The applicationof aeasonal to interannual climate forecasts basedon El Niño-Southern Oscillation (ENSO)events: Lessons from Australia, Brazil, Ethiopia,Peru, and Zimbabwe. Institute of InternationalStudies, University of California, Berkeley, WP99-3.

Parrish, R.H., C. S. Nelson, and A. Bakun. 1981.Transport mechanisms and reproductive successof fishes in the California Current. Biol. Oceanogr.1, 175-203.

Parrish, R.H., F.B. Schwing, and R. Mendelssohn.2000. Mid-latitude wind stress: the energysource for climatic shifts in the North PacificOcean. Fish. Oceanogr. 9, 224-238.

Pauly, D. 1989. An eponym for Reuben Lasker. Fish.Bull. U.S. 87, 383-384.

Penven P., C. Roy, G.B. Brundrit, A. Colin de Verdière,P. Fréon, A.S. Johnson, J. R.E. Lutjeharms andF.A. Shillington. 2001. A regional hydrodynamicmodel of upwelling in the Southern Benguela. S.Afr. J. Sci. 97, 472-475.

Peterman, R.M., and M. J. Bradford. 1987. Windspeed and mortality rate of a marine fish, thenorthern anchovy (Engraulis mordax). Science235, 354-356.

Peterson, W.T. and S.M. McKinnell. 2000. Oceanecology of juvenile salmonids along the NorthAmerican coast. PICES Press 8 (2), 25-26.

Pfaff, A., K. Broad and M. H. Glantz. 1999. Whobenefits from climate forecasts? Nature 397,645-646

Polovina, J.J, E. Howell, D.R. Kobayashi and M.P.Seki. 2001. The transition zone chlorophyll front,a dynamic global feature defining migration andforage habitat for marine resources. Prog.Oceanogr. 49, 469-483.

Polovina, J.J., G.T. Mitchum, N.E. Graham, M.P.Craig, E.E. DeMartini and E.N. Flint. 1994.Physical and biological consequences of a climateevent in the central North Pacific. Fish.Oceanogr. 3, 15-21.

Pulwarty, R.S. and T.S. Melis. 2001. Climate extremesand adaptive management on the ColoradoRiver: Lessons from the 1997-1998 ENSOevent. J. Environ. Manag. 63, 307-324.

Pulwarty, R.S., and K.T. Redmond. 1997. Climateand Salmon Restoration in the Columbia RiverBasin: The Role and Usability of SeasonalForecasts. Bull. Amer. Meteor. Soc. 78, 381-397.

Radovitch, J. 1979. Managing pelagic schooling preyspecies. In: Predator-Prey Systems in FisheriesManagement. Sport Fishing Institute,Washington, D.C.

Rayner, S., M. Houk, et al., 1999. InstitutionalIssues in Adoption of Probabilistic ClimateVariation Among US Water Managers. Societyfor Applied Anthropology 1999 Annual Meeting:Constructing Common Ground: Human andEnvironmental Imperatives, April 21-25,Tucson, AZ.


Roemmich, D., and J. A. McGowan. 1995. Climaticwarming and the decline of zooplankton in theCalifornia Current. Science 267, 1324-1326.

Roncoli, C. (ed.) 2000. Anthropology and climatechange: Challenges and contributions. PracticingAnthropology 22 (4 -special issue).

Rothschild, B.J. and T.R. Osborn. 1988. The effectsof turbulence on planktonic contact rates. J.Plankton Res. 10, 465-474.

Roy, C. and C. Reason. 1991. ENSO related modu-lation of coastal upwelling in the easternAtlantic. pp. 245-256. In McKinnell, S.M., R.DBrodeur, K. Hanawa, A.B. Hollowed, J.J.Polovina and C.-I. Zhang (eds.) Pacific ClimateVariability and Marine Ecosystem Impacts. Prog.Oceanogr. 49.

Roy, C., P. Cury and S. Kifani, 1991. Optimal envi-ronmental window and pelagic fish reproductivestrategies in upwelling areas. InternationalSymposium on Benguela Trophic Functioning,Resource Utilization from an EcosystemPerspective, Cape Town, South Africa, 8-13 Sep1991.

Sarewitz, D., R.A. Pielke Jr. and R. Byerly Jr. (eds.)2000. Prediction: Science, Decision Making,and the Future of Nature, Island Press,Washington, DC.

