UBC Faculty Research and Publications

Back to the future : advances in methodology for modelling and evaluating past ecosystems as future… Pitcher, Tony J. 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52383-FCRR_2004_12-1.pdf [ 5.46MB ]
Metadata
JSON: 52383-1.0074780.json
JSON-LD: 52383-1.0074780-ld.json
RDF/XML (Pretty): 52383-1.0074780-rdf.xml
RDF/JSON: 52383-1.0074780-rdf.json
Turtle: 52383-1.0074780-turtle.txt
N-Triples: 52383-1.0074780-rdf-ntriples.txt
Original Record: 52383-1.0074780-source.json
Full Text
52383-1.0074780-fulltext.txt
Citation
52383-1.0074780.ris

Full Text

                                                                                                                                                                                                   ISSN 1198-6727  Back to the Future: Advances in  Methodology for Modelling and Evaluating  Past Ecosystems as Future Policy Goals  Fisheries  Centre  Research  Reports 2004   Volume  12   Number 1   Fisheries Centre Research Reports  2004   Volume 12   Number  1     Back to the Future: Advances in Methodology  for Modelling and Evaluating Past Ecosystems As Future Policy Goals         Fisheries Centre, University of British Columbia, Canada ISSN 1198-6727  BACK  TO  THE  FUTURE:   ADVANCES  IN  METHODOLOGY  FOR  MODELLING  AND  EVALUATING  PAST  ECOSYSTEMS  AS  FUTURE  POLICY GOALS             Edited by  Tony  J. Pitcher      Sponsored by Coasts Under Stress    Fisheries Centre Research Reports 12(1) 158 pages ? published 2004  by   The Fisheries Centre, University of British Columbia 2259 Lower Mall Vancouver, B.C., Canada, V6T 1Z4      ISSN 1198-6727  F I S H E R I E S  C E N T R E  R E S E A R C H  R E P O R T S  V O L U M E  1 2  N U M B E R  1  2 0 0 4   BACK TO THE FUTURE: ADVANCES IN METHODOLOGY FOR  MODELLING AND EVALUATING PAST ECOSYSTEMS  AS FUTURE POLICY GOALS  Edited by Tony J. Pitcher   CONTENTS    Page Director's Foreword ........................................................................................................................3  Introduction to the methodological challenges in Back-to-the-Future research  Tony J. Pitcher .....................................................................................................................................4   A. Inputs to Models and Modelling  Synoptic methods for constructing models of the past   Johanna J. Heymans and Tony J. Pitcher .............................................................................. 11 What was the structure of past ecosystems that had many top predators?  Tony J. Pitcher..........................................................................................................................18 The problem of extinctions  Tony J. Pitcher ......................................................................................................................... 21 Challenging ecosystem simulation models with climate change: the ?Perfect Storm?  Tony J. Pitcher and Robyn Forrest......................................................................................... 29 Tuning ecosystem models to past data   Richard Stanford .................................................................................................................... 39 Dealing with migratory species in ecosystem models  Steve Martell.............................................................................................................................41 Estimating the effects of prey-predator vulnerability settings on Ecosim's dynamic function   Cameron Ainsworth .................................................................................................................45   Policy search methods for back to the future    Cameron Ainsworth, Johanna J. Heymans and Tony J. Pitcher........................................... 48 Environmental archaeology: principles and case studies   Trevor Orchard and Quentin Mackie ..................................................................................... 64 How traditional knowledge can contribute to environmental research and resource management    Bill Simeone..............................................................................................................................74   B. Evaluation and Policy Goals  Why we have to open the lost valley: criteria and simulations for sustainable fisheries  Tony Pitcher .............................................................................................................................78 Evaluating the ecological effects on exploited ecosystems using information theory   Johanna J. Heymans................................................................................................................87 Modifying Kempton?s biodiversity index for use with dynamic ecosystem simulation models    Cameron Ainsworth and Tony J. Pitcher ................................................................................91  An index expressing risk of local extinction for use with dynamic ecosystem simulation Models  Wai Lung Cheung and Tony J. Pitcher.................................................................................... 94 Back to the Future Methodology, Page 2  How do we value the restoration of past ecosystems?   Ussif Sumaila ........................................................................................................................ 103 Economic valuation techniques for Back-To-The-Future optimal policy searches  Cameron Ainsworth and Ussif R. Sumaila .......................................................................... 104 An employment diversity index used to evaluate ecosystem restoration strategies  Cameron Ainsworth and Ussif R. Sumaila .......................................................................... 108 Evaluating future ecosystems: a great step backward?   Nigel Haggan ........................................................................................................................ 109 Incorporating First Nations values into fisheries management: a proposal for discussion  Rashid Sumaila ...................................................................................................................... 112 Aboriginal Values   Simon Lucas........................................................................................................................... 114   C. Community and Workshop Inputs  How we carried out ?Back-to-the-Future? community interviews Cameron Ainsworth............................................................................................................... 116 The community workshop: how we did it and what we learned from the results  Melanie D. Power, Nigel Haggan and Tony J. Pitcher .........................................................125 Round-Table discussions from a Back-to-the-Future Symposium at UBC, February 2002:  Issues in Policy, Visualisation and Presentation  Melanie D. Power and Tony J. Pitcher .................................................................................129 Rapporteurs? report on discussion at the Back-to-the-Future Symposium, UBC, February 2002  Amy Poon and Yvette Rizzo ..................................................................................................135        ANNEX  Back-to-the-Future Symposium Programme, February 2002 ......................................................155       A Research Report from ?Back to the Future: the Restoration of Past Ecosystems as Policy Goals for Fisheries?  Supported by the Coasts Under Stress ?Arm 2? Project A Major Collaborative Research Initia tive of the Canadian Government  Fisheries Centre Research Reports 12(1) 158 pages ? Fisheries Centre, University of British Columbia, 2004      F I S H E R I E S  C EN T R E R ES E AR CH R EPO R T S  AR E A B S T R ACT E D I N  T H E FA O  A Q U AT I C S CI EN CES  AN D F I S H E R I E S  A B S T R A C T S  (AS FA ) IS SN  11 98- 67 2 7  Page 3, Fisheries Centre Research Reports 12(1), 2004 Director?s Foreword  Big Catc h for Hu m a ns  The fishers were not catching much when Jesus, sitting in one of the boats, encouraged Peter to cast the nets again in deeper water. Such a large amount of fish was caught that both boats began to sink. The disciples were astonished, but Jesus said to Peter "Do not fear, from now on you will be catching men." 1 Some might hold that this ?parable of the draught of fishes? is an early example of overfishing. There is a catch, so to speak, in sinking the darn boat, not to say in depleting all those Galilean fishes. But in fact, the parable means that if you fish in the right place with the right gear and information (divine in this case), your catch may surprise you. And indeed, Christianity, a really bright and shiny new idea at the time, ended up with an unexpectedly large catch of humans. (Yes, yes, there was a catch - a lot went very wrong later on!)  Back-to-the-Future (BTF), an integrative approach to restoration ecology of the oceans, is today another bright and shiny new idea, needing more supporters, that attempts to overcome the catch of overfishing. BTF uses past ecosystems as policy goals for the future. It harnesses an understanding of ecosystem processes and whole ecosystem simulation to insight into the human dimension of fisheries management. It includes new methods, reported in substance here, for quantitative descriptions of past ecosystems, for designing fisheries that meet criteria for sustainability and responsibility, and to evaluate the costs and benefits of fisheries in restored ecosystems. Alternative policy choices involve different trade-offs between conservation and economic value. Automated searches maximise values of objective functions, and the methodology includes analyses of model parameter uncertainty. Participatory workshops attempt to maximise compliance by fostering a sense of ownership among all stakeholders. Some challenges that have still be met include improving methods for quantitatively describing the past, reducing uncertainty in ecosystem simulation techniques and making policy choices robust against climate change. Critical issues discussed here include whether past ecosystems make viable policy goals, and whether desirable goals may be reached from today?s ecosystem.                                                          1 Bible, Luke 5: 1-11.   This report, covering new and adapted methodology devised to support Back-to-the-Future analyses and policy procedures, has been rather a long time in the making. This foreword has been in draft for over a year, and, in the event, turns out to be the last Director?s foreword (of 40 since 1993) that I have written for Fisheries Centre Research Reports .   The Fisheries Centre Research Reports  series publishes results of research work carried out, or workshops held, at the UBC Fisheries Centre. The series focusses on multidisciplinary problems in fisheries management, and aims to provide a synoptic overview of the foundations, themes and prospects of current research. Fisheries Centre Research Reports  are distributed to appropriate workshop participants or project partners, and are recorded in the Aquatic Sciences and Fisheries Abstracts . A full list appears on the Fisheries Centre's Web site, www.fisheries.ubc.ca. Copies of the reports are sent to meeting participants, and all papers are available for free download from our web site as PDF files. Paper copies of reports are available on request for a modest cost-recovery charge.   Tony J. Pitcher Professor of Fisheries  Director 1993- 2 0 0 3,  UBC Fisheries Centre  The D r a u g h t of Fishes , painted in 1515 by Raphaello Sanzio (1483-1520). Towards the end of his short life, Raphael moved briefly but spectacularly to Rome, where he initially helped to redecorate apartments vacated by the unsavory and detested Borgia Pope (Alexander 6 th). Ten full-size cartoons were commissioned from Raphaello by the urbane Medici Pope, Leo 10th, as designs for tapestries to hang in the Sistine chapel. The subsequent tapestries by Pieter van Aelst in Brussels (1519) were revolutionary in their use of light and shade, and can be seen today in the Vatican Museum. Note that the Vatican is visible on the lake shore, transposed by virtue of our painter?s benefactor to the shores of the Sea of Galilee in biblical times. Cranes in the foreground symbolize vigilance, while seagulls allude to the apostasy of the former regime. Victoria & Albe rt Museum, London , tempura on canvas, 399 x 440cm .  Back to the Future Methodology, Page 4   INTRODUCTION TO THE METHODOLOGICAL CHALLENGES IN ?B ACK-T O-T HE-F UTURE?  RESEARCH    Tony Pitcher Fisheries Centre, UBC    ABSTRACT   Many of the concepts in the Back-to-the-Future research process are new and so new methods, and modifications to existing methods, have been required for analysis, modelling and prediction of marine ecosystems and their fisheries.  Methodological issues have been encountered in describing and modelling past ecosystems, in devising an ecosystem approach to determine sustainable fisheries, in devising a rational basis for choosing appropriate restoration goals and in attempting to maximise consent and compliance through encouraging a sense of ownership of policy by stakeholders. This paper summarises these issues, and introduces each of the new methods later to be described in detail in papers in this report. Results from case studies of the BTF process are contained in a separate report.     INTRODUCTION  Back-to-the-Future (BTF) is a science-based restoration ecology aimed at the creation of truly sustainable food and wealth from capture fisheries and aquatic ecosystems (Pitcher e t al.  1999). The fisheries are embedded in aquatic ecosystems that, by quantitative analysis and with the agreement of stakeholders, trade-off wealth                                                            Pitcher, T.J. (2004) Introduction to the methodological challenges in ?Back-To-The-Future? research. Pages 4?10 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.  and food with a specified degree of retention of their unexploited biodiversity and trophic structure. Hence, BTF uses past ecosystem states as candidates for adoption as policy goals for the future (Figure 1, Pitcher 2001). In practice, the policy goals are subject to a number of practical constraints from species, habitat and climate changes (Haggan e t al.  2003). The six logical steps in the BTF process are outlined in Table 1 (Pitcher 1998, 2004a, Pitcher e t al.  2003).   Many new concepts have been developed as a part of the BTF research sponsored by Coasts Under Stress (CUS), and so it is not surprising that existing methods have not been adequate to express them. This report contains descriptions of the new methods that have been developed, along with papers of a general methodological nature from CUS research partners. BTF case studies and results are the subject of a separate publication.  The new methods can be divided into four groups: methods required to describe and model past ecosystems, ecosystem-based methods to determine sustainable fisheries, methods that set out a rational basis for choosing appropriate ecosystem restoration goals, and finally, practical techniques that attempt to secure compliance and consent through participation.  1 .   METHODS OF MODELLING PAST ECOSYSTEMS  The present-day ecosystem is represented by mass-balance and dynamic simulation modelling (at present using Ecopath  with Ecosim; Walters Figure 1.  Diagram illustrating the ?Back to the Future? concept for the restoration of past ecosystems. Triangles at left represent a series of ecosystem models, constructed at appropriate past times, where vertex angle is inversely related and height directly related to biodiversity and internal connectance. Time lines of some representative species in the models are indicated, where size of boxes represents relative abundance and solid circles represent local extinctions. Sources of information for constructing and tuning the ecosystem models are illustrated by symbols for historical documents ( p a p e r sheet symbol ), data archives ( tall  d ata sheet ), archaeological data ( trowel ), the traditional environmental knowledge of indigenous Peoples ( open balloons ) and local environmental knowledge ( solid balloons ). Alternative future ecosystems, restored ?Lost Valleys?, taken as alternative policy goals, are drawn to the right. (Diagram modified from Pitcher et al.  1999 and Pitcher 2001.) Page 5,  Fisheries Centre Research Reports 12(1), 2004  e t al.  1997) using techniques that have received a degree of approval by marine ecologists (e.g. Whipple e t al.  2000). This modelling is a far from trivial task, especially if fitting to time series of fisheries and survey data is undertaken. Moreover, highly migratory species like salmon, that exhibit lifetime shifts between different ecosystems, are included in ecosystem models with difficulty (see Martell 2004, this volume).   Models for past ecosystems are assembled using scientific archival data, archeological data, historical information, and local and traditional environmental knowledge. Scientific data derive mainly from published scientific papers, although material from unpublished reports and archives can often be valuable. Archaeological data has a similar set of sources (see Orchard and Mackie 2004, this volume). Historical information is gathered mainly from relevant books, letters, trade accounts and other historical documents, although, unlike science and archaeology, where searchable databases are the norm, finding and locating historic material can be quite hard. In some cases, translations are required. Local and traditional environmental knowledge, on the other hand, is rarely published and often has to be derived largely from oral sources through interviews and discussion held in coastal communities (see papers by Ainsworth, Simeon and Pitcher et al. 2002c, this volume).  Once found, all these data have to be assembled into a relational database together with evaluations of its scope and quality, to ease retrieval of relevant information for the models. (The CUS BTF project database will be described by Erfan in a later report.) Even so, a significant task is systematising the way in which information is collated for use in the models.  The reason is that, once documented, information has to be expressed in a form that can be used in building ecosystem model structure, in setting parameters, or in shaping dynamic responses to changes. Although presence and absence of a species is easily dealt with, the models require us to know actual biomasses, size and growth parameters, and items in the diet.   Information about the local fisheries, with analyses and surveys, and about local aquatic fauna and flora is relatively easily found, especially as an output of ?science workshops? comprised of research partners and local scientists with expert knowledge of the area and the taxonomic groups. One of the principal Table 1.  Stages in the ?Back to the Future? process for the restoration of fisheries and aquatic ecosystems.  Workshop phases are in italics. Modified from tables in Pitcher (1998) and in Pitcher et al.  (2003).  Stage Goals Steps 1 Model construction of past  and present aquatic ecosystems Assemble present-day mass-balance and ecosystem simulation model Assemble preliminary past models using compatible structure and parameters Search and score data archives, historical documents, archeological information Workshop of scientists knowled ge ab l e about system  Interviews for traditional environmental knowledge, and for fisher?s opinions and behaviour Assemble and standardize historical and interview scores database  Assemble and test suite of ecosystem simulation models  Workshop of scientists and managers to compare and stand ardise ecosystem models  (may need to return to this step after preliminary results) 2 Evaluation of ecological, economic and social benefits that could be gained from each system Determine sustainable fisheries with which to exploit reconstructed ecosystems (?Opening the Lost Valley?) Challenge model scenarios with uncertainty Challenge model scenarios with climate changes Ecosystem simulation scenarios under anticipated conditions Workshops to evalu ate policies with fishing communities Critique and evaluate ?Lost Valley? fisheries scenarios and adjust where required  Searches for optimal mix of fishing gears Determine Optimum Restorable Biomasses (ORBs) for ?Lost Valley? scenarios Quantify risks to ORB policies 3 Choice of system that maximises benefits to society Identify trade-offs among economic, ecological and social criteria  Ecological economic evaluations including analysis of risks Workshops with communities, manage rs, scientists, NGOs, and government  Participatory policy choice 4 Design of instruments to achieve this policy goal Model exploration of MPAs, effort controls, acceptable quotas, times and places for fishing  Evaluation of costs of the desired management measures 5 Participatory choice of instruments Community and stakeholder discussion and choice of instruments to achieve policy goals Workshops with communities, manage rs, scientists, NGOs, and government  Participatory policy choice 6 Adaptive management: implementation and monitoring  On-going monitoring, validation and improvement of model forecasts using adaptive management procedures On-going participatory guidance on instruments and policy goals Back to the Future Methodology, Page 6   problems here is data that has been gathered on either a very small or a very large scale compared to the area of focus (see Haggan 2004, this volume). Another issue often requiring a lot of work is the concordance of measurement units, since specialists on different taxa often work in very different fields. Scientists who generously make the relevant information available, often from a lifetime?s work on a group of organisms, are encouraged to publish a paper in one of the BTF reports so that they retain a recognised ownership of material that otherwise would easily vanish into model simulations.   For the CUS BTF project in Newfoundland and British Columbia, the output from an extensive process of consultation with the science community has been presented in detail in four reports (Ainsworth e t al.  2002, Pitcher e t al.  2002a, 2002b, Heymans 2003), where information essential to the modelling process, such as geographical scope, biomass, relative fishing mortalities, diets and other ecological information are assembled.  In the absence of local publications on these topics, as is often the case, interviews, conducted under suitable partnership agreements, are the best way to gather LEK and TEK information for use in the modelling. Ainsworth (2004, this volume) reports on method s used in interviews designed especially to gather material that can be used in ecosystem modelling for the CUS BTF project. A report on a community workshop is presented in Pitcher e t al.  (2002c, and see Power e t al.  2004, this volume).  For ease of comparison, the structure of the past and present ecosystem models should be similar, although of course biomasses and fluxes can be vastly different. Global extinctions of species cause some technical difficulties in modelling. When species have gone locally extinct (?extirpation?), this creates some difficulties (see Pitcher 2004d, this volume). Some practical solutions found in the CUS project are presented by Heymans and Pitcher (2004, this volume).    Another frequent problem is that reconstructions of the ancient past may suggest the presence of large numbers of top predators that are too numerous to be supported by what are thought to be realistic levels of forage organisms (Pitcher 2004c, this volume).   Representing changes in ecosystem structure over long periods of time represents a major challenge. Clearly, the effect of shifts in climate has to be accommodated in the forecasts as much as possible (see Pitcher and Forrest 2004, this volume). But early periods of depletion by human exploitation also had significant impacts on ecosystem structure and function. Recent reconstruction work by Jackson e t al.  (2001) shows what may be possible in this respect.   Ideally, the timing of the series of ecosystem models for BTF may depend on the locality, the dawn of quantitative documentary evidence, and major shifts in resource and ecosystem history such as the introduction of new fishing gears, damming of rivers and collapses of fish stocks. But because of the large amount of work involved in drawing up each ecosystem model, the gaps in time between a series of BTF models may be quite large. So an ideal choice of the time snapshots to use as BTF models is generally constrained by the resources available for the research. This raises a significant methodological problem in that failure to cover important changes that occurred within these time gaps can prejudice the choice of appropriate policy goals at the end of the BTF process. In the event, the choice of the time periods to model in a BTF analysis is something of a compromise.   In many cases, additional informative models might be drawn up for pre-modern humans in the late Pleistocene post-glacial era. Although such ancient ecosystems would be unlikely to ever become practical policy goals, they have the advantage of providing a ?pristine? baseline against which all more recent changes might be assessed. In fact, for some areas of the world only recently colonised by Europeans, such as Australia, New Zealand and the Pacific coast of America (Diamond 1997), models of ?pre-contact? ecosystems may serve this purpose well.   In models of the distant past, the estimation of the size and impacts of ancient fisheries presents many problems. Although the history of fishing technology is quite well known from archaeology and from traditional knowledge, its likely fishing power may be estimated, and ancient diets may be calculated, nevertheless, the size of the human populations that engaged in fishing is often hard to assess. Estimates of ancient human population sizes are often the subject of controversy among archaeologists and anthropologists. In one of the recent volumes from this CUS BTF project, Heymans (2003) presents an example of what may be done with ancient diets and fisheries. It is emphasised, however, that the aboriginal fisheries in the ecosystems are described only to provide an accurate picture of the ancient ecosystem, and they would not necessarily be chosen for a future restoration policy. This issue Page 7,  Fisheries Centre Research Reports 12(1), 2004  is discussed in more detail below.   Finally, many of these problems may be eased if we were able to run a past model forward to simulate its change into a more recent ecosystem. Performing this using Ecosim requires a great deal of data on fisheries and climate (see Stanford 2004, this volume), but has been possible for some ecosystems that have undergone rapid change, such as the Gulf of Thailand (Christensen 1998). Unfortunately, to date, attempts to do this with both BC and Newfoundland ecosystem models have been only partially successful (Heymans 2003). Heymans and Pitcher (2004, this volume,) summarise the construction of models of the past  in relation to the ecosystems researched for the CUS BTF project.  2 .  METHODS FOR DEVISING        SUSTAINABLE FISHERIES  A marine ecosystem restored to some semblance of its past state might be thought of as a ?Lost Valley?1: an ecosystem discovered complete with all of its former diversity and abundance of creatures (Pitcher 2004b, Pitcher e t al.  2004). The BTF process aims to describe a series of such ?Lost Valleys? as a set of potential restoration goals.    Since a ?Lost Valley? has to be fished sustainably, we have to ask how this might be achieved? Using the same fishing fleet as today in order to fish a restored ecosystem is generally not a viable option since massive depletion would soon ensue. Nor is it realistic to expect the fishing gear and methods of former times, including those of aboriginal fisheries, to be re-employed. Of course, some former fisheries might have attractively low by-catch, operating costs or ease of construction and use, so it is evident that some rational criteria for the selection and operation of sustainable fisheries need to be devised. The BTF process aims to devise such criteria. For example, a candidate fishery designed with the criteria could be challenged by assessing its conformity with the FAO Code of Conduct for Responsible Fisheries (FAO 1995) using a rapid appraisal technique (Pitcher 1999).  After applying the criteria in this way to design an ?ideal fishery? for a particular location, ecosystem simulations (using the Ecosim policy search interface; Walters e t al.  2002) can be used to find the relative fishing mortalities that should be used by each gear type in the ?ideal? fishery to                                                           1 We are grateful to Dr Daniel Pauly for suggesting this term in 2001. (See Pitcher e t al . 2004, this volume) achieve sustainable catches over a long time period, usually 100 years.  In addition, we may seek to challenge these results with climate changes that might realistically be expected for the locality in question, and in the face of uncertainty in the simulation modelling (see papers by Ainsworth et al.,  Pitcher and Forrest 2004, this volume ).   3.   METHODS FOR CHOOSING ECOSYSTEM        RESTORATION GOALS  Once we have snapshot of what a set of alternative restored ecosystems, complete with their sustainable fisheries, might look like, the remaining issue to solve is to find an objective way to choose a rational policy goal from among them. This may be done by comparing the benefits that will accrue to society from each alternative future represented by a fished ?Lost Valley? ecosystem. In order to show the full range of options that may be considered, included in this process is the present day ecosystem (albeit with fisheries designed to be sustainable), and perhaps an ecosystem even further depleted (Figure 1).  One fundamental way to evaluate the benefits of alternative restored ecosystem is the net present economic value of their fisheries, information that is readily estimated from the Ecosim simulations mentioned above.  A modification more in accord with ecological economics is to estimate present value using intergenerational equity calculations (see Ainsworth and Sumaila 2004a, this volume).  Purely economic considerations, however, are rarely considered sufficient for modern policy making. Therefore, in the BTF process we also estimate the relative impacts on biodiversity (see papers by Ainsworth and Pitcher, Heymans, and Chueng and Pitcher 2004, this volume) and social factors such as the likely number of jobs and their diversity (see Ainsworth and Sumaila 2004b, this volume). For a proper evaluation, the costs of restoration have to be considered alongside the benefits. This part of the evaluation system is not yet completed for the CUS BTF research and the issue is discussed further below.  4.   PARTICIPATORY AND ADAPTIVE         POLICY IMPLEMENTATION  Implementing a policy goal that has been chosen using any science-based process, including BTF, is, of course a much more difficult matter. When fishing communities and other essential stakeholders actively participate in the policy Back to the Future Methodology, Page 8   agenda, compliance and consent may be high (Hart and Pitcher 1998). For example, Haggan (2000) identifies 4 elements as critical to participation: recognition of the scope of the problem and our collective responsibility whether fishers, scientists, managers or policy makers; respect for different systems of knowledge; agreement to share knowledge in the interest of conservation and restoration; and, commitment to share in the benefits of restored systems.  In BTF the aim is to encourage a greater chance of success because a sense of ownership of the process is fostered and developed from the earliest stages of the work. The BTF process includes community participation in building models of the past (see Simeon 2004, this volume), in the choice of sustainable fisheries and in the evaluation of the costs and benefits of alternative restoration goals (see Power e t al . 2004, this volume, Pitcher e t al.  2002c). Moreover, the cognitive maps shaped by awareness of past abundance and diversity develop in BTF process may serve to assist consent and compliance with a restoration agenda (Pitcher and Haggan 2003). Participatory elements that are integral to three phases of the BTF process are summarised in Table 2.   Once management aims to make progress towards a specific BTF past state, the use of quantitative adaptive management (e.g., Walters 1986) is the wisest course, in order to try to avoid the disasters that a changing environment and imperfectly understood ecology can throw at any management plan.  CONCLUSIONS  Policy goals that reflect an approach of restoration ecology may be chosen using the BTF procedures outlined here and presented in more detail in subsequent papers in this volume. But a number of methodological challenges raised by BTF remain unresolved at this stage. The way in which historical information is turned into inputs for the ecosystem modelling could do with considerable improvement. Better semi-quantitative assessment of relative biomass, diet and sizes needs to be devised. Our historical data need a more rigorous and replicable transduction into the quantitative data needed for modelling.  For the CUS BTF research, a first step in this respect will be published by Ainsworth (2004) in the forthcoming ?results? volume.   BTF has an advantage in not relying exclusively on complex stock assessment (Walters 1998), although such work can help in the tuning of the ecosystem models. At present, the quantitative ecosystem modelling used for BTF to date relies almost exclusively on Ecopath  and Ecosim techniques. Yet many of the assumptions in this modelling system, while plausible, remain unvalidated. Of especial concern are the Ecosim ?vulnerability? parameters, to which specific results often appear very sensitive (see Ainsworth 2004, this volume). Moreover, these parameters not only shape predator-prey interactions (which  they do in an entirely credible fashion for a former evolutionary ecologist), but also pre-determine the scope for further biomass growth in relation to current abundance. For any series of ?time-shot? BTF ecosystem models, this creates a conflict between the need to compare the outcomes of various fisheries options while other parameters remain fixed, and setting parameters correctly for biomasses that were closer to unexploited levels in the past. These modelling problems have yet to be resolved.  As pointed out by Heymans and Pitcher (2004, this volume), past ecosystem models may resemble the actual past as a Picasso resembles reality. An important question is whether our comparative restoration policy scenarios can be made robust against such distortions. A deeper insight of the dynamics of ecosystems under change will be required before we can answer this question.   A broad participation by scientists, researchers, stakeholders, government, managers, NGOs and the public is critical for the success of any restoration policy that might be set up under the BTF banner. Yet we have barely scratched the surface of the deep issues raised by the need for this level of participation in the BTF policy searches and analyses. Nor have we enough experience of asking fishing communities to choose what kind of future they might wish to aim for. We are not yet sure how to convey the uncertainty in our work, which to many may seem arcane. Perhaps ?barefoot ecologists?, the equivalent of rural development generalists for Table 2 . Summary of integral participatory elements from local fishing communities in the BTF process. TEK = traditional ecological knowledge, LEK = local ecological knowledge. All stages are intended to work in concert with science-based decision making.   TEK: in model construction Model development phase LEK: in model construction  TEK/LEK/Community: model credibility and validation   Community choices: how to rebuild   Policy development phase Community choices: choice of best benefits to cost ratio for policy goal  Community choices: choice of acceptable and sustainable fisheries Operational phase Consent and compliance  Monitoring Page 9,  Fisheries Centre Research Reports 12(1), 2004  fisheries, as envisaged by Jeremy Prince might be able to help (Prince 2003).   The intention is to give BTF players a clear cognitive map of a future ecosystem that resembles one from the past, to which all may agree and aspire (see Pitcher 2004a). And so, to date, BTF analysis has not considered the costs of achieving each restoration, because this may divert attention from that ultimate goal. (Although it is noted that it may be logically argued that the true policy goal cannot be known until a full cost-benefit analysis is performed.)  The fundamental problem here is that estimating the costs of restoration may depend on precisely what techniques are adopted, and the actual instruments may themselves generate conflict (for example, MPAs set up adjacent to a traditional fishing community ? see Lucas 2004, this volume -  or reduced quotas for some sectors as fisheries are modified to become more sustainable). Again, these important issues remain to be resolved.  The logistics of mounting a quantitative, robust and credible BTF analysis are considerable. The sheer cost, in money and time, of assembling an inter-disciplinary team to gather, validate and analyse the historical, archaeological and ecological information needed for BTF is formidable. Moreover, like other synoptic work involving whole ecosystems, the scope of BTF work appears to be far outside the capacity of one graduate student thesis. In this project, it has therefore been gratifying to see modest financial support from Coasts under Stress  augmented by enormous enthusiasm and commitment from the team of graduate students, postdoctoral researchers and research partners who have helped with the research reported here.   REFERENCES  Ainsworth, C. (2004) How we ca rried out ?Back-to-the-Future? Community Interviews. Pages 116?124 in Pitcher, T.J. (ed.) Back to the Future: Ad vances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. (2004) Estimating  the Effects of Prey-predator Vulnerability Settings on Ecosim's Dynamic Function. Pages 45?47 in Pitcher, T.J.  (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. and Pitcher, T.J. (2004) Modifying Kempton?s Biodiversity Index for Use with Dynamic Ecosystem Simulation Models. Pages 91?93 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. and Sumaila, U.R. (2004a) Economic Valuation Techniques for Back-To-The-Future Optimal Policy Searches. Pages 104-107 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.  Ainsworth, C. and Sumaila, U.R. (2004b) An Employment Diversity Index Used to Evaluate Ecosystem Restoration Strategies. Page 108 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C., Heymans, J.J., Pitcher, T.J. and Vasconcellos, M. (2002) Ecosystem models  of Northern British Columbia for the time periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4): 41 pp. Ainsworth, C., Heymans, J.J. and Pitcher, T.J. (2004) Policy Search Methods for Back to the Future. Pages 48?63 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Cheung, W-L. and Pitcher, T.J. (2004) An  Index Expressing Risk of Local Extinction for Use with Dynamic Ecosystem Simulation Models. Pages 94?102 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Christensen, V. (1998) Fishery induced changes in a marine ecosystem: insight from models of the Gulf of Thailand. Journal of Fish Biology 53 (A): 128-142.   Diamond, J. (1997 ) Guns, Germs, and Steel: The Fates of Human Societies. Random House, London, UK. 480pp. FAO (1995) Code of Conduct for Responsible Fisheries. FAO, Rome, 41pp.  Haggan, N. (2000) Back to the Future and Creative Justice: Recalling and Restoring Forgotten Abundance in Canada?s Marine Ecosystems. Pages 83-99 in Coward, H., Ommer R. and Pitcher, T. (eds) Just Fish: Ethics in the Canadian Coastal Fisheries. ISER Books, St. Johns, Newfoundland. 304pp. Haggan, N., Pitcher, T.J. and Sumaila, U.R. (2003). Back to the Future in the Strait of Georgia.  In Droscher, T. and  Fraser, D.A. (eds) Georgia Basin/Puget Sound Research Conference Proceedings, Vancouver, BC. ( in press ). Hart, P.J.B. and Pitcher, T.J. (1998) Conflict, cooperation and consent: the utility of an evolutionary perspective on individual human behaviour in fisheries management. Pages 215-225 in Pitcher, T.J., Hart, P.J.B. and Pauly, D. (eds) Reinventing Fisherie s Management. Chapman and Hall, London, UK. 435pp. Heymans, J.J. (ed.) (2003 ) Ecosystem models of Newfoundland and southeastern Labrador: additional information and analyses for ?Back to the Future?. Fisheries Centre Research Reports 11(5): 79pp. Heymans, J.J. (2004) Evaluating the Ecological Effects on Exploited Ecosystems using Information Theory. Pages 87?90 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Heymans, J.J. and Pitcher, T.J. (2004) Synoptic Methods for Constructing Models of the Past. Pages 11?17 in Pitcher, T.J. (ed.) Back to the Futu re: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 157 pp.. Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J., Bradbury, R.H.,  Cooke, R., Erlandson, J., Estes, J.A.,  Hughes, T.P., Kidwell, S., Lange, S.B.,  Lenihan, H.S.,  Pandolfi, J.M.,  Peterson, C.H., Steneck, R.S.,  Tegner, M.J. and Warner, R.R. (2001) Historical Overfishing and the Recent Collapse of Coastal Ecosystems. Science 293: 629-637. Lucas, S. (2004) Aboriginal Valu es. Pages 114?116 in Pitcher, Back to the Future Methodology, Page 10   T.J. (ed.) Back to the Futu re: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Martell, S. (2004) Dealing with Migratory Species in Ecosystem Models. Pages 41?44 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Orchard, T. and Mackie, Q. (2004) Environmental Archaeology: Principles and Case Studies. Pages 64?73 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J. (1998) ?Back To The Future?: a Novel Methodology and Policy Goal in Fisheries. Pages 4-7 in Pauly, D., Pitcher, T.J. and Preikshot, D. (eds) Back to the Future: Reconstructing the Strait of Georgia Ecosystem. Fisheries Centre Research Reports 6(5): 99pp. Pitcher, T.J. (1999) Rapfish, A Rapid Appraisal Technique For Fisheries, And Its Application To The Code Of Conduct For Responsible Fisheries. FAO Fisheries Circular No. 947: 47pp. Pitcher, T.J. (2001) Fish eries Managed to Rebuild Ecosystems: Reconstructing the Past to Salvage The Future. Ecological Applications 11(2): 601-617. Pitcher, T.J. (2004a) Back To  The Future?: A Fresh Policy Initiative For Fisheries And A Restoration Ecology For Ocean Ecosystems. Phil.Trans.Roy Soc. ( in press ). Pitcher, T.J. (2004b) Why we have  to open the lost valley: criteria and simulations for sustainable fisheries, Pages 78?86 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J. (2004c) What was the structure of past ecosystems that had many top predators? Pages 18?20 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 157 pp.  Pitcher, T.J. (2004d) The problem of extinctions. Pages 21?28 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 157 pp. Pitcher, T.J. and Haggan, N. (2003) Cognitive maps: cartography and concepts for ecosystem-based fisheries policy. Pages 456-463 in Haggan, N., Brignall, C. and Wood, L. (eds) Putting Fishers' Knowledge to Work. Fisheries Centre Research Reports 11(1): 540pp. Pitcher, T.J. and Forrest, R. (2004) Challenging ecosystem simulation models with climate change: the ?Perfect Storm?. Pages 29?38 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J., Haggan, N., Preikshot, D. and Pauly, D. (1999) ?Back to the Future?: a method employing ecosystem modelling to maximise the sustainable benefits from fisheries. Pages 447-465 in Ecosystem Considerations in Fisheries Management, Alaska Sea Grant Anchorage, USA. Pitcher, T.J., Vasconcellos, M., Heymans, S.J.J., Brignall, C. and Haggan, N. (eds) (2002a)  Information Supporting Past and Present Ecosystem Models of Northern British Columbia and the Newfoundland Shelf.  Fisheries Centre Research Reports 10(1): 116 pp. Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) (2002b) Ecosystem models of Newfoundland for the time periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 74 pp. Pitcher, T.J., Power, M. and Wood, L. (eds) (2002c) Restoring the Past to Salvage the Future: Report on a Community Participation Workshop in Prince Rupert, BC. Fisheries Centre Research Reports 10(7) : 55 pp. Pitcher, T.J., Heymans, S.J.J., Ainsworth, C., Buchary, E.A., Sumaila, U.R. and Christensen, V. (2004) Opening The Lost Valley: Implementing A ?Back To Future? Restoration Policy For Marine Ecosystems and Their Fisheries. In Knudsen, E.E., MacDonald, D.D. and Muirhead, J.K. (eds) Fish in the Future? Perspectives on Fisheries Sustainability. American Fisheries Society, Bethesda, MD, USA. ( in press ). Power, M.D., Haggan, N. and Pitcher, T.J. (2004) The Community Workshop: how we did it and what we learned from the results. Pages 125?128 in Pitcher, T.J. (ed.) Back to the Future: Ad vances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Prince, J. (2003) The Barefoot Ecologist Goes Fishing. Fish and Fisheries 4(4): 359-371. Simeone, W. (2004) How traditional knowledge can contribute to environmental research and resource management. Pages 74?77 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Stanford, R. (2004)  Tuning Ecosystem Models to Past Data.  Pages 39?40 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Sumaila, U.R. (2004) How do we value the restoration of past ecosystems? Page 103 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Sumaila, U.R. (2004) Incorporat ing First Nations values into fisheries management: A proposal for discussion. Pages 112?113 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Sumaila, U.R, Pitcher, T.J., Haggan, N. and R. Jones, R. (2001) Evaluating the Benefits from Restored Ecosystems: A Back to the Future Approach. In R.S. Johnston and A. L. Shriver (eds.): Proceedings of the 10th International Conference of the International Institute of Fisheries Economics & Trade, Corvallis, Oregon, USA, 2000. Walters, C.J. (1986) Adaptive Management of Renewable Resources. MacMillan, New York, USA. 374 pp. Walters, C.J. (1998) Designing fisheries management systems that do not depend on accurate stock assessment. Pages 279-288 in Pitcher, T.J., Hart, P.J.B. and Pauly, D. (eds) Reinventing Fisheries Management. Kluwer, Dordrecht, Holland. 435pp. Walters, C.J., Christensen, V. and Pauly, D. (1997) Structuring dynamic models of exploited ecosystems from trophic mass-balance assessment. Reviews in Fish Biology and Fisheries 7: 139-172. Walters, C.J., Christensen, V. and Pauly, D. (2002) Searching for Optimum Fishing Strategies for Fishery Development, Recovery and Sustainability. Pages 11-15 in Pitcher, T.J. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries.  Fisheries Centre Research Reports 10(2), 156pp.  Whipple, S.J., Link, J.S., Garrison, L.P. and Fogarty, M.J. (2000) Models of predation and fishing mortality in aquatic ecosystems. Fish and Fisheries 1: 41-72.   Page 11,  Fisheries Centre Research Reports 12(1), 2004  SYNOPTIC METHODS FOR   CONSTRUCTING ECOSYSTEM  MODELS OF THE PAST   Johanna J. Heymans and Tony Pitcher Fisheries Centre, UBC    ABSTRACT   This paper gives a brief description of the steps that need to be taken when constructing a model of a past ecosystem. It is important to know what question is going to be asked of the model, as that affects all subsequent steps. To construct a model of the past it is important to know the area, time periods, species to include, what data is available, how to handle the calculation of Ecopath  parameters for species that have a different age structure from the present day, and finally how to make and test the assumptions needed in such a endeavor. Assumptions that have to be made lead to uncertainty, which may be examined using the emergent properties of the ecosystem.     INTRODUCTION  Models of the past are constructed for comparison with present day models.  They provide baselines for the emergent properties of these ecosystems. For the Coasts under Stress  Back-to-the-Future project we aim to assess the effect of long term trends in the social and environmental health of regional ecosystems on the environment and on human health. The question asked was:   ?How can local ecological and scientific knowledge help us to understand changes in environmental, community, and individual health in ways that will help develop better strategies for future ecological recovery??   In this paper the methodology of constructing models of the past will be illustrated by using two examples from the CUS BTF project: Northern British Columbia (including the Hecate Strait and Queen Charlotte Sound), and Eastern Newfoundland /Southeastern Labrador (NAFO Div. 2J3KLNO) (Figures 1 and 2), as defined in Pitcher et al. (2002) and Ainsworth e t al.  (2002).                                                               Heymans, J.J. and Pitcher, T.J. (2004) Synoptic Methods for Constructing Models of the Past. Pages 11?17 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp . METHODOLOGY USED IN CONSTRUCTING MODELS OF THE PAST  The steps involved in the construction of past models include: 1) defining the system, 2) choosing the time periods you want to model, 3) data gathering (on catch and biomass mostly), 4) which species to include, considering extinctions and incorporating species that are not well studied even at present, 5) calculating the energetic ratios for models of the past, and finally 6) making other assumptions for species where we have no better information.  Defining the system  To define a model of the past you have to define the boundaries of your ecosystem. The chosen system should preferably be contained in a natural or oceanographic feature, with a single climate. Generally a larger area is preferable as it increases the chances of having any historical information. By and large the international jurisdiction of the area does not matter, as the jurisdiction would have changed over the course of time. Usually the ecosystem is defined based on current knowledge of the system. For instance, in both the northern BC and Newfoundland models we defined the system based on current knowledge and more recent models constructed for these ecosystems (Heymans and Pitcher 2002a, and Ainsworth e t al.  2002).   In Newfoundland the area chosen (Figure 1) was similar to that used in prior models of the system (Bundy e t al. , 2000, and Heymans and Pitcher 2002). The area chosen included the DFO-NAFO divisions 2J3KLNO and incorporated the Labrador Current and the Grand Banks, as they Figure 1.  NAFO divisions of the east coast of Canada, showing the areas used in the CUS-BTF study (Divisions 2J, 3K, 3L, 3N and 3O).  Back to the Future Methodology, Page 12   are interconnected and for some species they are managed as a unit. The area chosen for the northern BC  model included both Hecate Strait and Queen Charlotte Sound (Figure 2). The area was chosen to answer particular questions, thus some inshore marine waters were included in the model area as salmon had to be included.  Choosing time periods   When choosing a time period it is advisable to choose times pre- and post pivotal gear changes or exploitation levels. Time periods pre- and post the start of formal recorded data are often important, and, finally, the time periods depend on the questions that are asked. For the CUS BTF project the question was to assess the longer term trends in the health of local and regional ecosystems.   The time periods chosen for the northern BC model were 1750, 1900, 1950 and the present day.  (Ainsworth et al.  2002). The 1750 model was chosen as it was pre-European contact, while 1900 was prior to large scale commercial fisheries and the resumption of whaling. By 1900 the number of First Nations people were drastically reduced, which had a positive effect on Steller sea lions, although the sea otters were already locally extinct by that time. The 1950 model incorporates the large scale purse seine fishery for herring, the collapse of pilchard and the start of DFO?s catch data series, while the present day (2000) model was initially based on a model constructed by Beattie (2001) but using more recent data.   In Newfoundland the time periods chosen were 1450, 1900-1905, 1985-1987 and 1995-1997 (Vasconcellos e t al.  2002). The 1450 model was pre-European contact, 1900 was prior to the large scale Grand Banks fisheries, but after the large scale whaling that took place in that area. The 1985-1987 model was based on the model constructed by Bundy e t al.  (2000) and was prior to the groundfish collapse based on reliable data, while the 1995-97 model was after the groundfish collapse but did not have the same quality of data as the 1985-87 model.  Data gathering  Information on past abundances, catches, etc. are not easy to obtain in normal scientific literature. However, building models of the present day gives a blueprint for models of the past. Information on past abundance and catches are generally found by looking at historical documents, or by using expert opinion of fisheries biologists on virgin population of key species. It is also possible for marine mammal or seabird experts to make ?educated guesses? on how many animals must have been in the system at a certain time. There are also archaeological and anthropological information available to assist with presence/absence of species, as well as the utilization of marine species by First Nations or European settlers. Finally, Traditional and Local Ecological Knowledge (TEK/LEK) can also be useful for models that are within their time frame, i.e. models that go back about 50 years.   Data for building the models of Newfoundland (Heymans and Pitcher 2002a, 2002b, Pitcher e t al.  2002a) were obtained mostly from historical documents and books that summarize changes in that ecosystem: (Lescarbot 1914, Howley 1915, Lewis and Doutt 1942, Wright 1951, Mercer 1967, Mowat 1984, Reeves et al.  1985b, Crosby 1986, Montevecchi and Tuck 1987, Cushing 1988, Pastore 1992, Hewitt 1993, Ryan 1994, Pope 1995, Turgeon 1995, Marshall 1996, Lear 1998, Hiller 2001, Cridland 1998, Whitridge 2001). For the calculation of the pristine population and catch of cod, a reconstructed time series obtained from (Hutchings and Myers 1995) was useful. The Internet was useful for obtaining information on historical populations. In Newfoundland the Heritage Site of Newfoundland and Labrador ( www.heritage.nf.ca) contains information on the fishing industry, First Nations and European Hecaa te Straa it QQ uu eenn  Chh aa rll otte Souu nn d Figure 2.  Map of the West coast of Canada showing the Hecate Strait and Queen Charlotte Sound, both in the study area. Page 13,  Fisheries Centre Research Reports 12(1), 2004  settlement in the area.  Data for building the models of northern BC (Ainsworth e t al.  2002) were obtained from historical documents, such as Hudson?s Bay Company records (Hammond 1993), as well as other historical records (Lord 1866, Chambers 1872, Anderson 1879, Dawson 1880, Mowat 1886, ANON 1892, Osgood 1901, Freeman 1904, Babcock 1910, Alexander 1912, Thompson 1916, Newcombe 1917, Muir 1935, Carrothers 1941, Akrigg 1975, Kenyon 1975, Jacobsen 1977, DeWhirst 1982, Vancouver 1984,  Reeves et al.  1985a, Webb 1988, Gregr 1999, 2002, Mackie  et al. , 2001). Data for building the model of 1950 was also obtained from interviews done in Prince Rupert and surroundings (s ee report in Pitcher et al.  2002). Both historical data and interview data for this system was collected in a database searchable on the web at: www.fisheries.ubc.ca/projects/btf/  (see Erfan, results volume).  Data on sport fish catches are rarely recorded in official ?catch statistics?, but can be considerable (e.g., Pitcher 2003, Pitcher and Hollingworth 2002). In Northern BC so me estimates have been made using interview and other techniques (Forrest 2002).   Data on catches made by First Nations are generally hard to find (e.g., Irwin 1984). However, in northern BC an estimate of salmon catches by First Nations were made by (Hewes, 1973) and assumptions had to be made for the catch of eulachon and marine mammals. In Newfoundland the catch of marine animals by First Nations was hard to calculate, as the Beothuk people of Newfoundland were extirpated by 1829. Assumptions had to be made about how much the people of Newfoundland would have eaten. With the help of an Ingeborg Marshall, an anthropologist from St. Johns, their consumption of marine resources were calculated by apportioning likely catches between marine mammals, salmon, and other marine resources ((Renouf 1999, Marshall 1996, Heymans 2003).  Which species should be included?  The species to be included usually depends on the question asked, what species have gone extinct, locally or globally, and what species migrate through the system. The question asked implies that some species would be important as single groups in one model vs. being able to combine them in other groups. For instance, in Newfoundland it was necessary to put Greenland cod and lobster into their own compartments, as the question asked pertained more to the human interaction and inshore system than to the offshore system.  Likewise, the importance of migratory species such as migratory salmon and transient killer whales become more important in the northern BC model, as these are important in the policy arena of that system. There are two other important considerations that need to be taken when deciding which species to include, namely extinctions and species that are not well studied.  Extinctions   Local and global extinctions make the inclusion of certain species very difficult. For comparison between emergent properties of ecosystem models it is important for the groups to have the same number of compartments. Similarly, simulations that span two different models would need all the compartments to be included in both models. Thus, it is important to include species that have become extinct during the course of the modeling exercise. These species are usually included by adding a very small biomass (1*10 -6 t.km-2) in the models where they are essentially extinct  (see Pitcher 2004, this volume).  An example of a local extinction in northern BC is the sea otter, which became extinct before the 1900 model. Pristine population estimates are given by (Kenyon, 1975), and were used for the estimation of sea otter biomass in 1750, but by 1900 and the subsequent models of 1950 and 2000 biomass was assumed to be 1*10 -6 t.km-2.   Three species have become extinct in Newfoundland since European contact: walrus and grey seals have become locally extinct, while the great auk is globally extinct (see Pitcher 2o04, this volume). No estimates of walrus or grey seal biomasses were available for 1450, but estimates of rookery sizes and whelping patches were given in the controversial book by (Mowat, 1984), which had to be used in lieu of any other data. By 1900 both these species were locally extinct in Newfoundland, and their biomass estimates were therefore assumed to be 1*10 -6 t.km-2.    The extinction of the great auk was easier to model. Although there were at least 100,000 nesting pairs of great auk in Newfoundland at the time of European contact, they were extinct by 1830 (Burke e t al.  2002, Sarjeantson 2001, Montevecchi and Kirk 1996). However, seabird biomass and impact is so small that they are usually grouped into functional groups. The great auk was therefore grouped with the piscivorous Back to the Future Methodology, Page 14   birds, and as such no assumption had to be made about their biomass, other than the assumptions made for bird biomass in general (see Pitcher 2004, this volume).  Species that are not well studied   In all ecosystem models there are some species that are very poorly studied. Incorporating them is usually problematic, and very little data are usually available for non-commercial species. Examples of these species are the rockfish in northern BC (a guild of over 30 species) and  Greenland cod in Newfoundland. There are many species of rockfish in northern BC, but until very recently, very little data was available on these species. In the present day model therefore, they were broken down into inshore rockfish, planktivorous and piscivorous rockfish. (Ainsworth et al.  2002, Foulkes in prep.). There are no historical estimates of biomass, production, etc. for these species, or for Greenland cod in Newfoundland, so their biomasses are estimated by Ecopath  by assuming that their P/B and Q/B ratios would be similar in the past as they are today.  Energetic parameters  The other parameters needed for constructing  Ecopath  models are also be different in models of the past. Parameters such as the P/B and Q/B ratios are often smaller in populations that have many more older fish, that produce and consume less than a population that consist mostly of younger smaller fish.   The P/B ratio is usually assumed to be:   P/B = F + M                       (1)   where F is fishing mortality and M is natural mortality. Fishing mortalities in most models of the past are generally small, but where estimates of catch and biomass are available, they should be added to the estimate of natural mortality calculated below (Palomares and Pauly 1998):   log M = 0.0066 ? 0.279 (log L ? ) +  0.65431 (log k) + 0.4631 (log T)            (2)  where L? is the population asymptotic length of the Von Bertalanffy growth function (and is usually  greater in populations of the past), k is the Von Bertalanffy growth parameter, and T is temperature in ?C.  The Q/B ratio is calculated from an empirical formula published by Palomares and Pauly (1998):  log Q/B = 7.964 ? 0.204 (log W ? )  ? 1.965T*  + 0.083A + 0.532h + 0.398d            (3)  where T* is the temperature in ?Kelvin, A is the tail aspect ratio (generally obtained from Fishbase ),  h = 1 for herbivores and 0 for all other groups, and d = 1 for detritivores and 0 for all other groups.   W? is estimated from the length weight relationship:  W = a+L b             (4)  where the a and b parameters are obtained from Fishbase .   Estimating natural mortality and consumption parameters for juveniles are more challenging, therefore in most instances these parameters for juveniles were assumed to be 1.5 x that of the adults. Sometimes it was not possible to estimate both the P/B and Q/B ratios for a group, and then the gross efficiency (GE) was assumed to be 0.2 and the P/B or Q/B was calculated by Ecopath .   Assumptions  Constructing models of the past involves making many assumptions for biomass, catch, etc.  Also, there is generally no data available on past diets and one has to assume that the diet in the past was similar to that of the present. Usually, when balancing the model the diet is the first parameter that is changed. Thus, starting with today?s diet and assuming that most species are generalists that would feed on similar species, the diet of the past is changed to balance the model. The assumption that past diets were not very different is vindicated by a paper showing that, in field studies on Georges Banks (Gulf of Maine), diets of many species changed in proportion as much as would be expected from the change in abundance and species composition (Link and Garrison 2002).   CONCLUSIONS  Constructing models of the past is not an exact science. Often the model obtained would seem closer to a Picasso painting than to reality (Figure 3, Heymans and Pitcher 2002a). In an abstract Picasso painting the parts of the whole are all present, but are not realistic in proportion or placement, and this creates the interesting reaction desired by the artist. In a painting by a Page 15,  Fisheries Centre Research Reports 12(1), 2004  great Renaissance master like Raphael, in contrast, things look as they do to the eye, although in fact subtle artistic artifice is employed to achieve this effect.     Similarly, an ecosystem model obtained by reconstructing the past incorporates most of the important groups and species that were present at the time period chosen, but the lack of information, and the quality of the available information influences the model. To counteract this problem it is advisable to describe the information and assumptions as well as possible, and to perform uncertainty estimations where possible. Testing for the effect of different input data on the emergent properties of the ecosystem is a valuable way of checking uncertainty. This needs to be done for the models of the Coasts Under Stress  BTF project. Additionally, putting the errors for the main parameters into the Ecopath  model can help later when the ecosystem model is used in simulation mode and the effects of parameter uncertainties on alternative polices can be checked. The aim eventually is to have ecosystem models of the past that encourage the familiar comfort of a Raphael rather than the shock of a Picasso.   REFERENCES  Ainsworth, C., Heymans, J.J., Pitcher, T.J. and Vasconcellos, M. (2002) Ecosystem models  of Northern British Columbia for the time periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4): 41 pp. Akrigg, A.G.P.V.H.B. (1975) Br itish Columbia Chronicle 1778-1846, Adventures by Sea and Land. Discovery Press.   Alexander, A.B. (1912) Prelim inary Examination of Halibut Fishing Grounds of the Pacific Coast. Department of Commerce and Labour,  Bureau of Fisheries Document No. 763: 13-54. Anderson, A.C. (1879) Return and Report of Inspector of fisheries for British Columbia for 1878. Victoria, B.C., Canada.  Anon (1892) Special Supplement of the Commercial relating to Vancouver Island, the Adjacent Coast and Northern Interior of British Columbia. Winnipeg, Canada.  Babcock, J.P. (1910) Pages 32-34 in Report of the Commissioner of Fisheries 1902-1909. Victoria, B.C., Canada.   Beattie, A.I. (2001) A New Model for Evaluating the Optimal Size, Placement and Configuration of Marine Protected Areas. M.Sc. thesis, Department of Resource Management and Environmental Science, University of British Columbia, Vancouver, Canada. 58pp. Bundy, A., Lilly, G.R. and Shelton, A.P. (2000) A mass balance model of the Newfoundland-Labrador Shelf. Canadian Technical Report of Fisheries and Aquatic Science 2310: 157 pp. Burke, C., Davoren, G.K., Montevecchi, W.A. and Stenhouse, I.J. (2002) Winging Back to the Future: An Historic Reconstruction of Seabird Diversity, Distribution  and Abundance in the Northwest Atlantic, 1500- 2000. Pages 27-37 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N.  (eds) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf.  Fisheries Centre Research Reports 10(1): 116 pp. Carrothers, W.A. (1941) The British Columbia Fisheries. University of Toronto Press, Toronto, 133 pp.  Chambers (1872) Queen Charlotte Islands. In: Chambers's Journal, March 1872. London, U.K.  Cridland, J. (1998) Late Prehistoric Indian Subsistence in Northeastern Newfoundland: Faunal Analysis of Little Passage Complex Assemblages from the Beaches and Inspector Island Sites. Masters of Arts thesis, Memorial University of Newfoundland, St. John's, Newfoundland.  Crosby, A.W. (1986) Ecological  Imperialism: The biological expansion of Europe, 900-1900. Studies in Environment and History, Cambridge University Press, UK. 368 pp.  Cushing, D. H. (1988) The provident sea. Cambridge University Press, Cambridge, UK. Dawson, G. M. (1880) On the Indians of Queen Charlotte Figure 3.   Paintings of attractive young women by Picasso and by Raphael. The Picasso creates a fascinating and shocking effect because all the right elements of the weeping woman (the artist?s mistress) appear in a pastiche, but are placed in unexpected locations. The Raphael, on the other hand, appears as a luminous and accurate representation of a smiling young woman  (t he artist?s wife), and whose artifice to deceive the eye is more subtle. The am is for ecosystem models of the past, perhaps initially comparable to a Picasso, to become more like a Raphael.  Pablo Picasso: Weeping Woman, 1937, etching/ aqua tint /d ry point. 49 x 69 cm, Museum of Modern Art, New York; Rapha e l lo Sanzo: School of Athens (de tail), 15 10, fresco, 8 x 5.5 m, Vatican Museum, Rome.   Back to the Future Methodology, Page 16   Islands. Geological Survey of Canada, Montreal.  DeWhirst, J. (1982) The origins of Nootkan Whaling: A definition of Northern and Central Nootkan Ecological Orientation for the Past four Millenia. Pages 1-11 in Anthropological Working Group,  Nortorf. Vol. 33.  Forrest, R. (2002) Estimating the sports catch in northern BC.  Pages 38-40 in Pitcher, T.J., Power, M.D. and Wood, L. (eds) (2002)  Restoring the past to salvage the future: report on a community participation workshop in Prince Rupert, BC. Fisheries Centre Research Reports  10(7): 56 pp.  Freeman, B.C. (1904) The Indians of Queen Charlotte Islands. In: The Methodist Young People's Forward Movement for Missions. Toronto, Canada.  Gregr, E. J. (1999) An analysis of historic (1908-1967) whaling records from British Columbia, Canada. Masters of Science thesis, University of British Columbia, Vancouver, Canada, 101 pp.  Gregr, E.J. (2002) Whales in Northern BC: Past and Present.  Pages 74-77 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N.  (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf.  Fisheries Centre Research Reports 10(1): 116 pp. Hammond, L. (1993) Marketing Wildlife: The Hudson's Bay Company and the Pacific Northwest, 1821-49.  Forest and Conservation History 37: 14-25.  Hewes, G.W. (1973) Indian Fisheries Productivity in Pre-Contact Times in the Pacific Salmon Area.  Northwest Anthropological Research Notes  7(2): 133-155.  Hewitt, K.W. (1993) The Newfoundland Fishery and State Intervention in the Nineteenth Century: The Fisheries Commission, 1888-1893. Newfoundland Studies 9(1): 58-80.  Heymans, J.J. (2003) First Nations Impact On the Eastern Newfoundland and Southern Labrador Ecosystem During Pre-Contact Times. Pages 4-10 in Heymans, J.J. (ed.) (2003) Ecosystem models of Newfoundland and southeastern Labrador: additional information and analyses for ?Back to the Future?. Fisheries Centre Research Reports 11(5): 79pp. Heymans, J.J. and Pitcher, T.J. (2002a) A model of the marine ecosystem of Newfoundland and Southern Labrador (2J3KLNO) in the time periods 1985-87 and 1995-97. Pages 5-43 in Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) Ecosystem models of Newfoundland for the time periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 74 pp. Heymans, J.J. and Pitcher, T.J. (2002b) A Picasso-esque view of the marine ecosystem of Newfoundland and Southern Labrador: Models for the Time Periods 1450 and 1900. Pages 44-74 in Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) Ecosystem models of Newfoundland for the time periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 74 pp. Heymans, J.J. (ed.) (2003 ) Ecosystem models of Newfoundland and southeastern Labrador: additional information and analyses for ?Back to the Future?. Fisheries Centre Research Reports 11(5): 79pp. Hiller, J.K. (2001) The Newfoundland Seal Fishery. Heritage Site of Newfoundland and Labrador, Memorial University of Newfoundland, Canada. Howley, J.P. (1915) Drawings by Shanawdithit. The Beothucks or Red Indians: The Aboriginal Inhabitants of Newfoundland. David Cantwell website www.cs.mun.ca/~david12/ Hutchings, J.A. and Myers, R.A. (1995) The biological collapse of Atlantic cod off Newfoundland and Labrador: An exploration of historical changes in exploitation, harvesting technology, and management. Pages 39-92 in Arnason, R. and Felt, L. (eds) The North Atlantic Fisheries: Successes, failures and challenges. Institute of Island Studies, Charlottetown, Prince Edward Island, Canada, Vol. 3.  Irwin, S.D. (1984) Hunters of the Sea. Pages 209-253 in The providers: Hunting and fishing methods of the North American Natives. Hancock House, Surrey, B.C., Canada. Jacobsen, J.A. (1977) Alaskan Voyage, 1881-1883, An Expedition to the Northwest Coast of America. E. Woldt (translator), The University of Chicago Press, USA.   Kenyon, K.W. (1975) The Sea Otter in the Eastern Pacific Ocean. Dover Publishing, New York, USA.  Lear, W.H. (1998) History of Fisheries in the Northwest Atlantic: The 500-Year Perspective.  Journal of Northwest Atlantic Fisheries Science, 23: 41-73.  Lescarbot, M. (1914) History of New France.  Pages 234-245 in The publications of the Champlain Society. The Champlain Society, Toronto, Canada. Vol. III.  Lewis, H.F. and Doutt, J.K. (1942) Records of the Atlantic walrus and the polar bear in or near the northern part of the Gulf of St. Lawrence.  Journal of Mammalogy  23(4): 365-275.  Link, J. S. and Garrison, L.P. (2002) Changes in piscivory associated with fishing induced changes to the finfish community on Georges Bank. Fisheries Research 55: 71-86. Lord, J. K. (1866) The Naturalist in Vancouver Island and British Columbia. Richard Bentley, New Burlington Street, London, U.K.  Mackie, Q., Orchard, T. and Steffen, M. (2001) Environmental Archaeology of the Late Pre-Contact/Contact Period in Gwaii Haanas. Pages 1-12 in 34th Annual Meeting of the Canadian Archaeological Association, Banff, Alberta, Canada. May 2001.  Marshall, I. (1996) A History and Ethnography of the Beothuk. McGill-Queen's University Press, Montreal, Canada. 640 pp.  Mercer, M.C. (1967) Records of the Atlantic Walrus, O dobenus rosmarus rosmarus,  from Newfoundland.  Journal of the Fisheries Research Board of Canada 24(12): 2631-2635.  Montevecchi, W.A. and Kirk, D.A. (1996) Great Auk ( Pinguinus impennis ). Pages 1-20 in Poole, A. and Gill, F. (eds) The Birds of North America. The American Ornithologists' Union and The Academy of Natural Sciences, Philadelphia, USA. Vol. 260.  Montevecchi, W.A. and Tuck, L.M. (1987) Newfoundland Birds: Exploitation, study, conservation. Publications of the Nuttall Ornithological Club, Cambridge, Massachusetts, USA. 21:  273 pp.  Mowat, F. (1984) Sea of Slaugh ter. Seal Books, McClelland-Bantam Inc., Toronto, Canada. 463 pp.  Mowat, T. (1886) Pages 248-257 in Report on the Fisheries of British Columbia, 1886. New Westminister, B.C., Canada.  Muir, M. (1935) Development of the Fishing Industry in B.C. up to 1900. In Paper Submitted in History II, Canadian History, Summer School, UBC, 1935. UBC Press, Vancouver, B.C., Canada.  Newcombe, C.F.G. and Fraser, C.M. (1917) Preliminary Report of the Commission on the Sea-lion question, 1915.  Contributions to Canadian Biology 1917.  Supplement to the 7th Annual Report of the Department of Naval Service Fisheries Branch. Osgood, W.H. (1901) Natural History of Queen Charlotte Islands, British Columbia. Page 25 in North American Fauna, no. 21., Government Printing Office, Washington, DC, USA. Palomares, M.L.D. and Pauly, D. (1998) Predicting food consumption of fish populations as functions of mortality, food type, morphometrics, temperature and salinity. Marine Freshwater Research 49: 447-453. Pastore, R.T. (1992) Shanawdithit's People: The Archaeology of the Beothuks. Atlantic Archaeology Ltd., St. John's, Newfoundland, Canada, 74 pp.  Pitcher, T.J. (2003) The compleat angler and the management of aquatic ecosystems. Pages 3-7 in Coleman, A.P.M. (ed) Regional Experiences for Glob al Solutions. Proceedings Page 17,  Fisheries Centre Research Reports 12(1), 2004  of the 3rd World Recreational Fisheries Conference, Darwin, Australia May 2002. Northern Territories Fisheries Report 67: 269 pp. Pitcher, T.J. (2004) The problem of extinctions. Pages 21?28 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J. and Hollingworth, C.E. (2002) Fishing for Fun: What?s the Catch?  Chapter 1, Pages 1-16 in Pitcher, T.J. and Hollingworth, C.E. (eds ) Recreational Fisheries: Ecological, Economic and Social Evaluation. Blackwell Science, Oxford, UK. 288pp. Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N. (2002a) (eds) Information Supporting Past and Present Ecosystem Models of Northern British Columbia and the Newfoundland Shelf. Fisheries Centre Research Reports 10(1): 116 pp. Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) (2002b) Ecosystem models of Newfoundland for the time periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 74 pp. Pitcher, T.J., Power, M. and Wood, L. (eds) (2002) Restoring the Past to Salvage the Future: Report on a Community Participation Workshop in Prince Rupert, BC. Fisheries Centre Research Reports 10 (7): 55 pp. Pope, P. (1995) Early Estimates: Assessment of Catches in Newfoundland Cod Fishery, 1660-1690. Pages 9-40 in Vickers, D. (ed.) Marine Resources and Human Societies in the North Atlantic Since 1500, St.John's Newfoundland, 20-22 October 1995, Memorial University of Newfoundland, Canada.  Reeves, R. R., Leatherwood, S., Karl, S. A. and Yohe, E. R. (1985a) Whaling results at Ak utan (1912-1939) and Port Hobron (1926-1937), Alaska.  Report to the International Whaling Commission 35: 441-457.  Renouf, M.A.P. (1999) Prehistory of Newfoundland hunter-gatherers: extinctions or adaptations? World Archaeology 30(3): 403-420.  Ryan, S. (1994) The Ice Hunters: A history of Newfoundland Sealing to 1914. Breakwater, St. John's, Newfoundland, Canada, 525 pp.  Thompson, W.F. (1916) A Report on Statistics of the Halibut Fishery of the Pacific: Their Bearing on the Biology of the Species and the Condition of the Banks. BC Commissioner of Fisheries,  6 Geo. 5.  S10-S12 pp. Serjeantson, D. (2001) The Great Auk and the Gannet: a Prehistoric Perspective on the Extinction of the Great Auk.  Int. J. Osteoarchaeol. 11: 43?55. Turgeon, L. (1995) Fluctuations in Cod and Whale Stocks in the North Atlantic During the Eighteenth Century. Pages 89-121 in Vickers, D. (ed.) Marine Resources and Human Societies in the North Atlantic Since 1500.  StJohn's Newfoundland, 20-22 October 1995, Memorial University of Newfoundland, Canada.  Vancouver, G. (1984) The Voyage of Discovery to the North Pacific Ocean and Around the World, 1791-1794. In Lamb W.K. (ed.), Volume 3. The Hakluyt Society, London, U.K., 451 pp. Vasconcellos, M., Heymans, J.J. and  Pitcher, T.J. (2002) Historical Reference Points For Models of Past Ecosystems in Newfoundland. Pages 7-12 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N.  (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf.  Fisheries Centre Research Reports 10(1): 116 pp. Webb, R.L. (1988) Commercial Whaling in the Pacific Northwest 1790-1967. University of British Columbia Press, Vancouver, B.C., Canada. 425 pp.  Whitridge, P. (2001) Zen Fish: A consideration of the discordance between artifactual and zoo-archaeological indicators of Thule Inuit fish use. Journal of Anthropological Archaeology 20: 3-72.  Wright, B.S. (1951) A walrus in  the Bay of Fundy: the first record.  The Canadian Field-Naturalist 65: 61-63.    For discussion following oral presentation of this paper, see page 136.    Back to the Future Methodology, Page 18   W HAT W AS THE STRUCTURE OF  PAST ECOSYSTEMS THAT HAD  MANY TOP PREDATORS?     Tony Pitcher Fisheries Centre, UBC   ABSTRACT   Analyses of the ancient past, from historical sources, archaeology and reconstructions, suggest the presence of large numbers of top predators where few are found today. In mass-balance ecosystem models, these animals are not generally able to be supported by what are thought to be realistic levels of forage organisms. This paper examines the logic of this issue, and suggests that the problem may be resolved by evidence from archaeology and stable isotope analysis. In the past, more species may have occupied the forage fish niches and the diet of top predators near to carrying capacity may have been wider due to intra-specific competition.    Historical sources (e.g., examples in Mowat 1984) and attempted reconstructions (e.g., coastal ecosystems: Jackson e t al.  2001; predatory fish: Myers and Worm 2003; sharks: Baum e t al.  2002; whales: Roman and Palumbi 2003) all suggest that past ecosystems had many more large and long-lived top predators than we find today. Analysis of archeological remains also often suggests large predatory species where few are present today, for example, bluefin tuna along the whole western Canadian coast (Tunnicliffe e t al.  2001, and see discussion page 139 this volume) and the North Sea (Mackinson 2001), and large old individuals of species in regions where they are represented by smaller, younger members today (e.g. cod and saithe at Skara Brae neolithic settlement, Orkney; Barret e t al.  1999, Childe 1931, Clarke 1977). Moreover, compared to the present day, fishery exploitation was low in the ancient ecosystems (e.g., aboriginal fisheries in Newfoundland, Heymans 2003, Lucas 2004, this volume).   The issue in question here is that, when such large amounts of top predators are inserted into a mass-balance ecosystem model, a very large amount of amount of prey organisms is required as food to maintain all these animals. The                                                            Pitcher, T.J. (2004) What was the structure of past ecosystems that had many top predators? Pages 18?20 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.  resulting biomass of forage animals is thought to be unrealistic compared to present day levels. We may ask if, in fact, this issue is some kind of artifact of the ecosystem modelling method, or a genuine conceptual problem?   On the modelling side, we may note that the P to B ratio of large old individuals of a species is far lower than the ratio characteristic of exploited populations today, and so adjustments in this respect are now routine in the creation of ecosystem models of ancient systems (see Heymans and Pitcher 2004, this volume, and e.g., Ainsworth e t al.  2002). Nevertheless, even with reduced P/B ratios, surprisingly large forage fish biomasses can still result.   Some simple answers to the problem offer themselves first.   