Bioeconomic Modelling and Valuation of Exploited Marine Ecosystems (Economy & Environment) [1 ed.] 1402040415, 9781402040412, 9781402040597

Citation preview

Bioeconomic Modelling and Valuation of Exploited Marine Ecosystems

Economy & Environment VOLUME 28

Scientific Advisory Board Scott Barrett, School of Advanced International Studies, Johns Hopkins University, Washington DC, U.S.A. Klaus Conrad, University of Mannheim, Mannheim, Germany David James, Ecoservices Pty. Ltd., Whale Beach, New South Wales, Australia Bengt J. Kriström, University of Umea, Sweden Raymond Prince, Congressional Budget Office, U.S. Congress, Washington DC, U.S.A. Domenico Siniscalco, ENI-Enrico Mattei, Milano, Italy / University of Torino, Italy

The titles published in this series are listed at the end of this volume.

Bioeconomic Modelling and Valuation of Exploited Marine Ecosystems by

J.C.J.M. van den Bergh Free University, Amsterdam, The Netherlands

J. Hoekstra National Institute of Public Health and the Environment, Bilthoven, The Netherlands

R. Imeson Free University, Amsterdam, The Netherlands

P.A.L.D. Nunes University of Venice, Italy

and

A.T. de Blaeij University of Utrecht, The Netherlands

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN-10 ISBN-13 ISBN-10 ISBN-13

1-4020-4041-5 (HB) 978-1-4020-4041-2 (HB) 1-4020-4059-8 (e-book) 978-1-4020-4059-7 (e-book)

Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. www.springer.com

Printed on acid-free paper

All Rights Reserved © 2006 Springer No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands.

CONTENTS

Preface

1. Background and summary

PART I TOOLS AND BASIC INSIGHTS

vii

1

13

2. Integrated assessment of marine ecosystem exploitation

15

3. Deterministic economic models of fisheries management and policy

29

4. Incorporating uncertainty in the economic analysis of marine ecosystem exploitation

53

5. Managing the fisheries: a synthesis of old and new insights

73

PART II BIOECONOMIC MODELING 6. Harvesting and conservation in a predator–prey system

91 93

7. Bioeconomic analysis of a shellfishery with habitat effects

117

8. Marine reserve creation for sedentary species with uncertain metapopulation dynamics

137

9. A spatial–temporal model of the interaction of shellfish and birds in a marine ecosystem

149

vi PART III

CONTENTS MONETARY VALUATION AND STAKEHOLDER ANALYSIS

183

10. Policy failure and stakeholder dissatisfaction in the Dutch wadden sea shellfishery

185

11. Stated choice valuation of multiple stakeholders in the Dutch wadden sea

209

12. The cost of exotic marine species: a joint travel cost – contingent valuation survey

229

References

241

Index

257

PREFACE

Economic dimensions have been somewhat neglected in research on nature policy and management of marine ecosystem exploitation. The current book aims to fill this gap. It examines the mismatches of natural, socio-economic, and regulatory processes and regimes in time and space. This involves dealing with the complexity and uncertainty that are inherent to the interaction of marine ecosystems and economic systems. The approach adopted is based on the idea that the design of sustainability policies requires an integration of insights from resource, environmental, and ecological economics on the one hand and marine biology and environmental sciences on the other hand. For this purpose, use is made of integrated assessment on the basis of theoretical and applied mathematical models. The book is the result of a large project, hosted by the Department of Spatial Economics of the Free University in Amsterdam, under the supervision of the first author. The project received funding from the Netherlands Organisation for Scientific Research (NWO), through a “priority program” titled “Sustainable Use and Conservation of Marine Living Resources,” which was initiated and coordinated by Prof. Wim Wolff of the University of Groningen. The work reported here has greatly benefited from external advice given by marine biologists who participated in this research program. Co-responsibility for the various chapters is as follows: van den Bergh – 1, 3, 5, 6, 7, 8, 10, 11, 12; de Blaeij – 11; Hoekstra – 2, 3, 6, 9; Imeson – 3, 4, 5, 7, 8, 10; and Nunes – 11, 12. Chapter 9 was co-authored by B. Ens. We are especially grateful to Peter Nijkamp for initial support, to Bruno Ens from Alterra for cooperation and feedback, and to Wim Wolff for taking the unusual initiative to bring natural and social scientists together in a single research program under the reliable umbrella of NWO.

vii

CHAPTER 1

BACKGROUND AND SUMMARY

1.1. MOTIVATION AND PURPOSE This study offers economic analyses of exploited marine ecosystems aimed at providing clear insights for policy and management of fisheries and other activities that affect these ecosystems. There is a large body of biological research and a smaller body of social science research aimed at understanding and managing marine ecosystems. So far, however, these have developed quite independently. The present study tries to provide various links between approaches and insights arising from the two bodies of research. In particular, it tries to move beyond traditional economic fishery analyses, in two respects. First, several theoretical and numerical models are offered that combine economic and ecological descriptions of fisheries. These models give special attention to spatial processes, to combining exploitation and conservation objectives, to sedentary fish species, and to the interaction between prey fisheries and predatory birds. Second, valuation and stakeholder concerns are addressed in empirical analyses employing both qualitative and quantitative approaches. The latter is done by using advanced methods of monetary valuation. In addition, the book aims to offer a general, advanced introduction through the inclusion in the first part of the book of short overviews of integrated assessment, economic modeling of fishery management, and approaches to incorporate various types of uncertainty in economic analyses. These introductions will allow readers of a variety of disciplines to create a sufficient background so as to be able to read later parts of the book. This first part also includes a chapter in which traditional and recent ideas on fisheries policy are critically discussed and synthesized. Many of the studies are motivated by and applied to a multifunctional marine ecosystem area in the Netherlands, known as the Wadden Sea (“Waddenzee”). This is an area of about 2500 km2 dominated by a shallow sea bordered by islands, extending along the North Sea coasts of The Netherlands, Germany, and Denmark. It is subject to various dynamic processes, caused by tidal channels and rivers, which affect sand, dunes, beaches, mud flats, and salt marshes in the area. Intensive shellfishing, notably on Blue mussels, cockles, and Spisula, has been an important economic activity in the Dutch coastal area for centuries. In the Wadden Sea, shellfishing has created a conflict between the fishing community and birds (Eider ducks and Oystercatchers) that forage on the same shellfish (Ens et al., 2004). This conflict is an important motivation for the analyses presented here. Despite this specific empirical context, the

1

2

1. BACKGROUND AND SUMMARY

resulting approaches, models, and policy insights are nevertheless sufficiently general and innovative to be of interest to a broad audience. There are many research results available for the Wadden Sea case study area. Beukema and Cad´ee (1996) provide an overview of unusual events that can be largely attributed to the sudden removal of large amounts of mussels and cockles from the Dutch Wadden Sea. The interdependence of shorebirds and shellfish has been documented in detail in Stillmann et al. (1996). A general account of the ecology of the area is offered by Wolff (1983). An evaluation of existing shoreline fishery policy is contained in Ministerie LNV (1998). Two chapters in this book deal with another application context, namely characterized by exotic, harmful algae in the North Sea. These result mainly from ship ballast water that originates from other parts of the world (van den Bergh et al., 2002). Also here use is made of particular biological insights and studies, as well as of insights from legal sciences. The economic literature on this topic is very scarce. Invasive, exotic species in general have received some attention from the perspective of dynamic modeling as well as cost-benefit analysis and monetary valuation (Perrings et al., 2000; Perrings, 2002). Perrings et al. (2002) argue that because the problem of invasive exotic species is primarily economic and of a “public good” (“bad”) nature it requires an economic solution. Fundamental causation factors are trends of increase in international travel, trade, and human population density. Biological studies have assessed the harmfulness of microalgae species in terms of causing damage to marine living resources and ecosystems. Integration of insights is urgently needed. This book develops a number of theoretical and empirical bioeconomic models of conflicts between fisheries and nature. The temporal dimension will receive special attention, since time scales of marine processes, economic processes, and policy cycles differ. Models will be developed that incorporate short-term and seasonal fluctuations and cycles, and long-term trends (Nijkamp and van den Bergh, 1998). The approach here can be regarded as integrating dynamic issues which have been addressed mostly in partial, isolated settings. These include the impact of discounting, intertemporal externalities, and myopic decision rules by fishermen, investors and regulators on longrun sustainability of the marine system and its uses. Particular attention will be given to the implications of alternative ways of modeling population dynamics of multispecies interactions (Hoekstra and van den Bergh, 2005), metapopulations (e.g. Brown and Roughgarden, 1997; Imeson and van den Bergh, 2004), and irreversible change due to destruction of shellfish (notably mussel) banks. The integration takes account of both multispecies fisheries, in particular shellfish, and interactions between shorebirds and shellfish populations. In addition to these biological components, bioeconomic models and analyses will include details of fisheries technology, behavior, price fluctuations, policy impacts, costs and benefits of private decisions, and externalities. Spatial dimensions will receive particular attention in a separate chapter. Both movement in space and locations of sedentary species are relevant here. The question is ultimately which policy-orientation – aimed at efficiency, precaution or experimentation – can contribute to long-run sustainability. The role of hidden objectives, externalities, interest groups pressure, conflicting objectives of different subgroups of fishers (e.g. associated with catching cockles and mussels), and other

1. BACKGROUND AND SUMMARY

3

stakeholders is important in understanding the difference between desirable and realized outcomes. Desirable outcomes need to satisfy sustainable fisheries, both from the perspectives of fishers and society at large. In order to arrive at integrated assessment, a wider environmental and spatial context is added to some of the fishery resource models. The result of this can be interpreted as consistent with fisheries management that focuses on ecological sustainability, as opposed to fisheries stock sustainability in more traditional economic models of fisheries.