Schaefer, K.M. 1998. Reproductive biology of yel-lowfin tuna (Thunnus albacares) in the easternPacific Ocean. Bull. Inter-Am. Trop. TunaComm. 21(5), 201-272.

Schneider, N. and A.J. Miller. 2001. Predicting west-ern North Pacific Ocean climate. J. Climate, 14,3997-4002.

Schwartzlose, R., J. Alheit, A. Bakun, T. Baumgartner,R. Cloete, R. Crawford, W. Fletcher, Y. Green-Ruiz, E. Hagen, T. Kawasaki, D. Lluch-Belda, S.Lluch-Cota, A. MacCall, Y. Matsuura, M.Nevarez-Martinez, R. Parrish, C. Roy, R. Serra,K. Shust, and M. Ward, and J. Zuzunaga. 1999.Worldwide large-scale fluctuations of sardineand anchovy populations. S. Afr. J. Mar. Sci. 21,289-347.

Serra, R. 1987. Impact of the 1982-83 ENSO on theSoutheastern Pacific fisheries, with an emphasison Chilean fisheries. pp. 24-29. In M. Glantz, R.Katz and M. Krenz (eds.) Climate Crisis: TheSocietal Impacts Associated with the 1982-83Worldwide Climate Anomalies. NCAR/ESIG,Boulder, Colorado, 105 pp.

Serra, R. 1991. Long - term variability of the Chileansardine. pp 165-172. In: T. Kawasaki, S. Tanaka,Y. Toba and A. Taniguchi (eds.) Proceedings ofthe International Symposium on the Long -Term Variability of Pelagic Fish Populations andtheir Environment. Pergamon Press, New York.

Serra, R., P. Cury and C. Roy. 1998. The recruitmentof the Chilean sardine and the optimal environ-mental window. pp. 267-274. In: M. Durand, P.Cury, R. Mendelssohn, C. Roy, A. Bakun, D.Pauly (eds.) Global versus local changes inupwelling systems. ORSTOM Editions, Paris,585 pp.

Sharp, G.D. and D.R. McLain. 1993. Fisheries, ElNiño-Southern Oscillation and upper-oceantemperature records: An eastern Pacific example.Oceanography 6, 13-22.

Sinclair, M. 1988. Marine Populations. An essay onpopulation regulation and speciation. WashingtonSea Grant Program. Univ. Washington Press,Seattle and London. 252 p.


Smith, P.E. 1985. A case history of an anti-El Niñoto El Niño transition on plankton and nektondistribution and abundances. pp 121-142 .In:W.S. Wooster and D.L. Fluharty (eds.) El NiñoNorth: Niño Effects in the Eastern SubarcticPacific Ocean. Washington Sea Grant Program,Univ. of Washington, Seattle.

Smith, P. 1990. Monitoring interannual changes inspawning area of Pacific sardine (Sardinopssagax). Calif. Coop. Oceanic Fish. Invest. Rep.31, 145-151.

Steele, J.H. 1974. The stucture of marine ecosystems.Harvard Univ. Press, Cambridge, Mass. 128 p.

Stern, P.C., and W.E. Easterling. 1999. MakingClimate Forecasts Matter. National AcademyPress, 175 pp.

Sugimoto, T., S. Kimura and K. Tadokoro. 2001.Impact of El Niño events and climate regimeshift on living resources in the western NorthPacific. Prog. Oceanog. 49, 113-127.

Sverdrup, H.U. 1953. On conditions for vernalblooming of phytoplankton. J. Cons. int.Explor. Mer 18, 287-295.

Tarte, S. 2001. The international politics of tunamanagement in the Western and Central Pacific.Conference on People and the Sea, Amsterdam,30 August – 2 September, 2001.

Tarte, S. 2002. Report on Pacific Climate andFisheries Workshop, Honolulu, Hawaii, 14-17November, 2001. Department of History/ Politics,University of the South Pacific, Suva, FIJI.

Tarazona, J. and E. Castillo. 1999. El Niño 1997-98y su impacto sobre los ecosistemas marino y ter-restre. Revista Peruana de Biología, VolumenExtraordinario, December 1999.

Thorrold, S.R., C. Latkoczy, P.K. Swart and C.M.Jones. 2001. Natal homing in a marine fishmetapopulation. Science 291, 297-299.