1 . There were not so many top predators. The high abundance of top predators may actually be a false impression, based on anecdotes of those impressed by local patches of high abundance? For example, in the accounts of the first European visits to Newfoundland (Pope 1997, Williams 1996) we find what at first sight appear to be  exaggerated references to cod so abundant that buckets full of the fish could be scooped up with little effort. Such reports may have been aimed, in part, at reassuring the late 15th Century financiers of expeditions to the New World that future gains would be considerable, as indeed they were. It is, however, a reasonable conclusion from the considerable archeological and documentary evidence that in ancient coastal ecosystems there were indeed large amounts of top predators, both in terms of species and in terms of large, old individuals, numbers and biomass. In addition  to the references cited above, the work of Jackson e t al.  (2001) is perhaps the most significant in this respect.   2 . A high biomass of forage fish is acceptabl e .  A second simple answer is that a high abundance of Figure 1.  Top predators, like this blue marlin, may have been so abundant in ancient ecosystems that a very large amount of prey forage fish was required to support them. This paper discusses the ecosystem modelling issues raised by this possibility.   Page 19,  Fisheries Centre Research Reports 12(1), 2004  prey needed to feed these top predators may be actually acceptable. Biomasses in excess of 40 tonnes per km2 are quite possible for small forage fish in upwelling or otherwise highly productive ecosystems (V. Christensen, pers. comm.). These fish may be highly productive, especially after a successful recruitment, feeding directly on blooms of phytoplankton and small zooplankton, with B/P ratios in excess of 3 in some cases. Although very high forage fish biomasses may be sporadic due to volatile recruitment, long-lived predators are presumably buffered against the fluctuations.  In some cases though, these simple answers may not be sufficiently convincing. Two more complex answers are discussed below.  3. T h e diet of abunda n t competing predators was broader in the ancient past . Populations of fish near their carrying capacity are not only comprised of large old individuals compared to exploited populations, but they are also characterised by high levels of intra-specific competition for food, space and other essential resources. This competition leads them to occupy all suitable habitat including the fringes of their normal range (MacCall 1990). For our purposes,  the concept may be extended from the physical habitat to elements of their trophic niche. Competition at high population densities may lead to less successful individuals eating all manner of unlikely prey at the fringes of the normal diet. Hence, for this reason the breadth of the top predator population?s diet may have been much wider than under ?normal? exploited conditions under which data on diets has been gathered today. In a mass-balance model of the ancient past, therefore, diet might be broadened to more species of likely prey animals, reducing the high biomasses of any one species required to support the abundant predators.  4. More forage fish species were present in the ancient past. A similar argument concerns the number of species of forage fish present ancient ecosystems. Where today forage fish often occur in single-species ?wasp-waist? ecosystems, in the past more species may have been present. According to Odum?s ratchet (Pitcher 2001), species with low P/B ratios become locally extinct first under the joint influence of climate and exploitation (Dulvy e t al.  2003, Christensen and Pauly 1997). Even today, several species of less abundant non-commercial small pelagic fish co-occur with dominant species such as herring, capelin and mackerel. In some areas, the biomasses of small non-commercial forage fish are not even surveyed (e.g. sand-lance in British Columbia). Hence, today?s species composition for this group of forage fish may not be a reliable guide to the food web that existed in ancient ecosystems. Since both #3 and #4 entail adjustments to the diet matrix of the mass-balance model, both arguments may need to be taken into account.  How can these issues be resolved? One approach is to look for archaeological evidence of the relative abundance of forage fish species (e.g., van Neera e t al.  1999). Here, care must be taken to apply a series of strict rules concerning the interpretation of archeological fish bones as being representative of what was present in the wild in ancient ecosystems (see Orchard and Mackie 2004, this volume). For example, values may be distorted by selective fishing, by taphonomic factors affecting relative preservation status, and, since forage fish are generally small, ineffective screening of middens for small bones (see discussion page 138). In some cases, accurate modern analyses based on bone collections that were made in the past may be prejudiced by inadequate preservation, provenance or stratigraphy (i.e., ?problems of collection, retention, curation and context?, see Leach and Davidson 2001). Figure 2.  Discovered after a violent storm in 1850, Skara Brae, Orkney Islands, Scotland is the best preserved Neolithic village in northern Europe and offers a unique window into the lives of the fishers and farmers who lived there between 5,100 and 3,450 BP. Photograph shows a house with a stone dresser (rear wall) around which are three tanks for preparing fish bait. Middens from the site contain bones from huge cod and saithe (Barrett e t al.  1999). Back to the Future Methodology, Page 20   Another helpful investigation would be to examine the breadth of ancient fish diets using stable isotope analysis on archeological remains.  Finally, it would be worthwhile to investigate the effect of the structure and breadth of the forage fish diet of top predators more rigorously and systematically using the Ecopath  auto-balancing facility (Kavanagh et al.  2004).   REFERENCES  Ainsworth, C., Heymans, J.J., Pitcher, T.J. and Vasconcellos, M. (2002) Ecosystem models  of Northern British Columbia for the time periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4): 41 pp. Barrett, J.H., Nicholson, R.A. and Cer?n-Carrasco, R. (1999) Archaeo-ichthyological Evidence for Long-term Socioeconomic Trends in Northern Scotland: 3500 BC to AD 1500. J.Archaeol.Sci. 26(4): 353-388. Baum, J.K., Myers, R.A., Kehler, D.G., Worm, B., Harley, S.J. and Doherty, P.A. (2002) Coll apse and Conservation of Shark Populations in the Northwest Atlantic. Science 299: 389-392. Childe, V.G. (1931) Skara Brae. H.M. Stationery Office Books, U.K.  27 pp. Christensen, V. and Pauly, D. (1997) Changes in models of aquatic ecosystems approaching carrying capacity. Ecol. Applic. 8 (1): 104-109. Clarke, D.V. (1977) The Neolithic Village at Skara Brae, Orkney: 1972-73 Excavations, an Interim Report. H.M. Stationery Office Books, U.K.  28 pp. Dulvy, N.K., Sadovy, Y. and Reynolds, J.D. (2003) Extinction vulnerability in marine populations. Fish and Fisheries 4(1): 25-64. Heymans, J.J. (2003) First Nations Impact On the Newfoundland Ecosystem During Pre-Contact Times. Pages 4-10 in Heymans, J.J. (ed.) Ecosystem models of Newfoundland and southeastern Labrador: additional information and analyses for ?Back to the Future?. Fisheries Centre Research Reports 11(5): 79pp. Heymans, J.J. and Pitcher, T.J. (2004) Synoptic Methods for Constructing Models of the Past. Pages 11?17 in Pitcher, T.J. (ed.) Back to the Futu re: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J., Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A.,  Hughes, T.P., Kidwell, S., Lange, S.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S.,  Tegner, M.J. and Warner, R.R. (2001) Historical Overfishing and the Recent Collapse of Coastal Ecosystems. Science 293: 629-637. Kavanagh, P., Newlands, N., Christensen, V. and Pauly, D. (2004) Automated parameter optimization for Ecopath  ecosystem models. Ecological Modeling ( in press ). Leach F. and Davidson, J. (2001) The use of size-frequency diagrams to characterize prehistoric fish catches and to assess human impact on inshore fisheries. Int. J. Osteoarchaeology 11 (1-2): 150-162. Lucas, S. (2004) Aboriginal Valu es. Pages 114?116 in Pitcher, T.J. (ed.) Back to the Futu re: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. MacCall, A.D. (1990) The Dynami c Geography of Marine Fish Populations. University of Washington Press, Seattle, USA. 153pp. Mackinson, S. (2001) Representing  trophic interactions in the North Sea in the 1880s, using the Ecopath  mass-balance approach. Pages 35-98 in Gu?n ette, S., Christensen, V. and Pauly, D. (eds) Fisheries impacts on North Atlantic ecosystems: models and analyses. Fisheries Centre Research Reports 9(4): 344pp. Mowat, F. (1984) Sea of Slaugh ter. Seal Books, McClelland-Bantam, Toronto, Canada. 463 pp.  Myers, R.A. and Worm, B. (2003) Rapid worldwide depletion of predatory fish communities. Nature 423: 280-283. Orchard, T.J. and Mackie, Q. (2004) Environmental Archaeology: Principles and Case Studies. Pages 64?73 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J. (2001) Fish eries Managed to Rebuild Ecosystems: Reconstructing the Past to Salvage The Future. Ecological Applications 11(2): 601-617. Poon, A. and Rizzo, Y. (2004) Rapporteurs? Report on Discussion at the Back-to-the-Future Symposium, UBC, February 2002. [See page 140.] P.ages 135?154 in Pitcher, T.J. (ed.) Back to  the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pope, P.E. (1997) The Many Landfalls of John Cabot. Univ. of Toronto Press, Canada. 208 pp Roman, J. and Palumbi, S.R. (2003) Whales Before Whaling in the North Atlantic. Science 301: 508-510. Tunnicliffe, V., O'Connell, J.M. and McQuoid, M.R. (2001) A Holocene record of marine fish remains from the Northeastern Pacific. Marine Geology 174: 197-210. Van Neera, W., Lougas, L. and Rijnsdorp, A.D. (1999) Reconstructing Age Distribution, Season of Capture and Growth Rate of Fish from Archaeological Sites Based on Otoliths and Vertebrae. Int. J. Osteoarchaeol. 9: 116?130. Williams, A.F. (1996) John Cabot and Newfoundland. Newfoundland Historical Society, St. John's, Newfoundland, Canada. 64pp.   For discussion following oral presentation of this paper, see page 149. Page 21,  Fisheries Centre Research Reports 12(1), 2004  THE PROBLEM OF EXTINCTIONS   Tony Pitcher Fisheries Centre, UBC   ABSTRACT   The extinction of species causes problems when, to enable comparison of emergent properties, a series of ecosystem models constructed through time must have a similar structure. Global extinctions are irreversible and approximate representations of such species in models of ancient ecosystems relies on historical and archeological information about their ecology, diet and growth. As a short-cut to preserve model structure, extinct species may sometimes be grouped with species of a similar function in ecosystem models. Local extinctions (?extirpations?), on the other hand, are potentially reversible by natural recolonisation, or by human re-introduction. Ecosystem modelling therefore needs to be able to capture this reversibility by explicitly including such species. Currently, it is especially difficult to model the effects of keystone species, such as sea otters, whose biomass level directly alters habitat structure.     Global extinctions of species, such as the great auk in the North Atlantic (Montevecchi and Kirk 1996), or Steller?s sea cow in the North Pacific (Anderson 1995), mean that there is little choice but to eliminate these species from future restoration goals in the Back-to-the-Future process. Local extinctions (= ?extirpations?), on the other hand, are potentially reversible by natural recolonisation or by human re-introduction. But for comparison between the emergent properties of the series of whole-ecosystem models in BTF, it is important for all of the models to have the same number of compartments, although of course biomasses and fluxes can be vastly different. Similarly, ecosystem simulations that span two or more different models need all the compartments to be included in both models. Extinction of species makes this comparison difficult. What can be done in ecosystem modelling, therefore, when species have become locally or globally extinct? How may these factors be accommodated in the suites of ecosystem simulation models used in BTF? Before answering these questions, I review marine species that have become globally or locally extinct in our two CUS BTF ecosystems.                                                             Pitcher, T.J. (2004) The problem of local extinctions. Pages 21?28 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp . GLOBAL EXTINCTIONS  The great auk, Alca impennis, was a large flightless, pelagic species of the Alcidae  (auks) in the North Atlantic, and the original recipient of the name penguin (pen-gwyn, meaning ?white head?, the winter plumage, in Welsh and Gaelic), a name later transferred to an entirely different order of Southern hemisphere birds ( Sp heniscidae ). Hunting by humans, usually at island breeding sites, rendered the piscivorous great auk extinct by 1844 (Figure 2). Although the bird had been eaten for thousands of years by coastal peoples, in the late 18th and early 19th Centuries great auks were harvested for food, feathers and eggs on an astounding scale. For example, during the Napoleonic wars, Britain mitigated a blockade of Grand Banks? cod by importing shiploads of great auks from the Figure 2 . The flightless North Atlantic ?Penguin?, Garefowl, Spearbill or Great Auk, Alca impennis , a 70cm, 5kg seabird once harvested by the shipload throughout the North Atllantic, and hunted to extinction by 1844. John J. Aud u bon, chromolithogra phic print, The Birds of America, 24 x 36. Sa n Joaquin County Pub lic Libra ry, US A.  Figure 1.  ? Actually, there were three arks. The one with dinosaurs and other extinct forms sank due to overcrowding. The one with marsupials was blown off course and landed in Australia.? A brave  attempt to explain extinctions and biogeography. See www.christianforums.com/t40474&page=2.   Back to the Future Methodology, Page 22   islands off Iceland. Soon, there was a boisterous and increasingly lucrative international trade in diminishing numbers of great auk eggs, skins and skeletal remains in late Victorian times. Despite unconfirmed reports of sightings up to the early 20th century, the financial incentives brought about by this trade likely ensured that there are no surviving colonies. For example, the last pair of birds seen in Iceland was killed for sale, together with their egg (3 rd June, 1844).  The classic study of the biology and demise of the great auk was published by Grieve (1885), and a recent book provides a thorough review (Gaskall 2000). Using data from archaeological remains in middens and from the sites of what appears to have been industrial-scale processing, Sarjeantson (1996) shows how the flightless great auk was wiped out, while the gannet, which was also exploited heavily but can fly, has avoided the same fate. Evidently, there was a population of at least a million birds in the North Atlantic before 1830 (Montevecchi and Ki rk 1996), and middens suggest a far greater population over a wide range from Florida to the Bay of Biscay throughout the North Atlantic, and even the Mediterranean,  in pre-historic times.   Although there is no quantitative data, fish species eaten by the great auk, which could dive to a depth of at least 10 metres, can be reasonably well deduced from some contemporary descriptions (see Grieve 1885). The diet likely consisted of pelagic fish such as capelin (Figure 3), herring and sandlanc e offshore, and large scuplins and juvenile cod when feeding inshore during the breeding season. For ecosystem modelling, metabolic parameters for this large bird might be taken as similar to the larger southern hemisphere penguins.   Hence, there is certainly data enough to include great auks explicitly in mass-balance models of ancient North Atlantic ecosystems and to make preliminary biomass estimates based on diet and the other Ecopath  parameters. But, in most cases, seabirds have such a small biomass and impact in marine ecosystem models that they are usually grouped into functional categories, such as invertebrate eaters, piscivores, inshore ducks and the like (Burke e t al.  2002). In fact, in the CUS BTF North Atlantic models to date, the great auk has been grouped with other piscivorous seabirds (Heymans e t al.  2002b, Davoren e t al.  2002). This means that, provided the great auk?s diet and metabolic parameters are represented in the appropriate functional group in the model, no special assumptions have to be made (Heymans e t al.  2002a). This device also has the advantage that the group structure of the series of BTF ecosystem models remains the same over time. But the trick has the disadvantage that the possible impacts of the great auk?s extinction on ecosystem structure cannot be explored. Since the great auk was clearly major predator of medium-sized fish, this would be an interesting topic to explore in the future.  In the 18th Century, the North Pacific was the location of two other dramatic global extinctions.  In 1741 on the Komandorski islands at the extreme west of the Aleutian chain, Steller found Figure 3.  The great auk eating an adult capelin. Few North Atlantic seabirds eat such large prey today. W. Imp, J. Gould and Whart 1840, coloured lithograph, 38.1 x 54.8 cm, J.H. Fleming Library, Ornithology Collections . Figure 4 . The extinct spectacled (= Pallas?) cormorant, P h a l acrocorax perspicillatus  a 5kg flightless bird found by Steller in the Komandorski islands in 1741.  Only 7 museum specimens of this North Pacific penguin-like bird survive and very little is known about it. Page 23,  Fisheries Centre Research Reports 12(1), 2004  a flightless spectacled cormorant ,  P ha lacrocorax perspicillatus (Figure 4).  He also discovered the Sea Cow, or rhytine, Hydrodamalis gigas, a large herbivorous sirenian (Figure 5).  Georg Wilhelm Steller, a stern, meticulous German, studied at the Un iversity of Wittenberg and then, after a spell as an army surgeon, worked in Russia at the Academy of Sciences in St Petersburg. Steller was 33 years old when he was employed as the naturalist on Janasson (?Vitus?) Bering?s 1741 expedition from the Tsarina Anna?s Russia to the region between Asia and America. Anna had emerged as Empress in 1730 from the turmoil following Peter the Great?s sudden death in 1725, and adopted the same expansionist agenda. Bering himself was a Dane serving in the Russian navy. The expedition was a tough call; among the hardships were scurvy, losing the other half of the expedition in a storm, shipwreck, over-wintering on what came known as Bering Island, and having to salvage wood to build a replacement ship1. Soon after the shipwreck, Vitus Bering died of scurvy that winter, along with half of his crew. But the tough naturalist Steller impressed the crew by searching out plants to treat scurvy2. By the next summer, the survivors began to hunt and eat the sea-cows and                                                           1 Only one man, Sava Starodubtsov, a Siberian carpenter, thought that he remembered how to build a ship. The 46 surviving crew depended on his knowledge for their lives.  2 Sven Waxell, one the ship?s officers, said that Steller, although stern, was "a great botanist and anatomist, well versed in natural science".  Steller saved the life of Waxell and his son. He named over 50 new species of animals  and plants on the expedition. they left the island with barrels of salted sea cow meat3. They also hunted and ate the large, flightless cormorant of which Steller wrote, "They weighed 12-14 pounds, so that one single bird was sufficient for 3 starving men."  Immediately after the expedition?s return, Siberian fur traders flocked to the Komandorski islands, trapping foxes and sea otters for fur. They used the sea cows, said to be similar to almond-flavoured veal, and the flightless cormorants as a living larder. Sea cow blubber was used for cooking and as lamp oil, the milk of slaughtered cows was made into butter, and the tough hide was used for shoes, belts and skin-covered boats. The animal soon became rare, and although an order prohibiting hunting of the sea cows4 was sent from St Petersburg to the Komandorski Islands on November 27th, 1755 (Domning, 1978), hunting seems to have continued. The last report of a sea cow being killed was in 1768.   The spectacled cormorant lasted longer, its last stronghold until 1850 being the small island of Aij Kamen (Stejneger 1889).  As well as an island, Bering got a sea named after him (on account of a filing error it seems, see Pitcher 1999). Steller ende d up lending his name to an eider duck, a jay, a sea-lion, a rock-trout, an eagle and the sea cow. Also, unexpectedly, his name was used posthumously for Stellerite, a kind of silicate crystal found on the Komandorski Islands in 1909. Hounded by the Tsarist bureaucracy for humane treatment of some prisoners, a drunken Georg Steller died a miserable death in a snow storm at Tjumen, a Siberian town to the east of the Ural mountains, in November 1748, only four years after the expedition. Fortunately his notes (written in Latin under the harsh conditions of the island shipwreck) were preserved, and were retrieved, edited and published by P.S.Pallas (1781) 5,                                                           3 A preserved sea cow carcass, and many other specimens, had to be left behind. 4 In 1754, an envoy of the Tsar wrote that sea cows were being exterminated at such a rate that they would soon be eradicated. Groups of two or three hunters from Kamchatka, the envoy wrote, were "inflicting huge waste and destruction".  5 And translated into German (Pallas 1781).  Figure  4.  The Sea Cow was discovered by Steller in 1741, all were killed by 1769. U p p e r  panel:  one of the few extant skeletons in the Helsinki natural history museum. Lower panel:  likely reconstruction of a Sea Cow (H ans Rothschaur, Germany). These sirenians were 7.5m long, weighed up to 11 tonnes, lived in herds close inshore, and appear to have eaten kelp and red algae.  Back to the Future Methodology, Page 24   himself a German naturalist of repute working in St Petersburg, with a cat and several birds, including that flightless cormorant, named after him.  Some believe that small colonies of Steller's Seacows still live in remote areas of the northern oceans. In 1962, the crew of a Russian whaler reported seeing six animals that resembled sea cows, feeding in the Gulf of Anadyr, north of Kamchatka. In 1977, a Kamchatkan fisher reported seeing a drifting animal that matched the description of a sea cow (M. Raynal; http://perso.wanadoo.fr/cryptozoo/dossiers/rhytine.htm). Possible  reports of sea cows before Steller might lend support to this idea. For example, in 1609 Henry Hudson reported animals that fit the description of sea cows near Novaya Zemlya . There are also reports from Greenland and other Arctic ocean sites But if these earlier reports from pan-arctic sites are correct, sea cow populations must have undergone a serious range collapse in the 17th Century before being described by Steller, or they would surely have been found by the many North Atlantic expeditions of the time. Sea cows were distinctive, large, impressive animals, forming obvious pair bonds, living inshore in small herds with juveniles, and Ste ller (1751) even reports them coming to the aid of stricken animals. If their pan-arctic demise was due to recent human predation, there would surely be Traditional Knowledge and Myth concerning these massive social animals among today?s native peoples of the arctic.   Archeological evidence places sea cows along the Pacific coasts of Asia and North America as far south as Japan and northern California. Their ease of capture and suitability for providing large amounts of human food would, like other North American megafauna, have rendered them susceptible to the ?clovis? hunting tools of first North Americans 12 to 15,000 years ago (Alroy 2001, Martin 1984).  Most of the sites of slaughter and butchering would today lie submerged as a result of rising sea levels after the ice age (see Josenhans e t al . 1997). It is interesting that the present coastal peoples of the Pacific North-West, whose DNA suggests that they arrived from Asia 6-8000 years ago (Morel 1997), have no knowledge or cultural memory of sea cows. It is likely then, that sea cows were wiped out by hunting very soon after boat-building humans  inhabited the Asian shores of the North Pacific 35,000 to 25,000 BP (E rlandson 2001). The abundant food (shellfish, finfish, marine birds ? including those flightless cormorants - and mammals) available from North Pacific kelp forests probably attracted early maritime people, and, it is thought, may have facilitated the earliest migrations of people from Eurasia to the Americas. It is possible that the whaling tradition of indigenous people of the North Pacific began with the over-harvest of the predator-naive and defenceless Steller?s sea cow, focusing thereafter on cetaceans that were more difficult to harvest (Domning 1972). What Steller discovered on the uninhabited Komandorski islands then, was a living remnant population of one of the Pleistocene megafauna.  There is sufficient historical information about sea cow diet, and reasonable inferences about metabolism may be made from extant sirenians, for us to attempt to model them explicitly in a mass-balance ecosystem model (Stejneger 1886). The animals seem to have lived mainly inshore, near to sources of fresh water (Domning 1976). Steller?s account indicates that the sea cow fed mainly on soft brown kelps and red algae, with a little sea grass. Anatomical adaptations to the sea cow?s mouth and gut seem to fit with this. The huge sea cow gut seems to have been an adaptation to digest large amounts of poorly masticated algae. There were no teeth, only horny lips and upper palate for rasping algae from the Figure 5.  18th century engraving of a Steller?s Sea Cow, Hyd rodamalis gigas,  being captured for food on Bering island by a ship?s crew in the mid-1700s.   Page 25,  Fisheries Centre Research Reports 12(1), 2004  rocks. Steller says that large amounts of torn and dislodged kelp floated around sea cow feeding sites. Metabolic parameters for dugongs, 3 metres in length, could be scaled to reflect the slower turnover and larger body size of the sea cow (Pitcher 1998). Sea cow predators would have been mainly killer whales and perhaps cold water sharks. A starting value for sea cow biomass in a model might come from the estimated 5000 population in the area around the Komandorski islands. Assuming an area of 100km by 50 km around the islands, this amounts to an average biomass of about one animal per km2 in inshore habitats, or about 7.5 tonnes per km2.  As yet, no-one has attempted to construct an inshore ecosystem model that contains sea cows grazing kelp. In fact, it seems that kelp canopies are remarkably resilient to cropping of the distal fronds (Steneck e t al.  2002). A multi-million dollar industry of canopy-cropping factory ships sustainably harvest kelp in California with little permanent damage to the kelp forests (Tegner and Dayton 2000). It is therefore unlikely that sea cow grazing of canopies deforested kelp beds. But the large quantities of kelp grazed would have dynamic effect on the kelp forest canopy structure, and would alter strategic cover and hence the survival of many inshore fish and invertebrates. And so, in contrast to most pelagic systems where floating phytoplankton comprises the food of higher trophic levels, these factors would make a sea cow/grazed kelp system structurally similar to many terrestrial ecosystems. Modelling the ecosystems of terrestrial game parks, or even dinosaur ecosystems, would make fascinating work in terrestrial or palaeo-ecology. Changes to the modelling framework to deal with habitat structural elements directly would be required, as discussed below.  LOCAL EXTINCTIONS: ABSENT BUT POTENTIALLY RESURGENT SPECIES  When species have become locally extinct (?extirpation? in conservationist language), one has to allow the possibility that they may return, either through natural migration or though active reintroduction.   An example of natural recolonisation is the humpback whale in the Strait of Georgia, British Columbia. More than 200 humpbacks were resident until wiped out by commercial whaling , a process that was complete by the 1920s  (Gregr 2002, Winship 1998, Merliees 1985). Humpbacks now seem to be making a slow return to the Strait (Gregr 2002). Hopefully, simulation models may be able to capture this process of recolonisation. On the other hand, in Newfoundland, almost a quarter of a million walrus were estimated to be resident before exploitation started in 1800 (Mercer 1967), but have shown no signs of returning. Grey seals in  Newfoundland have a similar status (see Heymans and Pitcher 2004, this volume). As with the globally extinct species discussed above, estimates of ancient biomass may be based on historical records of breeding sites, or, in the case of whaling, on records of whale kills.   Archeological remains of fish bones in middens show that Bluefin tuna, Th u nn us thynnus ,  were at one time distributed along the entire coast of British Columbia and Washington state (Tunnicliffe et al . 2001, and see discussion page 139). Traditional Environmental Knowledge concerning weather and seasons for the hazardous spearing of these fast, giant fish Figure 6.  Two hundred Humpback whales, Megap tera  novae an gliae , were common residents in the Strait of Georgia, BC, before commercial whaling wiped them out early in the 20th century. Nowadays, they may be slowly returning.  Figure 7.  A sea otter, Enhydr a lutris , eating a sea urchin. Sea otters were common residents along North Pacific coasts before being hunted for fur in the 18th and 19th centuries, and were wiped out in British Columbia. In recent years, sea otters have been re-established on Vancouver Island. Ecosystem modelling of sea otters is tricky because they are keystone species, altering the structure of inshore habitats.  Back to the Future Methodology, Page 26   suggest that they were  seasonal visitors to coastal habitats depending on weather and conditions (see Lucas, 2004, this volume). However, they appear to be entirely absent from the region today.   To accommodate dynamic ecosystem modelling, groups that are present early on, but are later absent, have to be included in some way. As mentioned above, unless the species has been grouped with species of similar function, it is important to include in all time periods species that have become locally extinct over period of the series of ecosystem models. One technique that has been used for the Newfoundland series of CUS BTF models (1750, 1900, 1987, 1995: Vasconcellos e t al.  2002), is to set biomass for the ?absent? periods to extremely low levels (zero cannot be used as it causes a software failure). For example, a value of 1*10 -6 tonnes/km 2 has been used for walrus in models of recent Newfoundland ecosystems. At this low level they are essentially extinct (Heymans and Pitcher 2004, this volume). This technique, however, can create some technical problems as, during simulations, the species may undergo an unexpected modelling resurgence if there is enough food for them to do so. It may be possible to ?hold them down? using a biomass forcing function in Ecosim (see Martell 2004, this volume and discussion page 149).   An example of active re-introduction of an extirpated species is the sea otter, Enhydra lutris ,  reintroduced to from Alaska to Vancouver Island in British Columbia in the 1990s (see Lucas 2004, this volume). Sea otters became extinct through hunting in BC before 1900 (Kenyon, 1975), but following reintroduction, today have a established a small but increasing biomass in few areas. Sea otter diet and metabolic parameters are well-known (e.g., Bodkin e t al.  1998, Reidman and Estes 1998) and it is not difficult to incorporate sea otters in ecosystem models (Ainsworth e t al.  2oo2).  The series of models for northern BC should ideally reflect the series of changes: abundant in the ancient past, absent after they were hunted to local extinction for their furs, and then re-introduced. But it is proving hard to include them explicitly in the models for every time period, and in models of restored BTF ?Lost Valley? ecosystems because, at very low biomass, they have ?plenty of food? and tend to undergo a modelling resurgence.    Problems in Modelling Keystone Species  An additional major problem for the BTF modelling here is that sea otters, however, are keystone species, causing large changes in habitat structure (Pitcher 1998, Simenstad e t al.  1978). They alter the type of kelp available as cover to a suite of juvenile fishes and invertebrates by foraging on kelp-eating sea urchins that themselves graze selectively (Riedman and Estes 1990). The consequence is that inshore kelp ecosystems with and without sea otters have very different habitat structure and a different fauna of inshore fishes and invertebrates (Steneck e t al.  2002).   When sea otters were extirpated in the Komandorski islands through hunting, this keystone mechanism may have helped to seal the fate of the sea cow: resurgent kelp-eating urchins would have competed for kelp as food (Anderson 1995).  The open canopy habitat known as ?kelp forest? appears to be dependent on the presence of  sea otters (Steneck e t al.  2002). Before human contact, predation by sea otters on urchins prevented overgrazing on kelp forests (Simenstad e t al.  1978, Estes e t al.  1998). In Alaska, Aleuts seem to have depleted sea otters as early as 2500 BP, causing the urchins to grow larger (Simenstad e t al.  1978). From 1700, fur traders hunted sea otters to the brink of extinction, and kelp forests were then destroyed from over-grazing by urchins released from sea otter predation. Then after 1900 in Alaska, legally-protected sea otter populations increased, and the resultant trophic cascade re-established the kelp forest. Recently, however, kelp forests have disappeared again as sea otter populations have fallen prey to killer whales (Estes e t al.  1998), that  have shifted their diet to otters from pinnipeds after the latter populations declined significantly. The reason for the pinniped declines is still open to debate (Rosen and Trites 2000).   The sea otter?s keystone effect is mediated Figure 8.   Print of an Aleut sea otter hunt at Sanak Island, Alaska.  Aleuts have been hunting sea otters for over 2500 years and devised a special whale-bone barbed dart that detaches from a shaft on contact with the otter. (See also Lucas 2004 , this volume.) Page 27,  Fisheries Centre Research Reports 12(1), 2004  through habitat change that in turn alters feeding opportunities and refuge from predators for inshore fish and invertebrate species (Estes e t al . 1989). Most of these chan ges are based on a living biomass acting as complex structured habitat, not on feeding interactions in a food web, and hence a purely trophic web model cannot simulate them. A routine to put ?non-trophic? mediation effects in Ecosim has been developed (Christensen and Walters 2003), but it is hard to fit the parameters for the interaction in anything other than a post-hoc fashion. In other words, keystone effects, like the sea otter, may be emulated in Ecosim, but not simulated.  The problem here is that spatial complexity and structure of habitats are not modelled explicitly in the EwE dynamic ecosystem system. For aquatic ecosystems this may be acceptable for the majority of cases, except where rooted macrophytes or coral reefs are involved, but it would be entirely unacceptable for most terrestrial ecosystems where plant architecture, both living and dead, provides a template of structured habitat for the vast majority of organisms. Alternative ecosystem modelling techniques, such as ?Atlantis? (Fulton e t al.  2003), may be more appropriate in representing the effects of ?plant architecture?.     CONCLUSIONS  Extinctions cause problems for dynamic ecosystem modelling. This paper has put forward some suggestions about how these issues may be tackled, but some fresh advances in ecosystem modelling techniques are needed before we can approach species extinctions with confidence. BTF is one of the few fisheries policy analysis systems to explicitly and quantitatively deal with the extinction issue (Pitcher 2002).    REFERENCES  Ainsworth, C., Heymans, J.J., Pitcher, T.J. and Vasconcellos, M. (2002) Ecosystem Models  of Northern British Columbia For The Time Periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4), 41pp. Alroy, J. (2001) A Multispecies Overkill Simulation of the End-Pleistocene Megafaunal Mass Extinction. Science  292: 1893-1896. Andersen, P.K. (1995) Competition, predation, and the evolution and extinction of Steller?s Sea Cow, Hyd rodamalis gigas . Marine Mammal Science 11 (3): 391-394.  Bodkin, J. L., Monson, D. H. and Esslinger, G. E. (1998) Mammals: Sea otter. In Okey, T. and Pauly, D. (eds) Trophic Mass-Balance Model of Alaska's Prince William Sound Ecosystem, for the Post-Spill Period 1994-1996. Fisheries Centre Research Reports 6(4): 144pp. Burke, C., Davoren, G.K., Montevecchi, W.A. and Stenhouse, I.J. (2002) Winging Back to the Future: An Historic Reconstruction of Seabird Diversity, Distribution and Abundance in the Northwest Atlantic, 1500?2000. Pages 27-37 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N. (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf. Fisheries Centre Research Reports 10(1): 116 pp. Christensen, V. and Walters, C.J. (2004) Ecopath  with Ecosim: methods, capabilities and limitations. Ecological Modelling ( in press ). Davoren, G.K.,Montevecchi, W.A. and Stenhouse, I.J. (2002) Seabirds. Pages 41-42 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N.  (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf.  Fisheries Centre Research Reports 10(1), 116 pp. Domning, D.P. (1972) Steller?s sea cow and the origin of North Pacific aboriginal whaling. Syesis 5: 187?189. Domning, D.P. (1976) An ecological model for Late Tertiary sirenian evolution in the North Pacific Ocean. Systematic Zoology 25:352-362.  Domning, D.P. (1978) Sirenian evolution in the North Pacific Ocean. University of California Publications in Geological Sciences 118: 1-176. Erlandson, J.M. (2001) Anatomically modern humans, maritime voyaging, and the Pleistocene colonization of the Americas. Pages 1?9 in Jablonski, N.G. (ed.) The First Americans: The Pleistocene Colonization of the New World. California Academy of Sciences, San Francisco. USA. Estes, J.A., Duggins, D.O. and Rathbun, G.B. (1989) The ecology of extinctions in kelp forest communities. Conservation Biology 3: 252?264. Estes, J.A., Tinker, M.T., Williams, T.M. and Doak, D.F. (1998) Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282: 473?476. Fulton, E.A., Smith, A.D.M. and Johnson, C.R. (2003) Effect of complexity on ecosystem models. Mar. Ecol. Prog. Ser. (in press ). Gaskell, J. (2000) Who Killed The Great Auk? Oxford University Press, UK. 228pp. Gregr, E.J. (2002) Whales in Northern BC: Past and Present.  Pages 74-77 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N. (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf. Fisheries Centre Research Reports 10(1): 116 pp. Grieve, S. (1885) The Great Auk or Garefowl: Its History, Archaeology and Remains. Jack, London, UK. Heymans, J.J. and Pitcher, T.J. (2002a) A Picasso-esque View of the Marine Ecosystem of Newfoundland and Southern Labrador: Models for the Time Periods 1450 and 1900. Pages 44-73 in Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) (200 2) Ecosystem Models of Newfoundland For The Time Periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 73 pp. Heyymans, J.J. and Pitcher, T.J. (2002b) A Model of the Marine Ecosystem of Newfoundland and Southern Labrador (2J3KLNO) in the Time Periods 1985-1987 and 1995-1997. Pages 5-43 in Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (eds) (200 2) Ecosystem Models of Newfoundland For The Time Periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Reports 10(5): 73 pp. Hoffecker, J.F., Powers, W.R. and Goebel, T. (1993) The colonisation of Beringia and the peopling of the New World. Science 259: 46-53. Josenhans, H., Fedje, D., Pienit z, R. and Southon, J. (1997) Back to the Future Methodology, Page 28   Early humans and rapidly-changing sea levels in the Queen Charlotte Islands, Hecate Strait, British Columbia. Science 277: 71-74. Kenyon, K.W. (1975) The Sea Otter in the Eastern Pacific Ocean. Dover Publishing, New York, USA.  