1.2. RESEARCH METHODOLOGY: MODELING, VALUATION, AND STAKEHOLDER ANALYSIS Integrated assessment is at the basis of the formal approaches adopted here to study management of marine ecosystems and their economic uses. Integrated assessment is multidisciplinary in nature. Many efforts have already been undertaken to bridge natural and social sciences, at both global scales (Rotmans and de Vries, 1997; Kelly and Kolstad, 1999) and regional scales (Holling, 1978; Costanza et al., 1993; Turner et al., 2000, 2003; van den Bergh et al., 2001, 2004). There exist basically three main approaches to bioeconomic modeling of interactions between marine ecosystems and economic-policy regimes, namely economic optimization, uncertainty modeling, and evolutionary modeling. The first approach focuses on dynamic optimization of social welfare, given deterministic models of single or multiple populations. The latter two approaches allow to deal with uncertainty and ecological resilience of disturbed ecosystems and communities, in different ways. Evolutionary models in particular deal with diversity, innovation, and selection. Each of these three types of models can contribute particular insights regarding policy and management for sustainable use of marine resources. In fact, each of these models may be related to particular policy perspectives, such as cost-benefit analysis (or economic efficiency) in the case of optimization models, precautionary approach (safe minimum standards) in the case of uncertainty models, and adaptive resource management (experimenting) in the case of evolutionary models. The traditional economic optimization analysis of exploitation of marine resources offers a variety of theoretical models and insights (see Clark, 1990, 1999; Bulte, 1997; Imeson et al., 2002; Chapter 3). These models are sometimes referred to as “bioeconomic” as they combine descriptions of biological and economic processes. On a theoretical level, economic models have been linked to age-structured (Clark, 1990) and multispecies models (May et al., 1979; Flaaten, 1988). Typically, the demand side has received less attention in these supply-oriented studies. The main purpose of the theoretical exercises has been to compare market (or better “laisser faire”) and socially optimal outcomes of fisheries, and to propose specific regulatory measures to bridge the gap between the two. Dynamic optimization has focused on dynamic adjustment of investment and fishery effort (capacity use) after some disturbance, but also on price fluctuations, technical change (e.g. vessels), and entry of new firms (e.g. fishers). The treatment of investment in the standard optimization framework has turned out to be a difficult problem, due to irreversibility of investments in vessels and infrastructure (Clark et al., 1979).

4

1. BACKGROUND AND SUMMARY

Fluctuating population and yield patterns require an analysis of risk strategies and compensation measures. Walters (1986) has adopted an uncertainty oriented “adaptive management approach” in which management problems are bounded by explicit and hidden objectives and practical constraints on action. In this approach models are used to make dynamic behavior explicit so that learning and improvement can take place. Here, uncertainty and its propagation through time is represented and translated into multiple hypotheses. It is argued that policies should not only be aimed at realizing a definite management situation, but also be devised so as to improve understanding of the marine–economy interactions and yet untested opportunities. This approach is closely linked to the notion of resilience, a sort of extended stability concept. This can be defined as the time necessary for a disturbed system to return to its original state (Pimm, 1984) or as the intensity of disturbance that a system can absorb before moving to another state (Holling, 1973). In the presence of complex, dynamic ecosystem management, according to the adaptive management approach one can best focus on policy learning through experiments with resource and ecosystem policy, and management. So far, however, the operationalization of these ideas has seen more convincing application to disturbed terrestrial than marine ecoystems, notably through Holling’s four-box model (Holling, 1986; Gunderson and Holling, 2002). For some types of questions evolutionary models oriented towards multidisciplinary and integrated research can produce complementary information to the deterministic and uncertainty approaches (van den Bergh, 2004). This approach allows for an analysis of technological progress (innovation) and disequilibrium and irreversible development of renewable resources, including fisheries (McGlade and Allen, 1987; Munro, 1997; Noailly et al., 2003). Selection processes that change diversity of strategies at the individual or public level are an obvious starting point for such models. Related to this are systems models in which a fishery is considered both part of the ecosystem and part of the economy, adopting a systems view as well (van den Bergh and Nijkamp, 1994). These types of models allow for a wider analysis of multiple use conflicts or reinforcement of negative impacts of economic activities on the unsustainability of marine system functions and uses. The approach can underpin loose theories of self-organization through the evolution of norms among local users of common pool resources (Hanna, 1997; Ostrom, 1990). In addition, there is some overlap between this and the previous approach, as illustrated by the use of evolutionary approaches to understand adaptive management (Rammel and van den Bergh, 2003; Janssen et al., 2004). The policy context from the economic perspective should involve two elements. First, there is a set of instruments that are potentially relevant in operationalizing ideal policies, or second-best policies given imperfections, like incomplete information, ineffective control, and market distortions, notably market power, externalities, and public goods. Second, the costs of management should be considered, as it is clear that an evaluation of concrete policies should also involve a weighing of social or sectoral benefits against transaction costs associated with a possibly complex and expensive monitoring system. These depend of course on the types of instruments, like taxes, fixed quota, or tradable quota traditionally analyzed, or more integrated policies focusing on land use and employment in coastal areas as well as economy-wide effects (van den Bergh and Nijkamp, 2000). Recent general insights on nature policy from an

1. BACKGROUND AND SUMMARY

5

ecological economics perspective are also useful in this context (van der Heide et al., 2002). In additional to formal analytical and numerical modeling, this book makes use of monetary valuation and stakeholder analysis techniques. Economic or monetary valuation is based in firm microeconomic theory (e.g. Freeman, 1993; Hanley and Spash, 1993). Monetary valuation assessments are based on comparing two states, that is, they focus on changes. These changes should preferably be small, notably when compared with income. Valuation of (marine) ecosystems as such does not fit in this framework. Instead, ecosystem valuation should be based on a specific ecosystem change scenario, where the change is the result of an exogenous (environmental, economic, or technological) factor or a newly implemented policy. A much noted study that aimed to estimate the monetary value of all ecosystems covering the world (Costanza et al., 1997), is inconsistent with the basic assumptions of monetary valuation theory (see Pearce, 1998; and various comments in Ecological Economics 1998, vol. 25(1)). Among others, this study lacked a realistic change scenario: indeed, the loss of all ecosystems of the world is beyond comprehension, and certainly inconsistent with the required small change to be valued. Monetary valuation techniques can, however, be used to estimate costs or benefits of ecosystem management scenarios. Monetary valuation of ecosystem changes has attracted much attention (Nunes and van den Bergh, 2001). This holds especially for terrestrial areas and inland wetlands used for recreation and agriculture. Nevertheless, certain types of marine ecosystems have seen a great number of applications as well, notably coral reefs (Cesar, 2000; Birkeland, 2004) and coastal wetlands (Turner et al., 2003). Open seas and other coastal areas in temperate zones have received much less attention (e.g. Nunes and van den Bergh, 2004). Monetary valuation makes use of a number of techniques. Travel cost models assess the value of nature areas or parks by linking recreation demand to the generalized travel costs, including loss of time. Hedonic price models relate property values (land, houses) to environmental indicators like clean air, noise, and access to nature. Both methods “reveal preferences”, that is, they analyze actual choices made by individuals in existing markets to estimate non-market values of environmental goods or services. Another class of methods, stated preference techniques, in particular contingent valuation, has received widespread attention. It asks individuals directly to assign, implicitly or explicitly, a value to a change scenario. This method can address a wide range of (hypothetical) environmental changes, including those that have not yet occurred. In addition, it allows the assessment of non-use values. A main disadvantage is that it suffers from various biases as a result of the hypothetical character of questions asked to respondents. Economic valuation can be criticized from many angles (see Gowdy, 1997; Blamey and Common, 1999). The main strength of it is that it can inform policy making with democratic choices made by citizens in the form of “monetary intensities”. The units of these intensities are then consistent with the monetary units of cost and benefit indicators that tend to dominate in public and political decision making. In the present book, the democratic context will be widened by performing valuation analyses involving different stakeholder groups, to account for the fact that stakeholders may hold distinct values, which in turn can have a unique relationship with certain characteristics of these stakeholders.

6

1. BACKGROUND AND SUMMARY

1.3. ORGANIZATION, SUMMARY, AND MAIN CONCLUSIONS OF THE BOOK The book is organized in three parts. Part I, presents basic tools and insights, Part II offers analyses that make use of various bioeconomic models, and Part III deals with monetary valuation and stakeholder analysis. Part I opens with Chapter 2, in which the integrated assessment of marine ecosystems is discussed. Integrated modeling takes the form of either scenario analysis or policy optimization for complex environmental problems. Climate change and acid rain have received much attention, but similar models exist for resource and ecosystem oriented analyses, often concerned with fisheries and forestry. Integrated assessment is a multidisciplinary exercise aimed at describing and analyzing a causal chain starting with socio-economic drivers, such as economic production, which causes physical or ecological change through polluting emissions or resource harvesting. Formal models are essential as the complexity of feedback quickly runs out of hand. But even in simplified versions, analytical models are required as intuition quickly falls short. Integrated assessment models can then best be regarded as computational laboratories. The discussion in this chapter revolves around four themes that are at the core of integrated assessment, namely dynamics, space, conflict, and uncertainty. These themes are confronted with a classification of models to see what combinations can be expected to be fruitful. The model classification is based on a number of oppositions, such as theoretical vs. empirical, prescriptive vs. descriptive, optimization vs. resilience, and analytical vs. numerical. The chapter ends with a subtle discussion of calibration and validation of models and trust in integrated assessment with models. The main conclusion is that integrated models enforce a consistent structure upon multidisciplinary knowledge, and can therefore explore the behavior of complex systems in a more reliable way than can mental models. It is concluded that in order to avoid incorrect use of model or model results, modelers should as much as possible make implicit assumptions and uncertainties transparent to users and decision makers. Chapter 3 presents a survey of deterministic bioeconomic models of fisheries management and policy as well as the main conclusions derived from these. Basic fishery models like the open-access and sole owner models are discussed and an interpretation of the obtained Golden Rule for the sole owner model is given. Ecological and economic processes that are of importance in describing the fishery are addressed. Ecological characteristics of the fish population and spatial characteristics of the marine environment can be integrated into bioeconomic models to reflect more realistic fish population dynamics. These characteristics include stock density, depensation, discrete time issues, metapopulation dynamics, age-related dispersal of fisheries, and ecological relations among species. Factors that will influence fishing behavior, such as the cost of investment and temporally fluctuating prices, and supply–demand interactions have also been integrated in bioeconomic models. Finally, various management policies have been studied using models, including quota, taxes, and minimum size rules. As will be shown, the strength of these models generally lies in their simplicity. They can be used to describe the rational behavior of profit optimizing fishermen under a wide range of assumptions regarding economic parameters of importance, fish population