Timmermann, A., J. Oberhuber, A. Bacher, M. Esch,M. Latif, and E. Roeckner. 1999. Increased ElNiño frequency in a climate model forced byfuture greenhouse warming. Nature 398, 694-697.

Trenberth, K.E. and J.W. Hurrell. 1994. Decadalatmospheric-ocean variations in the Pacific.Clim. Dyn. 9, 303 319.

Venrick, E.L., J.A. McGowan, D.R. Cayan and T.L.Hayward. 1987. Climate and chlorophyl a: long-term trends in the Central North Pacific Ocean.Science 238, 70-72.

Walters, C.J. 1989. Value of short-term forecasts ofrecruitment for harvest management. Can. J. Fish.Aquat. Sci. 46, 1969-1976.

Wada, T. and M. Kashiwai. 1989. Changes in growthand feeding ground of Japanese sardine withfluctuation in stock abundance. pp.181-190 In:T. Kawasaki, S. Tanaka, Y. Toba and A. Taniguchi(eds.), Long-term variability of pelagic fish popu-lations and their environment. 402p.

Ward, P., B. Kearney and N. Tsirbas, 2000. Sciencearrangements for the regional management oftuna fisheries. Marine Policy 24, 93-108.

Ware, D.M., and R.E. Thomson. 1991. Link betweenlong-term variability in upwelling and fish pro-duction in the northeast Pacific Ocean. Can. J.Fish. Aquat. Sci. 48, 2296-2306.


Appendix 1.

List of ParticipantsVERA AGOSTINI

School of Aquatic and Fishery SciencesBox 355020University of WashingtonSeattle, WA 98190tel: (206) [email protected]

JÜRGEN ALHEIT

Baltic Sea Research InstituteSeestr. 1518119 WarnemündeGermanytel: +49-381-5197 208fax: +49-381-5197 [email protected]

ANDREW BAKUN

IRI-The International Research Institutefor Climate PredictionNEW ADDRESS (after 1 July, 2002)Center for Sustainable FisheriesUniversity of Miami, RSMAS4600 Rickenbacker CausewayMiami, FL [email protected]

KENNETH BROAD

Research Asst. Professor, Div. of Marine Affairs & PolicyUniversity of Miami, RSMAS4600 Rickenbacker CausewayMiami, FL 33133tel: 305 361-4851fax: 305 [email protected]

MARY-ELENA CARR

Jet Propulsion Laboratory MS 300-3234800 Oak Grove Dr.Pasadena, CA 91101-8099tel: (818) 354-5097fax: (818) [email protected]

DAVE CHECKLEY, JR. Scripps Institution of Oceanography La Jolla, CA 92093-0218 USA tel: +01 (858) 534-4228fax: +01 (858) [email protected]

ROBERT COWEN

Director, Center for SustainableFisheries (CSF)University of Miami, RSMAS4600 Rickenbacker CausewayMiami, FL 33149tel: 305 [email protected]

JEROME M. COMCOWICH

IPRC - International Pacific Research CenterUniversity of Hawaii2525 Correa RoadHonolulu, HI [email protected]

JORGE CSIRKE

Chief, Marine Resources Service (FIRM)FAOVialle delle Terme di Caracalla00100 Rome, Italytel. +39 06 570 56506fax. +39 06 570 [email protected]

DAVE GILBERT

National Institute of Water &Atmospheric Research LtdP.O. Box 14-901WellingtonNew ZealandEmail: [email protected]: +64-4-386 0537fax: +64-4-386 0574Home telephone: +64-4-568 5360

MICHAEL GLANTZ

Senior ScientistESIG/NCARBox 3000Boulder, CO [email protected]

SAFWAN HADI, DR.Departemen Geofisika dan MeteorologiInstitut Teknologi BandungJl. Ganesha 10Bandung 40132Indonesiatel: 62-22-250-0409fax: 62-22-253-4139e-mail: [email protected]

MICHAEL HAMNETT

Pacific Basin Development CouncilSuite 1075Honolulu, HI 96813-5214tel: (808) 596-7229fax: (808) [email protected]

TOSHIO KATSUKAWA

Ocean Research Institute, University of Tokyo 1-15-1 Minamidai, Nakano-ku, Tokyo 164-8639, Japan tel: +81 3 5351 6511fax: +81 3 5351 6506 [email protected]