Lucas, S. (2004) Aboriginal Valu es. Pages 114?116 in Pitcher, T.J. (ed.) Back to the Futu re: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 157 pp. Martell, S. (2004) Dealing with Migratory Species in Ecosystem Models. Pages 41?44 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Martin, P.S. (1984) Prehistori c overkill: the global model. Pages 354-403 in Martin, P.S. and Klein, R.G. (eds) Quaternary extinctions: a prehistoric revolution. University of Arizona Press, Tucson, USA. Mercer, M.C. (1967) Records of the Atlantic Walrus, O dobenus rosmarus rosmarus , from Newfoundland. Journal of the Fisheries Research Board of Canada 24(12): 2631-2635. Merilees, W. (1985) The humpback whales of Strait of Georgia. Waters: Journal of the Vancouver Aquarium, 8: 24pp. Montevecchi, W.A. and Kirk, D.A. (1996) Great Auk ( Pinguinus impennis ). Pages 1-20 in Poole, A. and Gill, F. (eds) The Birds of North America. The American Ornithologists' Union and The Academy of Natural Sciences, Philadelphia, USA, Vol. 260. Morell, V. (1997) Genes may link ancient Eurasians, native Americans. Science 280: 520. Pallas, P.S. (1781) Erl?uterungen ?ber die im  ? stlichen Ocean zwischen Siberien und America geschehenen Entdeckungen. Neue nordische Beitrage zur physikalischen und geographischen Erd, 1. Pitcher, T.J. (1998) Pleistocene Pastures: Steller?s Sea Cow and Sea Otters in the Strait of Georgia. Pages 49-52 in Pauly, D., Pitcher, T.J. and Preikshot, D. (eds) Back to the Future: Reconstructing the Strait of Georgia Ecosystem. Fisheries Centre Research Reports 6(5): 99pp. Pitcher, T.J. (1999) Test ing the Cascade Hypothesis, Indigenous Peoples and the name of the Bering Sea. Pages 1-3 in Trites, A.W., P.A. Livingston, S. Mackinson, M.C. Vasconcellos, A.M. Springer and D. Pauly. Ecosystem Change and the Decline of Marine Mammals in the Eastern Bering Sea. Fisheries Centre Research Reports 7(1): 103pp. Pitcher, T.J. (2002) Restoring The Past To Salvage The Future:  Saving A Ship Of Fools. Pages 4-5 in Pitcher, T.J., Power, M.D. and Wood, L. (eds) (2002)  Restoring the past to salvage the future: report on a community participation workshop in Prince Rupert, BC. Fisheries Centre Research Reports  10(7): 56 pp.   Reidman, M.L. and Estes, J.A. (1990) The sea otter ( Enhydra lutris ): behavior, ecology and natural history. Biol. Rep. US Fish and Wild. Serv. 90(4). Riedman, M.L. and Estes, J.A. (1998) A review of the history, distribution and foraging ecology of sea otters. Pages 4-21 in Van Blarian, G.R. and Estes, J.A. (eds) The community ecology of sea otters. Springer Verlag, Germany. Rosen, D.A.S. and Trites, A.W. (2000)  Pollock and the decline of Steller sea lions: testing the junk-food hypothesis. Canadian Journal of Zoology 78: 1243-1258. Serjeantson, D. (2001) The Great Auk and the Gannet: a Prehistoric Perspective on the Extinction of the Great Auk. Int. J. Osteoarchaeol. 11: 43?55. Simenstad, C.A., Estes, J.A. and Kenyon, K.W. (1978) Aleuts, sea otters and alternate stable state communities. Science 200: 403-411. Steneck, R.S., Graham, M.H., Bourque, B.J., Corbett, D., Erlandson, J.M., Estes, J.A. and Tegner, M.J. (2002) Kelp forest ecosystems: biodiversity, stability, resilience and future. Environmental Conservation 29(4): 436?459. Stejneger, L. (1886) On the extermination of the great northern sea cow ( Rh ytina ). Bulletin of the American Geographical Society 4: 317-328. Stejneger L. (1889) Contribution to the history of the Pallas' cormorant. Proc. U. S. Natl. Mus. 12: 83-88. Steller, G.W. (1899) (1781) De Bestiis Marinus [The beasts of the sea. Translated 1899 by W. Miller and J. E. Miller]. Pages 180-201 in D. S. Jordan (ed.) The fur seals and fur seal islands of the North Pacific Ocean. Part 3. U.S. Government Printing Office, Washington, D.C., USA. Tegner, M.J. and Dayton, P.K. (2000) Ecosystem effects of fishing in kelp forest communities. ICES Journal of Marine Science 57: 576?589. Tunnicliffe, V., O'Connell, J.M. and McQuoid, M.R. (2001) A Holocene record of marine fish remains from the Northeastern Pacific. Marine Geology 174: 197-210. Vasconcellos, M., Heymans, J.J. and Pitcher, T.J. (2002) Historical Reference Points For Models of Past Ecosystems in Newfoundland. Pages 7-12 in Pitcher, T.J., Vasconcellos, M., Heymans, J.J., Brignall, C. and Haggan, N. (eds) (2002) Information supporting past and present ecosystem models of Northern British Columbia and the Newfoundland shelf. Fisheries Centre Research Reports 10(1): 116 pp. Winship, A. (1998) Pinnipeds and Cetaceans in the Strait of Georgia. Pages 53-57 in Pauly, D., Pitcher, T.J. and Preikshot, D. (eds) Back to  the Future: Reconstructing the Strait of Georgia Ecosystem. Fisheries Centre Research Reports 6(5): 99pp.   For discussion following oral presentation of this paper, see page 149.  Page 29, Fisheries Centre Research Reports 12(1), 2004   CHALLENGING ECOSYSTEM SIMULATION MODELS W ITH CLIMATE CHANGE:  THE ?P ERFECT STORM?   Tony Pitcher and Robyn Forrest Fisheries Centre, UBC   ABSTRACT   When ecosystem models of the past are constructed, appropriate climate regimes need to be incorporated. Likewise, the effects of possible future climate changes on ecosystem structure and function must be included in forecasts of sustainable fisheries in reconstructed ecosystem. This paper examines how these issues might prejudice the BTF policy process. We show examples of models driven by inter-annual climate indices or by direct indicators of primary production.     Alterations in ecosystem structure due to climate change represent a major challenge to Back-to-the-Future (BTF) investig ations. Climate changes that need to be addressed in BTF ecosystem simulations span time scales ranging from short inter-annual fluctuations to the major long-term shifts that result in ice ages. There are two aspects to the problem and each of them forms the basis of one of the most common criticisms of the BTF approach. First, the reconstruction of past ecosystems to use as future policy goals may be prejudiced if those past ecosystems existed under different climate regimes. Secondly, BTF relies on forecasts made by sustainably fishing restored past ecosystem states in which simulations are projected into the future ? the ?fished Lost Valley? scenarios ? and so these forecasts may not be viable unless likely climate change is taken into account. The two aspects of the problem differ fundamentally in scientific terms. Past climate changes are inherently knowable, can be estimated from reported observations, and these estimates, if poor, can be improved. Future climate changes, in common with all scientific forecasts such as weather forecasting, are unverifiable until the specified future time is reached, and so the best we can do is to project a series of likely scenarios, some of which may be more likely than others. In its most serious form the ?climate change? criticism goes something like this. Even if the climate of past times is well                                                            Pitcher, T.J. and Forrest, R. (2004) Challenging ecosystem simulation models with climate change: the ?Perfect Storm?. Pages 29?38 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.  understood and the ecosystem models of the past adjusted accurately to take account of those changes, past ecosystem states cannot be used for future policy goals. We can expect climate to induce differences among past ecosystems, the present day ecosystem from which we have to commence the reconstruction process, and the projected future. This pape r aims to analyse these issues and assess the degree to which the BTF process might, in practice, be prejudiced by them.  Types of climate change  Oceanographic influences on the living organisms in marine ecosystems are mediated ultimately through temperature and ocean circulation currents. Proximal factors driven by these changes affect thermocline depth and the upwelling of nutrients from sediments that determine phytoplankton production. Freshwater runoff, driven by rainfall, and ice melt, driven by temperature, can also have a profound influence Figure 1.  On November 1st 1991, the ?Perfect Storm? in the north-western Atlantic was accurately forecast by meteorologists ( top panel : composite radar picture, NOAA ) and many lives were saved, even though the swordfish vessel, the And rea Gail,  sank with all hands ( lower pane l: pre-production wate rcolour from film ) because they ignored the warnings (Junger 1997). The science of weather forecasting is pretty good these days, although, in October 15th 1987 the British meteorological office was blamed for failing to predict the most damaging storm (18 people died) to hit southern Britain since 1703 (26 th November, 8000 people died; Sutton 2003).  Likewise, it may be both encouraging and hazardous to attempt to forecast the state of marine ecosystems under the influence of inter-annual climate fluctuations, climate-induced regime shifts and one-off catastrophic events.  Back to the Future Methodology, Page 30  on inshore marine ecosystems. Changes in ocean currents are important in the physical dispersal of planktonic larval stages of fish and invertebrates. Temperature changes can affect fish physiology directly, but can also determine global wind patterns, that in turn affect ocean currents. Hence, a complex of climatic factors affects the templates of habitat offered by the marine environment to the suites of organisms that compose its ecosystems (Review: Barange 2002).  Time scales of climate change  These climate influences occur over a range of time and spatial scales. Seasonal climate changes are those with which we are most familiar and, especially in polar regions, can have a dramatic effect on the structure and functioning of marine ecosystems. In this paper the ecosystem modelling in which we are interested is based on annual changes in biomass, and so seasonal changes are not considered further here, although they can be incorporated into Ecosim modelling (S. Martell, unpublished). Inter-annual changes include more-or-less random fluctuations in temperature and ocean currents from year-to-year, whose variance is characteristic of a particular geographical location. It is this variance that is most likely to increase under the influence of a global warming trend.  Inter-annual changes also include major ocean forcing such as El Ni?o (male child), named because the main effects occur at ?navidad? (Christmas). Its primary effect is to shift the equatorial current in the tropical Pacific to a greater or lesser degree, with a time span for its effects of 6-18 months. Spring warming of the sea to the north of Indonesia causes the Eastward warm equatorial current to increase. This current then swings poleward off South America to displace and overlay the cold northerly Humboldt current, with origins in the Antarctic ice melt, whose upwelling normally drives exceptional marine production off Peru. Exactly what triggers El Ni?o to start is not yet known, although the Earth?s spin is reduced by the mass of less dense warm water. The opposite effect, La Ni?a, gave rise to the concept of ENSO (El Ni?o Southern Oscillation: Figure 2). Although based in the central and southern Pacific, ENSO?s influence extends to the North Pacific, Indian and Atlantic oceans. Records up to the 1970s indicated major ENSO events occurring about once every 15 years, but in the past two decades they have become up to three times as frequent.  Medium-term, quasi-cyclic changes occur over larger ocean regions on decadal time scales, for example the Pacific Decadal Oscillation (PDO: Figures 2, 3) with a period of 20-30 years. These are major shifts in currents and temperatures, Figure 2.  Temperature contour and water flow diagrams showing ENSO and PDO from a North Pacific perspective. Colour-scaled values are degree Celsius deviation from long-term  mean (from NOAA). Figure 3.  Recent changes in the Pacific Decadal Oscillation (PDO). Clear evidence  of decadal regime shift revealed by sea surface temperature anomalies in the North Pacific. (NOTE: coloured figure may not show up well in grey scale).  Panels show left-to-right, (a) 1970-1976, cool phase of PDO; (b) 1977-1983, warm phase of PDO; (c) 1999-2003, strong cool phase of PDO. These temperature changes are paralleled by sea level pressure and wind patterns. (Diagram from Peterson  and Schwing 2003.) Page 31, Fisheries Centre Research Reports 12(1), 2004   sometimes occurring rapidly between relatively stable periods.  A number of longer-term cycles have been suggested (e.g. Klyashtorin 2001), the most compelling of which is an approximately 62-year cycle in the Pacific, termed El Viejo/La Vieja [old man/old woman] (Chavez e t al.  2003) (Figure 4). Very long-term climate trends can lead to ice ages and consequent sea-level changes. In addition, there is good evidence for a dramatic human-made recent global warming trend (Figure 5).   Four questions are critical to the use of climate influences as a part of the BTF process. (1) Can we drive and/or tune past models using time series of climate or surrogate climate data? (2) Are these models stable? (3) Are the observed biomass dynamics realistic and do they emulate observed regime shifts? (4) Can we determine and ?lock on? to the appropriate state of ecosystem for the model of a past time period?  HOW DO CLIMATE CHANGES AFFECT  AQUATIC ECOSYSTEMS?  The vast majority of papers in the ?effects of climate on fisheries? literature describe climate impacts on a single species at single geographical location, and only a few deal with populations of the same species over a wider geographical area. As might be expected, a fair number of well-argued publications supported by solid data cover the impact of climate changes on fisheries recruitment. There are also a small number of synoptic, global-scale analyses of climate-induced changes to groups of fisheries of interest such as the small pelagics. There are very few attempts (e.g., Barange 2002) to deal with the integrated effects of climate on whole ecosystems, and even fewer attempt to compare ecosystem-scale effects over wide areas. This paper therefore includes a review of recent publications that shed light upon the ways in which climate changes alter fish Figure 4 . El Viejo/La Vieja marine climate regime analysis for the Pacific. In cooler conditions, anchovies dominate (La Vieja), while in warmer regime (El Viejo) sardines are abundant. Spatial SST and atmospheric circulation anomalies are shown for each regime (glo bes). Note that the eastern Pacific is out of phase with the central North and South Pacific. Some indices suggest rapid shifts (dashed line), whereas othe rs are gradual (solid line). Low sea surface height (TOPEX) and high chlorophyll (California Cu rrent) in the cool anchovy regime mean a shallow thermocline/nutricline. Associated basin-scale current sy stems support recent stronger California Current and a weaker Kuroshio Current. (Diagram from Chavez e t al.  2002).  Back to the Future Methodology, Page 32  populations and fisheries within an ecosystem context.   It seems that shifts in ocean climate regimes can alter ecosystem structure quite quickly, and these may be faster at lower trophic levels (Barange 2002, Hare et al. 1999). Changes in wind patterns affect oceanic circulation, salinity, thermocline depth and primary production; changes in the distribution and abundance of predators and prey influence fish, marine mammal and bird populations (Barange 2002). Sometimes changes can affect similar species within a single domain in opposite ways. Surprisingly, some hold that climatic regime shifts can have opposite effects on the same species in different ocean domains  (Benson and Trites 2002).  Fish growth is often affected directly if water temperature alters (e.g., halibut; Clark e t al.  1999), but there is usually little attempt to partition this effect into a direct metabolic influence and indirect effects mediated though the food web. Extreme temperatures can directly affect the physiology of migrating salmon (Hinch e t al.  2002).  It is well documented that climate shifts can have a serious impact on fisheries (e.g., Japan; Kawasaki and Omori 1995), especially when they coincide with overfishing, as in the classic collapses of the Monterey sardine in the 1950s and the Peruvian anchoveta in 1971 (see accounts in Pitcher and Hart 1981). And, more recently, recruitment of cod in the heavily overfished North Sea appears to be threatened by climate warming trends (O'Brien e t al.  2000).   Time series of catches and other data often suggest synchronous changes over large ocean basins, suggestive of climatic and oceanographic factors at work in determining abundance. For example, using catch time series several centuries long from the Mediterranean and adjacent Atlantic areas, Ravier e t al.  (2001) demonstrated 7-fold fluctuations in abundance, and synchronised 100-year and 20-year cycles, in the traditional tuna ?tonnara? trap fisheries that formerly caught bluefin tuna on their annual spawning migrations Moreover, coherent patterns observed across large regions of the Pacific demonstrate the strong role of climatic forcing in determining the size fish populations (Hollowed e t al.  2001). Catch records suggest that warmer years and regimes may lead to higher fisheries production (e.g., sablefish; King e t al.  2001) in higher latitudes (e.g., Beamish 1993).   Fishery catches, however, may be influenced by a number of factors and so other means have been explored to examine climate-linked changes. For example, nitrogen isotopes in lake sediments demonstrated large changes over 300 years of Alaskan sockeye salmon abundance related to climate (Finney e t al.  2000). Similar analyses Figure 5.  Global warming trend shown in average annual global  temperate anomalies, 1880-2000, from land and sea. Source: National Climatic Data Center, NOAA, USA. 0 20 40 60 80 100 120 140Simulated yearsFigure 6.  Simulated climate influence time series constructed from 62 and 20-year cycles (sine waves), and ENSO anomaly (triangular probability distribution). Note that co incidence of two cycles can lead to ?rapid shift of regime?, and to ?stable plateau? periods. Authors? simulations show that these effects depend on relative wavelengths and starting point. Page 33, Fisheries Centre Research Reports 12(1), 2004   spanning 2,200-years reveal very large shifts in abundance resulting from climatic forcing, far exceeding the decadal-scale variability recorded from catches during the past 300 years. For example, salmon declined from 2100?1200 BP, but were more abundant from 800?100 BP (Finney e t al.  2002). On equally long time scales, the abundances of 1200 years of Pacific sardine and Northern anchovy off the California coast (Baumgartner e t al.  1992) alternate with the salmon, giving some clues as to the ocean mechanisms at work (Finney e t al.  2002). The regime of high clupeid abundance (2000?800 BP) is confirmed by archeological studies (Tunnicliffe et al.  2001).   Alheit and Hagen (1997) describe an example of long-term climate forcing of European herring and sardine populations. In the Skagerrak, since the 10th century, there were nine boom periods for inshore herring fisheries, each lasting several decades. Otherwise, the herring fishery was very small. Some other European herring fisheries coincide (English Channel and the Bay of Biscay), whereas others (Norwegian herring and sardines) alternate with these periods, apparently driven by negative anomalies in the North Atlantic Oscillation index (more sea-ice in the Arctic, cold European air and water temperatures, fewer westerly winds). In Norwegian waters, the North Atlantic Oscillation index relates to recruitment of North East Arctic cod (God ?  2003), while sea surface temperature is linked to Barents Sea capelin and Norwegian spring spawning herring stocks, although heat flux, ice cover and heat transport are also important variables (Stiansen e t al.  2002).    Climate influences on recruitment are often a very important mechanism. In the North Sea, 22 years of data on climate during larval stages explained more than 70% of recruitment variability leading to models that could forecast recruitment in the summer of the spawning year (Svendsen et al.  1995). In coho salmon climate factors determine cohort strength; faster growing fish better survive the first winter at sea when upwelling nutrients have led to better plankton feeding conditions in the previous summer and autumn (Beamish and Mahnken 2001). Rodhouse (2001) shows how squid recruitment is correlated with synoptic oceanographic data. For example, in the eastern Pacific coastal upwelling system catches in a squid fishery for Dosidicus gigas  are linked to the El Ni?o (ENSO) cycle. Twenty-fold fluctuations in mackerel recruitment in the Gulf of St. Lawrence were related to copepod abundance, which was negatively related to climate expressed as freshwater discharge (Runge e t al.  1999). Many prawns recruit like this too. 12131415161716381688173817881838188819381988YearSummer Temperaturedegrees CelsiusFigure 7 . Summer temperatures in northern British Columbia as reconstructed from pine tree rings.  Dotted horizontal line shows mean. Note recent warming trend. Data from Szeicz and MacDonald (1995). 0. 61 . 01 . 40 5 1 0 1 5 20Simulation yearsBiogenic silica indexFigure 8.   Twenty-year time series of data used to drive primary production in dynamic ecosystem model forecasts of Lake Malawi. Data is based on published time series of biogenenic silica in lake sediments, in randomized order, normalised to unit mean, and the variance adjusted iteratively to fit likely extreme lake biomass values. Data from Johnson e t al.  2001. Back to the Future Methodology, Page 34  The recruitment ? climate relationship may be quite complex. For example, the spring phytoplankton bloom can vary by up to 6 weeks in Newfoundland, driven by the amount of colder, fresher water from glacial runoff.  First-feeding cod larvae have a precise dietary requirement: the nauplii of a copepod Calanus finmarchicus  in the spring bloom. 'Match/mismatch' feeding conditions drive cod recruitment success, and global warming may prejudice recovery of depleted cod stock by creating long runs of ?mismatch? years (Conover e t al.  1995). In a similar way, climate-driven fluctuations of sardine, hake and mackerel populations in the northern part of the California current appear to be linked to specific diatoms required by sardine larvae (McFarlane and Beamish 2001).  In the Pacific North-west, the PDO cycle has a strong influence on sockeye, pink, chinook, and chum salmon, herring and halibut, especially juveniles (Clark e t al.  1999, Beamish and Bouillon 1993, Mantua e t al.  1997). The El Viejo/La Vieja cycle describes an alternation between warm eastern boundary currents favouring sardine and colder conditions favouring anchovy regimes. Moreover, in this system the transitions between different regimes are relatively abrupt, but may be out of phase in different parts of the Pacific (Figure 4).   Other less obvious ecological effects are sometimes found. For example, a long-term trend for warmer water has stabilized the water column in Lake Tanganyika, reducing mixing, so that primary production is reduced by 20% and fish production by 30% (O?Reilly e t al.  2003).  Fish species with life spans comparable to or exceeding the duration of adverse conditions may weather out the adverse period of a cycle, but at low population sizes, cascade effects can impede population growth in the good period. Fish with life spans shorter than the duration of adverse conditions can only be managed by linking catches to the environmental conditions, preferable using a delayed response (MacCall 2002).   DRIVING ECOSYSTEM MODELS  Clearly, the effect of climate change has to be accommodated in forecasts using ecosystem simulation models as much as possible. To do this, primary production, and other parameters of Clim ate seriesModelled Biomasses1750                                      1800                  1850210Figure  9.  Example of 46-compartment whole ecosystem model driven by 100 years of a marine climate index based on tree ring data. Annual values of climate series shown at bottom panel and were used to drive primary production. Modelled biomass changes relative to starting values shown above and below the unity line in upper panel. Starting model is a reconstruction of Northern British Columbia as it may have been in 1750. Marine climate data from Gedalof and  Smith (2001). Page 35, Fisheries Centre Research Reports 12(1), 2004   ecosystem models, such as stock-recruitment relationships, may be driven in a variety of ways.  Driving models with forcing functions   Although precise forecasts of inter-annual climate changes may never be possible, randomized selections of such data, or functions that emulate past climate changes, can be used to drive forecasts on the basis of likely scenarios. Forcing functions may be based upon empirical inter-annual variation, decadal or longer-period oscillations, or climate proxies such as a local upwelling index. Longer term climate cycles may be included in the forcing function, like the 62-year ?La Vieja/El Viejo? alternation between warm/cold eastern boundary current sardine/anchovy regimes (Chavez e t al.  2003). All these factors can fairly easily translated into a driving variable for ecosystem modelling (e.g., Figure 6). The algorithm could be modified to take account of  large ENSO events that may trigger PDO shifts (Peterson and  Schwing 2003).  Residuals between historically measured biomasses and simulated biomasses can be minimized in Ecosim models by comparing fits of a range of climate forcing functions. Climate forcing of modelled phytoplankton production may be sufficient, but in some cases climate forcing of recruitment parameters may also be useful for some fish species.  D riving models with climate data  Rather than use mathematical surrogates, tree rings can supply long historic records of inter-annual temperature changes in a region (e.g., Figure 7, northern British Columbia; Szeicz and MacDonald 1995). In the deep sea, growth rings of bamboo coral have been used in a similar fashion (Koslow and Thresher 1996). Sea surface temperature data has been used to drive a biomass model for Japanese sardine (Noto and Yashuda 2003). Sea temperature anomalies successfully improved biomass fits in an Ecosim model of the English Channel (Stanford 2004). In small pelagic fish in upwellings world-wide, production rate rather than biomass seems to be the best correlate for climate regimes (Jacobson e t al.  2001) and hence the best variable to force in single species models. Multi-species modelling driven by temperature time series suggests that species respond differently to climate depending on their position in the food web (Jurado-Molina and Livingston 2002).   Lehodey (2001) has modelled the spatial effect of warmer and cold waters. In the tropical Pacific, ENSO affects a cold tongue of upwelling water that favours high production adjacent to warm unproductive pools. A spatial production model of skipjack tuna uses spawning area, larval and juvenile transport, adult tuna temperature preferences and forage fish prey driven by primary production. Observed movements of skipjack confirm the model results, which show ENSO driving an out-of-phase pattern between the western Pacific region and the cold tongue.   In some cases, biogenic silica deposits in sediments, which track the abundance of diatoms, may accurately reveal the past annual changes in primary production (e.g., Lake Malawi; Johnson e t al.  2001), and such data has been used to drive forecasts in ecosystem simulation models (Figure 8).  Figure 9 illustrates an ecosystem model, representing a past time (1750) in Northern British Columbia, with phytoplankton production driven by a 10-year time series of marine climate data (transformed tree ring data; Gedalof and Smith 2001). Figure 10 shows separate plots of some of the groups in the model: large climate-driven changes in coho and juvenile halibut Figure 10.  Biomasses of some of the individual groups in the ecosystem model driven by climate time series shown in Figure 9. Top left panel: phytoplankton; Top right panel: adult ha libut; lower left panel: coho salmon; lower left panel: juvenile halibut.  Back to the Future Methodology, Page 36  emulate those expected from the literature (Mantua e t al.  1997, Clark e t al.  1999).  Information about specific past times from climate time series might be used to adjust each of a series of BTF ecosystem models to the appropriate contemporary regime conditions. This has not yet been attempted, however, because there is a logical complication. Starting values for the ecosystems of ?sustainably fished Lost Valley? analyses would have to be re-adjusted to a regime appropriate for starting the rebuilding process. Past climate data could at least enable past models to avoid major fluctuations compared to the present. In addition, the problem might be minimised by approximating an average ecosystem state over a period of 10-20 years.    THE ?C AST OF PLAYERS?  TECHNIQUE : A SUGGESTION  To emulate changes in species composition in an ecosystem model as climate changes, the modeling system could perhaps be modified to use a ?cast of players?, members of which might be brought on-stage and off-stage when conditions are appropriate (Pitcher 2004). In the ?on-stage? condition a species would play its full part in the food web of functioning model, acting as a predator, prey and competitor, and as an actor in any mediation processes, according to its model parameters and diet matrix. When ?off-stage? the species would play no part in the model dynamics. On-stage and off-stage conditions would be set, for example, by the value of an external time series, such as water temperature or a climate index. By bringing on stage members of a food web at different times and temperatures, a large number of intermediate ecosystems might be modelled. The technique could also be used to emulate species re-introductions, or recolonisations, following local extinctions (see Pitcher 2004, this volume).  Some archaeological data sets may provide useful test-beds for the ?cast of players? technique. For example, in the Cueva de Nerja, Andalusia, Spain (Figure 11), human middens reveal the fish that early Mediterranean people were eating over a 9000-year sequence (Morales e t al.  1994, Rosello-Izquierdo and Morales-Muniz 2001). Early in the sequence, from about 14000 BP, the human diet consisted of a sparid fauna similar to the present, but, during a pluvial period at the end of the last Ice Age between 11000 BP and 9000 BP, humans were eating large cod and haddock, a fauna typical of Norway today. The midden fish bones the show that, by 8000 BP, a typical Mediterranean fauna had returned. The shift from Mediterranean to Nordic and then back to Mediterranean ecosystems might be emulated using the ?cast of players? driven by ancient temperature or climate proxies. Stratigraphic archaeological data could be used to ?tune? the process.  If successful in a trial such as the above, the ?cast of players? technique could also be used to forecast the consequences of global warming in a marine ecosystem by including a set of species from adjacent warmer ocean areas as well as those present today, for example, and then driving the actors on-stage and off-stage with a trend in temperature or climate factors.   CONCLUSIONS  We are now in a position to provide some preliminary answers to the four questions raised in the introduction to this paper. (1) It is certainly possible to make credible attempt to drive and/or tune models of the past using time series of climate or surrogate climate data. (2) Only a few climate-driven whole-ecosystem models have Figure 11.  Top left: the first excavations taking place at Nerja, Adalusia, in the 1920s. Top right: human skull from the cave dated at 18,0000BP; Lower left: cave art of food fish, possibly Pagrus ; Lower right: more cave art food. (From www.cuevadenerja.es). Page 37, Fisheries Centre Research Reports 12(1), 2004   been constructed, all of which have appear to be stable, but far more will have to be done before we are sure of their stability. (3) Whole-ecosystem models have yet to emulate observed regime shifts, and the validity of their biomass dynamics needs more investigation. (4) We do not yet know if we can determine and ?lock on? to the appropriate state of ecosystem for the model of a past time period. Clearly, it is early days for climate-driven whole-ecosystem modelling.   Climate cycles mean that fisheries management aimed at rebuilding stocks may have to use a much longer planning horizon has been typical. MacCall (2002) suggests that, during adverse periods little rebuilding may occur even if fishing is halted, while in favourable periods, depleted populations of large predators allow smaller unfished competitors to thrive, again inhibiting population growth. Consequently, rebuilding of apex predators may require a century or more. Again, for policy work based on whole-ecosystem modelling, these issues need systematic investigation using the new climate-driven simulations.  Climate-change affects species mix and shifts centres of production.  Everett (1997) warns that,   ?The positive effects of climate change, such as longer growing seasons, lower natural winter mortality, and faster growth rates in higher latitudes, may be offset by negative factors such as changes in established reproductive patterns, migration routes, and ecosystem relationships. Serious consequences could occur where these factors interact with pervasive over-fishing, diminishing nursery areas, and extensive coastal pollution.?  Barange (2002) advises that multi-disciplinary research is required to understand the challenges of climate change. Moreover, as in other areas of fishery management, suitable actions consequent upon accurate forecasts of the effects of climate fluctuations may be hard to implement as policy. All sorts of human constraints may apply, such as lack of understanding, failure to appreciate uncertainties, and unanticipated reactions depending on unequitable benefits (El Ni?o, Broad e t al.  2002).  The ?Perfect Storm? in November 1993 was formed when two independent meteorological  phenomena occurred together. The storm was accurately forecast and its likely impacts well understood, but when it arrived, those caught up in its fury were both astonished and ill-prepared. It is to be hoped that forecasts of the effects of directional climate change - global warming - on natural ecosystems in the sea will not catch us so unprepared. While recent effort has been rightly focused on the disastrous effects of an era of outrageous and uncontrolled overfishing, it is sobering to realise that climate may bring about equally devastating changes. Finney e t al.  (2002) warns that   ? an unprecedented shift to a very low productivity regime, lasting centuries, can occur even without the influence of fisheries and other anthropogenic impacts?.  Today, with both overfishing and climate shift independently caused by human actions, we may have unwittingly set the stage for a ?Perfect Storm? of changes in the ocean ecosystems.   REFERENCES  Alheit, J. and Hagen, E. (1997) Long-term climate forcing of European herring and sardine populations. Fisheries Oceanography 6(2): 130-139. Barenge, M. (2002) Influence of climate variability and change on the structure, dynamics and exploitation of marine ecosystems. Pages 57-82 in Hester, R.E. and Harrison, R.M. (eds) Global Environment Change. Royal Society of Chemistry, Cambridge, UK.  Baumgartner, T.R., Soutar, A. and Ferreira-Bartrina, V. (1992) Reconstruction of the history of Pacific sardine and northern anchovy populations over the last two millennia from sediments of the Santa Barbara Basin, California. CalCOFI Rep. 33: 24-40. Beamish, R.J. (1993) Climate and exceptional fish production off the west coast of North America. Canadian Journal of Fisheries and Aquatic Sciences 50: 2270-2291. Beamish, R.J. and D.R. Bouillon (1993) Pacific salmon production trends in relation to climate. Canadian Journal of Fisheries and Aquatic Sciences 50: 1002-1006.  Beamish, R. J., and Mahnken, C. (2001) A critical size and period hypothesis to explain natural regulation of salmon abundance and the linkage to climate and climate change. Progress in Oceanography 49: 423-437. Benson, A.J. and Trites, A.W. (2002) Ecological effects of regime shifts in the Bering Sea and eastern North Pacific Ocean. Fish and Fisheries 3(2): 95-113. Brayne, M. (2003) The Greatest  Storm: Britain's Night of Destruction, November 1703. Sutton, UK. 288 pp. Broad, K., Pfaff, A.S.P. and Glantz, M.H. (2002) Effective and equitable dissemination of seasonal-to-interannual climate forecasts: Policy implications from the Peruvian fishery during El Ni?o 1997-98. Climatic Change 54(4): 415-438. Chavez, F.P., Ryan, J., Lluch-Cota, S.E. and ?iquen, M.C. (2003) From anchovies to  sardines and back: multidecadal change in the Pacific Ocean. Science 299: 217-221. Clark, W.G, Hare, S.R., Parma, A.M., Sullivan, P.J. and Trumble, R.J. (1999) Decadal changes in growth and recruitment of Pacific halibut ( Hippoglossus stenolepis ). Canadian Journal of Fisheries and Aquatic Sciences 56: 242-252.  Conover, R.J., Wilson, S., Harding, G.C.H. and Vass, W.P. (1995) Climate, copepods and cod: some thoughts on the long-range prospects for a sustainable northern cod fishery. Clim. Res. 5(1): 69-82. Everett, J.T. (1997) Fisheries and climate change: the IPCC second assessment. Pages 109-115 in Hancock, D.A., Smith, D.C., Grant, A. and Beumer, J.P. (eds) Developing Back to the Future Methodology, Page 38  and sustaining world fisheries resources. The state of science and management. Proc. 2nd World Fisheries Congress, CSIRO, Australia. 797pp. Finney, B.P., Gregory-Eaves, I., Sweetman, J., Douglas, M.S.V. and Smol, J.P. (2000) Impacts of climatic change and fishing in Pacific salmon abundance over the past 300 years. Science 290: 795-799. Finney, B.P., Gregory-Eaves, I., Douglas, M.S.V. and Smol, J.P. (2002) Fisheries producti vity in the northeastern Pacific Ocean over the past 2,200 years. Nature 416: 729-733. Gedalof, Z. and Smith, D.J. (2001) Interdecadal climate variability and regime-scale shifts in Pacific North America. Geophysical Research Letters 28(8): 1515-1518.  God?, O.R. (2003) Fluctuation in stock properties of north-east Arctic cod related to long-term environmental changes. Fish and Fisheries 4(2): 121-137. Hare, S.R., Mantua, N.J. and Francis, R.C. (1999) Inverse Production Regimes: Alaska and West Coast Pacific Salmon. Fisheries 24(1): 6-15.  Hinch, S.G., Standen, E.M., Healey, M.C. and Farrell, A.P. (2002) Swimming patterns and behaviour of upriver-migrating adult pink ( Oncorhynchus gorbuscha ) and sockeye ( O. ner k a ) salmon as assessed by EMG telemetry in the Fraser River, British Columbia, Canada. Hydrobiologia 483: 147-160. Hollowed, A.B., Hare, S.B. and Wooster, W.S. (2001) Pacific basin climate variability and patterns of Northeast Pacific marine fish production. Progress in Oceanography 49: 257-282. Jacobson, L.D., De Oliveira, J.A.A., Barange, M., Cisneros-Mata, M.A., Felix-Uraga, R., Hunter, J.R., Kim, J.Y., Matsuura, Y., ?iquen, M., Po rteiro, C., Rothschild, B., Sanchez, R.P., Serra, R., Uriarte, A. and Wada, T. (2001) Surplus production, variability, and climate change in the great sardine and anchovy fisheries. Can. J. Fish. Aquat. Sci. 58(9): 1891-1903.  Johnson, T.C., Barry, S.L., Chan, Y. and Wilkinson, P. (2001) Decadal record of climate variability spanning the last 700 years in the southern tropics of East Africa. Geology 29(1): 83-86. Junger, S. (1997) The Perfect Storm: A True Story of Men Against the Sea. Norton, NY, USA. 227 pp. Jurado-Molina, J. and Livingston, P. (2002) Climate-forcing effects on trophically linked groundfish populations: implications for fisheries management. Canadian Journal of Fisheries and Aquatic Sciences 59(12): 1941-1951. Kawasaki, T. and Omori, M. (1995) The impacts of climate change on Japanese fisheries. Pages 523-528 in Beamish, R.J. (ed) Climate Change and Northern Fish Populations. Can. Spec. Pub. Fish. Aquat. Sci. 121. King, J.R., McFarlane, G.A. and Beamish, R.J. (2001) Incorporating the dynamics of marine systems into the stock assessment and management of sablefish. Progress in Oceanography 49: 619-639. Klyashtorin, L.B. (2001) Climate change and long-term fluctuations of commercial catches. The possibility of forecasting. FAO Fish. Tech. Pap. 410: 86 pp.  Koslow, J.A. and Thresher, R. (1 996) Climate and fisheries on the southeast Australian continental shelf and slope. CSIRO, Hobart, Australia. 81 pp.  