1. BACKGROUND AND SUMMARY

7

dynamics, and management policies. Yet, they have their limitations as well as they are rather general in nature and do not capture incomplete information about ecosystem dynamics and fishing behavior at the local or global level. This means that management recommendations regarding the fisheries should not solely be based on these types of models. Chapter 4 considers how stochastic population dynamics have been incorporated in models for the purpose of analyzing marine and fisheries systems and policies. There are several sources of uncertainty that affect fisheries management: uncertainty due to the complexity of the fish population, uncertainty a result of uncertain economic processes, and imperfect information regarding ecological and economic systems. Moreover, these types of uncertainty are generally interrelated. In some cases stochasticity has been represented through economic parameters, but the overriding majority of models encountered address stochasticity by way of population dynamics. This chapter immediately makes apparent the danger of deterministic models used in fisheries management, which arises from the fact that no attention is given to how uncertain fish population dynamics really are. The result of stochastic models focused on fisheries generally is that fish population management should be more conservative. Nevertheless, examples also exist where uncertainty is so large, that it is better to either stop fishing at all, or fish the fish population down to unsustainable levels, depending on ones particular objectives. The final chapter in this first part, Chapter 5, brings together old and recent insights on how to manage fisheries in complex environmental, spatial, and economic settings. It discusses the difficulty of successful fisheries management and highlights the fact that it is not just a simple case where the government adopts a centralized approach to fisheries management telling the fishing community how much to catch. Nor is fisheries management necessarily about incorporating all users into the fisheries management process. Rather fisheries management depends on issues such as formulating a common objective and addressing the economic and ecological spatial and temporal scales at all levels where fisheries management needs to be carried out. Attention is given, among others, to centralized control vs. market interventions by governments, and to managing fisheries by way of co-management. The Common Fisheries Policy in the EU is evaluated against the lessons obtained. One cannot conclude that a centralized management regime will always lead to fisheries mismanagement, nor that collaborative approaches will always give rise to sustainable fisheries. But centralized governmental control has often failed to create sustainable fisheries and healthy socio-economic conditions within the fishing sector, which sometimes was the result of introducing property rights regimes in places where there was little experience with fisheries management in the first place. A long-term management plan having the support of all user-groups involved can generate a common understanding regarding conservation measures in the fisheries and facilitate compliance. It is in the interest of both government agencies and user-groups to create sustainable fisheries and sustainable fishing populations, and to minimize conflicts between all user-groups involved. The degree of power sharing between the government and user-groups depends on the nature of the fishery in question, the amount of uncertainty the fisheries are shrouded in, and the characteristics of the user-groups involved. Finally, fisheries management needs to be

8

1. BACKGROUND AND SUMMARY

formulated addressing the economic and ecological, as well as the temporal and spatial scales of the problem appropriately. Furthermore, responsibilities of all user-groups and government agencies need to be outlined at all levels. The case of fisheries management in the EU has addressed several pitfalls a badly formulated management plan may face. The objectives of the common fishery policy in terms of capacity reduction are not supported or difficult to execute at lower levels, partly due to the objectives in terms of desired fish population levels having been disputed. Furthermore, the issue of scale has been addressed incorrectly. Once targets were agreed upon centrally at the EU level, the national governments were expected to manage their fishing sector nationally, despite the fact that fishermen had the freedom to fish in all European waters. Part II of the book is devoted to bioeconomic modeling. A quite broad typology of models is illustrated, and the result, though certainly not complete, is quite representative of the sorts of models one can encounter in the broad literature on fisheries economics and marine ecosystem studies. Chapter 6 deals with a very fundamental issue in the context of nature policy, namely the optimal conservation and exploitation in a predator-prey system. The fundamental trade-off between conservation and exploitation is central here and made explicit in an intertemporal objective (social welfare) function. The dynamics of the system are represented by a predator-prey system. Predators (birds) and humans (fishers) compete for prey (shellfish). The behavior of the system is studied and conditions for optimal control are deduced. Various optimal harvest strategies are identified for particular ecosystem and economic parameters, such as cases in which birds survive or not. The approach path toward an optimal regime is shown qualitatively for different types of optimal harvest regimes. The type of solution depends on economic parameters, such as the maximum harvest rate, the discount rate, and the cost of fishing, as well as on ecological parameters such as the predator’s search and handling time of prey. After a transition period the system reaches an end state, which is characterized by one of three possible fishing scenarios: not fishing, fishing maximally, or fishing at a singular harvest rate. For some combinations of parameters, several end states can exist. The initial conditions will then determine which end state is optimal. The approach path towards an equilibrium end state will usually consist of some period of bang-bang control followed by a singular harvest rate. Several possibilities are shown graphically. Chapter 7 presents a bioeconomic analysis of a shellfishery. This type of fishery typically has received little attention from economists. Our motivation for this study was determined by the need to deliver insights for the Wadden Sea area, which is characterized by sedentary fish species. A deterministic bioeconomic model of a fishery targeting a sedentary fish population is developed. The model possesses metapopulation characteristics typical of a shellfish population, as well as other sedentary marine populations. Fishing is modeled as having an impact on the shellfish population that goes beyond inflicting immediate fishing mortality through harvesting the resource. Fishing is also assumed to have an impact on the habitat of the fish population, which is delayed. This impact can be negative, for example, due to habitat destruction, or positive due to freeing up space for future generations to recruit to. The policy implication depends on the net effect of these opposing forces. Chapter 8 follows with a discussion of marine reserve creation for sedentary species with uncertain population dynamics. It presents a stochastic bioeconomic sink-source

1. BACKGROUND AND SUMMARY

9

model for a marine reserve based on a shellfishery with metapopulation dynamics. Marine reserves can serve to protect fish species from being overfished. Spill-over effects from a protected reserve to adjacent fishing grounds can improve fish-stock levels and benefit the fishery in the long run (Guichard et al., 2004). The chapter starts out with explaining the reasons for marine reserve creation and addresses the complexity of marine reserve design. One of the concerns is that closing of areas to the fishery could result in fishing efforts being intensified on the remaining fishing grounds. This may undo any potential spill-over effects from the reserve to the fishing ground and can result in increased depletion of the fishing population on the fishing ground. The question is whether the threat of additional closure of areas to the fishery, due to uncertain future requirements regarding fish population conservation, will affect fishing behavior. Whether the fishing community considers it more desirable to preserve additional stock in the source or in the sink depends on which stock is considered to be more valuable, as well as on the population dynamics of the stock in both locations. It turns out that fishermen are willing to close additional areas if this reduces uncertainty regarding future conservation requirements. It is therefore not straightforward to assume that fishing pressure on the open fishing grounds will increase as a result of establishing a marine reserve. Chapter 9 address the important theme of spatial modeling of fisheries in marine ecosystems. Computer models of ecosystems are an invaluable tool in developing management policies of renewable resources, such as fisheries. This chapter presents such a model, which takes the form of a multispecies, multiuser, spatial temporal model of an ecosystem consisting of birds (resembling oystercatchers and eiders), shellfish (resembling mussels and cockles), and (potentially) fisheries. In order to analyze management policies, one first needs to understand the behavior and sensitivities of the underlying ecosystem model. Therefore, the behavior of the ecosystem model without fishing is examined by performing a sensitivity analysis according to a Plackett–Burman design. The behavior of the system is characterized by an equilibrium. When parameters slightly deviate from those in the base case, the system can experience predator-prey cycles with a period of 15–20 years. The sensitivity analysis shows that average population levels of birds are very sensitive to parameters that directly govern the amount of shellfish in the ecosystem in relation to what birds need to survive. These include parameters such as the average mass and survival into adulthood of shellfish and the daily energy intake a bird needs to survive. The parameter that represents the efficiency with which birds choose to emigrate to better feeding grounds is also important. Unfortunately, this is a parameter whose value is not measured accurately in the field. Finally, the model shows that to survive in the field, a bird roughly needs four times as much shellfish – in energy terms – than its pure metabolistic energy requirement. This is due to, among other things, search (in)efficiency. This means that a policy directed towards the conservation of a certain bird population should leave much more shellfish unexploited than the amount that represents the bird population’s energy need. Part III of the book is devoted to monetary valuation and stakeholder analysis, implying that the studies reported here tend to have a more applied character than the ones in the previous part. Here, Chapters 10 and 11 address the case of shellfisheries in the Dutch Wadden Sea, while Chapters 12 and 13 are concerned with the consequences of harmful, invasive species in the North Sea.

10

1. BACKGROUND AND SUMMARY

Chapter 10 opens with a broad, historical in-depth analysis of policy failure and stakeholder dissatisfaction in the context of management of a complex ecosystem, namely the Dutch Wadden Sea. Managing the shellfishery in the Wadden Sea clearly illustrates the problems associated with managing a complex ecosystem with multiple stakeholders. Here, exploitation of shellfish and conservation of birds and nature at wide are at conflict. First the objectives of the various stakeholders at the start of the current Wadden Sea fisheries management plan are assessed, and whether these have been met. Next, it is examined why the co-management arrangement has failed. Factors that are thought to be essential in promoting marine ecosystem sustainability are identified, as well as possible scenarios that have been suggested for the future of the sea and Coastal Fisheries Policy (SCFP) and whether they promote ecosystem sustainability. The main problems encountered within the co-management arrangement are the lack of communication between stakeholders and the lack of a common objective shared by all stakeholders. The latter circumstance has produced an additional problem for the SCFP: disagreement about which policies are appropriate. This has led to a situation in which the government, after the first review of the SCFP in 1998, chose a policy option that could be seen as offering a compromise between the objectives of the fisheries sector and the conservationists, rather than an adaptive policy that would change the fisheries policy according to the objectives set out by the SCFP. A main conclusion is that the threat of intensive shellfishery has been treated without sufficient precaution. Chapter 11 offers a demand-side perspective by performing a stated choice valuation analysis of alternative management scenarios for the Dutch Wadden Sea. Multiple stakeholders are included in the analysis. This is in fact the first valuation study of shellfishery policy in the Dutch Waddenzee. This theme is politically very sensitive and, until recently, appeared very high on the agenda of the Dutch government. The motivation for the study is a structural shortage of shellfish resources for both the cockle fishery and large populations of birds that need cockles as food. The study values different fishery policy scenarios across four groups of stakeholders. The latter include local residents, tourists, policy makers, and natural scientists. As a method, the stated choice experiment was used, because it has the advantage that strategic behavior is minimized, and that trade-offs between and values of attributes can be assessed. The results indicate various differences in preferences among stakeholders. All groups prefer the policy that restricts fishing to half the current fishing area. Tourists and local residents dislike the “extreme scenario” in which fishing is banned. Tourists and local residents favor the “rotation” policy measure, whereas the policymakers, and more so the natural scientists, dislike this measure. The latter believe that this fishery policy may destroy the ecosystem in the Wadden Sea. Finally, all respondents prefer a higher bird population than in the current situation. In addition, the influence of individual motivational profiles is examined, notably the warm glow profile in willingness-topay responses. In fact, the chapter offers one of the first attempts to incorporate a combination of attitudinal information and warm glow in a valuation exercise using the stated choice method. The results confirm the presence of a robust warm glow effect, which turns out to vary across the stakeholders under consideration. The final chapter of the book, Chapter 12, provides insight into the social costs of exotic marine species. For this purpose, travel cost and contingent valuation studies of coastal recreation in the Netherlands have been undertaken. HABs can be responsible