HAK GYOON KIM

Director, Marine Environment, Oceanography& Harmful Algal Bloom ResearchDepartmentNational Fisheries Research andDevelopment Institute (NFRDI)408-1 Shirang-ri, Kijang-up Kijang-gun,Busan 619-902, [email protected]

JIN-YEONG KIM

National Fisheries Research &Development Institute408-1, Sirang-RiGijang-EupGuang-GunBusan 619-902, Koreatel: (051) 720-2270fax: (051) [email protected]


SUAM KIM

Dept. of Marine BiologyPukyong National University599-1 Daeyeon3 - Dong Nam-GuPusan 608-737, Koreatel (051) 620-6268fax (051) [email protected]

R. MICHAEL LAURS

Director, Honolulu Laboratory Southwest Fisheries Science Center National Marine Fisheries Service(NOAA) 2570 Dole Street Honolulu, HI 96822-2396 tel: 808-983-5300fax: [email protected]

PATRICK LEHODEY

Senior Fisheries Scientist Oceanic Fisheries Programme Secretariat of the Pacific Community BP D5 98848, Noumea cedex New Caledonia tel: (687) 26 20 00fax: (687) 26 38 18 [email protected] web: http://www.spc.int/oceanfish/

SALVADOR E. LLUCH-COTA

Centro de Investigaciones Biologicas delNoroeste, S.C.PO Box 128 La Paz, BCSMexico 23000[Courier delivery:Centro de Investigaciones Biologicas delNoroeste, S.C.Km 0.5 Acceso a la Telefonica s/nEl Conchalito23096 La Paz, B.C.S. CR-23001Mexico]tel: (52-112)53633 ext 3432fax: (52-112)[email protected]

HSUEH-JUNG LU

Department of Fisheries ScienceNational Taiwan Ocean University2, Pei-Ning Road, Keelung, Taiwan20224, R.O.C.tel: +886-2-2462-2192 ext.5033fax: [email protected] [email protected]

LORENZ MAGAARD

IPRC - International Pacific Research CenterUniversity of Hawaii2525 Correa RoadHonolulu, HI 96822tel: (808) 956-7509fax: (808) [email protected]

ALEC MACCALL

Fishery BiologistSanta Cruz LaboratoryNational Marine Fisheries Service110 Shaffer Rd.Santa Cruz, CA 95060 ztel: (831) 420-3950fax: (831) [email protected]

JOHN MARRA

Lamont-Doherty Earth Observatory ofColumbia University 61 Route 9W Palisades, NY 10964-8000 tel: 1-845-365-8891 fax: 1-845-365-8150 [email protected]

OLIVIER MAURY

IRD (Institut de Recherche pour leDéveloppement)SFA BP 570Victoria, Iles [email protected] or [email protected] & fax: (248) 22 47 42

JAY MCCREARY

Director, IPRC - International PacificResearch CenterUniversity of Hawaii2525 Correa RoadHonolulu, HI [email protected]

SKIP MCKINNELL

Assistant Executive SecretaryNorth Pacific Marine ScienceOrganizationc/o Institute of Ocean SciencesP.O. Box 6000Sidney, B.C., CanadaV8L 4B2tel: 1-250-363-6826fax: [email protected]

ARTHUR J. MILLER

Research Oceanographer Scripps Institution of OceanographyNierenberg Hall, Room 439University of California, San Diego La Jolla, CA 92093-0224 tel: (858) 534-8033fax: (858) [email protected] http://meteora.ucsd.edu/~miller

RAGHU MURTUGUDDE

CSS Bldg, Room 2201Univ of MD, Collge Park, MD 20742tel: (301) [email protected]

I-HSUN NI

Department of Environmental Biology& Fishery ScienceNational Taiwan Ocean University2, Pei-Ning Road, Keelung, Taiwan,R.O.C.tel: (886-2) 2463-2341fax: (886-2) 2462-2192 ext 5013 [email protected]

DONALD OLSON

Univ. of Miami, RSMAS4600 Rickenbacker Cswy.Miami, FL 33149tel: 305 361 4074fax: 305 361- [email protected]

ROBERT J. OLSON

Inter-American Tropical Tuna Commission8604 La Jolla Shores DriveLa Jolla, CA 92037-1508 tel: (858) 546-7160fax: (858) [email protected]