Lehodey, P. (2001) The pelagic ecosystem of the tropical Pacific Ocean: dynamic spatial modelling and biological consequences of ENSO. Progress in Oceanography 49: 439-468. MacCall, A.D. (2002) Fishery-management and stock-rebuilding prospects under conditions of low-frequency environmental variability and species interactions. Bulletin of Marine Science 70(2): 613-628. McFarlane, G.A. and Beamish, R.J. (2001) The re-occurrence of sardines off British Columbia characterises the dynamic nature of regimes. Progress in Oceanography 49: 151-165. Mantua, N.J., Hare, S.R., Zhang, Y., Wallace, J.M. and Francis, R.C. (1997) A Pacific interdecadal climate oscillation with impacts on salmon production. Bulletin of the American Meteorological Society 78(6): 1069-1080. Morales, A., Rosello, E. and Canas, J.M. (1994) Cueva de Nerja (prov. Malaga): a close look at a twelve thousand year ichthyofaunal sequence from southern Spain. Pages 253- 262 in van Neer, W. (ed.) Fish Exploitation in the Past. Ann. Sci. Zool. Mus. Roy. L?Afrique Centrale, Tervuren, Belg. 274. Noto, M. and Yasuda, I. (2003) Empirical biomass model for the Japanese sardine, S a rdinops melanostictus , with sea surface temperature in the Kuroshio extension. Fisheries Oceanography 12: 1-9. O'Brien, C.M., Fox, C.J., Planque, B. and Casey, J. (2000) Climate variability and North Sea cod. Nature 404: 142. O'Reilly, C.M., Alin, S.R., Plisni er, P., Cohen, A.S. and Mckee, B.A. (2003) Climate change decreases aquatic ecosystem productivity of Lake Tanganyika, Africa. Nature 424: 766-768.  Peterson, W.T. and Schwing, F.B. (2003) A new climate regime in northeast pacific ecosystems. Geophysical Research Letters 30(17):  1896., 4pp.  Pitcher, T.J. (2004) ?Back To The Future?: A Fresh Policy Initiative For Fisheries And A Restoration Ecology For Ocean Ecosystems. Phil. Trans. Roy. Soc. ( in press ). Ravier, C., and Fromentin, J-M. (2001) Long-term fluctuations in the eastern Atlantic and Mediterranean bluefin tuna population. ICES J. Mar. Sci. 58: 1299-1317. Rodhouse, P.G. (2001) Managing and forecasting squid fisheries in variable environments. Fisheries Research 54(1): 3-8. Rosello-Izquierdo, E. and Morales-Muniz, A. (2001) A new look at the fish remains from Cueva de Nerja. Fish Remains Working Group, Paih ia, New Zealand, October 2001. Runge, J.A., Castonguay, M., De Lafontaine, Y., Ringuette, M. and Beaulieu, J-L. (1999) Covariation in climate, zooplankton biomass and mackerel recruitment in the southern Gulf of St Lawrence. Fisheries Oceanography 8(2): 139-149. Szeicz, J.M. and MacDonald, G.M. (1995) Dendroclimatic reconstruction of summer temperatures in northwestern Canada since AD 1638 based on age-dependent modeling. Quaternary Research 44: 257-266. Stanford, R. (2004) Ecosystem Modelling and Management of English Channel Fisheries. Fisheries Centre Research Reports ( in press ). Stiansen, J.E., Loeng, H. Svendsen, E., Pettersson, L., Johannessen, J., Furevik, T., Handegaard, N.O. and Fredo, O. (2002) Climate-fish relations in Norwegian waters. Fisken og Havet [Fisheries and Oceans] 12.  Svendsen, E., Aglen, A., Iversen, S.A., Skagen, D.W. and Smestad, O. (1995) Influence of climate on recruitment and migration of fish stocks in the North Sea. Pages 641-653 in Beamish, R.J. (ed.) Climate Change and Northern Fish Populations. Can. Spec. Pub. Fish. Aquat. Sci.  121. Tunicliffe, V., O?Connell, J.M. and McQuoid, M.R.A. (2001) Holocene record of marine fish remains from the Northeastern Pacific. Mar. Geol. 174: 197?210.   For discussion following oral presentation of this paper, see page 142.   Page 39, Fisheries Centre Research Reports 12(1), 2004  TUNING ECOSYSTEM  MODELS TO PAST DATA   Richard Stanford Fisheries Centre, UBC   ABSTRACT   This note sets out how whole ecosystem simulation models may be tuned using past surveys or fisheries assessment outputs.    The question is often asked as to whether these Ecopath-with-Ecosim  models could actually influence decision-making? Their usefulness for policy is strongly connected to the accuracy of their outputs. Are the predictions that they make reliable and robust? The process of tuning is intended to enable the models to reflect reality.  Ecosystems have enormous complexity and the grim reality for modelers is that capturing this is an impossible task (Oreskes e t al.  1994). To predict with absolute certainty what the future holds is beyond the capabilities of computer simulations. Conversely, if the models cannot simulate what has already known to have occurred, they are not reliable enough to be used as a predictive tool.  Hence, the essence of ?tuning? an Ecopath  with Ecosim model is to run the model through time and compare it?s estimates to observed time-series data.   Tuning is an iterative process through which group by group the model moves towards the a better representation of the actual ecosystem. The first stage is to have a balanced base Ecopath  model that you are confident reflects the time period you have modeled. Simple diagnostic checks on the model, such as setting fishing mortality to zero, or increasing it ten-fold, give an early indication of the validity of the results. Certain groups with a high Ecotrophic Efficiency (EE), whose abundance is controlled primarily by fishing, will rapidly increase if fishing pressure is suddenly reduced. The modeler, being aware of the system, will be able to ascertain whether this increase is reasonable and modify the basic input parameters if necessary.    The aim of my English Channel ecosystem model was to predict into the future using a range of                                                            Stanford, R. (2004) Tuning Ecosystem Models to Past Data.  Pages 39?40 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. policies. Sufficient data were available to build an accurate contemporary model of the English Channel (Stanford 2002, 2004). In order to have confidence in these predictions it was necessary to build a past model, which would act as an anchor point from which to extrapolate to the present day.  A number of the commercially exploited stocks had been assessed since 1973 and this was the year designated for the earlier model. This model was constructed and run from 1973 to 1995 using stock assessment data for fishing mortality. Where this was not available, estimates were provided from experts or similar stocks so that for each exploited functional group there were fishing mortality data.  Where the biomass estimates of the model significantly differed from stock assessment data the English Channel model required modification through one or more of three ways:  1. T h e basic input parameters entered into Ecopath could be changed . Fishing mortality may cause the EE to be close to 1 and an increase in fishing will cause the group to decline. If the 1973-1995 time-series data indicated that the group was more resilient than predicted by the model, increasing the starting biomass or the estimate for production/biomass will dull the impact of fishing.   2. V u ln e rability settings can be altered .  Vulnerabilities are a measure of whether the system is predator or prey controlled (top down or bottom up). Increasing the vulnerability towards 1 means that predators control the system and that prey are constantly available for capture. A group that consumes prey with a high vulnerability is likely to increase rapidly if its predation or fishing mortality is reduced. Conversely the increase will be moderated by lower vulnerabilities. There are three stages to modeling vulnerabilities. The first is to accept the default value of 0.3 for all groups. Secondly, vulnerabilities can be set according to the trophic level of the prey and finally the best method is to assign vulnerabilities on a group-by-group basis that enable the model to closely replicate time-series data (see Ainsworth 2004, this volume).   3. Forcing functions can be used.  Ecopath  with Ecosim has the capability of including effects not generated by trophic interactions or fishing. These were particularly significant in the English Channel model because the English Channel is located at the boundary of two bio-geographical regions (Southward and Boalch 1988). Variations in temperature will allow different species to be successful (see Pitcher and Forrest 2004, this Back to the Future Methodology, Page 40  volume). Hence, a warmer climate will have positive impacts on sole ( Solea solea) stocks (Henderson 1994) and negative effects on cod ( G a d us morhua ) (O'Brien e t al.  2001); (Planque and Fox 1998). Although sole fishing mortality increased between 1973 and 1995 their biomass simultaneously increased as warmer temperatures had a positive effect on recruitment. Forcing functions do not fix the biomass so a rapid increase in fishing or predation pressure will still reduce sole biomass.     One of the major problems with tuning the model is knowing what to change. Regarding sole, there is significant evidence that there is a correlation between temperature and recruitment, although the exact mechanism of this depends on the region (Rijnsdorp e t al.  1992, Philippart e t al.  1996, Henderson 1994). Hence it was justifiable to include this in the model. For other groups that had not been studied so extensively it was difficult to know which data to trust. Using Virtual Population Analysis will mean that past data becomes more accurate with each new assessment. Conversely, past estimates for a group may be based only on an expert?s guesstimate or current techniques such as acoustic surveys may mean that the contemporary estimates are better. Consequently, although time-series data may suggest a change is necessary, high confidence in your contemporary model may mean that it is not changed.   It is at this stage of tuning that the pedigree screen in Ecopath  is valuable. This gives an indication of what data can be trusted and provides a basis for tuning. For example, in the English Channel it is known that the abundance of sharks (blue sharks Prionace glauca , porbeagle L amna nasus  and tope Ga l eorhinus galeus ) has decreased. This is attributable to both a reduction in prey species and increasing fishing mortality. The pedigree screen in Ecopath  indicated that there was little confidence for most of the data for this group. Hence I could legitimately modify biomass, P/B and the vulnerability until the model predicted the decline in abundance that the literature seemed to indicate.   Tuning cannot overcome all of the inadequacies in a model. It can identify where functional groups may need to be sub-divided, particularly if temperature influences recruitment. Comparing the model?s output to stock assessment data is a valuable exercise that can bring confidence to both the modeler and the policy maker that the results are realistic. In response to the original question in the Prince Rupert workshop concerning their value to decision-making (see Pitcher e t al.  2002), we would affirm that yes these models are valuable when tuned to data and that enhancing the time-series data can only increase their predictive power .   REFERENCES   Ainsworth, C. (2004) Estimating the Effects of Prey-predator Vulnerability Settings on Ecosim's Dynamic Function. Pages 45?47 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.  Henderson, P.A. and Seaby, R.M.H. (1994) On the factors influencing juvenile flatfish abundance in the Lower Severn Estuary, England. Netherlands Journal of Sea Research. 34. 321-330. O'Brien, C.M., Fox, C.J., Planque, B. and Casey, J. (2000) Fisheries: Climate variability and North Sea cod. Nature 404: 142-144. Oreskes, N., Shrader-Frechette, K. and Belitz, K. (1994) Verification, validation and confirmation of numerical models in the earth sciences. Science 263: 641-646. Philippart, C.J.M., Lindeboom, H.J., van der Meer, J., van der Veer, H.W. and Witte, J.I.J. (1996) Long-term fluctuations in fish recruit abundance in the western Wadden Sea in relation to variation in the marine environment. ICES Journal of Marine Science 53: 1120-1129. Pitcher, T.J. and Forrest, R. (2004) Challenging ecosystem simulation models with climate change: the ?Perfect Storm?. Pages 29?38 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J., Power, M. and Wood, L. (eds) (2002) Restoring the Past to Salvage the Future: Report on a Community Participation Workshop in Prince Rupert, BC. Fisheries Centre Research Reports 10(7) : 55 pp. Planque, B. and Fox, C.J. (1998) Interannual variability in temperature and the recruitment of Irish Sea cod. Marine Ecology Progress Series 172: 101-105. Rijnsdorp, A.D., Van Beek, F.A., Flatman, S., Millner, R.M., Riley, J.D., Giret, M. and De Clerck, R. (1992) Recruitment of sole stocks, Solea solea,  in the Northeast Atlantic. Netherlands Journal of Sea Research 29: 173-192. Southward, A.J. and Boalch, G.T.  (1988) Aspects of long term changes in the ecosystem of the western English Channel in relation to fish populations. Pages 415-447 in Wyatt, T. and Larraneta, M.G. (eds) Long term changes in fish populations. Vigo, Spain. Stanford, R. (2002) The English Channel: A mixed fishery, but which mix is best? M.Sc. Thesis. Resource Management and Environmental Studies. University of British Columbia. Stanford, R. (2004) Ecosystem Modelling, Management and Climate in the English Channel. Fisheries Centre Research Reports (in prep).    For discussion after the oral presentation of this paper, see page 150.   Page 41, Fisheries Centre Research Reports 12(1), 2004  DEALING W ITH MIGRATORY SPECIES  IN ECOSYSTEM MODELS   Steve Martell Fisheries Centre, UBC   ABSTRACT   This paper sets out the logic for dealing with migratory species in the Ecopath -with-Ecosim whole ecosystem simulation modelling framework. Examples are provided from salmon and hake populations.     As a technical convenience, Ecopath  models are bounded by arbitrary borders that allow the user to ?define? the system. This ?box? should be large enough so that interactions within  the system add up to a larger flow than the interactions between  the system and the ecosystems outside the box. In almost all cases, it is not possible to define such an area that includes the entire life history of all groups in the model. Furthermore, some groups only use a portion of the box, and never interact with other groups in the model (e.g., A in Figure 1), whereas another group?s distribution may overlap with the defined ecosystem model (e.g., B in Figure 1). Neither example poses a significant problem when building an Ecopath  model. In case A, simple accounting of trophic interactions determines which prey is consumed and which predators consume the group that has a limited distribution. In case B, the fraction of the stock that resides within the model area is used as the biomass input. But what potential problems arise as we move from static pictures of the ecosystem to dynamic changes over time?  Biomass dynamics can have profound effects on the distributions of species, Ecosim is a biomass dynamics model that uses the Ecopath  inputs to calculate initial states. If the ellipse in example A, figure 1, represents the                                                           Martell, S. (2004) Dealing with Migratory Species in Ecosystem Models. Pages 41?44 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. entire distribution of the species in question, then there are no real or potential problems in calculating biomass within the system over time. In example B, it is possible to assume that the fraction of the total stock remains constant over time, and there is little or no exchange across the boundary. At first this assumption may sound crazy, but consider species such as abalone that have limited mobility and small dispersal distances.  A more realistic, and worrisome, case is represented in example C, where the distribution of a species changes over time. Here the area over which the stock is distributed is a function of stock size. When stock is reduced in abundance, though fishing activities, or perhaps increased predation, the range collapses to a smaller area of more favorable habitat. This phenomenon has been observed in many fish A BC DFigure 1. Four examples of ecosystem models, where model boundaries are represented by rectangles, and ovals represent distribution of a group in the model. A) Here the group is only partially distributed in the entire ecosystem, B) the distribution overlaps with ecosystem boundaries, C) distribution overlaps, but may collapse outside boundaries as stock is reduced, and D), arrows represent a complicated life history trajectory, where the gray circles might represent an area of importance such as spawning grounds, or where the fishery takes place . Back to the Future Methodology, Page 42  stocks around the world (e.g., Atlantic cod off the east coast of Canada, Rose 1999). Such range collapses might involve the species leaving the defined ecosystem.  Example D in Figure 1 represents one of the more difficult issues to be represented in Ecosim. In this example, only a portion of the life history trajectory is within the defined ecosystem, and important events such as spawning or targeted fisheries occur both in and outside the define ecosystem. Pacific salmon are probably the best example of a species with a highly migratory life history, where fisheries occur in the oceanic, coastal and freshwater environments. Furthermore, there is another ecosystem that salmon play a functional role in, such as food for bears and eagles, and providing nitrogen through decaying bodies (Watkinson, 2001). Juvenile salmon spend one or more years rearing in freshwater and are subject to variable mortality rates due to competition with other stream inhabitants or anthropogenic impacts such as logging or urban development. Modellers of salmon using Ecosim need therefore to deal with this issue (Stanford 2002). Users of Ecosim should be aware that the stock recruitment relationship represented by split pool dynamics, assumes that everything that happens outside the box remains constant over time. There are however, some built in tools that can be used to represent variation in stock-recruitment production, or the effect of hatcheries. More on this later.  There are two phases to the ?Back to the Future? (BTF) approach. First is to reconstruct several ecosystem models, usually representing the present day, some time period in the past that might represent an unfished state, and one or more ?intermediate? states between pristine and present. The second phase is to simulate how one should optimally utilize the resources of an unfished ecosystem and compare results of such simulations to present day states. The reconstruction phase might include fitting Ecosim models to time series data to help parameterize the model. The second phase simply makes forward projections to explore alternative management policies. Each phase challenges the modeler with different problems for dealing with migratory species or populations that are only partially represented in the ecosystem model. In the reconstruction phase, time series information is required about changes in abundance within the model area. For example, if the distribution is changing over time, then what fraction of the total stock, or total catch, at each year were within the defined system? Often the data lack the spatial resolution that would allow total catch, or biomass to be partitioned among spatial areas. This should be taken into consideration when defining the boundaries of the system.   Species with complicated life history trajectories, where only part of the life history is represented in the model, are even more problematic. For example, dramatic changes in abundance may be a result of mortality that occurred outside the defined ecosystem, yet we search for mortality Figure 2.  An example of estimating the fraction of a large migratory hake stock that enters the Canadian zone each year. The fraction of hake that enters Canadian waters is a function of sea level height, and we can use sea-level height to predict the fraction of the total stock (solid line) to generate a time series of biomass (dashed line) that enters the defined ecosystem. Page 43, Fisheries Centre Research Reports 12(1), 2004  agents within the ecosystem to explain the observed declines. These problems carry forward into the simulation phase of BTF in addition to deciding how to represent life-history trajectories that occur outside the defined ecosystem. We cannot assume the freshwater phase of salmonid production remains constant over time! Ecosim, at its present stage of development, is not capable of explicitly modeling dynamic changes that might occur outside the system. Despite this limitation, there are alternative solutions for dealing with migratory species, or populations that share boundaries between neighboring ecosystem.  One of the most obvious options is to simply do nothing. That is, just assume that what happens outside the ecosystem remains constant over time, and assume that stocks that overlap the defined system are disconnected. Such assumptions may be valid for reasons such as limited dispersal, or because the biomass pool is simply too small to be of importance to modeling questions. An alternative option is to increase the scale of the model such that the entire distribution, or life history trajectory is included in the model (e.g. turn cases B and C in Figure 1 into case A). Exercising this option may be tricky for groups that have long distance migrations, as expanding to such large scales may introduce more problems with data. Having to add additional groups that live outside the previously defined ecosystem, may also require increased participation and substantial increase in the scale of the project.  There are a couple of alternative options for dealing with migratory species in reconstructed dynamic models using Ecosim. One such option is to impose a time series of biomasses on the ecosystem, where this time series is estimated independently of Ecosim. For example, biomass for a particular group could be estimated using single species models (incidentally, this should be done anyway to generate a fishing rate time series to drive fishing mortality in Ecosim), then read into Ecosim1. Also, the time series should be corrected for the fraction of the total stock that is within the defined ecosystem. As an example, Pacific hake populations off the west coast of Vancouver Island are part of a large migratory stock that winters in southern California and some fraction of the total stock migrates into Canadian waters in late spring-early summer. The stock is assessed every three years using information from fishery independent surveys,                                                           1 Note that this time series should be scaled to Ecopath units (e.g. tonnes/km-2), and use the ??1? option for the data type code in the *.csv file. and the proportion present in the Canadian zone is a function of sea level height which is correlated with water temperature (Dorn 1995). Here the southern boundary of the ecosystem model is the Canada-US border, and the objective is to correct the assessment predictions to reflect hake biomass inside Canadian waters. Figure 2 presents a logistic relationship between sea level height and proportion in the Canadian zone. This logistic equation can be applied to the total stock to estimate a time series of hake biomass present inside the defined ecosystem (dashed line in bottom panel), this time series can then be read in as a forced biomass pool.  ?Egg forcing? is an option for forcing split pool dynamics frequently used to represent enhancement programs such as salmon hatcheries. Salmon hatcheries, in some cases, have more than doubled smolt output into the marine environment, and in Ecosim this is just represented by a doubling of egg production in the forcing function. A time series of hatchery releases scaled to wild salmon production is read into Ecosim using the standard *.csv file and specified shape number. The shape number is then applied to egg production for the salmon group. A similar option could also be used to represent other disturbances that might have occurred in the freshwater phase of salmonids (e.g., set egg production to near zero to represent the catastrophic impacts of the Fraser canyon slide that nearly destroyed Fraser River sockeye stocks). This could apply to entrainment of fish larvae in cooling towers for nuclear power plants. The options are endless, but just require some time series data and a known scale of the effect of juvenile production. These time series effects on fish production could also be implemented in forward projections, where alternative hypotheses about the magnitude of the effect can be explored. For example how might the removal of hydroelectric dams affect eulachon populations in the Columbia River?  Another interesting simulation issue related to salmonids, anadromous fishes, and groups that move between two distinctly different ecosystems is how to connect the two systems. As an example, consider the life history of Pacific salmon, where the marine phase involves complicated migrations, consumption, predation, and variability in annual survival rates. These exact same processes also occur in the freshwater phase, where adult salmon are food resources for scavengers/predators such as bears and eagles (Watkinson 2001 and references therein). The remaining adults that survive the predator gauntlet are also responsible for egg production Back to the Future Methodology, Page 44  and recruitment, while juveniles remain in the freshwater environment for up to a year or more and face other challenges. If the ecosystem model only represents the marine phase of the salmonid life history, we might simply proceed with policy exploration using new high-tech gear that reduces by-catch and conclude it is safe to proceed with such developments. The new policy works great in the model and in practice, but grizzly bears are starving and going extinct in many watersheds. Oops! Clearly we should consider how our policy affects neighboring ecosystems, and the question is how do we do this?  For neighboring ecosystem models that share a couple of groups (consider a near-shore versus off-shore ecosystem, where one group forages and spawns near-shore during the summer months) it is a simple matter to combine two Ecopath  models. In nature, species interactions can be direct (i.e. predation), or indirect (i.e. competition). In Ecopath , direct interactions are specified by setting a non-zero value in the diet matrix for predator j on prey i, and indirect interactions are specified when two groups share the same resource. It is possible to carry out the same mass balance exercise for two independent Ecopath  models that are loaded into the same file. In other words, you can have two independent models of 10 groups each, or one model with 20 groups. When you balance these models, parameter estimates are the same if the two diet matrices are independent of each other. To connect the two ecosystems, to represent a group that moves between the two systems, simply recalculate the diet composition for that group, where some proportion P comes from one model, and 1-P comes from the adjacent model. Such an exercise has already been shown to work quite well for the Prince William Sound model (Okey and Pauly 1998), where the ecosystem was sub-divided into nearshore and offshore components. Since predator-prey interactions are specified in the Ecopath  diet matrix, there is no problem moving into Ecosim and representing a group moving between the two systems. With a little programming experience, it is also possible to integrate Ecosim with other models that represent the dynamics of neighboring ecosystems.   For example, suppose we had a terrestrial model for salmon recruitment in the freshwater environment that includes predation and population dynamics of bears and eagles (Watkinson 2001). The input to this model is the number of adult salmon entering the river. Within each annual time step, we can pass predicted adult abundance of salmon from Ecosim to the terrestrial model, where bear and eagle dynamics are updated partly based on how much salmon was available. The terrestrial model then returns the number of juvenile salmon to Ecosim 1-2 years later, where Ecosim graduates the juveniles into adults and the process repeats for N years. Such a framework would provide insights about the affects of harvesting salmon in the marine environment on bears and eagles in the terrestrial environment.   REFERENCES  Dorn, M.W.  (1995) The effects of age composition and oceanic conditions on the annual migration of pacific whiting, Me rl uccius productus . CalCoFi Rep. 36: 97-105. Rose, G.A. and Kulka, D.W. (1999) Hyperaggregation of fish and fisheries: how catch-per-effort increased as the northern cod ( G ad us morhua ) declined. Can. J. Fish. Aquat. Sci. 56 (Suppl. 1): 118-127. Watkinson, S. (2001) Salmon Carcasses, Nature?s Nitrogen Pill. Masters Thesis, University of British Columbia. Vancouver, Canada. Okey, T.A. and Pauly, D. (eds) (1998) A Trophic Mass-Balance Model of Alaska?s Prince William Sound Ecosystem, for the Post-Spill Period. Fisheries Centre Research Reports  6(4).  144pp. Stanford, R. (2002) How to Model Salmon. Page 36 in Pitcher, T.J., Power, M.D. and Wood, L. (eds) (2002)  Restoring the past to salvage the future: report on a community participation workshop in Prince Rupert, BC. Fisheries Centre Research Reports  10(7): 56 pp.     For discussion after the oral presentation of this paper, see page 148.  Page 45, Fisheries Centre Research Reports 12(1), 2004  ESTIMATING THE EFFECTS OF PREDATOR-P REY VULNERABILITY SETTINGS ON ECOSIM? S DYNAMIC FUNCTION   Cameron Ainsworth Fisheries Centre, UBC   ABSTRACT   In the context of foraging arena theory, prey vulnerabilities are proportional to the flux of prey from safe refuges to feeding areas, where they are subject to predation. In the absence of empirical data, Ecosim modelers may use an approximation method to estimate prey vulnerabilities. Four such methods are evaluated in this report in their ability to permit Ecosim to generate predictions of abundance that resemble stock assessment time-series. The first method is to scale prey vulnerabilities proportionately to the trophic level of their predators (?predator control? hypothesis). The second is to scale vulnerabilities proportionately to the trophic level of the prey (?prey control?). The third is to apply a flat vulnerability to all groups. The fourth method customizes group vulnerabilities according to logical rules. Four Ecosim models are used to compare the assumptions. The results fall marginally in the favour of prey control. Three out of four models show improved dynamic functioning under this assumption and biomass trends are improved in 18 out of 32 functional groups (compared to 12 groups for predator control). Prey control was therefore adopted for all Back-to-the-Future applications. Ideally, each predator-prey combination should receive its own independent score. This will be addressed in later revisions of ?Lost Valley? policy search methodology.    Central to the dynamic function of Ecosim are the input prey vulnerabilities to predators. Vulnerability, a concept rooted in foraging arena theory, describes the flux of prey from safe refuges to feeding areas, where they are subject to predation (Walters e t al. , 1997). The vulnerability parameter (v) is assumed by Ecosim to be proportional to the relative time spent feeding and hiding. Figure 1 shows a schematic representation of Ecosim vulnerabilities.  The vulnerability parameter is defined in Ecosim on a logarithmic scale from 0.01 to 1.0. Low prey vulnerability indicates bottom-up control; high                                                            Ainsworth, C. (2004) Estimating the Effects of Prey-predator Vulnerability Settings on Ecosim's Dynamic Function. Pages 45?47 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. vulnerability indicates top-down (Lotka-Volterra) control. Christensen e t al.  (2000) warn that strict bottom-up control in Ecosim tends to produce unrealistically smooth changes in prey and predator biomass that fail to propagate through the food web, while strict top-down control may cause rapid oscillations in biomass and unpredictable simulation behaviour (see also Mackinson 2002).   Further, Cheung e t al.  (2002) suggest that using the blanket assumption (applied to all groups) of top-down control (>0.5 vulnerabilities) will generate a complex response surface with many optima; they found it difficult to find a global maximum when searching for optimal fisheries.  Moreover, Martell e t al.  (2002) found that low blanket vulnerabilities impart on the system a high degree of resiliency to fishing effects. Models based on this assumption, they suggest, will return unreliably optimistic policy recommendations. The default setting in Ecosim describes a mixed condition (on the low end of the vulnerability spectrum as established by convention), where all prey vulnerabilities are set to 0.3. Cheung e t al.  (2002) report that a consensus emerged at the FAO/Fisheries Centre Ecopath  workshop that scaling vulnerabilities in proportion to trophic level (TL) was more realistic than the blanket assumption. Here we test these methods as well as a more customized approach, which involves assigning group vulnerabilities according to logical rules.   Walters (p ers. comm .) suggests that each predator-prey combination should ideally receive its own unique vulnerability since anti-predator defenses (e.g. behavioural, structural) may provide differential protection against various modes of predator attack. In lieu of vulnerability estimates derived from data, modelers may employ a shortcut - scaling vulnerabilities proportionately to either predator or prey trophic U navailab le p reyBi - ViA vailabl e p reyViV(Bi- Vi)vViP redat orBjaij ViBjU navailab le p reyBi - ViA vailabl e p reyViV(Bi- Vi)vViP redat orBjaij ViBj Figure 1.  Ecosim vulnerabilities in the context of foraging arena theory. Vulnerability (v) describes the exchange rate between vulnerable prey biomass pool (Vi) and invulnerable pool (Bi-Vi). (a) describes predator (i) search rate for prey (j). Bj is predator biomass pool. Source: Walters e t al.  1997. Back to the Future Methodology, Page 46  level (TL). While Cheung e t al.  (2002) were the first to try the latter method (repeated by Mackinson e t al.  (2002), Martell e t al.  (2002) and others), the former is tried here for the first time. These two techniques make different assumptions about trophic interactions. The former ?predator control? assumption contends that a prey species will be more vulnerable to high TL predators than low TL predators. The alternate hypothesis, ?prey control?, implies that low TL prey is more vulnerable to predators than high TL prey.   This paper examines whether prey or predator control hypotheses enable Ecosim to predict a biomass trend that more closely resembles a time-series of biomass from stock assessment, and whether either technique improves on the default (all vs =0.3) assumption. Finally, we test the ability of a more customized vulnerability regime to recreate known biomass trends.   To test these issues, I have used four Ecopath   models of past times from various authors along with time-series abundance estimates of their (commercial) functional groups: 1970 Bay of Biscay (Ainsworth e t al. , 2001), 1950 Strait of Georgia (Dalsgaard e t al. , 1998), 1973 English Channel (Stanford 2002) and 1950 northern British Columbia (Ainsworth e t al. , 2002).   METHODS  Using default Ecosim settings for all four models,  I first set the vulnerability of prey groups in proportion to their predators? trophic level (predator control), and then in proportion to their own (prey control). In the Ecosim interface (under the ?flow control? tab), vulnerabilities are entered vertically for predator control and horizontally for prey control. For both trials, the range of vulnerabilities was set from 0.8 for high TL groups to 0.2 for low TL groups.   A simulation was run for each model, under each hypothesis. The biomass trend, obtained from Ecosim?s output CSV file, was compared to stock assessment records with a non-parametric Spearman?s correlation test.  Northern BC-0.8-0.6-0.4-0.200.20.40.60.81TransientSalmonCohoChinookHerringPacificOcean PerchFlatfishHalibutPacific CodSablefishLingcodV er t ica l (pr eda tor ) v u lnHor izon ta l (pr ey ) v u lnRu lesFla t  (0 .3 )Straight of Georgia-0.4-0.200.20.40.60.81HerringspawnHerring juv.(0-2y)Chinook (3-4y) CohoLingcodBay of Biscay-1 .5-1-0.500.51L. DeepwaterM. DeepwaterL. PelagicM. PelagicAnchovySardineXL. DemersalHigh TLXL. DemersalLow TLL. DemersalCrabsEnglish Channel-1-0.8-0.6-0.4-0.200.20.40.60.81SolePlaiceCodHakeAnglerfishMackerelScad  Figure 2 . Correlation of biomass outputs from four Ecosim models with time-series stock assessment under four assumptions of prey vulnerability. Dark bars show predator control; stippled bars show prey control; white bars show logical rules; shaded bars show all vs = 0.3. Crossbars show correlation needed for significance at >= 0.05.  Page 47, Fisheries Centre Research Reports 12(1), 2004  RESULTS  Figure 2 shows the correlation of stock assessment information with the biomass trend predicted by Ecosim. Dark bars show the correlation under predator control, light bars show correlation under prey control. Significance level at a=0.05 is indicated by crossbars for each functional group.  Prey control vulnerabilities allow Ecosim to generate a biomass trend that more closely resembles stock assessment information in 18 out of 32 functional groups studied; predator control vulnerabilities perform better for 12 groups and the two methods perform equally well for 2 groups. Prey control generates a closer overall correlation in all models except northern BC, where only 3 functional groups correlate better under prey control and 6 groups correlate better under predator control.  CONCLUSION  In most cases, prey control vulnerabilities allow Ecosim to predict an index of relative abundance that more closely conforms to stock assessment than the alternate hypothesis, predator control. Unfortunately, the BC model (which is the subject of CUS BTF applications) does not perform better under this assumption. However, we judge predator control to be less supportable, since it requires that prey know which predator is attacking them. Although evolution has probably equipped them with this ability to some extent, in the absence of supportive data it is safer to assume that a prey has adjusted its transfer rate to protect itself equally from all likely predators encountered.  Ideally, we would examine each predator-prey combination individually; this is the next step to improve the dynamic function of the BC model. Avdin and Friday (2001) found that vulnerabilities in the lower order prey groups were most critical to the simulation; this is where fine-tuning should begin.   Consequently, vulnerabilities for all Ecosim models used in the Lost Valley policy search (Ainsworth et al . 2004, this volume, Ainsworth, 2004a and 2004b) were set to prey control (vulnerabilities proportional to prey TL in the range 0.2-0.8 by convention).    ACKNOWLEDGEMENTS   I would like to thank Dr Carl Walters and Dr Villy Christensen for their input and discussion. REFERENCES  Ainsworth, C. (2004a) BTF in Northern BC. 1: Forecasts for Fisheries in Present and Restored Ecosystems. Fisheries Centre Research Reports (Results Volume) in prep.  Ainsworth, C. (2004b) BTF in Northern BC. 2: Policy Goals from Past and Present Ecosystems. Fisheries Centre Research Reports (Results Volume)  in pre p .  Ainsworth, C., Heymans, J.J. and Pitcher, T.J. (2002) Ecosystem Models of Northern British Columbia for the Time Periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4): 41pp. Ainsworth, C., Ferris, B., LeBlond, E. and Gu?nette, S. (2001) The Bay of Biscay, France; 1998 and 1970 models. Pages 217-313 in Gu?nette, S., Christensen, V., Pitcher, T.J. and Pauly, D. (eds) Fisheries Impacts on North Atlantic Ecosystems: Models and Analyses. Fisheries Centre Research Reports 9(4):  344pp.  Ainsworth, C., Heymans, J.J. and Pitcher, T.J. (2004) Policy Search Methods for Back to the Future. Pages 48?63 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Aydin, K. and Friday, N. (2001) The Early Development of Ecosim as a Predictive Multi-Species Fisheries Management Tool. Document SC/53/E3 presented to the IWC Scientific Committee July 2001 (unpublished). 8pp. Christensen, V., Walters, C.J. and Pauly, D. (2000) Ecopath  with Ecosim: a User?s Guide, October 2000 Edition. Fisheries Centre, University of British Columbia, Vancouver, Canada and ICLARM, Penang, Malaysia, 130pp. Cheung, W-L., Watson, R. and Pitcher, T.J. (2002) Policy Simulation of Fisheries in the Hong Kong Marine Ecosystem. Pages 46-53 in Pitcher. T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports. 