1. BACKGROUND AND SUMMARY

11

for important damages to marine ecosystems, such as the red tides that cause a massive destruction of fish and bottom-living animals. They can also cause thick foams with repelent odors and lead to the discoloration of the beach water, with important damages to beach recreation. This chapter presents a monetary valuation study of a marine protection program. It focuses on the prevention of HABS along the coastline of the Netherlands. It entails the construction of a ballast water disposal treatment in the Rotterdam harbor and the implementation of a monitoring program of the water quality in the open sea along the North-Holland beaches. The valuation study is based on a questionnaire undertaken at Zandvoort, a famous Dutch beach resort. The economic value of the marine protection program includes non-market benefits associated with beach recreation, human health, and marine ecosystem impacts. Both contingent valuation and travel cost methods are used. These valuation techniques have not yet been applied to value HABs damages. The valuation results indicate that the protection program makes sense from an economic perspective as long as its cost is between 225 and 326 million Euro. Taken all together, the studies offer a broad range of tools, applications, and insights which contribute to understanding the interaction between economic activities and ecological functions provided by complex marine ecosystems. In addition, they offer policy guidelines to enhance sustainable management of these ecosystems.

PART I

TOOLS AND BASIC INSIGHTS

13

CHAPTER 2

INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

2.1. INTRODUCTION In the last decades numerous integrated models have been developed that combine elements of the natural sciences with elements of the socio-economic sciences with the aim to formulate policies to effectively and efficiently combat environmental problems. Most integrated modeling studies aim at scenario analysis or policy optimization for environmental problems, such as climate change and acid rain. But many bioeconomic models developed for fishery, forestry, or other renewable resource management purposes also integrate biology or physics with economic behavior. To allow comparison among different areas of application, this chapter will concentrate on models developed not only for fisheries and marine ecosystems but also for the problem of climatic change. Harris (2002) argues that integrated models are necessary to address the major environmental problems that society faces today, such as climatic change, loss of biodiversity, acid rain, and sustainable natural resource management. Bailey (1997) advocates that policymakers must incorporate other knowledge as well for their decisions then just the results of integrated assessment models. Mono-disciplinary insights from economics and ecology should not be neglected. Sophisticated decision makers should use even less formal information from discussion with scientists, policymakers, and stakeholders, to motivate policies. Shackley (1997) notes that the prime function of models in the policy process is to function as a learning device rather than as a truth-machine that predicts the future. In particular, an integrated model can facilitate communication between scientific disciplines and between scientist and policymakers. The remainder of this chapter is organized as follows. Section 2.2 presents the general structure of integrated models, and discusses their advantages and disadvantages. Section 2.3 discusses four themes that play a major role in integrated modeling. Section 2.4, then, describes a number of important characteristics of integrated models in relation to the aforementioned themes. Section 2.5 discusses the trustworthiness of models as a prerequisite for use in policy formulation. A final section concludes.

2.2. CHARACTERISTICS, ADVANTAGES, AND DISADVANTAGES OF INTEGRATED MODELING Integrated models are multi-disciplinary by definition. They describe a causal chain starting with socio-economic drivers like economic production that causes physical or 15

16

PART I: TOOLS AND BASIC INSIGHTS

Profits and damages

Resource use

Socio-economic system

Natural system - greenhouse gas concentrations - air temperature - precipitation

- human population - economic production - capital

Policy response

Policy implementation, mitigation and adaptation

- environmental policy - resource policy

Problem identification and signalling

Figure 2.1. A schematic overview of the interaction and feedbacks between the economic and the natural system as addressed by integrated environmental models.

ecological change through polluting emissions or resource harvesting. A change in a socio-economic variable, such as economic production, is linked to a change in physical variables, such as greenhouse gas concentrations. The state of the natural system alters as a consequence. The state of the natural environment is expressed by endogenous variables such as air temperature and stock size. The changes in the natural system feed back to the socio-economic system. For example, global warming causes changes to the economic system through, among others, a decrease in agricultural production. In the context of fisheries management, catch reduces the fish stock, which can cause an increase in harvest costs. The feedback to the socio-economic system is indirect when it leads to a policy response to mitigate or adapt to the change. In the context of climate change, impacts such as sea level rise and loss of agricultural production can stimulate greenhouse gas emission reduction policies (mitigation) or can lead to building dikes and a change of crops used in agriculture (adaptation). In the context of fisheries, quota and marine reserves can be implemented to avoid overfishing or the extinction of related species. Figure 2.1 sketches the conceptual framework of a general integrated model. It reflects that socio-economic developments force a change in the physical system that feeds back to the economic system either through economic damages and changes or

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

17

through a policy response. Not every integrated model will incorporate all feedbacks of Figure 2.1. Usually, the complete physical and ecological system is not modeled. Instead, models are restricted to the part that is considered relevant for the issue being studied. Sometimes only the (forward) causal chain of effects is described by coupling mono-disciplinary models. The output of an economic model then serves as input to an ecological model. Obviously, many feedbacks are not modeled in this case. Simulating a (wide) range of (policy) scenarios can then create insight into the effects of policy measures or just in how model variables interact. This contributes to a better understanding of the problem at hand, which in turn may help to formulate appropriate solutions. However, the effects of many feedbacks that are included in real integrated models are omitted in this way. An integrated model shows the potential effect of policy measures. In addition, modeling exercises are sometimes able to identify where gaps in knowledge exist. Integrated modeling has the advantage that it systematically combines theories from different disciplines. In order to synthesize all those theories, simplifications are inevitable, which will not always be welcomed by mono-disciplinary scientists (van den Bergh et al., 2004). This is not a problem as long as one realizes that integrated models are not meant to replace, but to complement highly detailed mono-disciplinary models. Integrated modeling forces scientist to be consistent and explicit about theories and assumptions that they use from the different scientific disciplines. Those assumptions need to be communicated well to users to avoid incorrect use of results (Schneider, 1997). Rotmans (1998) mentioned the following advantages of integrated models: they can explore the interactions and feedbacks of the system, they are a consistent framework to structure scientific knowledge, and they are a tool for communicating complex scientific issues to decision makers and other scientists. He also sees some weaknesses: a high level of aggregation, inadequate treatment of uncertainties and accumulation of various types and sources of uncertainty, and limited calibration and validation. Integrated models are used in many stages of the policy process. The use and characteristics of the models vary depending on the stage of the policy process. The former Dutch minister of the environment, Winsemius (1986), identifies four stages in the political process: recognition, policy formulation, policy implementation, and control. Especially in the first two stages, integrated models can play an important role. In the first stage, models can be aimed at clarifying the issue and getting it on the political agenda. In the second stage, integrated models can help searching for effective and efficient policies. In general, one and the same model is not successfully used in both stages. Integrated models that are developed for signaling environmental problem are usually relatively simple. But they are able to project the effects of overfishing or ever increasing greenhouse gas emissions. Models used in the second stage need much more detail and accuracy to analyze what kind of policy measures cure the problem. Van Dalen et al. (2002) analyze in detail the role of models and how they are used in the environmental policy cycle.

2.3. THEMES IN INTEGRATED MODELING Four themes play a major role in the integrated modeling of fisheries and climate change. These themes are: dynamics, space, conflict, and uncertainty. Various chapters

18

PART I: TOOLS AND BASIC INSIGHTS

of this book will focus on one or more of these themes. Here the themes are introduced and briefly discussed.

2.3.1. Dynamics Dynamics points at the changes of variables over time. The problems described by integrated models involve the future effects of current policy decisions and socio-economic behavior. Accumulation of pollutants and delays in the reaction of the biophysical system are common in environmental problems. Irreversibility is another aspect usually associated with environmental problems. Often environmental damages are irreversible – think of the extinction of a species. Adopting a policy to mitigate environmental damages can change society irreversibly as well – think of a specific choice of energy infrastructure to control global warming. Issues such as (the stability of) equilibria, multi-stable states, optimal control, path-dependence, and lock-in are associated with environmental problems. Concern about possible feedbacks is important and one of the reasons to use integrated models. All of these elements demand to be studied in a dynamic context. A comprehensive overview of the importance of dynamics for environmental problems is given by Lines (1995) who discusses aspects such as equilibria, attractors, chaos and catastrophe theory, evolutionary systems, and irreversibilities and the problem of discounting. The models in Part II are explicitly dynamic and focus on optimal control, equilibrium analysis, and delays and feedbacks.

2.3.2. Space Space is a complicating factor that often at first is ignored for reasons of simplicity and later is incorporated to add more realism to the model. Local interactions by definition need a spatial approach. Jansen and de Roos (2000) show the importance of the introduction of space because it can change the complex population dynamics in an ecosystem. From a combined economic and environmental angle, trade and transport are important reasons to adopt a spatial treatment. In the shellfisheries management problem as modeled in Chapters 6 and 7 space is ignored whereas in Chapter 8 and 9 it is added, albeit in different ways. Another example, considering climate change is the famous DICE model (Nordhaus, 1992) that later evolved to the multi-regional RICE (Nordhaus and Yang, 1996). RICE focuses on the regional differences in costs in CO2 -reduction measures. In almost any environmental problem space is important. In fisheries management explicit treatment of space is relevant in order to deal with issues like marine reserves, nursery rooms, and meta-populations.