IAN PERRY

Fisheries & Oceans CanadaPacific Biological Station3190 Hammond Bay RoadNanaimo, B.C. V9R 5K6Canadatel: (250) 756-7137fax: (250) [email protected]

GUILLERMO PODESTA

University of Miami, RSMAS4600 Rickenbacker CausewayMiami, FL [email protected]


JEFFREY J. POLOVINA

Honolulu Laboratory Southwest Fisheries Science Center National Marine Fisheries Service(NOAA) 2570 Dole Street Honolulu, HI 96822-2396 tel: 808-983-5300 fax: [email protected]

RENATO QUINONES

Department of OceanographyUniversity of ConcepcionCasilla 160-C, ConcepcionChiletel: +56-41-203861fax: [email protected]

RUBEN RODRIGUEZ-SANCHEZ

CICIMAR Apdo. Post. 592La Paz, B.C.S.23000, Mexicostreet address: Playa El Conchalito s/nLa Paz, Baja California Sur23000, Mexicotels: (52-612) 253-44 / 253-66/ 303-50 fax: (52-612) [email protected]

CLAUDE ROY

IRD (France)French-South African IDYLE ProjectDept. of OceanographyUniversity of Cape TownRondebosch 7701 South [email protected]

RODOLFO SERRA

Instituto de Fomento PesqueroDivision de Evaluacion de PesqueriasCasilla 8-V, ValparaisoChiletel: +56-32-212-630fax: [email protected]

FRANK SCHWING

Pacific Fisheries Environmental Laboratory1352 Lighthouse AvenuePacific Grove, CA 93950-2097tel: (831) 648-8515fax: (831) [email protected]

EILEEN SHEA

The East-West Center, Burns Hall 2062University of Hawaii1601 East-West RoadHonolulu, HI USA 96848-1601tel: 808-944-7253fax: [email protected]

JEFFREY SHIMA

Biological Sciences Victoria University PO Box 600 Wellington New Zealandtel: +64 4 463 6494fax: +64 4 463 5331 [email protected]

TAKASHIGE SUGIMOTO

Ocean Research Institute, University of Tokyo 1-15-1 Minamidai, Nakano-ku, Tokyo 164-8639, Japan tel: +81 3 5351 6511fax: +81 3 5351 6506 [email protected]

KAZUAKI TADOKORO

Frontier Research System for Global Change31173-25 Showa-machi Kanazawa-ku, Yokohama-cityKanagawa 236-0001, Japantel: +81-45-778-5602fax: [email protected]

SANDRA TARTE

Department of History/PoliticsUniversity of the South Pacifictel: 679 212577fax: 679 301 [email protected]

SIMON THORROLD

Biology Department MS # 35Woods Hole Oceanographic InstitutionWoods Hole, MA 02543tel: 1 508 289-3366 (Office)

1 508 289-2985 (Lab)fax: 1 508 [email protected]

SIDNEY THURSTON

Associate Program Director, Austral-Asia RegionNOAA Office of Global ProgramsLevel 14, Hibiya Central Building1-2-9 Nishi-Shimbashi Minato-kuMinato-KuTokyo 105-0003, Japantel: 813-5532-7249fax: [email protected]

NEIL WARD

IRI-The International Research Institutefor Climate Prediction101 Monell Bldg.Lamont-Doherty Earth Observatory,Columbia UniversityPalisades, NY 10964-8000tel: (845) 680-4446fax: (845) [email protected]

CHANG IK ZHANG

Permanent address:Dept. of Marine Production Mgmt.College of Fisheries SciencesPukyong National University599-1 Diamonding, Nam-goPusan 608-737, Koreatel: (82-51) 620-6124fax: (82-51) [email protected] address:Visiting Fisheries ScientistAlaska Fisheries Science CenterNMFS/NOAABIN C15700; Bldg. 47600 Sand Point Way, NESeattle, WA 98115-0070tel: 206-526-4250fax: [email protected]

The IRI was established as a cooperative agreement between U.S. NOAA Office of Global Programs and Columbia University.

INTERNATIONAL RESEARCH INSTITUTE

FOR CLIMATE PREDICTION

Columbia Earth Institute

61 Route 9W

Palisades, NY 10964-8000 USA

Phone: 845-680-4468

Fax: 845-680-4866

http://iri.columbia.edu

Documents

The International research Institute for Climate prediction