10(2): 156pp. Dalsgaard, J., Wallace, J.S., Salas, S. and Preikshot, D. (1998) Mass-Balance Model Reconstructions of the Strait of Georgia: the Present, One Hundred, and Five Hundred Years Ago Pages 72-89 in Pauly, D., Pitcher, T.J. and Preikshot, D. (Eds) Back to the Future: Reconstructing the Strait of Georgia Ecosystem. Fisheries Centre Research Reports 6(5): 99pp. Mackinson, S. (2002) Simulating Management Options for the North Sea in the 1880s. Pages 73-82 in Pitcher. T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports 10(2): 156pp. Martell, S., Beattie, A., Walters, C.J., Nayar, T. and Briese, R. (2002) Simulating Fisheries Management Strategies in the Strait of Georgia Ecosystem using Ecopath  and Ecosim. Pages 16-23 in Pitcher, T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports 10(2): 156pp. Stanford, R. (2002) The English Channel: A mixed fishery, but which mix is best? M.Sc. Thesis. Resource Management and Environmental Studies. University of British Columbia.  Walters, C.J., Christensen, V. and Pauly, D. (1997) Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7(1): 39-172.   Back to the Future Methodology, Page 48  POLICY SEARCH METHODS FOR  BACK- TO- THE-F UTURE    Cameron Ainsworth,  Johanna J. Heymans and  Tony J. Pitcher Fisheries Centre, UBC   ABSTRACT   Using the policy search routine in Ecosim we identify the pattern of exploitation that would allow us to gain the most benefit from restored ?Lost Valle y ? ecosystems of northern BC and Newfoundland. The policy search determines the fishing mortalities for each gear type that will maximize its objective function over a 50 year-simulation. Five objective functions are considered: ecological, economic and social, as well as a mixed objective and a conservative ?portfolio log utility function? that resists altering the ecosystem far from its baseline. The ecological function increases the abundance of slow-growing groups, the economic function maximizes rent from the system, and the social function maximizes fishery employment. A mixed objective function combines economic, social and ecological priorities. The portfolio log-utility function combines these priorities as well, but includes a risk aversion algorithm. Using the mandated rebuilding routine, constraints were included in ecological and mixed objective runs for northern BC models to prevent extinctions. Four time periods are evaluated as starting points for the optimization in each ecosystem (1750, 1900, 1950 and 2000 for northern BC and 1450, 1900, 1985 and 1995 for Newfoundland); the most valuable of these represents possible restoration goals. Three fleets are considered in their ability to harvest the restored system. The ? lost valle y ? fleet includes twelve and sixteen fisheries in northern BC and Newfoundland, respectively. These allow a minimal level of bycatch and discards. The ? no recreational ? fleet omits the sport fishery and the ? no trawlers?  fleet omits groundfish trawl and shrimp trawl. We confirm that the search routine has identified the optimal policy by conducting additional trials using random fishing mortalities a starting point rather than Ecopath  baseline values. The restored systems are subjected to 100 years of simulated fishing including a 50-year (dynamic) fishery development phase and a 50-year (steady-state) equilibrium phase. Seven valuation techniques examine the resulting harvest profile and ecosystem condition to measure the success of each restoration period, fleet and harvest objective. Economic valuation considers the conventional and intergenerational net present value of the harvest profile. Ecological valuation measures biodiversity of the restored system based on the Q90 statistic, the                                                            Ainsworth, C., Heymans, J.J. and Pitcher, T.J. (2004) Policy Search Methods for Back to the Future. Pages 48?63 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.   change in ascendancy throughout the system, and the occurrence of local extinctions. Social valuation determines the total number of jobs generated by each restoration/harvest scheme, and the employment diversity across fishing sectors.    INTRODUCTION  Harvesting the Lost Valley  The methods presented here  describe an attempt  to determine the combination of gear types and fishing mortalities will allow us to optimally exploit the restored Lost Valley ecosystems of northern BC and Newfoundland under a variety of policy objectives. The general principles and methodology of the Lost Vall e y  approach is described in Pitcher (2004b, this volume) and in Pitcher et al.  (2004).  Briefly, we have constructed Ecopath  with Ecosim (EwE) ecosystem models of northern British Columbia and Newfoundland to represent the marine environment as it appeared in the distant past, the recent past and the present. These periods are here referred to as Lost Valley ecosystems. In order to evaluate these as possible restoration goals for the future, we determine the optimal pattern of fishing mortality per gear sector (using the policy search routine in Ecosim) that will generate the greatest benefit in a specified harvest objective. Five objectives are tested that, together, span the spectrum of human use versus conservation. In order to explore which combination of gear types (see Ainsworth 2002) should be used to harvest the restored system, we conceptualize an idealized lost valley fleet, which includes only responsible fisheries (Pitcher 2004b this volume, Pitcher 2004a, Pitcher e t al.  2004). We also test two abbreviated versions: one that includes no recreational sector and one that includes no trawle rs .   We apply the optimal fishing patterns to the restored system, simulating 100 years of harvest (including a 50-year dynamic period and a 50-year steady-state equilibrium). The resulting harvest profiles are then valued in economic, ecological and social terms.   By quantifying the benefits that each historic period has to offer as a restoration goal, we can judge what costs are justified in achieving restoration. Future work (Ainsworth, in prep ) will look at strategies for achieving restoration in northern British Columbia, and determine how far into the past we should restore to maintain cost-effectiveness. Page 49, Fisheries Centre Research Reports 12(1), 2004   METHODS  The 1750, 1900, 1950 and 2000 Ecopath  base models of northern BC used here are based on Ainsworth e t al.  (2002); the 1450, 1900, 1985 and 1995 models of Newfoundland are based on Pitcher e t al.  (2002). For both the west and east coasts, this exercise will test the ability of our three proposed responsible fleets to harvest each of our restoration goals (based on four historical periods), in order to maximize five harvest objectives (Figure 1). A total of sixty runs were conducted for each ecosystem.  For the northern BC models, we verify that the policy search has found the optimal combination of fishing mortalities under its objective by conducting additional searches using random fishing mortalities (F) as the starting point for the optimization procedure, rather than Ecopath  base values. Ideally, the search algorithm will locate the global maximum on the response surface curve regardless of the starting values of F.   Ecosim parameterization  Most Ecosim parameters were left as default.  Appendix A Tables A1, A2 and A3a and b detail parameters used to initialize Ecosim including specific run information, group information and juvenile/adult linking parameters. The juvenile/ adult linking parameters for cod and American plaice in Newfoundland were different for 1985 and 1995; the values for 1985 were used for the 1900 and 1450 models.   Prey vulnerabilities to predators (Table A4a and b) were set in proportion to prey trophic level, where the lowest trophic level prey receives a vulnerability of 0.2 to its predators and the highest trophic level prey receives a vulnerability of 0.8. The prey vulnerabilities were varied for the different Newfoundland time period models (Table A4b). Scaling vulnerabilities proportional to prey trophic level rather than predator trophic level was chosen for reasons discussed in Ainsworth (2003).   Initializing the Policy Search  Fleets  The Lost Valley fleet chosen for the west coast includes groundfish trawl, shrimp trawl, shrimp trap, herring seine, halibut longline, salmon freezer troll, salmon wheel, live rockfish, crab trap, clam dredge, aboriginal and recreational fisheries. The Lost Valley  fleet chosen for the east coast includes bottom and shrimp trawls, recreational and First nations fisheries, cod traps, capelin seine, longline, midwater trawl for redfish, traps for lumpfish, snow crab, inshore crabs and lobster, salmon wheels, pole and line, clam dredges and urchin diving.  Retained bycatch on the west coast occurs in all fleets except salmon wheel, live rockfish, clam dredge and aboriginal fisheries, while cod longline, cod traps and redfish midwater trawls retain bycatch on the east coast. Discards were assumed to be minimal, only groundfish/shrimp trawls and clam dredge produce discards on both coasts.   Percentages listed in Appendix B Tables B1a and b (catch) and B2a and B2b (discards) refer to the proportion of total group biomass caught in the first year of the optimal policy exploration; these values will change throughout the simulation. Group biomass for each time period is listed in Ainsworth e t al.  (2002) for the west Canadian coast and in Pitcher e t al.  (2002) for the east coast.   Generally, the fisheries were set to initially catch 2.5% of the total biomass of their target groups annually, and 0.5% or 0.25% of retained bycatch groups. In northern BC, major discards were set at 1.25% of group biomass, while minor discards were set to 0.25% or 0.025% of group biomass. In Newfoundland discards were all set at 2.5% of group biomass for fish and 0.1% for birds. Catches and discards vary between time periods in proportion to system biomass. As the policy exploration progresses these values are free to change.   The baseline values of fishing mortality should have no impact on the final policy. However, we had to use initial Fs small enough to avoid having to rebalance the model for each of our trial fisheries (thereby affecting the search results), and large enough so that the routine?s outputs (which are multipliers of the base F) remain small 1750/14501900/19001950/19852000/1995Period (BC/NFLD)LV fleetLV no rec.LV no trawlFleetEcologicalHarvest objectiveEconomicSocialMixedPort. Log.   Figure 1 . Policy searches. The optimal fishing policy is determined for each period, under each fleet and harvest objective. Sixty runs were conducted for each ecosystem. Re-trials with random F starting points validate the optima. Back to the Future Methodology, Page 50  for convenience. Since the policy search routine was designed to accommodate much larger initial fishing mortalities (as would be seen when evaluating any real-world fishery for example), the output multipliers deliver an uninformative ?>60? string, when optimal Fs are greater than sixty times the baseline value1. Careful choice of baseline fishing mortalities can circumvent this software limitation.  For both coasts there are three fleets tested in the present analysis. First, the Lost Valley fleet, secondly the Lost Valley fleet minus recreational gear and finally the Lost Vall e y  fleet minus trawlers (shrimp trawl and groundfish trawl on both coasts). The aboriginal fishery was held constant, omitted from the policy search for all fleets and objectives. In ?no recreational fishery? trials, the recreational fishery was removed from the base model and omitted from the policy search. Similarly, shrimp trawl and groundfish trawl were removed from the model and omitted from the policy search for the ?no trawl? trials.  Policy objectives  Five Lost Valley policy objectives were considered: ecological, economic, social, mixed and portfolio log utility optimization. These are discussed in the following sections. Since the search routine does not normally attempt to preserve species biodiversity, we entered into the ecological and mixed objective runs of the northern BC ecosystem a constraint (using mandated rebuilding) that there should be no extinctions. The portfolio log utility optimization, in only a few cases, would recommend harvest policies that included extinction of vulnerable groups. A constraint was added to prevent this (see below). For the economic and social optimization runs, no such constraints were included ? extinctions were allowed under these objectives. The optimization procedure was not constrained for any of the Newfoundland models.   Mandated rebuilding  The mandated rebuilding routine was designed to allow users to identify fishing policies that would facilitate the rebuilding of a depleted stock. In this exercise we do not try to increase stock size, but use the routine (in the BC models) to prevent                                                           1 In preliminary work, the EwE code was modified to return numerical multipliers beyond sixty times. However, this version of the code was abandoned when a more fundamental bug was discovered in the policy search routine that limited the number of fleets that could be examined.  Unfortunately, the next version of EwE, which corrected the more severe bug, did not include the maximum-multiplier fix. extinctions by setting the biomass goal to one times the Ecopath  base level. This novel procedure works well to maintain a steady abundance in protected groups. Although in ecological and mixed objective runs many functional groups tended towards extinction, it was possible in all cases to identify a key group, which when protected, allowed the run to proceed without any extinctions. The smallest mandated rebuilding weight that would stop extinctions was used, so as not to disturb the optimum policy any more than necessary.  Initially, with the BC trials, we tried to prevent extinctions ecological and mixed runs by increasing the biomass/production (B/P) ratio of key groups. As explained below, the ecological objective (present in the mixed objective run as well) increases the biomass of functional groups with high B/P ratios. Groups prone to extinction would then have an inflated importance in the policy search. However, this technique was rejected for mandated rebuilding since there was no single set of B/P values found that would stave off extinctions when commonly applied to all models.  Software difficulties  To prevent the policy search program from becoming unstable, it was sometimes also necessary to use mandated rebuilding to prevent groups from exploding or going extinct. The economic optimization runs were particularly prone to instability, 8 out of 12 economic runs in northern BC required restraint on problem functional groups to allow the program to operate. Two out of 12 social runs required manipulation. We gave mandated rebuilding a low priority in the policy search: enough to allow the search routine to function, but not enough to stop extinctions (since we did not wish to perturb the outcome any more than was necessary). In northern BC, the migratory group, transient salmon , in particular was prone to exploding in abundance under most policy objectives causing a computer crash. It was often necessary to restrict its growth to a factor of about eight times the baseline in order to avoid crashes. The problem in modeling migratory species has to do with the diet matrix. When groups feed primarily out of the study area, their food source is not subject to systemic fluctuations in productivity. In times of low system productivity, biomass of the migratory group is inappropriately bounded only by top-down control. For a complete discussion on the problems of migratory species in Ecosim modeling refer to Martell (2004, this volume). Less often than transient salmon, it was also Page 51, Fisheries Centre Research Reports 12(1), 2004  necessary to manage skates and juvenile/adult turbot to allow the policy search to complete itself.   Mandated rebuilding was not used in the Newfoundland policy exploration; unstable runs are indicated in Appendix B Table B4b. Social runs proved the most problematic for the Newfoundland trials, with all the social runs in 1450 and 1985 becoming unstable. Unstable runs resulted in either huge oscillations of biomass, a collapse in biomass (especially salmon, shortfin squid and large and small crabs), or an explosion of biomass (adult Greenland halibut in the 1450 model).  Policy objectives  Ecosystem  Under the ecosystem policy objective, the search seeks to maximize the occurrence of long-lived species. Pristine and unfished ecosystems have been characterized as having many large slow-growing animals (Odum 1969). Therefore, using a high biomass/production (B/P) ratio across functional groups as a surrogate to describe this condition, the ecosystem policy objective suggests an exploitation profile that will increase the abundance of slow-growing functional groups. Cheung e t al.  (2002) were the first authors to use this technique. The B/P ratios used in the present exercise are listed in Appendix A Table A5a and b. However, the ecological objective does not necessarily preserve species diversity; it will sacrifice high turnover groups (e.g. predators, competitors) in favour of the long-lived animals. Therefore, (in the BC models) we used mandated rebuilding to protect against extinction of any functional group under this policy objective.  Economic  The economic objective seeks to maximize total rent from the system. Under this objective, high value fisheries will be favoured at the expense of low value fisheries, even to the extent of causing extinctions among detrimental groups (e.g. predators, competitors). We do not expect this run to preserve biodiversity. Economic valuation methodology is presented in Ainsworth and Sumaila (2004a).  Social  The social optimization will increase the number of jobs by eliminating fisheries with a low number of jobs per catch value in favour of more labor-intensive gears. Appendix B Table B3a and b give the jobs per catch value used for initialization. At 15 jobs per catch value unit (an estimate), the recreational fishery of northern BC employs three times as many people as the next highest fishery. Relative values were estimated by expert opinion (Pitcher, p e rs. comm .).  Mixed  The mixed objective combines ecological, economic and social elements. The search routine attempts to maximize the total objective function (the weighted sum of all components). Mackinson (2002) tried a similar mixed objective function on a model of the North Sea. He found that the relative improvement in ecosystem criteria consistently failed to match the relative improvement of social and economic criteria and it did not improve markedly as a higher relative weight was given. However, that author used much smaller relative weightings for ecology than the present paper (i.e. the largest relative weighting he applied was 10, 1 and 1 for ecological, social and economy). Zeller and Freire (2002) likewise found that the relative improvement over baseline of ecology was quite invariant to the weighting given to the ecological objective. Buchary e t al.  (2002) also found that a 1,1,1 mixed search for ecology, economics and social benefit results in an optimal policy that is very similar to their social optimization. These authors used a low relative weighting, with the ecological function receiving the same weight in the policy search as economy and social (i.e. 1, 1 and 1 for ecology, economy and social functions).   However, it is evident that entering equal weightings in the EwE  software panel does not result in an equal improvement in criteria over Ecopath  baseline. Since there is no intrinsic comparability between the three objective functions, then the relative weightings used to parameterize the search are meaningless and so  a 1:1:1 ratio between the three objective functions does not imply that the policy search will increase all objectives evenly. Rather, only the relative improvement in each field over baseline is significant. We therefore adjust the weightings iteratively, based on the overall figure given by the completed search, so that each factor influences changes in the overall figure by an equal (or the desired) amount. This technique has been used in the LV work reported in Pitcher et al.  (2004).   From the baseline condition, we find that in general, a much higher relative weighting must be given to ecology in order to achieve an equal improvement among mixed factors (see Pitcher Back to the Future Methodology, Page 52  2004b, this volume). In this paper, the relative weighting of the three fields were determined in such a way as to minimize variance between the overall improvement values of the three functions. It turned out that a relative weighting of 1, 1 and 100 for economic, social and ecological priorities was found to consistently produce the most equal increase as measured from the final line of multipliers in the policy search. This ratio was therefore adopted for all BC runs. Variance of the relative ec0logical, economic and social improvements for BC runs are presented in Table B4a.   The Newfoundland models required a relative weighting of 0.1, 0.1 and 100 for economic, social and ecological weightings to obtain an equal increase in each priority. Thus, the ecological priorities had to be three orders of magnitude higher than the social and economic priorities to get similar outcomes for these three functions (as opposed to the two orders of magnitude required by the BC models). The only Newfoundland model that did not conform to the 0.1:0.1:100 ratio was 1450 (see Appendix B Table B4b).   The very high value required for ecological improvement in both ecosystems suggests that it is more difficult to manage an increase in the B/P surrogate than it is to increase rent, for example (i.e., Ecosim must structure virtually the entire strategy towards ecological gain in order to produce a minimal increase in average B/P of the system). Although the relative weightings required to levy an equal improvement across criteria will be model-specific. The relative insensitivity of the ecological function is also noted by Mackinson (2002), Zeller and Freire (2002), Buchary e t al.  (2002) and Pitcher e t al.  2004.   Portfolio Log Utility  The recently devised portfolio log utility function attempts to account for the inherent uncertainty in changing the system far from its base state. Christensen e t al.  (2000) and Christensen and Walters (2004) provide a more detailed description of this Ecosim subroutine. Policies that promise the greatest benefit tend to carry with them the greatest risk, since the extreme combination of fisheries required to manipulate the ecosystem into a hyper-productive state will change the system far from its present condition. Such a policy may, for instance, involve destroying competitors and predators of the most valuable species, as is done in agriculture.   In portfolio log utility the user enters three parameters. Prediction variance describes the amount of uncertainty associated with changing the ecosystem far from its baseline. A high value will increase the discounting rate (reducing the net present value of future benefits), and make large returns unappealing when they require drastic manipulation of the ecosystem. Existence value defines the worth one assigns to the continued existence of functional groups: assigning a high value to this parameter will maintain a diverse biological ?portfolio? in economics terms. Finally, users enter a coefficient that modifies the net present value from the system (the sum of profits from all functional groups, discounted over time). A high value of this can make risky policies worthwhile.  For some runs with the BC models there was a precarious balance between receiving a policy recommendation that included extinctions, and receiving a flat line (zero change from base state). To fine-tune these runs we added a very small prediction variance, from 0.02 to 0.003. This fix helps prevent extinctions by devaluing daring portfolio choices. Only the lowest existence value that would still prevent extinctions was used. With the Newfoundland models this was not a problem. We only used existence values without having to use prediction variance. The existence values used for the Newfoundland models ranged from 0.01 to 0.1 (Appendix B Table B4b).   The portfolio log utility trials are very stable. Runs change slowly from the base state, and are not subject to the same wild fluctuations in biomass often seen when using the other policy objectives. This is the most conservative method. We do not expect high returns from the system compared to ecological, economic, social or mixed runs.  Verification of optimal policy  For the northern BC models, we next repeated each optimization 25 additional times, using random fishing mortalities as the starting point for the optimization, rather than Ecopath  base values. Ideally, each replication should result in the same optimum fishing pattern (i.e., locate the global maximum on the response surface). However, prior investigations revealed that random F starting points do not necessarily allow the search routine to converge on the same maximum. Rather, the resultant ?optimal? fishing mortalities seem to cluster around common peaks, indicating that the search can stall on local maxima of the response surface.   In the CUS BTF results report for BC models Page 53, Fisheries Centre Research Reports 12(1), 2004  (Ainsworth et al.  2004), a two-way analysis of variance tests whether the 25 treatments have generated a statistically similar pattern of fishing mortalities. Results from the second factor, gear type, are discarded since we expect fishing mortality to vary between gear types. If the random F runs are shown to be dissimilar, this may indicate that the policy search routine has identified two or more local maxima for a given scenario, or alternatively, that the search routine has identified a single, broad peak (i.e. a plateau) where major variation in the harvest pattern yields an equivalent improvement over baseline.   Multidimensional scaling (MDS) offers a method to differentiate between these possibilities. Using SPSS v.10.0 statistical software, MDS is performed on a subset of runs (chosen to demonstrate the potential of this analysis in describing the shape of the response surface). MDS reduces all factors affecting scenario performance (i.e. fishing mortalities per gear type) to two dimensions, allowing us to sketch the shape of the optimal peak and/or detect the presence of local maxima. Such an approach may be used to judge the robustness of a harvest recommendation for management; however, more random F runs would be required to fully explore the shape of the response surface.  If a recommended harvest policy resides on a narrow peak, than any variation from the specified optimal fishing pattern may result in sub-optimal harvests. If however, the identified maximum resides on a broad peak, than deviations in fleet-effort structure may still result in a near-optimal manipulation of the ecosystem. The latter situation may represent a more robust goal for management than the former.  Valuation indices  Having determined the optimal combination of fishing mortalities per gear type that will maximize our five objective functions for each restored period, we then simulated a 100-year harvest regime (50 years dynamic and 50 years equilibrium) under each of our 3 idealized fleet structures. The resulting harvest profile was evaluated using two economic measures: conventional and intergenerational net present value (Sumaila and Walters 2003, 2004; Sumaila 2001, Sumaila and Bawumia, 2000). Economic valuation methodology is discussed in Ainsworth and Sumaila (2004, this volume).   The ecological success of the restoration/harvest scheme was determined using three valuation measures: the Q-90 statistic, system resilience and presence of local extinctions. Based on Kempton?s Q index (Kempton and Taylor 1976), the Q-90 statistic is a measure of biodiversity that concerns species evenness. It looks at the slope of the cumulative species abundance curve between the 10 and 90 percentiles (see Ainsworth and Pitcher 2004, this volume, for methods). A second index involves measuring the resiliency of the system to fishing using ecosystem redundancy from network analysis (see Heymans 2004, this volume, for methodology and theory). The third measures the risk of local extinctions in composite functional groups (see Cheung and Pitcher 2004, this volume).   Social valuation measures include relative number of jobs created and employment diversity. Relative number of jobs created by an optimal plan is calculated as the product of total catch value (i.e. all simulation years summed) and the gear-specific jobs per unit catch value (Tables B3a and b). Employment diversity across fishing sectors is calculated after Atteran (1986). Ainsworth and Sumaila (2003b) describe how this index was applied to BTF methodology.   Using Kendall?s coefficient of concordance (W; Kendall 1962), we finally determine the ability of each restoration period, fleet structure and harvest objective to maximize these economic, ecological and social valuation measures. Specific expectations are discussed below.  All valuation results will be presented in Ainsworth e t al.  (2004) for British Columbia and Heymans et al.  (2004) for Newfoundland.   DISCUSSION  Ecosystem value will depend mainly on what period is restored. The pre-contact systems have in them the greatest biomass of valuable commercial groups; we therefore expect this period to permit the most valuable fisheries ? scoring high in the economic analyses. On both coasts, models of the recent past represent a more depleted state than do models of the distant past; these will not be able to generate as much economic benefit.   Since the conventional model of discounting places most value on the immediate future, we expect also that the pre-contact and 1900 runs will do especially well under this valuation scheme. These simulations start at a high level of biomass and Ecosim can fish down the natural capital, generating immediate revenue and leaving the system in a depleted (but more Back to the Future Methodology, Page 54  productive) state. Intergenerational discounting, however, will not favour the immediate profit as strongly; it will be content to leave more natural capital in the sea and maintain high harvests farther into the future. Therefore, although the pristine states (pre-contact and 1900) should always produce greater revenues than the more depleted systems (1950 and 2000 in BC; 1985 and 1995 in Newfoundland), the difference will be more apparent under conventional discounting than under intergenerational discounting because of the relative shape of harvest profiles. The more recent time periods will require rebuilding in order to generate maximum monetary returns. Their harvest profiles will slope upwards (or slope downwards less sharply than distant past periods); therefore, they will score proportionately better under intergenerational discounting.  Of the three fleets tested ( Lost Valle y , no recreational , no trawl ), we expect the Lost Valley fleet to generate the most valuable harvest of the restored system for two reasons. First, the additional gear types allow the search routine to probe for the best policy with improved dexterity. Since the policy search is at liberty to minimize any of its fleets, allowing more gear types can only enhance the search routine?s ability to manipulate the ecosystem into its most commercially valuable condition. Secondly, the CUS BTF models (at this stage) do not consider the problems of trawl damage, ghost fishing, or any other deleterious gear effect. In the simulation, there is no ecological or economic benefit associated with preserving habitat, and nothing is to be gained by restricting damaging fisheries (except perhaps a coincidental reduction in discards). Similarly, ecologically responsible fleets that omit damaging gear types will not be credited with their full ecological benefit. Future efforts to model the system spatially will allow us to include these considerations.   We expect the mixed objective function to yield exploitation profiles similar to the ecological runs. Our preliminary efforts have confirmed the findings of other researchers that the ecological objective is the most difficult to maximize ? the policy search must virtually disregard the other objective functions in order to increase the ecological criteria. For example, a typical ecological run will rarely exceed a 10% improvement in the B/P surrogate over 50 years, under even the most vigorous attempts to do so. Rent and jobs, on the other hand, regularly exceed a seven times improvement on the economic and social objective functions. Where improving the ecology involves a slow restructuring of the ecosystem (and a sacrifice in catch), the economic and social functions need only to redistribute fishing effort to increase rent or jobs. This is especially true since the economic and social functions were not constrained by the requirement to avoid extinctions. Further, the search routine will be hard pressed to improve the B/P ratio of the already under-exploited pre-contact and 1900 models. As an objective, it is easier to disassemble the ecosystem, particularly one that is under-exploited, than it is to build it.   ACKNOWLEDGEMENTS   Thanks to Drs Daniel Pauly, Carl Walters and Villy Christensen at the UBC Fisheries Centre, who provided advice and content included in this report.   REFERENCES  Ainsworth, C. (2002) Can we split the ECOSIM fisheries by license type? Page 36 in Pitcher, T.J., Power, M.D. and Wood, L. (eds) (2002) Restoring the past to salvage the future: report on a community participation workshop in Prince Rupert, BC. Fisheries Centre Research Reports  10(7): 56 pp.   Ainsworth, C. (2004) Estimating the Effects of Prey-predator Vulnerability Settings on Ecosim's Dynamic Function. Pages 45?47 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. and Pitcher, T.J. (2004) Modifying Kempton?s Biodiversity Index for Use with Dynamic Ecosystem Simulation Models. Pages 91?93 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. and Sumaila, U.R. (2004a) Economic Valuation Techniques for Back-To-The-Future Optimal Policy Searches. Pages 104-107 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Ainsworth, C. and Sumaila,U.R. (2004b) An Employment Diversity Index Used to Evaluate Ecosystem Restoration Strategies. Page 108 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.. Ainsworth, C.,  Heymans, J.J. and Pitcher, T.J. (2002) Ecosystem Models of Northern British Columbia for the Time Periods 2000, 1950, 1900 and 1750. Fisheries Centre Research Reports 10(4): 41pp. Ainsworth, C., Heymans, J.J., Cheung, W-L. and Pitcher, T.J. (2004) Evaluating Ecosystem Restoration Goals and Sustainable Harvest Strategies in Northern BC. Fisheries Centre Research Reports ( results volume ).  Attaran, M. (1986) Industrial Diversity and Economic Performance in U.S. Areas. Annals of Regional Science.  20(2): 44-55. Buchary, E., Alder, J., Nurhakim, S. and Wagey, T. (2002) The Use of Ecosystem-based Modelling to Investigate Multi-species Management Stategies for Capture Fisheries in the Bali Strait, Indonesia. Pages 25-32 in Pitcher. T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Page 55, Fisheries Centre Research Reports 12(1), 2004  Capture Fisheries. Fisheries Centre Research Reports. 10(2): 156 pp. Cheung, W-L., Watson, R. and Pitcher, T.J. (2002) Policy Simulation of Fisheries in the Hong Kong Marine Ecosystem. Pages 46-53 in Pitcher. T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports. 10(2):  156pp. Cheung, W-L. and Pitcher, T.J. (2004) An Index Expressing Risk of Local Extinction for Use with Dynamic Ecosystem Simulation Models. Pages 94?102 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Christensen, V. and Walters, C.J. (2004) Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling (in press ). Christensen, V., Walters, C.J. and Pauly, D. (2000) Ecopath  with Ecosim Version 4, Help system. Univ. of British Columbia, Fisheries Centre, Vancouver, Canada and ICLARM, Penang, Malaysia. Heymans, J.J., Ainsworth, C., Cheung, W-L. and Pitcher, T.J. (2004) Back to the Future Policy Search Results on the East Coast of Canada. Fisheries Centre Research Reports ( r esults volume ). Heymans, J.J. (2004) Evaluating the Ecological Effects on Exploited Ecosystems using Information Theory, Pages 87?90 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Kendall, M.G. (1962) Rank Correlation Methods. 3rd ed., Griffin, London. 199pp. Kempton, R.A., and Taylor, L.R. (1976) Models and Statistics for Species Diversity. Nature 262: 816-820. Layard, P.R.G., Glaister, S. and Layard, R. (1994) Cost Benefit Analysis. Cambridge University Press, UK, 2nd Ed. 497 pp. Odum, E.P. (1966) The strategy of ecosystem development. Science 104: 262-270. Pitcher, T.J. (2004a) ?Back To The Future?: A Fresh Policy Initiative For Fisheries And A Restoration Ecology For Ocean Ecosystems. Phil. Trans. Roy. Soc. (i n press) . Pitcher, T.J. (2004b) Why we have to open the lost valley: criteria and simulations for sustainable fisheries, Pages 78?86 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Pitcher, T.J., Heymans, S.J.J., Ainsworth, C., Buchary, E.A., Sumaila, U.R. and Christensen, V. (2004) Opening The Lost Valley: Implementing A ?Back To Future? Restoration Policy For Marine Ecosystems and Their Fisheries. In Knudsen, E.E., MacDonald, D.D. and Muirhead, J.K. (eds) Fish in the Future? Perspectives on Fisheries Sustainability. American Fisheries Society, Bethesda, MD, USA. (in press ). Pitcher, T.J., Heymans, J.J. and Vasconcellos, M. (2002) Ecosystem models of Newfoundland for the time periods 1995, 1985, 1900 and 1450. Fisheries Centre Research Report 10(5): 74 pp. Mackinson, S. (2002) Simulating Management Options for the North Sea in the 1880s. Pages 73-82 in Pitcher. T.J. and  Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports. 10(2): 156pp. Martell, S. (2004) Dealing with Migratory Species in Ecosystem Models. Pages 41?44 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Sumaila, U.R. (2001) Generational cost benefit analysis for evaluating marine ecosystem restoration. Pages 3-9 in Pitcher, T.J., Sumaila, U.R. and Pauly, D. (eds) Fisheries Impacts on North Atlantic ecosystems: Evaluations and Policy Exploration. Fisheries Centre Research Reports 9(5). 94 pp.  Sumaila, U.R and Bawumia, M.  (2000) Ecosystem justice and the  marketplace. Pages 140?153 in Coward, H., Ommer, R. and Pitcher, T.J. (eds) (2000) Just Fish: the Ethics of Canadian Fisheries. Institute of Social and Economic Research Press, St John's, Newfoundland, Canada. 304pp. Sumaila, U.R. and Walters, C.J. (2003) Intergenerational Discounting. Pages 19-25 in Sumaila, U.R. (ed.) (2003) Three Essays on the Economics of Fishing. Fisheries Centre Research Reports 11(3): 33pp. Sumaila, U.R. and Walters, C.J. (2004) Intergenerational discounting: a new intuitive approach. Ecological Economics (in press ). Walters, C.J., Christensen, V. and Pauly, D. (2002) Searching for Optimum Fishing Strategies for Fishery Development, Recovery and Sustainability. Pages 11-15 in Pitcher, T. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports. 10(2): 156pp. Walters, C.J., Christensen, V. and Pauly, D. (1997) Structuring dynamic models of exploited ecosystems from trophic mass-balance assessments. Reviews in Fish Biology and Fisheries 7: 139-172 Zeller, D. and Freire, K. (2002) A Preliminary North-East Atlantic Marine Ecosystem Model: the Faroe Islands and ICES Area Vb. Pages 39-45 in Pitcher, T.J. and Cochrane, K. (eds) The Use of Ecosystem Models to Investigate Multispecies Management Strategies for Capture Fisheries. Fisheries Centre Research Reports 10(2): 156 pp.   For discussion after the oral presentation of this paper, see page 153.     Back to the Future Methodology, Page 56  APPENDIX A ECOSIM PARAMETERS   Table A1.   Run Information for both ecosystems.  Duration of simulation (years) 50 Integration steps (per year) 100 Relaxation parameter [0,1] 0.5 Discount rate (% per year) 5 Equilibrium step size 0.003 Equilibrium max. fishing rate (relative) 3 Number of time steps for averaging results 5 Table A2.   Group Information for both ecosystems   Maximum relative feeding time 2 Feeding time adjustment rate 0.5 Fraction of 'other' mortality sensitive to changes in feeding time 1 Predator effect on feeding time 0 Density dependant catchability 1 QBmax/Qbo 1000 Table A3a .  Stage (Juvenile/adult linking parameters) for northern BC.  Herring Piscivorous rockfish Turbot Flatfish Halibut Sablefish Lingcod Pollock Pacific Ocean Perch Pacific Cod Min. time as juv. (rel. to orig. setting) 1 1 1 1 1 1 1 1 1 1 Max. time as juv. (rel. to orig. setting) 1.0001 1.0001 1.0001 1.0001 1.0001 1.0001 1.0001 1.0001 1.0001 1.0001 Recruitment power parameter 1 1 1 1 1 1 1 1 1 1 Weight (g) at transition to adult group 1 1 1 1 1 1 1 1 1 1 Age (year) at transition to adult group (tk) 2.1 16 4.5 4.5 10 4.5 4 2.3 16 2.3 Wavg / Wk (Av. adult weight / weight at transition) 2 2.7 2 2 1.357 1.88 3.684 3.597 2.7 1.725 K of the VBGF (/year) 0.47 0.05 0.243 0.243 0.08 0.3 0.263 0.373 0.88 0.27 Base fraction of food intake used for reproduction 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 Fraction of increase in food intake used for growth 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 Table A3b.   