2.3.3. Conflict Environmental problems are really a conflict between polluters and victims. Economists for this reason use the term (negative) externalities, meaning direct physical effects caused by one agent onto the production or welfare of another. Integrated environmental

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

19

models deal with conflicts either directly or indirectly. Some models start from the notion that the interests of several stakeholders need to be balanced, giving rise to the use of stakeholder participation in integrated assessments (Gough et al., 1997). In integrated models the conflict of interests can be either implicit or explicit. In the case of climate change the conflict is implicit. On one side, we have the present economic interests of those who emit greenhouse gasses, and on the other side those in the future, who are the victim of climatic changes. For instance, we have the interests of the coal mining industry vs. the need for stabilized sea level rise of low-lying island states. Chapter 6 presents an integrated model of the conflict of interest between profitseeking fishers and nature conservationists and provides for analytical solutions to finding the optimal trade-off.

2.3.4. Uncertainty Models can explicitly represent uncertainties by introducing a stochastic term in the model and assuming an associated probability distribution. A distinction can be made between uncertainty and risk. We talk about risk if a probability distribution is known so that the probability of an uncertain event can be calculated. For risk analysis, stochastic models are needed. These models incorporate a probabilistic element. Uncertainty is an integral part of almost every environmental problem. Pindyck (2002) recognizes two types of uncertainty. The first stems from the economic subsystem, i.e. it is impossible to predict the future costs and benefits of environmental damages and policies exactly. In the case of climate change this means that even if we knew the exact temperature change a century from now, we would be unable to know the exact damages this causes in terms of loss in agriculture yields, extra flooding etc. The second type stems from the ecological or physical subsystem. The evolution of an ecosystem, given known policies, cannot be estimated without error. In the context of fisheries management this means that stock levels next year cannot be predicted exactly even if quota and other measures are fully respected. Van Asselt and Rotmans (2002) use a different classification. They characterize two main types of uncertainty. The first is variability, which is an attribute of the system under study. It stems from stochastic processes in nature, such as shellfish recruitment in marine ecosystems and the weather. They can be the result of unpredictability of chaotic systems or result from some fundamentally random process. In socio-economic systems the behavior of economic actor is not completely determined. For example, the outcome of the negations following the Kyoto protocol is impossible to predict. Also technological breakthroughs fall under this heading. We can expect technological development but when and how is impossible to foresee. The second type of uncertainty is limited knowledge. This is a property of the analyst (or the scientific community) analyzing the system. In general, it reflects that we just do not know everything exactly. This can be due to a lack of observations, measurement errors, non-identification of certain processes or effects, or simply regarding them as unimportant. Any integrated environmental model has to deal with uncertainty. There is always some element that is not modeled but changes over time. This exogenous disturbance has some uncertain influence on the system. Integrated models very often use parameters

20

PART I: TOOLS AND BASIC INSIGHTS

and fudge factors whose value is difficult to estimate so that errors remain. The knowledge of the involved processes has often not matured to a level that uncertainties in the structure of the model can be excluded. Because of the broad multi-disciplinary approach very often it is necessary to parameterize processes. This means the formulation of a process is simplified, leaving out detail that creates some unknown error. Some non-linear models can behave chaotic. In chaotic systems, we have the fundamental problem of not knowing exactly the initial conditions. Measurement error, however small, will make the system unpredictable. Various conceptual and model techniques allow addressing uncertainty. Rotmans and van Asselt (2001) advocate a pluralistic type of uncertainty analysis whereby a specific worldview underlies the choice of models to assess the most salient uncertainties. Chapter 4 discusses specific approaches to dealing with uncertainty in integrated models of fisheries and marine ecosystems.

2.4. IMPORTANT CHARACTERISTICS OF INTEGRATED MODELS Here, we will discuss important characteristics of integrated models that relate to the four themes of the previous section. These characteristics are formulated in a dichotomous manner: r r r r r r

Theoretical vs. empirical, Prescriptive vs. descriptive, Optimization vs. resilience, Simple vs. complicated, Analytical vs. numerical, A focus on economics vs. on natural sciences.

Table 2.1 shows the degree to which these characteristics play a role in the aforementioned themes. The table must be interpreted as follows, “+”/(“−”) denotes that models possessing the characteristic are often (rarely) used to analyze explicitly the specific theme.

2.4.1. Theoretical vs. empirical models Theoretical models are relatively simple, general representations of reality. The aim is to capture just enough detail to be able to make general statements. Often analytical results can be obtained. The structure of the models is motivated by a theoretical concept. Examples are carrying capacity in ecological models or rational profit maximizing behavior in economic models. Because of the models’ simplicity results are difficult to implement directly. Theoretical models aim at a general understanding of the problem and are hoped to stimulate creativity when in a real-life system policies must be devised. The advantage of theoretical models is that their concepts can easily be understood and communicated to policymakers. Nevertheless, theoretical models need to be confronted with data to become trustworthy.

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

21

Table 2.1. Degree of importance of model characteristics in the themes

Theoretical Empirical Prescriptive Descriptive Simple Complicated Analytical Numerical Economics Natural sciences

Dynamics

Space

Conflict

Uncertainty

+ + + + + + + + + +

+ + + + − + − + + +

+ − + − + − + + + −

+ + + + + + + + + +

Empirical models on the contrary are specific and actually “filled with” data. Parameters are estimated and the simulations of the model are ideally compared with observations. This constitutes a test for the credibility of the model. These models are much more geared toward a specific problem. Therefore their solutions are easier to implement directly. For example, empirical models may regard the specific problem of overfishing of mussels in the Dutch Wadden Sea instead of the general problem of economic overfishing in a theoretical model. Usually the level of aggregation is much higher in theoretical models. This has the advantage that results of a more general nature are obtained. At the same it has the disadvantage that the assumptions are not in line with all specific cases. For example, the parameters in the ecological model in Chapter 9 are within the range of empirical values found in the scientific literature. Nevertheless, it is very unlikely that the modeled area is fully consistent with any particular, real estuary.

2.4.2. Prescriptive vs. descriptive models A modeling approach is normative or prescriptive if it answers the question: What should policymakers do in order to reach a particular goal? The preferences of the policymaker must be explicitly modeled then. The method is goal oriented and works well when there is a clear single criterion that needs to be maximized (Makowski, 2000). Profits are an example of such a clear single criterion. The traditional economic approach in resource exploitation is a prescriptive method that searches for optimal management strategies (see Chapter 3). This is the approach of Chapters 6–8 in Part II of the book. Within the modeled world the prescribed solution is the best solution given the modeled preferences. However, the real world will differ from the model and when the solution is sensitive to parameter or structural errors a far from optimal policy may be the result. Clark (1990 – original version 1976) is a classic book that describes the prescriptive framework for bioeconomic models of fishery management.

22

PART I: TOOLS AND BASIC INSIGHTS

The descriptive approach, which is mainstream in the natural sciences, models the systems behavior without explicitly referring to some best policy. The interactions of the system variables are modeled; the preferences of the policymaker are not. Policy measures are regarded as input with which one can experiment. The modeling approach answers questions of the “what . . . if ” type, such as: “What happens to global warming if the Kyoto protocol is ratified and is successfully complied with?” or “What happens to the eider population if the cockle fishery is banned?” Descriptive models can be seen as performing policy evaluation and prescriptive models as performing policy optimization. Optimization is complex and computationally expensive. Hence, prescriptive models incline to describe the economic and natural science part of the model in relatively simple terms (Kelly and Kolstad, 1999). Descriptive models tend to become more prescriptive if a systematic testing and evaluation of many different scenarios is applied. Typically, optimization will start with some base policy, evaluate it and systematically improve it until no better solution can be found. Obviously, the optimization result relies heavily on the underlying model. That model is often simple because of the computational effort involved. Therefore, it is a fruitful approach to evaluate optimal policies found by relatively simple models with more complex descriptive models. In general, a better understanding of environmental problems and their solutions can be realized if potential policies are evaluated with several models that differ in terms of degree of complexity and detail of modeled processes.

2.4.3. Optimization vs. resilience Because of the usually somewhat simpler description of the dynamics in optimization models, they generally focus on an optimal equilibrium. However, especially ecological systems can experience complex behavior characterized by multiple equilibria or chaotic attractors and sudden bifurcations from one seemingly steady state into another (Holling, 1973, Janssen et al., 2004). Systems that exhibit this sort of behavior are in need of control policies that do not aim at optimization but aim at avoiding catastrophic system change i.e. by keeping the system resilient. The concept of resilience has two main variants (Perrings, 1998). The first is due to (Pimm, 1984) and is concerned with the time taken for a system to return to the initial state after a disturbance. The second is subscribed to (Holling, 1973) and focuses on the magnitude of a disturbance before a system changes (irreversibly) from one state into another. Both concepts are useful when looking at the interaction of an economic and ecological system. It is important to know how far a system can be pushed without flipping to some undesired state but it is also useful to know how much time it takes for a disturbed system to return to its desired optimum. Both approaches assume some sort of stable state. Ludwig et al. (1993) refine concepts of stability and resilience for exploited ecosystems. They consider systems with multiple equilibria, some stable, some not, and describe the dynamics of such a system. Such a system can very quickly move from one stable equilibrium to another when one variable is slowly increasing or decreasing

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

23

and surpasses a threshold value. Very often this trajectory is irreversible. In order to keep the system sustainable, managers must control the position of the system in statespace close enough to the desired attractor. But also far enough away from the unstable area where the system may suddenly and irreversibly change to an undesirable new equilibrium or attractor. As also noted by Walters (1986) and Carpenter et al. (1999), in not fully known or stochastic systems sustainability cannot be achieved by seeking a fixed policy. In order to keep the system in the desired domain it may be necessary to experiment and to continually adjust ones policy. In order to find the right adjustments, the manager has to explore and learn about the system. Walters (1986) calls this adaptive management. However, these explorations may move the system out of the desired domain and therefore require careful consideration.