Stage (Juvenile/adult linking parameters) for Newfoundland Cod American plaice Greenland halibut  1985 1995 1985 1995  Min. time as juv. (rel. to orig. setting) 1 1 1 1 1 Max. time as juv. (rel. to orig. setting) 1.0001 1.0001 1.0001 1.0001 1.0001 Recruitment power parameter 1 1 1 1 1 Weight (g) at transition to adult group 1 1 1 1 1 Age (year) at transition to adult group (tk) 7 7 5 5 9 Wavg / Wk (Av. adult weight / weight at transition) 1.247 1.051 3.427 2.299 2.000 K of the VBGF (/year) 0.07 0.07 0.099 0.099 0.025 Base fraction of food intake used for reproduction 0.3 0.3 0.3 0.3 0.3 Fraction of increase in food intake used for growth 0.8 0.8 0.8 0.8 0.8 Weight at transition 1.927 2.353 0.104 0.095  Adult weight 2.403 2.474 0.358 0.218  Page 57, Fisheries Centre Research Reports 12(1), 2004   Table A4a.   Flow control in northern BC.   Functional Group 1750 1900 1950 2000 Seals, sea lions 0.80 0.80 0.80 0.80Transient salmon 0.50 0.50 0.50 0.50Coho salmon 0.76 0.76 0.76 0.76Chinook salmon 0.71 0.71 0.71 0.71Small squid 0.57 0.57 0.57 0.57Squid 0.71 0.71 0.71 0.71Ratfish 0.57 0.57 0.57 0.57Dogfish 0.64 0.64 0.64 0.64Juvenile pollock 0.53 0.53 0.53 0.53Pollock 0.60 0.60 0.60 0.60Forage fish 0.46 0.46 0.46 0.46Eulachon 0.48 0.48 0.48 0.48Juvenile herring 0.47 0.47 0.47 0.47Adult herring 0.51 0.51 0.51 0.51Juvenile POP 0.48 0.48 0.48 0.48Adult POP 0.52 0.52 0.52 0.52Inshore rockfish 0.72 0.72 0.72 0.72Juvenile picivorous rockfish 0.53 0.53 0.53 0.53Adult picivorous rockfish 0.59 0.59 0.59 0.59Juvenile planktivorous rockfish 0.49 0.49 0.49 0.49Adult planktivorous rockfish 0.62 0.62 0.62 0.62Juvenile turbot 0.73 0.73 0.73 0.73Adult turbot 0.77 0.77 0.77 0.77Juvenile flatfish 0.51 0.51 0.51 0.51Adult flatfish 0.53 0.53 0.53 0.53Juvenile halibut 0.70 0.70 0.70 0.70Juvenile Pacific cod 0.44 0.44 0.44 0.44Adult Pacific cod 0.79 0.79 0.79 0.79Juvenile sablefish 0.54 0.54 0.54 0.54Adult sablefish 0.66 0.66 0.66 0.66Juvenile lingcod 0.70 0.70 0.70 0.70Adult lingcod 0.77 0.77 - - Shallowwater benthic fish 0.62 0.62 0.62 0.62Skates 0.65 0.65 0.65 0.65Large crabs 0.45 0.45 0.45 0.45Small crabs 0.47 0.47 0.47 0.47Commercial shrimp 0.35 0.35 0.35 0.35Epifaunal invertebrates 0.20 0.20 0.20 0.20Infaunal carnivorous invertebrates 0.23 0.23 0.23 0.23Infaunal invertebrate detritivores 0.20 0.20 0.20 0.20Carnivorous jellyfish 0.23 0.23 0.23 0.23Euphausiids 0.25 0.25 0.25 0.25Copepods 0.20 0.20 0.20 0.20Macrophytes 0.20 0.25 0.25 0.25Phytoplankton 0.20 0.23 0.23 0.23  Table A4b.   Flow control in Newfoundland.  Functional Group 1500 1900 1985 1995 Walrus 0.53 0.53 0.48 0.53Cetaceans 0.71 0.72 0.66 0.67Grey seals 0.79 0.80 0.73 0.79Harp Seals 0.73 0.73 0.68 0.74Hooded Seals 0.80 0.80 0.80 0.80Ducks 0.45 0.45 0.42 0.45Piscivorous Birds 0.78 0.76 0.69 0.75Planktivorous Birds 0.58 0.58 0.53 0.53Cod (> 40 cm) 0.68 0.68 0.67 0.71Cod (? 40 cm) 0.60 0.60 0.60 0.63American plaice (> 35 cm) 0.56 0.56 0.56 0.55American plaice (? 35 cm) 0.54 0.54 0.56 0.59Greenland Halibut (> 65 cm) 0.79 0.79 0.75 0.77Greenland Halibut (? 65 cm) 0.75 0.75 0.68 0.73Yellowtail Flounders 0.48 0.48 0.44 0.48Witch flounder 0.45 0.45 0.42 0.46Winter flounder 0.47 0.47 0.43 0.45Skates 0.75 0.75 0.68 0.73Dogfish 0.70 0.70 0.63 0.67Redfish 0.62 0.62 0.56 0.58Transient Mackerel  0.66 0.66 0.60 0.64Dem. BP Pisc. (>40 cm) 0.77 0.77 0.71 0.75Dem. BP Pisc. (? 40 cm) 0.67 0.68 0.63 0.61Demersal Feeders (> 30 cm) 0.54 0.54 0.49 0.51Demersal Feeders (? 30 cm) 0.52 0.52 0.48 0.48Small Demersals 0.47 0.47 0.44 0.47Lumpfish 0.59 0.59 0.54 0.55Greenland cod 0.67 0.71 0.64 0.69Salmon 0.76 0.76 0.69 0.74Capelin 0.51 0.51 0.47 0.49Sandlance 0.50 0.50 0.46 0.48Arctic cod 0.54 0.54 0.50 0.51Herring 0.52 0.52 0.48 0.49Transient Pelagics 0.70 0.71 0.65 0.68Small Pelagics 0.55 0.55 0.51 0.50Small Mesopelagics 0.54 0.54 0.50 0.50Shortfin squid 0.69 0.69 0.64 0.69Arctic Squid 0.52 0.52 0.48 0.48Large Crabs (> 95 cm) 0.43 0.43 0.40 0.43Small Crabs (? 95 cm) 0.47 0.47 0.43 0.46Lobster 0.43 0.43 0.40 0.43Shrimp 0.31 0.31 0.30 0.31Echinoderms 0.20 0.20 0.20 0.20Polychaetes 0.20 0.20 0.20 0.20Bivalves 0.20 0.20 0.20 0.20Other Benthic Invertebrates 0.20 0.20 0.20 0.20Large Zooplankton 0.34 0.34 0.32 0.28Small Zooplankton 0.20 0.20 0.20 0.20Phytoplankton 0.30 0.30 0.30 0.30Back to the Future Methodology, Page 58   Table A5a.   Biomass/production ratios for BC *.   2000 1950 1900 1750Sea Otters 7.69 7.69 7.69 7.69Mysticetae 50.00 50.00 50.00 50.00Odontocetae 25.00 50.00 25.00 25.00Seals, sea lions 16.67 16.67 16.67 16.67Seabirds 10.00 10.00 10.00 10.00Transient salmon 0.40 0.40 1.61 1.93Coho salmon 0.36 0.36 0.94 0.86Chinook salmon 0.46 0.46 2.75 2.73Small squid 0.17 0.17 0.17 0.17Squid 0.17 0.17 0.17 0.17Ratfish 10.10 10.10 5.03 5.03Dogfish 10.10 10.10 7.14 9.09Juvenile pollock 0.94 0.94 4.35 4.35Pollock 3.80 3.80 6.49 6.54Forage fish 0.70 0.70 1.70 1.68Eulachon 0.70 0.70 1.67 1.67Juvenile herring 0.46 0.46 0.85 0.85Adult herring 1.46 1.46 1.25 1.26Juvenile POP 1.49 1.49 2.96 2.96Adult POP 6.94 6.94 4.41 4.41Inshore rockfish 5.26 5.26 5.49 5.49Juvenile picivorous rockfish 3.83 3.83 3.83 3.83Adult picivorous rockfish 27.03 27.03 27.03 27.03Juvenile planktivorous rockfish 3.83 3.83 3.83 3.83Adult planktivorous rockfish 14.71 14.71 14.71 14.71Juvenile turbot 3.03 3.03 3.03 3.03Adult turbot 4.55 4.55 4.55 4.55Juvenile flatfish 0.52 0.52 2.62 2.62Adult flatfish 1.05 1.05 3.89 3.89Juvenile halibut 1.67 1.67 8.62 10.10Adult halibut 2.50 2.50 11.90 14.93Juvenile Pacific cod 0.51 0.51 3.88 3.88Adult Pacific cod 0.76 0.76 5.75 5.75Juvenile sablefish 1.67 1.67 3.66 3.66Adult sablefish 3.62 3.62 5.43 5.46Juvenile lingcod 0.83 0.83 2.57 2.57Adult lingcod 1.25 1.25 3.33 3.82Shallowwater benthic fish 0.67 0.67 3.76 3.76Skates 3.23 3.23 6.67 6.67Large crabs 0.67 0.67 0.67 0.67Small crabs 0.29 0.29 0.29 0.29Commercial shrimp 0.09 0.09 0.18 0.18Epifaunal invertebrates 0.69 0.69 0.69 0.69Infaunal carnivorous invertebrates 0.50 0.50 0.50 0.50Infaunal invertebrate detritivores 0.74 0.74 0.77 0.77Carnivorous jellyfish 0.06 0.06 0.06 0.06Euphausiids 0.16 0.17 0.17 0.17Copepods 0.04 0.04 0.04 0.04Corals and sponges 100.00 100.00 100.00 100.00Macrophytes 0.19 0.19 0.19 0.19Phytoplankton 0.01 0.01 0.01 0.01 *Ecological objective ma ximizes B/P surrogate   Table A5b.   Biomass/production ratios for NFLD.   1450 1900 1985 1995Walrus 16.6 16.6 16.6 16.6Cetaceans 20 10 10 10Grey seals 16.6 16.6 16.6 16.6Harp Seals 9.8 9.8 9.8 9.8Hooded Seals 9.2 9.2 9.2 9.2Ducks 4 4 4 4Piscivorous Birds 4 4 4 4Planktivorous Birds 4 4 4 4Cod (> 40 cm) 4.6 10.4 2.4 3.4Cod (? 40 cm) 4.8 4.2 0.6 0.6American plaice (> 35 cm) 12 12 4.4 11.4American plaice (? 35 cm) 8 8 1.6 2.4Greenland Halibut (> 65 cm) 17 29.8 3.4 10.2Greenland Halibut (? 65 cm) 13.2 39.8 1.2 2.6Yellowtail Flounders 3.2 3.2 1.8 3.2Witch flounder 4.2 4.2 1.8 2.8Winter flounder 3.8 3.8 3.8 3.8Skates 4.2 9 2.8 3.2Dogfish 6.2 6.2 5.2 5.2Redfish 8.8 8.8 2 6.8Transient Mackerel  1.8 1.8 3.4 3.4Demersal BP Piscivores (>40 cm) 10.2 10.2 1.6 4.8Demersal BP Piscivores (? 40 cm) 6.8 6.8 6.8 6.8Demersal Feeders (> 30 cm) 6.4 6.4 3.6 4.4Demersal Feeders (? 30 cm) 4.4 4.4 4.4 4.4Small Demersals 1.8 1.8 1.8 1.8Lumpfish 8.8 8.8 8.8 8.6Greenland cod 9.8 9.8 6 1.6Salmon 3.6 3.6 1.6 1.6Capelin 1.4 2 0.8 0.8Sandlance 1 1 0.8 0.8Arctic cod 1.8 1.8 2.4 1.8Herring 2 2 1.8 1.8Transient Pelagics 5.4 5.4 2.4 2.4Small Pelagics 1.6 1.6 1.6 1.6Small Mesopelagics 0.8 0.8 0.8 0.8Shortfin squid 1.6 1.6 1.6 1.6Arctic Squid 2 2 2 2Large Crabs (> 95 cm) 2.6 2.6 2.6 2.6Small Crabs (? 95 cm) 2.6 2.6 2.6 1.6Lobster 2.6 5.2 2.6 2.6Shrimp 0.6 0.6 0.6 0.6Echinoderms 1.6 1.6 1.6 1.6Polychaetes 0.4 0.4 0.4 0.4Bivalves 1.8 1.8 1.8 1.8Other Benthic Invertebrates 0.4 0.4 0.4 0.4Large Zooplankton 0.2 0.2 0.2 0.2Small Zooplankton 0.2 0.2 0.2 0.2Page 59, Fisheries Centre Research Reports 12(1), 2004  APPENDIX B POLICY SEARCH PARAMETERS   Table B1a.    Lost Valley catch for BC *  Groundfish Trawl Shrimp Trawl Shrimp Trap Herring Seine Halibut Longline Salmon Freezer Troll Salmon Wheel Rockfish Live Crab Trap Clam Dredge Aboriginal Recreational Transient salmon      2.5 2.5    2.5  Coho salmon      2.5     2.5 2.5Chinook salmon      2.5     2.5 2.5Ratfish 0.25 0.25           Dogfish 0.25 0.25    0.25       Pollock 0.25            Eulachon  2.5         2.5  Juvenile herring    2.5         Adult herring    2.5         Adult POP 2.5            Inshore rockfish 2.5    0.25 0.25  2.5    0.25Adult picivorous rockfish 2.5     0.25      0.25Adult planktivorous rockfish 2.5     0.25       Juvenile turbot     0.25        Adult turbot 0.25 0.25   2.5        Juvenile flatfish     0.25        Adult flatfish 2.5 0.5   0.25        Juvenile halibut     2.5       0.25Adult halibut     2.5      2.5 0.25Adult Pacific cod 2.5    0.25        Adult sablefish 0.25    0.25        Adult lingcod 0.25    0.25   2.5    2.5Shallow water benthic fish  0.25 0.25 0.25         Skates 0.25 0.25   2.5        Large crabs 0.25        2.5    Small crabs         0.25    Commercial shrimp  2.5 2.5          Epifaunal invertebrates          2.5    *Percentages indicate the fraction of the total group biomass caught in the first year of the policy exploration. The Ecopath  description is available in Ainsworth e t al.  (2002). 2.5% of total biomass is caught for target species, 0.25% or 0.5% of total biomass is caught in retained bycatch. Back to the Future Methodology, Page 60      Table B1b.  Newfoundland Lost Valley Catch as a percentage of the biomass of each group *  Group Name Bottom trawl Shrimp trawl Recreational First Nations Cod trap Capelin Longline Redfish Lumpfish trap Snow crab traps Inshore crab traps Lobster traps Salmon Pole and line Bivalves Urchins Walrus       0.25                         Cetaceans       0.01                         Grey Seals       0.25                         Harp Seals       0.25                         Hooded Seals       0.25                         Cod > 35cm 2.5   0.25   2.5   2.5 0.25                 Cod < 35 cm 0.25 2.5     0.25     0.25                 American plaice > 35cm 2.5           2.5 0.25                 American plaice < 35cm 0.25 2.5                             Greenland halibut > 40cm 2.5           2.5 0.25                 Greenland halibut < 40cm 0.25 2.5         0.25                   Yellowtail Flounder 2.5           2.5 0.25                 Witch flounder 2.5           2.5 0.25                 Skates 2.5 0.25         2.5                   Dogfish 2.5 0.25         2.5                   Redfish 2.5             2.5                 Transient mackerel     0.25                           L. D. Bentho-pelagic Pisc. 2.5           2.5 0.25                 S. D. Bentho-pelagic Pisc. 0.25 0.25         0.25                   L.Dem.Feeders  2.5           2.5                   S.Dem.Feeders 0.25 0.25         0.25                   O.S.Demersals 0.25 0.25                             Lumpfish 0.25 0.25             2.5               Greenland cod         2.5                       Salmon     2.5                   2.5       Capelin           2.5                     Herring 0.25                               Transient Pelagics                           2.5     Small Pelagics 0.25 0.25 0.25                           Shortfin squid 0.25                               Large Crabs                    2.5             Small Crabs                      2.5           Lobster                       2.5         Shrimp 0.25 2.5                             Echinoderms                               0.25Bivalves                             2.5    * Percentages indicate the fraction of the total group biomass caught in the first year of the policy exploration. Ecopath  description isavailable in Pitcher et al.  (2002). 2.5% of total biomass is caught for target species, 0.25% or 0.25% of total biomass is caught inretained bycatch.  Page 61, Fisheries Centre Research Reports 12(1), 2004         Table B3b.  Jobs per catch value for Newfoundland.  Gear Jobs/catch value Bottom trawl 0.4 Shrimp trawl 0.6 Recreational 15 First Nations 0.1 Cod trap 2 Capelin 0.4 Cod long-line 1.3 Redfish 0.6 Lumpfish trap 5 Offshore crab traps 1 Inshore crab traps 5 Lobster traps 5 Salmon 0.2 Pole and line 1 Bivalves (clams etc.) 10 Sea urchins 10 Table B2a.  West coast discards. Percentages indicate the fraction of total biomass caught in the first year of the policy exploration. Major sources of bycatch are set at 1.25% of group biomass, minor bycatch is 0.25% or 0.025%.   Group Name Groundfish Trawl Shrimp Trawl Salmon Freezer Troll Clam Dredge Seabirds   0.025  Small crabs 1.25 1.25  0.25 Epifaunal invertebrates 1.25 1.25  0.25 Infaunal carnivorous invertebrates 1.25 1.25  0.25 Infaunal invertebrate detritivores 1.25 1.25  0.25 Corals and sponges 1.25 1.25  0.25 Table B3a.  Jobs per catch value for northern BC.  Fleet Jobs/catch value Groundfish Trawl 0.4 Shrimp Trawl 0.6 Shrimp Trap 5 Herring Seine 4 Halibut Longline 1.3 Salmon Freezer Troll 2 Salmon Wheel 0.2 Rockfish Live 5 Crab Trap 5 Clam Dredge 5 Aboriginal* - Recreational 15  *Policy search did not include aboriginal fleet. Table B2b.  East coast discards. Percentages indicate thefraction of total biomass caught in the first year of thepolicy exploration. Major sources of bycatch are set at1.25% of group biomass, minor bycatch is 0.25% or0.025%.  Group Name Bottom trawl Shrimp trawl Bivalves Echinoderms 2.5 2.5 2.5 Polychaetes 2.5 2.5 2.5 Bivalves 2.5 2.5  Other Benthic Invertebrates 2.5 2.5 2.5 Back to the Future Methodology, Page 62     Table B4a . Value weight settings for fleets, years and  policy objectives in northern BC.  *Bold values indicate that mandatedrebuilding was required to prevent computer crashes.  ** Numbers in parentheses indicate the biomass goal of the policy searchrelative to the Ecopath  baseline. Fleet Period # Objective Policy Search Parameters       Ecological Economic Social Mandated Rebuilding* Variance (2) of mixed MR protected groups** Lost Valley 1750 1 Ecological 1 0 0 0     2 Economic 0 1 0 0     3 Social 0 0 1 0     4 Mixed objective 100 1 1 0.1 0.309 Juv/ad turbot (1)   5 Portfolio Log Utility Existence value = 0.1      1900 6 Ecological 1 0 0 5  Juv/ad turbot (1)   7 Economic 0 1 0 0     8 Social 0 0 1 0     9 Mixed objective 100 1 1 5 0.192 Juv/ad turbot (1)   10 Portfolio Log Utility Existence value = 0.1      1950 11 Ecological 1 0 0 10  Skates (1)   12 Economic 0 1 0 5  Juv/ad turbot (1)   13 Social 0 0 1 0     14 Mixed objective 100 1 1 10 0.316 Skates (1)   15 Portfolio Log Utility Existence value = 1 Prediction variance = 0.005  2000 16 Ecological 1 0 0 7     17 Economic 0 1 0 5  Transient Salmon (1)   18 Social 0 0 1 0     19 Mixed objective 100 1 1 10 0.679 Transient Salmon (1)   20 Portfolio Log Utility Existence value = 10 Prediction variance = 0.02 No Recreat. 1750 21 Ecological 1 0 0 0     22 Economic 0 1 0 10  Transient Salmon (0.5)   23 Social 0 0 1 0     24 Mixed objective 100 1 1 0.1 0.194 Juv/ad turbot (1)   25 Portfolio Log Utility Existence value = 1      1900 26 Ecological 1 0 0 5  Juv/ad turbot (1)   27 Economic 0 1 0 10  Transient Salmon (0.5)   28 Social 0 0 1 0     29 Mixed objective 100 1 1 50 0.304 Juv/ad turbot (1),  Skates (1.5)   30 Portfolio Log Utility Existence value = 1      1950 31 Ecological 1 0 0 0     32 Economic 0 1 0 10  Juv/ad turbot (1)   33 Social 0 1 0 10  Skates (1)   34 Mixed objective 100 1 1 5 0.268 Skates (1)   35 Portfolio Log Utility Existence value = 1      2000 36 Ecological 1 0 0 2  Skates (1)   37 Economic 0 1 0 10  Transient Salmon (1)   38 Social 0 0 1 1  Skates (1)   39 Mixed objective 100 1 1 20 0.247 Skates (1)   40 Portfolio Log Utility Existence value = 1 Prediction variance = 0.003 No Trawlers 1750 41 Ecological 1 0 0 0     42 Economic 0 1 0 5  Transient Salmon (1)   43 Social 0 0 1 0     44 Mixed objective 100 1 1 0 0.218    45 Portfolio Log Utility Existence value = 0.1      1900 46 Ecological 1 0 0 1  Transient Salmon (1)   47 Economic 0 1 0 5  Skates (1)   48 Social 0 0 1 0     49 Mixed objective 100 1 1 2 0.097 Juv/ad turbot (1)   50 Portfolio Log Utility Existence value = 0.1      1950 51 Ecological 1 0 0 0     52 Economic 0 1 0 0     53 Social 0 0 1 0     54 Mixed objective 100 1 1 10 0.258    55 Portfolio Log Utility Existence value = 0.1      2000 56 Ecological 1 0 0 0     57 Economic 0 1 0 0     58 Social 0 0 1 0     59 Mixed objective 100 1 1 0 0.278    60 Portfolio Log Utility Existence value = 0.1     Page 63, Fisheries Centre Research Reports 12(1), 2004    Table B4b.   Value weight settings for fleets, years and policy objectives in Newfoundland. *Mandated rebuilding was not used with theNewfoundland models; some species went extinct. **Group biomass increased or decreased more than twice. Increased indicated by + anddecreased indicated by - ***Unstable indicates that ecosystem never stabilized over the 50 year time span.  Fleet Period # Objective Policy Search Parameters       Ecological Economic Social Mandated Rebuilding* Variance (?2) of mixed Large change in group biomass** Lost Valley 1450 1 Ecological 1 0 0 0  Salmon (-) G. halibut (+)   2 Economic 0 1 0 0  Many (+), many (-)   3 Social 0 0 1 0  Unstable***   4 Mixed objective 100 1 0.5 0 0.646 Skate, sf squid (-) halibut (+)   5 Portfolio Log UtilityExistence value = 0.05      1900 6 Ecological 1 0 0 0  Salmon (+) short fin squid (-)   7 Economic 0 1 0 0  Large and small crabs (-)   8 Social 0 0 1 0  Crabs, transient pelagics (-)   9 Mixed objective 100 0.1 0.1 0 0.195 Salmon (+) short fin squid (-)   10 Portfolio Log UtilityExistence value = 0.05      1986 11 Ecological 1 0 0 0     12 Economic 0 1 0 0  Salmon (-) short fin squid (+)   13 Social 0 0 1 0  Unstable***   14 Mixed objective 100 0.1 0.1 0 0.181    15 Portfolio Log Utility Existence value = 0.1   1996 16 Ecological 1 0 0 0     17 Economic 0 1 0 0  Salmon (-)   18 Social 0 0 1 0  Salmon (-)   19 Mixed objective 100 0.1 0.1 0 0.304    20 Portfolio Log Utility Existence value = 0.1  No Recreational 1450 21 Ecological 1 0 0 0  Skate, sf squid (-) halibut (+)   22 Economic 0 1 0 0  Many (+), many (-)   23 Social 0 0 1 0  Unstable***   24 Mixed objective 100 1 0.1 0 0.095 Skate, sf squid (-) halibut (+)   25 Portfolio Log Utility Existence value = 0.1      1900 26 Ecological 1 0 0 0  Salmon (+) short fin squid (-)   27 Economic 0 1 0 0  Large and small crabs (-)   28 Social 0 0 1 0  Large and small crabs (-)   29 Mixed objective 100 0.1 0.1 0 0.196 Salmon (+) short fin squid (-)   30 Portfolio Log Utility Existence value = 0.01      1986 31 Ecological 1 0 0 0     32 Economic 0 1 0 0  Short fin squid (+), many (-)   33 Social 0 1 0 0  Unstable***   34 Mixed objective 100 0.1 0.1 0 0.177    35 Portfolio Log UtilityExistence value = 0.05      1996 36 Ecological 1 0 0 0  Salmon (-)   37 Economic 0 1 0 0  Many (+), many (-)   38 Social 0 0 1 0  Many (+), many (-)   39 Mixed objective 100 0.1 0.1 0 0.191 Salmon (-)   40 Portfolio Log UtilityExistence value = 0.05  No Trawlers 1450 41 Ecological 1 0 0 0  Salmon (-) G. halibut (+)   42 Economic 0 1 0 0  Many (+), many (-)   43 Social 0 0 1 0  Unstable***   44 Mixed objective 100 1 0.1 0 0.335 Salmon (-), G. halibut (+)   45 Portfolio Log Utility Existence value = 0.1      1900 46 Ecological 1 0 0 0  Salmon (+) short fin squid (-)   47 Economic 0 1 0 0  Large and small crabs (-)   48 Social 0 0 1 0  Large and small crabs (-)   49 Mixed objective 100 0.1 0.1 0 0.145 Salmon (+) short fin squid (-)   50 Portfolio Log UtilityExistence value = 0.05      1986 51 Ecological 1 0 0 0     52 Economic 0 1 0 0  Short fin squid (+), many (-)   53 Social 0 0 1 0  Unstable***   54 Mixed objective 100 0.1 0.1 0 0.039    55 Portfolio Log Utility Existence value = 0.1      1996 56 Ecological 1 0 0 0     57 Economic 0 1 0 0  Many (+), many (-)   58 Social 0 0 1 0  Salmon (-)   59 Mixed objective 100 0.1 0.1 0 0.251    60 Portfolio Log Utility Existence value = 0.1     Back to the Future Methodology, Page 64  ENVIRONMENTAL ARCHAEOLOGY: PRINCIPLES AND CASE STUDIES   Trevor J. Orchard and Quentin Mackie  Department of Anthropology,  Universities of Toronto and Victoria   ABSTRACT   Archaeological data are most commonly applied towards understanding past human activities. However, these data can include environmental information such as animal and plant remains which offer insight into past environmental history. This paper outlines general introductory principles of environmental applications in archaeology, including the character of archaeological data, the preservation of environmental remains, and problems of interpretation arising from the ?cultural filter? through which these remains necessarily have passed. We conclude by noting problems and prospects in environmental archaeology, leading to two case studies which demonstrate the value and potential of archaeological analyses to the reconstruction of past ecosystems. The first case study explores the period of European contact in Gwaii Haanas (Queen Charlotte Islands, British Columbia), a time characterized by rapid and substantial environmental changes. In particular, archaeological evidence is described that relates to the extirpation of the sea otter during the maritime fur trade and the resulting impact on ecologically related species such as abalone, sea urchin, and kelp-dependent fish. The second case study examines prehistoric fish use in the Aleutian Islands. Specifically, size reconstruction of Pacific cod specimens recovered from Aleut archaeological sites shows the harvesting of fish that exceed the size of those commonly encountered by modern commercial fisheries. Together, these case studies demonstrate that archaeological analysis can provide a picture of the past environment that is not readily available through other sources of data.    INTRODUCTION TO ENVIRONMENTAL ARCHAEOLOGY  Archaeological faunal remains provide a useful, if imperfect, record of the past environment. The majority of archaeological faunal remains enter a site's deposits through direct human action, though a portion of such remains may result from the activities of scavengers and other animals, or may enter a site as a secondary by-product of the                                                            Orchard, T.J. and Mackie, Q. (2004) Environmental Archaeology: Principles and Case Studies. Pages 64?73 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.   targeted resources (Erlandson and Moss 2001, Lyman 2002, Moss and Erlandson 2002, Orchard 2001b). As humans tend to harvest resources from a wide variety of niches, these deposits often provide a broad view of the environments available to a site?s inhabitants. The anthropogenic nature of archaeological deposits, however, means that faunal remains from archaeological sites can be seen as a culturally filtered sample of the environment from which the site residents obtained their resources. Despite this bias, however, the abundance, accessibility, visibility and broad scope of archaeological faunal deposits make them a particularly useful environmental record, especially when compared to typically rare and limited natural faunal deposits. This is particularly true for marine mammals and fish, which have vanishingly small probabilities of ending up in accessible paleontological deposits. The value of archaeological sites as sources of environmental history has been recognized in a number of recent projects and texts (Amorosi e t al.  1997; Cannon 1995; Grayson 1984, 2001; Orchard 2001b; Reitz e t al.  1996; Reitz and Wing 1999; Sandweiss 1996). Of particular interest and relevance to the case studies outlined below, are papers that discuss and exemplify the role that zooarchaeological analysis can play in wildlife management (Amorosi e t al.  1996; Lyman 1996). The following are some simple analytical techniques or domains which have promise for answering questions about paleo-ecology.   Addressing bio-diversity is most straightforward through the creation of a species list from identified remains. Such lists from shell-bearing archaeological sites ? which typically offer the best preservation of bone ? can run into hundreds of taxa. From such lists, local ecological niches can be identified and past biodiversity compared to the present. Of particular interest are indicator or keystone species with very narrow niches or specific environmental tolerances, or whose presence or absence is a strong predictor of other species. Sea Otter probably fills such a role in near-coastal marine ecologies, as discussed below (and see Pitcher 2004, this volume).  Another area of inquiry includes changes in faunal ?demographics?. Some species will have undergone historic change in population structure or growth and development as a result of changing human or animal predation patterns or intensity. For example, species which are under heavy predation may exhibit a flattened population structure, with fewer mature individuals and more immature individuals than Page 65, Fisheries Centre Research Reports 12(1), 2004  might be expected. Potentially, population information can be derived from:  ? Shellfish, especially bivalves, through the study of annual growth rings, size, and growth rates. Such studies have been commonly undertaken in archaeology, sometimes showing a decline in average size with apparent increased predation (Ham and Irvine 1975; Wessen 1988, Claassen 1998). Shellfish incorporate seasonal and annual growth rings, and relatively complete shells, especially bivalve shells, can be thin-sectioned and these rings examined.  This has the potential to illuminate both the cultural and natural history of an area by tracking changing predation pressures, water temperatures, and so forth.  ? Fish, through the study of size and age structure of the population. The main sources of data would be otoliths (ear bones) whose rings track age and growth rates, and scales , which can preserve surprisingly well in archaeological sites. Rockfish otoliths are the largest and most robust amongst likely fish remains to be found (Wigen pers. comm.). Fish vertebrae can also be aged using x-ray densitometry. Key indicator fish skeletal elements can be correlated via regression equations to length and body mass. For example, rockfish size can be accurately estimated using the diameter of the atlas (Wigen pers. comm.); Pacific cod using dimensions of the quadrate and other mouth elements (see below).  ? Mammals and birds: as they have different reproductive strategies than the above, may be predated upon differently, and as their remains may be rarer in absolute terms, it is more difficult to be confident in one?s ability to draw conclusions. An interesting exemplary species is the sea otter, whose population is known to have declined to extirpation by ca. 1830. Knowing the temporal parameters and outcome of this increased predation it would be of use to see if this was archaeological visible through changing age structure of recovered remains (see below). Also, juvenile mammal remains can be aged, teeth can reveal information about dietary stress, and stable-isotope analysis can show changes in long-term diet. For example, preliminary, unpublished results from Haida Gwaii suggest that prior to ca. 10,000 years ago, black bears consumed little or no marine protein, in stark contrast to their present day habits.   Archaeological data can provide a very alluring source of Palaeonvironmental data for other historical sciences, but the use of these data should be well-informed. The following discusses some interpretive constraints in environmental archaeology, with emphasis on BC coastal processes.  Cultural choice: the faunal record at an archaeological site is a product of culturally-mediated choice. It is not a microcosm of the natural ecology, but a reflection of the human niche in that ecology. In spite of this, it is important to remember that not all the taxa were taken directly by humans: some came in as incidentals, stomach contents, etc. Furthermore, humans can only select from what is actually available, although trade and exchange can widen their catchments considerably. This ?cultural filter? must always be accounted for. Hence, for example, a finding that 70% of the fish bones in a particular site are from herring tells us more about human taste in food than the absolute or relative abundance of herring in the environment.  Differential preservation of environmental remains: some classes of evidence, such as large land mammal bones and shells preserve relatively well, while other remains, such as delicate fish elements, crustaceans, small land mammal and bird bones preserve less well. Some environmental information, such as terrestrial plants, marine plants, and fungi only preserve in special situations. Further, preservation may be spatially heterogeneous across the site (Stein 1996). Therefore, the diversity and proportions of environmental remains in the present do not necessarily bear a 1:1 correspondence with the material when deposited. Furthermore, most of these taphonomic processes unfold over time, meaning that the actual remains found in an archaeological site is a complex function of time, inherent durability, soil chemistry, and site sampling strategies. All of these factors need to be accounted for when attempting interpretation of environmental remains, whether for cultural or natural historical ends.  A major interpretive consideration is the amount of material that must be excavated to produce a reliable sample size and representative taxonomic diversity. At Crescent Beach, a shell midden near Vancouver, the relationship between diversity of fish taxa and size of sample (expressed as number of identified skeletal elements, NISP) is clear. After ca. 750 to 1,000 elements of all fish taxa are recovered, there is virtually no increase in taxonomic diversity (Driver 1993: 93). Achieving these sample sizes of archaeological fish remains is fairly common. However, fish tend to be among the more numerous faunal categories, and if similar numerical relationships hold for birds and mammals, then it may become an issue whether taxonomic diversity is fully represented at any Back to the Future Methodology, Page 66  given site.  The case studies outlined below exemplify some of the methods which can be applied to the gathering and analysis of environmental data from archaeological sites, as well as the types of results that may be obtained. There are many more methods that could be, or have been, applied to these cases (see, for example, Dincauze 2000), and the results presented are in some cases preliminary.  Together, the case studies demonstrate some of the problems and prospects of an archaeological contribution to marine environmental reconstruction and management.  CASE STUDY 1:   GWAII HAANAS  The period of European Contact in Gwaii Haanas National Park Reserve/Haida Heritage Site, Haida Gwaii (Queen Charlotte Islands, British Columbia), was one of rapid and dramatic change for the Haida (Acheson 1998; Duff and Kew 1958). Similarly, the current ecology and environment of Gwaii Haanas has been profoundly influenced by historic-period environmental changes, beginning with the first European contact in 1774 (Blackman 1990). In particular, activities related to the maritime fur trade such as the rapid extirpation of sea otter populations, had a dramatic impact on the local environment. The removal of sea otters, for example, is known to have allowed the spread of sea urchins, which in turn limits the growth of kelp forests and their associated ecosystems (Bodkin 1988; Breen e t al.  1982; Duggins 1981; Estes and Palmisano 1974; Pace 1981). Similar changes are known to have resulted from the introduction of non-indigenous species, such as deer (Vourc'h e t al.  2001), rats (Bertram and Nagorsen 1995; Drever 1997; Taylor et al.  2000), and raccoons (Hartman and Eastman 1999); and from modern industrial harvesting of timber and other resources (Forest 2001; Grzybowski and Slocombe 1988). Furthermore, European contact introduced diseases and changed settlement patterns which lead to mass human depopulation of Gwaii Haanas and the sequential amalgamation of small villages of 2 to 3 houses into larger villages (Acheson 1998). By 1890, all the surviving Haida had settled in the villages of Skidegate and Masset on Graham Island to the north of Gwaii Haanas (Blackman 1990), and thus the Gwaii Haanas human ecology had also been greatly altered.  Despite the importance of this period in Haida culture history, relatively little work has attempted to document or address these issues. Rather, most archaeological work in Gwaii Haanas has focussed on early Holocene occupations or on general site inventory (Fedje e t al.  1996a,b, 2001; Fedje and Christensen 1999; Hobler 1978), with contact-period archaeology limited to excavations at only a very few sites (Abbott and Keen 1993; Acheson 1998; Duff and Kew 1958; MacDonald and Cybulski 1973). Of greatest relevance to the current case study is a project carried out by Acheson (1998), which revealed the wealth of environmental data available from sites dating to the last 2,000 years, recovering remains of 165 separate faunal taxa, representing a wide range ecological niches, from small scale excavation at 18 archaeological sites. Acheson?s work, however, was not intended to address issues of environmental reconstruction, and only three of his excavated sites included historic period deposits (Acheson 1998). Although scholars in other disciplines have examined the Gwaii Haanas ecology from a current perspective while acknowledging historic changes (eg. Forest 2001; Grzybowski and Slocombe 1988), no one has specifically used archaeological data to examine the pre-contact to early contact period environment of the region.  Thus, though it is possible to speculate about many of the factors that are likely to have caused environmental changes in Gwaii Haanas since the time of first European contact, the pre-contact environment itself is largely unknown. Examination of environmental data from archaeological sites dating to the late pre-contact/early contact periods provides a unique window into this period of environmental change. Aside from providing a better understanding of the context in which the Haida people lived prior to European contact, knowledge of the "natural" pre-contact environment is a useful tool for the management of the relatively recently established Gwaii Haanas National Park Reserve/Haida Heritage Site. Although Parks Canada?s mandate includes the environmental management of the region, the question remains as to which environment to manage, the pre-contact environment prior to the impact of European activities, the current environment, or that of some intervening period. Greater knowledge of a pre-contact environmental via archaeological environmental data would contribute to such management issues. In addition, demonstrating the inherent role of Haida food harvesting in the long-term ecological structure of the Gwaii Haanas region may provide evidence for aboriginal use-rights within the park reserve.  In order to investigate the potential for environmental archaeological work, a pilot project was conducted in June of 2000. As Page 67, Fisheries Centre Research Reports 12(1), 2004  indicated above, much of the recent archaeological work in Gwaii Haanas has consisted of an extensive program of site survey, the results of which have been compiled in a Parks Canada database. This database contains information on the locations of all the known sites in Gwaii Haanas, the types of deposits found at each site, the dates of the sites when known, and the artifacts found or recovered at each site, providing a basis for the identification of sites with high potential for containing the information that we wished to recover. Specifically, we were interested in examining sites that: were occupied during the late pre-contact to early contact transition, and thus had dates or artifacts that indicated this period; contained shell midden deposits and thus had a high potential for the preservation of environmental remains; each represented a different set of environmental conditions in the form of exposed, protected and intermediate locations. Thus, the study sites (Figure 1) were selected from the database prior to the beginning of our field season.  Prior to excavation, each site was examined and tested via surface exposures, deposits in windfalls, cutbanks and other natural exposures, and through probe and auger testing. Such testing served primarily to verify the presence of preserved environmental remains in the form of shell midden deposits. Based on this testing one site, 1221T on the East coast of Lyell Island, was eliminated from our sample due to inadequate shell midden deposits. This site was replaced with site 740T on East Copper Island, another exposed site. Soil probes and augers were also used to aid in the placement of excavation units. Such subsurface sampling techniques have been shown to provide a reasonably good picture of the distribution of subsurface deposits (Stein 1986; Casteel 1970). Auger samples were also collected in some cases, and may be used as a supplemental source of environmental data (see Cannon 2000; Casteel 1970). Excavation units were placed judgmentally based on the results of soil probing and augering, with 1m by 1m units excavated in 10 cm arbitrary levels. To facilitate the recovery of environmental data, all material was water-screened through 1/8 inch mesh, with all bone, a representative sample of shell, and any other environmental remains, such as floral remains and fish scales, collected. In addition, column samples were collected from one wall of each unit following excavation, as column samples have been shown to provide a representative sample of environmental remains from an excavation unit (Casteel 1970, 1976a). All artifactual material was also collected, as were carbon samples for dating purposes when available, and each site was mapped with a total station.  The final analysis of materials from this pilot project is incomplete, and will form part of the ongoing Gwaii Haanas Environmental Archaeological Project being conducted as a component of the doctoral research program of the senior author. Nevertheless, preliminary results suggest that faunal remains from small-scale investigations can provide a picture of the past environmental characteristics of a site's local region, and can map regional environmental differences between sites in different ecological niches (see Table 1) (Mackie e t al.  2001). This is Table 1.  Number of Taxa Recovered Per Site (from Mackie e t al.  2001). Totals are for all three sites, therefore columns do not add up. Site Vertebrate Taxa Invertebrate taxa Total1  1134T (Protected) 14 21 35 923T (Semi-Protected) 10 10 20 740T (Exposed) 23 24 47 Totals1 31 36 67 Back to the Future Methodology, Page 68  particularly evident in the differences in the diversity of fish taxa between the protected site (1134T) and the exposed site (740T) as illustrated in Figure 2. Unsurprisingly, the protected site is dominated by salmon remains, contains the majority of the terrestrial-based avifauna, and was the only site to contain terrestrial mammal (Black bear). In contrast, the exposed site contained the greatest diversity of taxa, including  a wide variety of fish (13 taxa), numerous remains of marine birds, and the greatest quantity of sea otter remains. Similarly, California mussel comprises the majority of invertebrate remains from the exposed and semi-exposed sites, whereas the protected site contains primarily Butter clam, Littleneck clam and small mussel (probably edible Mussel: M y tilus trossulus ) (Mackie e t al.  2001). Also of considerable interest is the small but intriguing correlation between the presence of sea otter in the assemblages and the presence of related taxa such as sea urchins, abalone, and kelp-dependent fish. As seen in Table 2, a strong presence of sea otter remains is loosely correlated with a near absence of abalone and sea urchin and an abundance of kelp dependent fish at the exposed (740T) and semi-exposed (923T) sites, while the opposite pattern is evident at the protected site (1134T). The well documented relationship between kelp and sea urchin grazing (Duggins 1981; Pace 1981) provides an ecological link between sea otter predation on sea urchins and the presence of nearshore, kelp dependent communities of fish. It is important to note, as well, that the low density and resulting low weight of sea urchin shell yields low proportions for sea urchin when compared to other invertebrate remains from each assemblage. However, the difference in proportion between 1134T (0.82) and the other two sites, 740T (0.03) and 923T (0.01), is relatively quite significant. Slightly differing dates at these sites (Mackie et al.  2001; Orchard 2001a) suggests that this pattern may map the shift from a pre-fur trade to a post-fur trade environment.  