2.4.4. Simple vs. complicated models In the first stage of the policy process, i.e. signaling the problem, models can be relatively simple. It is important to understand what happens, why this is so, and to communicate this to policymakers. Simple models are easy to interpret and understand. After the signaling process those who benefit from the status quo quite often argue that the problem may not be so serious because many processes have been modeled far too simple. In reaction, modelers will incorporate more processes and detail. Even though every newly incorporated process potentially improves the model and can deliver new insights, simple models often provide similar results as to the effect of policy measures. Nevertheless, complicated models can show surprising effects of feedbacks that are otherwise ignored Rotmans (1998) argues that a distinction must be made between complicated models and complex models. Complicated models incorporate a large variety of processes and detail but do not generate complex behavior. The latter is characterized by strange attractors, multiple stable states, irreversibilities, and bifurcations. If incremental changes in driving forces in complicated models lead to incremental changes in the simulated projections the models behave pseudo-linear and should not be regarded as complex at all. Complex models are non-linear by definition and may contain few processes, but are characterized by certain incremental change causing a considerable change in output. More complicated models may help to improve understanding of a certain problem. They sometimes make it possible to simulate more accurately potential policies because processes are less aggregated. But more complicated models will not necessarily produce better forecasts. Adding more processes means more parameters and consequently more room for error in parameter estimation or in model structure. Much more effort is needed to build and test a more complicated model, which does not always translate to better projections. Models should therefore be kept as simple as possible. To quote (Dowlatabadi, 1995): “Keep it as simple as possible, iterate on details only where needed to address policy relevant questions and permitted by available knowledge.” Nevertheless, it is still difficult to know when to stop adding detail to a model.

24

PART I: TOOLS AND BASIC INSIGHTS

2.4.5. Analytical vs. numerical models If models are simple, i.e. contain few equations that are not too non-linear, analytical solutions can sometimes be derived. Within the neo-classical optimizing tradition in economics analytical results are frequently obtained from theoretical models (see Chapter 3). In other cases models are usually too complicated, i.e. with too many equations and very non-linear. Then a translation of the model into computer code followed by using numerical simulation techniques is needed to generate results. A disadvantage of complex models is that solutions are very likely to contain errors due to computer coding or solving the model. Moreover, solutions only apply to the specific values of the parameters for which the model is being solved. The main advantage is that in order to be able to find solutions, one does not have to simplify the problem so much that agreement with reality becomes disputable. Furthermore, it is usually relatively simple to extend the model slightly so as to incorporate more detail in response to additional insight or advice from others. Analytical results are stronger in the sense that the results are general and independent of specific numerical values of parameters. They solely depend on specified functional relationships between variables. Unfortunately, analytical results can only be obtained from relatively simple models. Most models that resemble reality accurately become quickly too complicated to be able to derive analytical solutions.

2.4.6. A focus on economics vs. on natural sciences Some integrated models place relatively much emphasis on the economic part of the model, while others stress the natural science part. Most models focused on economics are general equilibrium models or bioeconomic resource extraction models based on neo-classical approaches. Often these models are prescriptive. The models that are focused on the natural sciences usually are dynamic system models describing the processes based in physics or biology. Where the focus is depends on the question one tries to answer with the model. MERGE (Manne and Richels, 1992), for example, is a general equilibrium model with a relatively complex economy and a simple climate part. Its aim is to find costeffective emission strategies. DIALOGUE (Visser et al., 2000) is a model that focuses on the natural sciences part of the climate change problem. The economic dimension of climate change is modeled in an extremely simple way. Its aim is to illustrate the physical impacts of climate change, such as sea level rise, given greenhouse gas emission scenarios.

2.5. TRUST IN AND CALIBRATION AND VALIDATION OF MODELS Models have no meaning unless they are credible or trustworthy. Trust in a model can only be achieved if the user shares the assumptions and structure represented by the mathematical equations. Ideally, these should be completely transparent when model outcomes are presented.

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

25

For numerical models parameter values need to be estimated. Many parameter values can be obtained from experiments or measurements. But very often container parameters remain that describe processes at a very aggregated scale. Ideally, these parameters must be tuned or calibrated, for example, to adequately (possibly qualitatively) replicate a specific historical pattern. The accuracy of parameter estimation (calibration) will be limited by the amount and quality of the available data. The practise of calibration and validation depends on the scientific discipline. In the natural sciences, models are usually loosely tuned or calibrated whereas economists and econometricians estimate parameters more systematically. This is not so because economics forms a far more rigorous and thorough scientific discipline but can be attributed to the more complicated and complex models that are widely used in the natural sciences. Regression techniques that are generally used by econometricians are often not applicable to these kinds of models. Of course systematic calibration approaches are advocated and used in the natural sciences (Heemink, 1991; Janssen and Heuberger, 1995; Spear, 1997; Finley et al., 1998). The trustworthiness of the model is further established by validating or testing it against another historic data set than the one used for calibration (Refsgaard and Hendriksen, 2004). The accuracy of computer model calculations is usually examined by plotting model calculations and observations in the same graph. The match between them is evaluated by inspection. This is a subjective approach without a firm statistical fundament but can be very informative nevertheless. Many statistical techniques, such as residual analysis and goodness-of-fit tests can be applied. However, a model validated in this way does not imply that the simulation results are an accurate prediction (given exogenous future developments) but only that the results are somewhat more certain and trustworthy than those from a model that is not validated against observations. Conducting validation as well as calibration puts a large demand on data. It is usually hampered by the imbalance between the complexity of the model and the availability of the data. Often, long-time series are needed that have not been measured in the past. Unfortunately if data exists it often addresses a different temporal or spatial scale than those of the model results. In many circumstances adequate data is not available and the modeler has to deal with what he can get. Large climate models (GCM) and large general equilibrium models suffer from this lack of data and are usually not thoroughly validated. Many smaller models are not validated either. Not only must the time series be long, but also the data needs to be at the right scale. Model output should be compared with empirical data at the scale of the lowest resolved model elements (Schneider, 1997) In the literature the meaning of the notion of verification and validation is confusing. Validation loosely means that some test is performed – using observations – that contributes to trust in the model. Oreskes et al. (1994) make a case for using the right semantics. They claim that models of natural systems are impossible to verify and validate due to the openness of such systems. A system is considered closed if its true conditions can be predicted exactly, such as in mathematics. Systems where the true behavior cannot be computed due to uncertainties in parameter values and boundary conditions are open. Environmental systems represented through integrated models are open. For instance, the only way to verify whether the population size of

26

PART I: TOOLS AND BASIC INSIGHTS

oystercatchers projected by the model of Chapter 9 is accurate is to stop fishing and measure the population size for many years. Since all relevant integrated models describe open systems, the validation that any model truly represents the real world system is impossible. This applies equally to economic models (Sterman et al., 1994). Note that from a philosophy of science viewpoint models can only be invalidated or refuted, because models are no more than the mathematical and computer coded representation of scientific hypotheses (Popper, 1959). In addition, there is the associated problem of whether the computer code is a true representation (within accuracy bounds) of the conceptual model expressed as a system of mathematical equations. Pragmatically, this means that one has to ascertain that the code contains no bugs. Tests to ensure this are, for example, that the model must simulate an analytically known or trivial solution. In practise one can never be sure of errorless code. It is almost sure that many large complicated models contain programming errors. Therefore counterintuitive results should be looked at sceptically. A useful and pragmatic approach to validation of models is to confront model output with measurement data as well as expert judgement and sensitivity analysis. Janssen and Heuberger (1995) discern three major aspects (see also Schneider, 1997): 1. Assessment of the ability of the model to reproduce the real world system behavior – does the model output resemble observations and measurement data? 2. Assessment of the suitability of the model for the intended use – does the model answer the research or policy question? 3. Assessment of the robustness or sensitivity of the model – does the assessment of the problem and its various solutions differ greatly when parameters have extreme but acceptable values? Finally, for a proper use of models in the policy arena Parson (1995), Mensink (2000) and Harris (2002) provide the advice that a model should address the decision power and the preferences of the policymaker for which it is intended. So a model that computes the optimal global greenhouse gas emissions is of no use for a policymaker that must decide about national emission reductions but can be useful for an international negotiator, provided the model correctly assumes his preferences. Or to take an example in fisheries management: if a regulator cannot influence the effort exerted by fishers then a model that maximizes national welfare by prescribing optimal effort is useless. The model should perhaps instead describe optimal tax or quota. In other words, the model must describe measures that the decision maker can control and implement.

2.6. CONCLUSION The great advantage of integrated modeling is that multi-disciplinary knowledge is structured consistently. Integrated models help to describe the logical consequence of assumptions, data, and the described knowledge of processes. Thus, models can explore the behavior of complex systems more reliably than mental models, provided the assumptions are well-understood. A model should be aimed at incorporating decision

2. INTEGRATED ASSESSMENT OF MARINE ECOSYSTEM EXPLOITATION

27

variables at the level of the policymaker for whom the model is developed. This means that modeling should focus on the questions policymakers have, may have or should have depending on the stage of the problem in the political cycle. In the recognition phase, a model is a learning device that creates insight into the problem. Models used in the policy formulation stage must explicitly be able to evaluate the effect of potential policies and must be directed at the power of the policymaker. To prevent misconceived use of their model, modelers must be well aware of their implicit assumptions and make these transparent. Uncertainties play a key role and must be communicated to decision makers. Finally, very often the environmental problem is composed of several subproblems that raise different research questions. A set of multiple models geared to these specific questions generally creates more insight into the problem and provides more incentive for creative solutions than one comprehensive model.

CHAPTER 3

DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY

3.1. INTRODUCTION1 A vast amount of literature is devoted to integrated analysis of management strategies to control fishing industries and to sustain the stock of exploited fish populations. The rise in interest in managing the fisheries is due to a globally perceived decrease in ocean and sea productivity as a result of overfishing and mismanagement of the fishing resources. An increase in fishing power due to improved technologies, a rise in world population, and a lack of knowledge about the characteristics of the exploited fish species, have all contributed to fish populations world-wide having become or running the risk of becoming eroded. Following the influential papers by Gordon (1954) and Schaefer (1954), a wide array of more or less integrated models possessing both biological and economic characteristics has been created to analyze the management and policy of fisheries, providing a valuable picture of how the economic and environmental characteristics of fisheries interact. Although several of these models have been used empirically to model existing fisheries, the purpose of this chapter is not to compare their empirical strength, since this would be virtually impossible for two reasons. First, applied models usually highlight different aspects of the fishery to be managed. Second, separate fisheries modeled are very difficult to compare due to their distinct nature. Rather, the purpose of this chapter is to address the contributions of various models to describe and understand a range of economic, biological, and interactive processes that can potentially occur in a fishery and the way these can be managed. The bioeconomic models addressed in this overview look at the fishery as possessing one of two kinds of ownership characteristics: open-access and sole-ownership. When considering the fishery as an open-access resource, there is no restriction on the number of fishermen entering the fishery and the extent to which they are allowed to exploit the fishery. Sole-ownership means that the fishery is considered to be “private property” and managed by a sole-owner. For a more subtle analysis of ownership and property rights in the fisheries, covering also common property regimes, see Hanna (1997). Other approaches to modeling fisheries have been employed. Although these are beyond the scope of this survey, they are certainly valuable to promote an understanding of complex marine-economy interactions. In particular uncertainty and evolutionary 1

Reprinted from R.J. Imeson, J.C.J.M. van den Bergh and J. Hoekstra (2002). Integrated models of fisheries management and policy. Environmental Modelling and Assessment 7(4):259–271.

29

30

PART I: TOOLS AND BASIC INSIGHTS

modeling are relevant in this context. Walters (1986), for example, has adopted an uncertainty oriented “adaptive management approach” in which management problems are bounded by explicit and hidden objectives and practical constraints on action. In this approach, models are used to make dynamic behavior explicit so that learning and improvement can take place. In addition, evolutionary modeling oriented toward multidisciplinary and integrated research can generate information that is complementary to insights from deterministic and uncertainty approaches (Allen and McGlade, 1987; Munro, 1997). The organization of the remainder of this chapter is as follows. In Section 3.2 the most basic open-access and sole-owner model and their optimal solution will be discussed. In the subsequent sections additions to or changes in the basic model and how they affect the models and their optimal solution will be studied. Model elements that take fishery dynamics into consideration are discussed in Section 3.3. Model elements that take economic behavior into consideration are considered in Section 3.4. Model elements relating to management policy are dealt with in Section 3.5. Section 3.6 concludes.

3.2. BASIC FISHERY MODELS 3.2.1. The open-access and sole-owner model One of the simplest models in the economic theory of the fisheries was developed by Gordon (1954) and is commonly known as the Gordon–Schaefer model. This model deals with finding a static equilibrium solution in an open-access fishery and has served as the basis of many more bioeconomic models. An open-access fishery is defined as a fishery that is completely unregulated and can be exploited by anyone wishing to do so. The profit provided by the open-access fishery resource fished upon by the fishermen is given by: πt = TRt − TCt = pCt − cE t

(1)

where tdenotes the time, TRt the total revenue of fishing, TCt the total costs of fishing, E t the amount of fishing effort exerted by the fishermen, Ct the catch, p the price of effort, and c the cost of effort exerted. p and c are assumed to be constant. Catch is expressed as: Ct = q E t X t

(2)

where at time t, q denotes the catchability coefficient andX the size of the fish population. The term q E is also referred to as the fishing mortality rate. The dynamics of the population size is determined by: dX t = G(X t ) − Ct (3) dt where G (X ) denotes the natural growth-rate of the population (In the remainder of this chapter the time index t will be omitted unless necessary). The growth curve commonly used in the Gordon–Schaefer model is:   X G(X ) = r X 1 − (4) K

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 31 (A)

TSC

TC

TR

A MSY C B

Effort

(B)

A MSY

B

Effort

Figure 3.1. (A) The Total Revenue (TR) and Total cost (TC) curve. (B) Relationship between effort and fish population. B, bioeconomic equilibrium; A, maximum sustainable profit, and MSY, maximum sustainable yield.

where K is the maximum carrying capacity of the population under consideration, and r the growth coefficient. Setting dX t /dt equal to 0 and graphing out catch in relation to effort exerted, a static yield-effort curve can be derived which describes the following relationship between sustainable yield and effort at equilibrium:  q  C = qEK 1 − E (5) r

In an open-access, fishery effort is expected to move to equilibrium where the economic forces affecting fishermen and the biological productivity of the resources are in balance. This equilibrium is generally known as the bioeconomic equilibrium, as it is determined by both biological and economic parameters, and is defined by the economic condition π = 0. When exerting an effort level B, in Figure 3.1A, the bioeconomic equilibrium is the point where the TR parabolic figure intersects the TC function. Note that TR is parabolic since it is dependent on catch, which has a parabolic relationship with effort as determined by Equation (5).

32

PART I: TOOLS AND BASIC INSIGHTS

No level of effort higher or lower than the one exerted at bioeconomic equilibrium can be maintained. In the case of a higher effort exerted, total costs of fishing would exceed total revenues, forcing some fishermen to withdraw from the fishery and reducing the level of effort back to bioeconomic equilibrium. In the case of lower effort exerted, additional fishermen would be attracted by the profit they can earn and effort would subsequently increase. The maximum sustainable yield (MSY) occurs when the fish-population growthrate reaches a maximum. It is the maximum that can be caught on a sustainable level without reducing the long-term stock, and is obtained by exerting that level of effort at which total sustainable revenue is maximized. Clearly any bioeconomic equilibrium point to the right of the MSY is inefficient. More fish can be caught against lower cost when exerting a level of effort left of the MSY, because the fish population is larger (see Figure 3.1B). In some cases the aim of fishery regulation has been to achieve MSY by measures such as increasing the cost of fishing or shortening the fishing season. Economists have argued, however, that the policy to achieve MSY is generally economically inefficient when taking into account that the private costs of fishing for each fisherman is influenced by other fishermen and the specific fishing conditions (Plourde, 1970; Brown, 1974). Private costs to each fisherman can increase as a result of, for example, vessel congestion on the fishing grounds and, thus, increased vessel operating costs to each fisherman. These private costs lead to an uncompensated loss of welfare to the fishermen. The total social cost (TSC) of fishing is the costs of fishing taking into consideration the private costs to the fishermen. In Figure 3.1A, the TSC of fishing yields the social bionomic equilibrium C, where the fishermen will exert a lower level of effort. Note that point C lies closer to the MSY than point B does. It is clear that not taking private costs of fishing into account, fishery regulation aimed at achieving MSY may overshoot its mark and lead to an inefficient bionomic equilibrium to the left of MSY. In an open-access setting all participating fishermen earn zero economic profits. If, however, property rights were allocated to a sole-owner who is a profit maximizer, be it a single fisherman or a government managing body, positive economic profits would accrue from the fishery. Surprisingly the sole-owner will not catch as many fish as he can sustainably do but will fish less (point A) than at the MSY level. In fact, for any growth curve of a shape similar to the one in the Gordon–Schaefer model and for any non-negative linear cost function of effort he will never fish beyond the MSY level. In an intertemporal setting the sole-owner maximizes the present value (PV) of his stream of profits accrued from the fishery over a certain time horizon subject to Equation (3). Price p, effort cost c, and catchability q are again constant and set. The PV is given by:

PV =

T

πt e−δt dt

(6)

0

with πt as given by Equation (1), T denoting the time horizon, and δ the discount rate. The optimal fish population level X ∗ that maximizes the present value of the fishery is derived by applying the maximum principle where Equation (6) is the profit function to

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 33 be maximized subject to Equation (3) (Clark, 1976). Using this approach X ∗ is given by the formula: G ′ (X ∗ ) −

c′ (X ∗ )G(X ∗ ) =δ p − c(X ∗ )

(7)

where, c(X ) =

c qX

(8)

Equation (7) is also known as the Golden Rule equation. Note that c(X ) in Equation (8) is the unit harvesting cost when the population level is X . This can be derived from Equation (2), using the fishing costs cE as defined in Equation (1), and setting catch C equal to 1. The optimal effort level to be applied now follows the so-called “bang-bang control”, defined as: ⎧ E if X t > X ∗ ⎪ ⎪ max ∗ ⎨ G(X ) if X t = X ∗ E t∗ (9) ∗ ⎪ q X ⎪ ⎩ 0 if X t < X ∗

which means that the optimal steady-state X ∗ as determined from Equation (7), should be attained as fast as possible and effort should be directed, as defined above, to reach and stay at that level (Clark, 1990). The occurrence of a “bang-bang control” is due to the fact that this model constitutes a linear optimization model (because the resulting Hamiltonian is linear in effort). If the price would instead of being constant vary with catch we would have a non-linear optimization model, which is much harder to solve for an optimal fishing policy. In such a case, the optimal approach to the steady-state population level would reflect a more gradual transition than a “bang-bang control” (see, for example, Bulte and van Kooten, 1999; Horan and Shortle, 1999).