In addition to these interesting faunal results, radiocarbon dates from the sites and the recovery of contact-period artifacts (Mackie e t al.  2001; Orchard 2001a) supports the occupation of the selected sites during the targeted time period, thus providing support for the utilized methodology. This is further evidenced by the absence, in the recovered faunal assemblages, of any introduced species, confirming that the recovered assemblages date prior to the major environmental changes discussed above. Though patterns in the data are clearly present, the small sample size and the potentially conflicting effects of varying exposure and varying temporal period may bias these results. An increased sample size resulting from ongoing work should clarify this issue. Generally, then, the pilot project demonstrated the potential of small-scale archaeological excavation to contribute to Table 2.  Sea Otter, Sea Urchin, and Ecologically Related Taxa (Derived from Mackie e t al.  2001). Taxon  740T (Exposed) 490?40 to 390?501  923T (Semi-Protected) 150?50 1134T (Protected) 430?70 to 60?60 Sea Otter (% mammal by NISP) 57.1 57.1 0 Sea Urchin2 (% invert. by weight) 0.03 0.01 0.82 Abalone3 (% invert. by weight) 0 0.59 0 Nearshore/Kelp ForestFish4 (% fish by NISP) 21.2 35.7 14.1 Nearshore/Kelp ForestFish (# identified taxa) 3 1 1  1Radiocarbon age ranges include marine reservoir correctedshell dates. 2Sea urchin is one of the primary food sources of sea otters(Estes and Palmisano 1974; Estes et al.  1978; Breen et al.1982). 3Abalone density has also been inversely correlated with seaotters (Cooper e t al.  1977). 4A variety of fish taxa are dependent upon or ecologicallyrelated to kelp forests, and are thus tied into the sea otterecological web. In the Gwaii Haanas assemblages such fishinclude greenling (Estes and Palmisano 1974), rockfish(Bodkin 1988), and cabezon (Bodkin 1988). Page 69, Fisheries Centre Research Reports 12(1), 2004  environmental reconstruction during the targeted late pre-contact to early contact time period. The ?cultural filter? through which these data have passed is important, but the underlying ecological relationships show through. This confirms the availability of a wealth of environmental data in sites that are known, through the presence of early European trade goods, ethnohistoric records, and radiocarbon dates, to have been occupied through the early contact period (Mackie et al.  2001; Orchard 2001a).  CASE STUDY 2:  ALEUTIAN ISLANDS                                     PACIFIC COD  The reconstruction of the live size of animals represented by archaeological remains can provide useful information for both culture historical and environmental reconstructions. Our second case study examines the potential of such an approach in the context of environmental reconstruction through the synthesis of a project which was aimed at examining fish size in prehistoric Aleut sites as related to Aleut subsistence and to ecological change (Orchard 2001b). The Aleutian islands of southwest Alaska form a particularly interesting illustration of the potential of environmental archaeology, as they represent a relatively unique environmental context. It is this unique setting and the isolation of the archipelago that makes it particularly useful as a ?cultural laboratory? (McCartney 1975: 288; cf. Black 1981; Corbett e t al.  1997a; McCartney and Veltre 1999; Yesner and Aigner 1976). The project outlined here, completed as the M.A. thesis of the senior author (Orchard 2001b), involved the analysis of faunal assemblages from 5 sites in the central and western Aleutian archipelago (see Figure 3). This includes two sites on Shemya Island (ATU-021 and ATU-061), one site on Buldir Island (KIS-008), and two sites on Adak Island (ADK-009 and ADK-011). For the most part, the results of the excavations at these sites, all conducted by members of the Western Aleutian Archaeological and Paleobiological Project, remain unpublished. The exception is site KIS-008 on Buldir Island, which has generated several publications (Corbett et al.  1997b; Lef?vre e t al.  1997; Bouchet et al.  1999), as well as a single publication from site ADK-011 on Adak Island (Bouchet e t al.  2001).  Regression analysis provides a technique for the statistical comparison of the live size of fish, either length or weight, to the size of skeletal elements. This technique has been widely applied to fish taxa and has demonstrated the strong correlation that exists between fish size and skeletal element size (Casteel 1974, 1976b; Crockford 1997; Desse and Desse-Berset 1996; Enghoff 1983; Leach e t al.  1996; Owen and Merrick 1994; Rojo 1986; Smith 1995). The case study involved the use of regression analysis to estimate the size (length and weight) of fish specimens from six of the most prevalent taxa encountered in the archaeological samples under consideration. The analysed taxa included Atka mackerel (Pl e urogrammus monopterygius ), greenling (H e xa g rammos  sp.), Irish Lord ( Hemilepidotus  sp.), Pacific cod ( Ga dus macrocephalus ), rockfish ( S e bastes  sp.), and walleye pollock ( T h e ra g ra chalcogramma ). For each taxon comparative specimens of known live length and weight were used to generate regression formulae that compare these size measurements to measurements of a selection of skeletal elements (Orchard 2001b). These formulae, which produced strong correlations Figure 3.  Map of Aleutian Islands study area with site locations (modified from Lantis 1984). Back to the Future Methodology, Page 70  with r2 values generally greater than 0.90 (Orchard 2001b), were then used to generate size estimates from the same measurements of archaeological skeletal specimens.  The estimated sizes of Pacific cod are particularly noteworthy in the context of discussions of environmental reconstruction and fisheries management. Archaeological Pacific cod specimens ranged up to and beyond the size ranges commonly encountered by modern commercial fisheries (see Figure 4). Of the total MNI1 of 215 Pacific cod, 27 exceed 90 cm in length and 14 exceed 100cm in length. In comparison, published maximum sizes of Pacific cod range from 1 meter (Hart 1973) to 118 centimeters (Vinnikov 1996). Reported size ranges of commercial catches include 7 to 110 cm from the eastern Bering sea (Bakkala 1984), and 27 to 97 cm from Canadian catches (Foucher 1987). It is also telling that the largest specimen in the University of Victoria comparative collection, which was derived largely from modern commercial fisheries specimens, is only 88cm in length. In addition to the general size of Pacific cod specimens, there is some indication from the archaeological remains of a decrease in the size of Pacific cod over time (Table 3). Though the mean lengths show no consistent temporal trend across assemblages, the maximum lengths  show a fairly consistent decrease over time (also see Orchard 1998). However, when the mean lengths and the proportion of individuals larger than 100cm in length are considered, site KIS 008 appears to stand out from the general trend (Table 3). Both the generally large size of Pacific                                                           1 Note that MNI values were determined using a combination of the traditional MNI approach (White 1953) and the additional data available from regression-estimated lengths (see Orchard n.d.). cod from Aleutian sites and the apparent temporal trend provide insight into the structure of past populations of Pacific cod in the region. Generally, archaeological fish size profiles, such as those for Pacific cod presented in figures 4 and 5, may provide insight into ancient fish population structures, and when combined with established dates for the archaeological deposits, can reveal long term trends and variation in commercially important stocks. In a consideration of similar archaeological data for Atlantic cod, Amorosi and colleagues suggest that ?zooarchaeology . . . would appear to have an important role in lengthening the observational series of environmental managers, perhaps warning of critical threshold discontinuities before the resource crash (rather than after, as in the case of the Atlantic cod)? (1996: 151). Thus, the cod length data presented here may have some utility in the management of the Pacific cod fishery. This is further evidenced by the utilisation of aspects of this methodology in the assessment of Steller sea lion prey consumption as it relates to North Pacific commercial fisheries (Zeppelin et al.  2001).   CONCLUSIONS  The two case studies presented above are unified in their use of archaeological faunal assemblages to help answer questions about past environmental conditions and changes. The first case study demonstrates that small-scale regional archaeological testing can provide faunal samples that reflect local ecological variation, and thus can be helpful in the reconstruction of local environmental histories. In addition, this case demonstrates that predicted changes in local ecology as a result of sea otter extirpation are visible in archaeological faunal samples. The Table 3.  Temporal patterns in Aleutian Islands Pacific cod (from Orchard 2001b).    Site  Radiocarbon Dates Mean Length (mm) Max. Length (mm) Proportion > 100cm (%) ATU 061 2570 ? 140 to 3096 ? 155 687 1250 10.00 ATU 021 1700 ? 70 to 1980 ? 60 746 1198 9.38 ADK 009 1040 ? 70 to 1240 ? 90 726 1122 4.62 ADK 011 180 ? 60 to 440 ? 40 (<2490 ? 50) 704 1048 1.96 KIS 008 220 ? 60 to 390 ? 80 807 1073 14.29 13001200110010009008007006005004003003020100Le ngthFrequencyLength DistributionC o mbined Sample of Pac if ic Cod f ro m All SitesMNI = 215Figure 4.  Size distribution of archaeological Pacific cod individuals from all five sites in the Aleutian Islands study area. Page 71, Fisheries Centre Research Reports 12(1), 2004  second case study demonstrates that the detailed reconstruction of fish size from archaeological faunal assemblages can provide data relevant to reconstructing the history of commercially important fish species, data which may play a role in current management plans for those species. Generally, these case studies exemplify the ability of archaeological data to make a useful contribution to the reconstruction of past environments and to the documentation of environmental changes.   ACKNOWLEDGEMENTS   For the Gwaii Haanas Environmental Archaeology Pilot Project we would like to acknowledge the field crew, namely Daryl Fedje and Ian Sumpter of Parks Canada, Tommy Greene of the Haida Nation and Cynthia Lake and Martina Steffen of the University of Victoria, and the support of Barb Wilson and Brian Reader of Parks Canada, and the Archipelago Management Board. In addition, Rebecca Wigen of Pacific Identifications was responsible for the faunal identification, and the project was generously funded by the BC Heritage Trust and by Parks Canada. For the Aleut fish use project Susan Crockford provided substantial support and expertise, Debbie Corbett provided access to the Aleutian Islands faunal assemblages, and the National Marine Mammal Lab, Seattle, provided comparative specimens and financial support for analysis. Finally, we would like to thank Nigel Haggan and Melanie Power for the invitation to participate in the Back to the Future Symposium. Our participation was funded by the Coasts Unde r Str ess  MCRI grant.   REFERENCES   Abbott, D.N., and Keen, S. (1993) Report on Excavations around Totem Pole Bases at Anthony Island. A Royal British Columbia Museum Heritage Record. Royal British Columbia Museum, Victoria, Canada. Acheson, S. (1998) In the Wake of the ya'?ats' xaatg?ay [?Iron People']: A study of changing settlement strategies among the Kunghit Haida. BAR International Series 711. Oxford: BAR. Amorosi, T., Buckland, P., Dugmore, A., Ingimundarson, J.H. and McGovern, T.H. (1997) Raiding the Landscape: Human Impact in the Scandinavian North Atlantic. Human Ecology 25(3): 491-518. Amorosi, T., Woollett, J., Perdikaris, S. and McGovern, T.H. (1996) Regional Zooarchaeology and Global Change: Problems and Potentials. World Archaeology 28(1): 126-157. Bakkala, R.G. (1984) Pacific Cod of the Eastern Bering Sea. International North Pacific Fisheries Commission Bulletin 42: 157-179. Bertram, D.F. and Nagorsen, D.W. (1995) Introduced Rats, Rattus spp., on the Queen Charlotte Islands: Implications for Seabird Conservation. The Canadian Field-Naturalist 109: 6-10. Black, L.T. (1981) Volcanism as a Factor in Human Ecology: The Aleutian Case. Ethnohistory 28(4): 313-340. Blackman, M.B. (1990) Haida: Traditional Culture. Pages 240-260 in Suttles, W. (ed.) Handbook of North American Indians, Volume 7: Northwest Coast. Smithsonian Institution.Washington:  Bodkin, J. L. (1988) Effects of Kelp Forest Removal on Associated Fish Assemblages in Central California. Journal of Experimental Marine Biology and Ecology 117: 227-238. Bouchet, F., Lef?vre, C., West, D. and Corbett, D. (1999) First Paleoparasitological Analysis of a Midden in the Aleutian Islands (Alaska): Results and Limits. Journal of Parasitology 85: 369-372. Bouchet, F., West, D. Lef?vre, C. and Corbett, D. (2001) Identification of Parasitoses in a Child Burial from Adak Island (Central Aleutian Islands, Alaska). Comptes Rendus de l'Academie des Sciences, Series III, Sciences de la Vie 324: 123-127. Breen, P. A., Carson, T.A., Foster, J.B. and E. A. Stewart. E.A. (1982) Changes in Subtidal Community Structure Associated with British Columbia Sea Otter Transplants. Marine Ecology ? Progress Series 7: 13-20. Cannon, A. (1995) The Ratfish and Marine Resource Deficiencies on the Northwest Coast. Canadian Journal of Archaeology 19: 49-60. Cannon, A. (2000) Assessing Variability in Northwest Coast Salmon and Herring Fisheries: Bucket-Auger Sampling of Shell Midden Sites on the Central Coast of British Columbia. Journal of Archaeological Science 27: 725-737. Casteel, R.W. (1970) Core and Column Sampling. American Antiquity 35: 465-467. Casteel, R.W. (1974) A Method for Estimation of Live Weight of Fish from the Size of Skeletal Elements. American Antiquity 39: 94-98. Casteel, R.W. (1976a) Fish Remains from Glenrose. Pp. 82-87 in Matson, R.G. (ed.) The Glenrose Cannery Site. National Museum of Man, Mercury Series, Archaeological Survey of Canada Paper No. 52. Ottawa: National Museum of Man. Casteel, R.W. (1976b) Fish Remains in Archaeology and Paleo-environmental Studies. Academic Press, London, UK.  Claassen, C. (1998) Shells. Cambridge University Press, UK. Cooper, J., Wieland, M. and A. Hines, A. (1977) Subtidal Abalone Populations in an Area Inhabited by Sea Otters. The Veliger 20(2): 163-167. Corbett, D.G., Lef?vre, C. and Siegel-Causey. D. (1997a) The Western Aleutians: Cultural Isolation and Environmental Change. Human Ecology 25(3): 459-479. Corbett, D.G., Lef?vre, C., Corbett, T.J., West, D. and Siegel-Causey, D. (1997b) Excavations at KIS-008, Buldir Island: Evaluation and Potential. Arctic Anthropology 34(2): 100-117. Crockford, S.J. (1997) Archeological Evidence of Large Northern Bluefin Tuna, Thunnus thynnus, in Coastal Waters of British Columbia and Northern Washington. Fishery Bulletin 95: 11-24. Deese, J., and Desse-Berset, N. (1996) Archaeozoology of Groupers (Epinephelinae) Identification, Osteometry and Keys to Interpretation. Archaeofauna 5: 121-127. Dincauze, D. (2000) Environmental Archaeology: Principles and practice. Cambridge: Cambridge University Press. Drever, M.C. (1997) Ecology and Eradication of Norway Rats on Langara Island, Queen Charlotte Islands. Masters Thesis, Simon Fraser University. Ann Arbor: UMI. Driver, J.C. (1993) Zooarchaeology in British Columbia. BC Studies 99: 77-105. Duff, W., and Kew, M. (1958) Anthony Island, A Home of the Haidas. British Columbia Provincial Museum of Natural History and Anthropology Report for 1957: 37-64. Duggins, D.O. (1981) Sea Urchins and Kelp: The Effects of Short Term Changes in Urchin Diet. Limnological Oceanography 26(2): 391-394. Enghoff, I.B. (1983) Size Distribution of Cod ( Gad us morhua  L.) and Whiting ( Merl an gius merlan gus  (L.)) (Pisces, Gadidae) from a Mesolithic Settlement at Vedb?k, North Zealand, Denmark. Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening 144: 83-97. Erlandson, J.M. and Moss, M.L. (2001) Shellfish Feeders, Carrion Eaters, and the Archaeology of Aquatic Adaptations. American Antiquity 66(3): 413-432. Estes, J.A., and Palmisano, J.F. (1974) Sea Otters: Their Role in Structuring Nearshore Communities. Science 185: 1058-1060. Back to the Future Methodology, Page 72  Estes, J.A., Smith, N.S. and Palmisano, J.F. (1978) Sea Otter Predation and Community Organization in the Western Aleutian Islands, Alaska. Ecology 59(4): 822-833. Fedje, D.W. and Christensen, T. (1999) Modeling Paleoshorelines and Locating Early Holocene Coastal Sites in Haida Gwaii. American Antiquity 64(4): 635-652. Fedje, D. W., McSporran, J.B. and Mason, A.R. (1996a) Early Holocene Archaeology and Paleoecology at the Arrow Creek Sites in Gwaii Haanas. Arctic Anthropology 33(1): 116-142. Fedje, D. W., Mackie, A.P., McSporran, J.B. and Wilson, B. (1996b) Early Period Archaeology in Gwaii Haanas: Results of the 1993 Field Program. Pages 133-150 in  Carlson, R. and Dalla Bona, L. (eds) Early Human Occupation in British Columbia. UBC Press. Vancouver, Canada Fedje, D. W., Wigen, R.J., Mackie, Q., Lake, C.R. and Sumpter, I.D. (2001) Preliminary Results from Investigations at Kilgii Gwaay: An Early Holocene Archaeological Site on Ellen Island, Haida Gwaii, British Columbia. Canadian Journal of Archaeology 25: 98-120. Forest, M.S.E. (2001) Ecological Sustainability on Haida Gwaii. Doctoral Dissertation, University of Oregon. Ann Arbor: UMI. Foucher, R.P. (1987) Length Composition of Pacific Cod ( Gad us macrocephalus ) from Commercial Landings by Canadian Trawlers, 1974-85. Canadian Data Report of Fisheries and Aquatic Sciences, No. 621. Nanaimo: Department of Fisheries, Fisheries Research Branch, Pacific Biological Station. Grayson, D.K. (1984) Quantitative Zooarchaeology. Orlando: Academic Press, Inc. Grayson, D.K. (2001) The Achaeological Record of Human Impacts on Animal Populations. Journal of World Prehistory 15(1): 1-68. Grzybowski, A.G.S., and Slocombe, D.S. (1988) Self-organization Theories and Environmental Management: The Case of South Moresby, Canada. Environmental Management 12(4): 463-478. Ham, L.C. and Irvine, M. (1975) Techniques for Determining Seasonality of Shell Middens from Marine Mollusc Remains. Syesis 8: 363-373. Hart, J.L. (1973) Pacific Fishes of Canada. Ottawa: Fisheries Research Board of Canada. Hartman, L.H. and Eastman, D.S. (1999) Distribution of introduced raccoons Procyon lotor  on the Queen Charlotte Islands: implications for burrow-nesting seabirds. Biological Conservation 88: 1-13. Hobler, P.M. (1978) The Relationship of Archaeological Sites to Sea Levels on Moresby Island, Queen Charlotte Islands. Canadian Journal of Archaeology 2: 1-13. Lantis, M. (1984) Aleut. Pp. 119-135 in David Damas (ed.) Handbook of North American Indians, Vol. 5: Arctic. Smithsonian Institution, Washington, USA. Leach, B. F., Davidson, J.M., Horwood, L.M. and Anderson, A. (1996) The Estimation of Live Fish Size from Archaeological Cranial Bones of the New Zealand Barracouta Th y rsites atun . Tuhinga: Records of the Museum of New Zealand Te Papa Tongarewa 6: 1-25. Lef?vre, C., Corbett, D.G. West, D. and Siegel-Causey.  D, (1997) A Zooarchaeological Study at Buldir Island, Western Aleutians, Alaska. Arctic Anthropology 34(2): 118-131. Lyman, R.L. (1996) Applied Zooarchaeology: The Relevance of Faunal Analysis to Wildlife Management. World Archaeology 28(1): 110-125. Lyman, R.L. (2002) Taphonomic Agents and Taphonomic Signatures. American Antiquity 67(2): 361-365. MacDonald, G.F. and Cybulski, J.S. (1973) Haida Burial Practices. Archaeological Survey of Canada, Mercury Series Paper No. 9. Ottawa: National Museum of Man. Mackie, Q., Orchard, T. and Lake, C. (2001). The Environmental Archaeology Pilot Project in Gwaii Haanas. Report to British Columbia Heritage Trust, Re: BC Heritage Trust Grant 00-78, December 2001. Report on file, British Columbia Heritage Trust. McCartney, A.P. (1975) Maritime Adaptations in Cold Archipelagos: An Analysis of Environment and Culture in the Aleutian and Other Island Chains. Pp. 281-338 in W. Fitzhugh (ed.) Prehistoric Maritime Adaptations of the Circumpolar Zone. The Hague, Paris: Mouton. McCartney, A.P. and Veltre, D.W. (1999) Aleutian Island Prehistory: Living in Insular Extremes. World Archaeology 30(3): 503-515. Moss, M.L. and Erlandson, J.M. (2002) Animal Agency and Coastal Archaeology. American Antiquity 67(2): 367-369. Orchard, T.J. (1998) A Case Study in Faunal Analysis: Analysis of Pacific Cod ( Gad us macroceph alus ) Remains from Shemya Island. Poster presented at the 8th International Congress of the International Council for Archaeozoology, August 1998, Victoria. Orchard, T.J. (2001a) Environmental Archaeology in Gwaii Haanas. Canadian Zooarchaeology 19: 2-8. Orchard, T.J. (2001b) The Role of Selected Fish Species in Aleut Paleodiet. M.A. Thesis, Department of Anthropology, University of Victoria, Canada. Orchard, T.J. (n.d.) The Use of Statistical Size Estimations in Minimum Number Calculations. Unpublished manuscript in the possession of the author. Owen, J.F. and Merrick, J.R. (1994) Analysis of Coastal Middens in South-Eastern Australia: Sizing of Fish Remains in Holocene Deposits. Journal of Archaeological Science 21: 3-10. Pace, D. (1981) Kelp Community Development in Barkley Sound, British Columbia Following Sea Urchin Removal. Proceedings of the International Seaweed Symposium 8: 457-463. Pitcher, T.J. (2004) The problem of extinctions. Pages 21?28 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp. Reitz, E., Newsom, L. and Scudder, S. (1996) Case Studies in Environmental Archaeology. Plenum Press. New York, USA. Reitz, E.J. and Wing, E.S. (1999). Zooarchaeology. Cambridge: Cambridge University Press, UK. Rojo, A. (1986) Live Length and Weight of Cod ( G ad us morhua ) Estimated from Various Skeletal Elements. North American Archaeologist 7: 329-351. Sandweiss, D.H. (1996) Environmental Change and Its Consequences for Human Society on the Central Andean Coast. Pages 127-146 in Reitz, E.J., Newsom, L.A. and Scudder, S.J. (eds) Case Studies in Environmental Archaeology. Plenum Press, New York, USA. Smith, P. (1995) A Regression Equation to Determine the Total Length of Hake ( Me rl uccius merluccius ) from Selected Measurements of the Bones. International Journal of Osteoarchaeology 5: 93-95. Stein, J.K. (1986) Coring Archaeological Sites. American Antiquity 51: 505-527. Stein, J.K. (1996) Geoarchaeology and Archaeostratigraphy: view from a Northwest Coast Shell Midden. Pages 35-54 in Reitz, E.J., Newsom, L.A. and Scudder, S.J. (eds) Case Studies in Environmental Archaeology. Plenum Press, New York, USA. Taylor, R.H., Kaiser, G.W. and Drever. M.C. (2000) Eradication of Norway Rats for Recovery of Seabird Habitat on Langara Island, British Columbia. Restoration Ecology 8(2): 151-160. Vinnikov, A.V. (1996) Pacific Cod ( G ad us macrocephal us ) of the Western Bering Sea. Pages 183-202 in Mathisen, O.A. and Coyle, K.O. (eds) Ecology of the Bering Sea: A Review of Russian Literature. University of Alaska Fairbanks, USA. Vourc?h, G., Martin, J-L., Duncan, P., Escarr?, J. and  Clausen, T.P. (2000) Defensive Adaptations of T h u j a plicata  to Ungulate Browsing: A Comparative Study Between Page 73, Fisheries Centre Research Reports 12(1), 2004  Mainland and Island Populations. Oecologia 126: 84-93. Wessen, G.C. (1988) The Use of Shellfish Resources on the Northwest Coast: The View from Ozette. Research in Economic Anthropology, Supplement 3: 179-207. White, T.E. (1953) A Method of Calculating the Dietary Percentage of Various Food Animals Utilized by Aboriginal Peoples. American Antiquity 18: 396-398. Yesner, D.R. and Aigner, J.S. (1976) Comparative Biomass Estimates and Prehistoric Cultural Ecology of the Southwest Umnak Region, Aleutian Islands. Arctic Anthropology 13(2): 91-112. Zeppelin, T.K., Call, K.A. and Orchard T.J. (2001) Using fish bones to estimate length of prey consumed by Steller sea lions (Eumetopias jubatas ) in the Bering Sea and Gulf of Alaska. Poster presented at the 14th Biennial Conference on the Biology of Marine Mammals, November 28 to December 3, 2001, Vancouver, BC, Canada.   For discussion after the oral presentation of this paper, see page 138. Back to the Future Methodology, Page 74  HOW TRADITIONAL KNOWLEDGE CAN CONTRIBUTE TO ENVIRONMENTAL RESEARCH AND RESOURCE MANAGEMENT   Bill Simeone   Al aska Department of Fish and Game   Over the last three years I, along with my colleague Dr James Kari, have worked with First Nations in Alaska documenting their traditional knowledge of salmon. The objectives of this research are to provide fisheries biologists with information that could be useful in resource management and improve communications between First Nations and biologists. One of the problems is that within the scientific and management communities there is considerable uncertainty as to how traditional knowledge can contribute to scientific research. In this paper I outline four ways that traditional knowledge can contribute to environmental research and resource management. These are: 1) Traditional knowledge has a chronological depth which far surpasses written historical sources; 2) Traditional knowledge includes observations of the environment that are usually far more detailed than those collected by scientists; 3) Traditional management systems are community based; and 4) Traditional knowledge stems from a belief system that is ecological in nature.   Traditional knowledge can be divided into three analytical components: knowledge, practice and belief (Berkes 1998:13-14). The knowledge base includes such basic information as species identification, taxonomies, species behavior and distribution, and life histories. This knowledge has two significant attributes: it has considerable time depth and it is often very detailed. Collected over generations, traditional knowledge provides information that is not available anywhere else. The earliest written records relating to western Canada and Alaska go back to the 18th century and are often limited in time and space. As a result scientists today have short chronologies on which to build predictions or management plans. In contrast, the historical narratives and oral traditions of First Nations extend well past the earliest arrival of Europeans and often contain precise information about the environment and                                                            Simeone, W. (2004) How traditional knowledge can contribute to environmental research and resource management. Pages 74?77 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.   environmental change. For example, oral traditions often contain information about catastrophic environmental events such as floods, earthquakes, volcanic eruptions, and unusual weather, as well as descriptions of extinct flora and fauna (cf. Cruikshank 1981).  Traditional knowledge includes considerable detail. Hunters and fishers acquire extensive knowledge of the environment because of the variety of activities they undertake in all seasons of the year. Their dependence on animals and plants for food, clothing, and tools requires a detailed knowledge of when and where resources are available and the environmental processes that affect their availability. This breadth of knowledge is reflected in traditional classification systems that are often much more extensive than those provided by science. Learning how First Nations classify natural systems provides us with a more detailed and nuanced view of the environment.  Ahtna Athabaskans, a First Nation living in Alaska, have developed an accurate and complete taxonomy of all fish species found in the Copper River Basin and gained knowledge of salmon distribution, salmon life histories, and behaviour.   Figure 1. Map of the Copper River area, Alaska, USA. Page 75, Fisheries Centre Research Reports 12(1), 2004  The Ahtna lexical inventory for fish is a good example of local people?s ability to accurately describe local fauna. In the Ahtna language there are terms for 19 species of fish, including all 14 species found in the Copper River Basin, and inventoried by the Alaska Department of Fish and Game (ADF&G). The additional five species exist outside the basin and are known to Ahtna through trade. The Ahtna taxonomy for fish is divided into two empirical categories, tsabay, which are fish other than salmon, and the more general term used for the class Pisces, and ?uk?ae, a term referring both to salmon in general and sockeye in particular.  For the term ?uk?ae there is considerable lexical embellishment revealing extensive and specific knowledge of salmon ecology. For example, the Ahtna language includes terms covering almost every phase in the life cycle of salmon. Salmon alevin, are ?uk?ae yiige (salmon?s spirit); salmon fingerling are referred to as ?uk?ae ggaay (little salmon); little salmon fry headed down stream are called ?u??uli (those that are swimming past); spawning fish are tazdlaexi (those that are swimming in water), and dead salmon are called tu?taeni (the one that is dead in water). Female salmon are referred to as K?unn?i (the roe one), and male fish are tl?ets?i (the milt one). Seasonal variations of fish are also noted. Full sized, prime early Figure 2.  Andy Tyone of Gulkana pulling a chinook salmon out of his fish wheel on the Copper River, Alaska. Fish wheels were introduced to the Ahtna at the beginning of the 20th century and are now the preferred method for catching salmon on the Copper River. Each fishwheel is registered and receives a number from the Alaska Department of Fish and Game. All fish wheels are home made, usually out of logs and lumber. The Ahtna have a tradition that no metal is to be used in the construction of a fish wheel because the salmon are believed not to like metal. Figure 3 . Processing salmon on the Copper River. Today some Ahtna keep fish camps but others bring their fish home to process them. On the left side of the photograph is a smoke house made from logs and chicken wire. Ba? or drying salmon can be seen hanging. Using the traditional method, the salmon are first covered with dust and placed in pits for one or two days and then soaked in the river. This removes some of the grease and makes them easy to handle. The heads are then removed and left to soak while the carcass is split and the backbone removed from the meat. The fish are hung for a week or more until they are dried and then bundled up and stored in a cache. Back to the Future Methodology, Page 76  running sockeye are called nulaeggi (island swimmer), and late running sockeye are named dak?aay (that which is ridged, humped). Late running sockeye in Tonsina Lake, located in the lower Copper River drainage, are called tsiis luugge? (ocher salmon), and whitefish caught in late fall at freeze up are nen?ten luugge? (frozen ground fish). The comprehensiveness of these terms indicate that Ahtna have long been aware of the various phases in the life cycle of the salmon.  Ahtna have recognized and named 21 distinct salmon populations that emanate from particular home streams. The best known of these, recognized by biologists and Ahtna alike, are natae? luugu? ?roasted salmon fish,? the large sockeye bound for Tanada Lake, located in the Wrangell Mountains at the head of the Copper River. These populations are similar to the salmon stocks identified by biologists of the Alaska Department of Fish and Game, but whereas biologists differentiate between stocks that spawn at different locations within the same system, Ahtna do not. Biologists, for example, consider sockeye bound for Tanada Lake as two separate stocks, one that spawns at the outlet of the lake and one that spawns in the lake, but Ahtna classify all sockeye from Tanada Lake as natae? luugu?.  First Nations have put their knowledge of the environment into practice by developing successful management strategies. Traditional management systems are community based. Management is in the hands of the resource users who adhere to the rules in response to social pressure, cultural mores, and/or ideological conviction rather than government or administrative authority (Feit 1988: 74). The advantage of such systems is that they are designed around a common set of values that everyone understands and accepts. Decisions are not made at a distance or from the top down but locally. One key to implementing successful management strategies is to have the users understand and accept the goals and objectives of the resource managers. For this to happen the users have to have a stake in management.  The ?self management? systems developed by First Nations involve both an understanding of ecological processes and a code of ethics that govern human-environmental relationships. These ethical standards stem from a belief system, or worldview, that is ecological in nature. From this perspective everything in the environment is linked, there is no separation of society from nature. The individual is considered part of a complex web of relationships that includes both human society and the natural environment. Behaviour in all relationships, whether with humans or animals, is guided by a set of principles that stress cooperation, restraint, and balance. Animals are considered powerful actors who freely give themselves to humans, if humans treat them appropriately. Proper Figure 4.  A processed sockeye salmon ready for hanging in the smoke house. A stick is used to hold the meat open so that is will not curl up and leave a raw space where flies can lay their eggs. The meat and backbone are left attached until they are completely dried, then the backbone is removed and stored separately.  Page 77, Fisheries Centre Research Reports 12(1), 2004  treatment involves the sustainable use of animals, maintaining a clean habitat, and taking only what you need without waste.  To avoid waste Ahtna carefully gauge their harvest against the capacity to process the fish. Once this capacity is reached the harvest is suspended, so that fish are not unnecessarily caught and spawning fish can escape. Ahtna are also concerned with catching the right kinds of salmon. To make ba? or dried fish, Ahtna select salmon based on their sex and reproductive condition, preferring male salmon to females because the former are larger and fatter. As one Ahtna elder remarked:   That what he used to do, he [we] keep more males?just throw em back in river. Sometime he [we] take em all, sometime he let the female go. That?s why he used to have a lot of fish long time ago.  Kata?ile?i , (spawning salmon) they let them go.  In the past when female salmon were caught in a dip net or trap they were released, but modern fishing technology has altered this practice. Fishwheels run during the night when no one is around, so people are obliged to keep all of the fish they catch. As Ahtna elders note, old fishing practices were in place to ?save everything,? that is to ensure a sustained yield.  In summary, traditional knowledge can contribute to basic scientific research and to resource management. First Nations have detailed knowledge of their environment and an understanding of long-term ecological processes. Their knowledge provides a time depth that is unsurpassed in the North in its continuity and can help explain ambiguities found in other kinds of evidence that can be incorporated into research. While traditional management systems are rooted in an understanding of the human-nature relationship different from science they can provide us with insights that could spark alternative explanations about the natural world.   To gain understanding means sharing information, which requires creating venues where all parties can feel comfortable sharing information (cf. Pinkerton 1990: 335). Effective communication requires acknowledging that local people do have valuable information or insights, and that scientists and managers have legitimate views and concerns. The objective is to build relationships with local people so that managers and locals can develop common goals.    REFERENCES  Berkes, F. (1999) Sacred Ecology: Traditional Ecological Knowledge and Resource Management. Taylor and Francis, Philadelphia,USA.  Cruikshank, J. (1981) Legend and landscape: Convergence of oral and scientific traditions in the Yukon Territory. Arctic Anthropology 18(2): 67-94. Feit, H. (1988) Self-Management in the North American Arctic: The Case for Co-anagement. In Traditional Knowledge and Renewable Resource Management. Pages 72-91 in Freeman, M.M.R and Carbyn, L.N. (eds) University of Alberta, Edmonton, Canada. Pinkerton, E.W. (1990) Summary and Conclusions Pages 317-337 in Dyer, C.L. and McGoodwin, J.R. (eds) Folk Management in the World's Fisheries: People, Problems, and Policies. Stanford University, USA.   For discussion after the oral presentation of this paper, see page 144.            Back to the Future Methodology, Page 78  W HY W E HAVE TO ?O PEN THE LOST VALLEY?:  CRITERIA AND SIMULATIONS FOR SUSTAINABLE FISHERIES    Tony J. Pitcher Fisheries Centre, UBC   ABSTRACT   This paper examines why and how sustainable fisheries might be opened in a restored marine ecosystem in the  ?Back to the Future? (BTF) approach, termed ?Opening the Lost Valley? (LV). A sequential list of nine criteria for designing LV fisheries includes historical gear types, conservation, community and cultural values. Sustainability is estimated by maximizing ecological, social and economic objective functions, moderated by a set of rules ensuring both sustainability and social acceptance. Pyramids of trophic flows, a surrogate diversity index and biomass profile diagrams provide comparison with present day ecosystems.  An example LV analysis is presented for the North Sea restored to its 1880 condition. Optimizing an equal balance of economic, social and ecosystem objectives results in larger fisheries than adopting ecosystem objectives alone, and larger catches entail trade-offs of conservation with depletions of some ecosystem components. Model uncertainty resides principally in ?top-down? or ?bottom-up? trophic control parameters that govern predator-prey interactions. Process uncertainty mainly lies in responses to climate change.    Imagine a restored ecosystem. All the grief and pain of fisheries being closed to get there. Then the goal is achieved and the fisheries are  opened again. In the fishing ports, laid-up fishing vessels are de-rusted, repaired, gear refurbished and the fleets sets off for the first open season in many years. Naturally, huge catches are made. But this situation does not last long, and the depletions of the past are soon repeated because of the huge overcapacity of the fishing fleet (Figure 1). In an ecosystem restored to some state resembling the past under the BTF process, it is clear that we cannot use today?s fleet. This paper examines a way to design sustainable fisheries to use in a restored future.  A marine ecosystem restored to some semblance of its past state might be thought of as a ?Lost                                                   Pitcher, T.J. (2004) Why we have  to open the lost valley: criteria and simulations for sustainable fisheries. Pages 78?86 in Pitcher, T.J. (ed.) Back to the Future: Advances in Methodology for Modelling and Evaluating Past Ecosystems as Future Policy Goals. Fisheries Centre Research Reports 12(1): 158 pp.   Valley?1,  an ecosystem, like Arthur Conan Doyle?s Lost World (Figure 2, Doyle 1912), discovered complete with all of its former diversity and abundance of creatures. This paper describes how we might achieve sustainable fishing in a restored                                                  1 We are grateful to Dr Daniel Pauly for suggesting this term in 2001. Although I think the Conan Doyle reference is the most appropriate, ?Lost Valley? is also the title of a Max Brand cowboy novel from the 1950s, and is now the name of several remote ski and dude ranch resorts in USA. Figure 2. Cover (left) of the first 1912 edition of ?The Lost World? by Sir Arthur Conan Doyle (1859-1930, right), creator of the detective Sherlock Holmes. This book, in which explorers discover an intact ecosystem of dinosaurs from the Jurassic,  was one of a series of stories about Professors Summerlee and Challenger, whose characters were based on real life Professors William Rutherford and Sir Robert Christison from Edinburgh University.  Another character in the stories, Lord John Roxton, was based on Roger Casement, a British diplomat executed for treason in 1916 because he persuaded the Germans in the First World War to allow Irish nationalists to fight on their side. The ?Lost Valley? term used in BTF combines the ?Lost World? term with the title of an earlier Conan Doyle novel ?The Valley of Fear? (1911).   Large Reef Fish Biomass & CatchHong Kong Model0.00.20.40.60.81.00 2 4 6 8 10 12 14 16 18 20 22 24yearsBiomass, t/km200.000020.000040.000060.000080.0001Catch, t/km2 /yearBiomass, tonnes/km2  Catch tonnes/km2 /year Si mu l ati o n Ye ar s  Figure 1.  Biomass of one group (large reef fish; left axis) from an ecosystem simulation model of Hong Kong. Biomass recovers during a 5-year no-take period (shaded), only to be rapidl y depleted