3.2.2. Interpretation of the Golden Rule for the sole-owner model To gain a better understanding into the dynamics of the Golden Rule equation, it is useful to compare the optimal steady-state X ∗ with the static solution given by point A in Figure 3.1A. Equation (7) shows that the optimal fish population level varies with the discount rate δ and thus is not necessarily identical to point A. The reason is that when the value of fish caught in the initial (non-equilibrium) state is higher than the discounted value of its offspring, it is worthwhile to catch more fish than in point A. When the discount rate is zero, the future is as valuable as the present and the optimal fish population level X ∗ coincides with point A. This can be seen by noting that Equation (7) then becomes: ( p − c (X ∗ )) G ′ (X ∗ ) − c′ (X ∗ ) G (X ∗ ) = 0

(10)

34

PART I: TOOLS AND BASIC INSIGHTS

which can be rewritten as: d {( p − c(X ∗ ))G(X ∗ )} = 0 (11) dX So X ∗ is the solution to finding the maximum of ( p − c(X ∗ )) G(X ∗ ), which is equivalent to maximizing the difference between the total revenue curve and the total cost line in Figure 3.1A, which occurs at point A. To see what happens when the discount rate approaches infinity, which is equivalent to placing no value on future catch, we rewrite Equation (7) as (Pearce and Turner, 1990): 1 d{ p − c(X ∗ )G(X ∗ )} · = p − c(X ∗ ) (12) δ dX Equation (12) now reduces to p = c(X ∗ ). This situation is equivalent to total revenue being equal to total cost. Profit is fully eroded, and the corresponding optimal population level is equal to point B in Figure 3.1A, which would be reached in an open-access situation. A situation of open-access is thus identical to an extreme case of myopic behavior. Equation (12) helps us to interpret the meaning of Equation (7) more clearly: If we reduce the stock by harvesting a small amount X , there will be an immediate gain equal to the right side of the Equation. There will, however, be a loss of future sustainable profit ( p − c(X )G(X ), the present value of which is equal to the left side of the equation [Note that c(X ) is a decreasing function ofX ]. To promote better understanding of the results this section will continue to illustrate the relationship between the Golden Rule equation and two well-known rules for optimality utilizing a resource with costless extraction. In the case where prices are not constant over time, which will be discussed further in Section 3.4.1, Equation (7) is replaced by: c′ (X t∗ )G(X t∗ ) d p/dt dG + = δt − ∗ dX t pt − c(X t ) pt − c(X t∗ )

(13)

The extra term on the right is the change of the value of the asset. When the unit cost of effort is assumed to be constant, the equation reduces to: dG d p/dt + = δt dX t pt

(14)

This is a well-known formula in capital theory stating that the marginal productivity of the resource plus the capital gain should equal the discount rate. A special case of this formula results when the growth of the resource is taken to be zero. In this case, Equation (14) is transformed into the so-called Hotelling rule (Hotelling, 1931), the fundamental equation of non-renewable resource exploitation: d p/dt = δt pt

(15)

The interpretation of this equation is that the speed of resource exploitation should be such that the price of the (non-renewable) resource in question changes at a rate equal to the discount rate.

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 35

3.3. MODEL ELEMENTS RELATING TO FISHERY DYNAMICS 3.3.1. Stock density In many fisheries fishermen will not encounter the same amount of fish at all locations. This means that the catch should be regarded as being dependent on the density of the fish population. This density dependency can be reflected in a variable catchability coefficient q in Equation (2) (Clark, 1985): q(X ) =

θρ(X ) X

(16)

where ρ(X ) ≥ 0. The concentration profile ρ(X ) represents the maximum density of the fish population observed in a certain region where fishing takes place. It is a non-decreasing function of X , meaning that as the fish population is depleted the concentration of the fish in the region observed will go down or remain constant. θ is a constant factor detailing the effort exerted by the fishing gear used by the fishermen over a certain region. The general assumption is that areas with the highest concentration profiles are depleted first (Clark, 1985): fishermen want to maximize their catch rates. So generally, for most fish species, if X decreases so will ρ(X ), as shown in Figure 3.2. Knowing how ρ(X ) behaves as a fish population X decreases is very important. The standard Gordon–Schaefer model assumes a linear relationship between the two and therefore the catchability coefficient used in Equation (2) is a constant. This is clearly a unique situation: not many species of fish have such a concentration profile. A concave or constant concentration profile (a horizontal line) as X decreases can be very treacherous: the fish concentration at a location does not change much as the fish population is decreased but then suddenly drops. For a constant concentration profile the fishermen would not notice that the fish population is decreasing: they perceive an unchanging fish concentration up until the moment the fish population is wiped out to extinction, making it impossible to take corrective action. Although this situation

Concentration

Stock X

Figure 3.2. Concentration decreases as the fish population X decreases.

36

PART I: TOOLS AND BASIC INSIGHTS

seems somewhat unrealistic, schooling fish are known to have a concentration profile that is practically horizontal up to a very low and critical level of population. Deriving the bioeconomic equilibrium X for the Gordon–Schaefer model in the open-access situation with condition 16, it is found to occur at the fish population level where: c ρ(X ) = (17) θp The zero-profit condition in an open-access situation with or without Equation (16) remains the same. The open-access situation discussed in Section 3.2.1, however, has a profit function independent of the catchability coefficient. This in contrast to the open-access situation with Equation (16), which has a profit function dependent on the catchability coefficient and thus dependent on the concentration profile of the fish encountered. If the fish concentration is so low that ρ(X ) < c/θ p, fishing is not worthwhile because total revenues exceed the total cost of effort exerted by the fishermen.

3.3.2. Depensation The natural growth curve of the fish population in the Gordon–Schaefer model is taken as being convex over all values of X . Nevertheless this does not have to be the case. The situation depicted in Figure 3.3 shows a curve with a negative growth-rate, A, turning positive beyond a critical threshold level of the fish biomass. This model is said to show critical depensation, meaning that when the fish population has been fished down to a certain level it is doomed for extinction. Non-critical depensation as in curve B Growth rate

A B

Biomass X

Figure 3.3. Critical depensation (A) and non-critical depensation (B).

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 37 occurs when the growth rate remains positive, initially increasing and then decreasing. The effect is that when a fish population is fished down to a certain level it can take a long time before the stock is back to its original level. In fact, when the population falls below a certain level it will be driven to extinction unless fishing effort is severely reduced. It can be shown that depensation (in some case critical depensation) is most likely to occur for fish showing concentration profiles (see above) that are concave or constant over the biomass (X ) range. In a critical depensation model under open-access or sole-ownership exploitation it is possible that the fish population level will be driven to extinction. All that is required is that the optimal population level is below the minimum viable population size. The possibility of extinction due to critical depensation is aggravated by the possibility of a sudden collapse related to causes exogenous to fishing, for example, a severe storm. Amundsen and Bjorndal (1999) have modeled the effect a possible exogenous sudden collapse of the fish population level has on optimal management policies. They considered both a fishing policy that allows for extinction and one that does not. Maximizing net present value of a fishery before and after a potential collapse, they show that possible extinction after a collapse due to critical depensation can well be part of an optimal harvesting scheme. This is the case if the present value is largest in a fishing scenario that allows for extinction, which depends on factors such as the size of the collapse of the population and the possibility of a collapse. Their Golden Rule equation applicable to a fishery where the population will be driven to extinction after a collapse is given by Equation (18): G ′ (X ∗ ) −

c′ (X ∗ )G(X ∗ ) =δ+λ p − c(X ∗ )

(18)

where λ is the rate of collapse. It is easy to show that the left-hand side is a negative function of the stock. Therefore, the higher the rate of collapse, the lower will the optimal steady-state fish population be. Consider the case where λ is equal to 0. In this case Equation (18) reduces to the standard Golden Rule equation given by Equation (7). If λ is equal to infinity, it is optimal to fish everything down to the point where profit is fully eroded. The future of the fishery in this case has no value, which is as can be expected when certain doom of the fish population is imminent.

3.3.3. Discrete time growth models The response of the fish population to external forces is not always instantaneous. This is the case, for example, when there is a delay in recruitment, i.e. when it takes a number of years for the young fish to become part of the adult fish stock that can be exploited. For such situations continuous time models as above can be replaced by discrete time models, which offer a much simpler way to model periodic and seasonal fluctuations. If we consider recruitment X k+1 from year k to year k + 1 then the discrete time version of Equation (3) becomes: X k+1 = σ (X k − Ck ) + G(X k − Ck ) = σ (Sk ) + G(Sk )

(19)

38

PART I: TOOLS AND BASIC INSIGHTS

where Ck and Sk denote catch and escapement in year k, respectively, and σ the natural survival rate [so σ (Sk ) represents surviving adults from year k]. The escapement is a general term used in fishery terminology to describe the remainder of the stock after fishing has taken place. Equation (19) can be seen as a combination of two elements. The first, σ (Sk ), can be seen as a change in the current stock due to natural and fishing mortality. The second, G(Sk ), is a function describing the relationship between stock in year k and recruitment in year k + 1. Note that G(·) does not have the same interpretation as in Section 3.2 where it was a function reflecting the growth of the whole fish population. Ck is bounded by: 0 ≤ Ck ≤ X k

(20)

When the fishery targets a fish population aged n and onwards, fish recruitment is said to take place with a delay of n years (after spawning of the adult fish), where n is the age at which the young fish matures. This form of recruitment is commonly known as delayed recruitment (recruitment described by Equation (4) is known as direct recruitment). In this case Equation (19) changes to: X k+1 = σ (Sk ) + G(Sk−n ) A discrete form of the net present value of the revenues is as follows: k ∞   1 π (X k , Ck ) 1+δ k=0

(21)

(22)

where δ denotes annual discount rate. If we consider the delayed recruitment case, maximizing 22 subject to 20 and 21 gives the following discrete version of the Golden Rule Equation (7) from which the optimal values C ∗ and S ∗ can be derived (Clark, 1990):

n  dπ/dX + dπ/dC 1 ′ ∗ =1+δ (23) G (S ) · σ+ 1+δ dπ/dC

It can be shown that the optimal approach to the steady-state escapement level S ∗ is not a “bang-bang” approach, due to the fact that the optimization model misses a certain linearity. By taking n = 0 and σ (Sk ) = 0, Equation (23) reduces to an equation determining the optimal steady-state values for C ∗ and S ∗ for the direct recruitment model: G ′ (S ∗ ) ·

dπ/dX + dπ/dC =1+δ dπ/dC

(24)

By setting σ (Sk ) equal to zero, there are no surviving adults from year to year. Instead, the entire adult population is, somewhat unrealistically, made up of recruits. This model, however, has the nice property that the optimal approach to the steadystate solution S ∗ is a “bang-bang” approach (the optimization model does possess the required linearity in the control variable). Determining the best-fit value of the survival rate σ (which can be a function) and the shape of the recruitment function G(·) to be used in the model, is a question of

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 39 testing different (functional) forms for a particular fishery that is studied. Bjorndal (1988) has performed such a test in a study of optimal management of the North-Sea herring. He found that the Ricker recruitment function, X t ea(1−X t /b) , and a survival rate σ possessing depensation characteristics gave the best fit on the data. His findings are not surprising since herring is a schooling fish, which are known to show depensation at low population levels (they have a progressively increasing rate of birth over the range of low stock densities). Fishing mortality at low population levels is therefore compensated by higher birth rates, which gives rise to a recruitment level that is virtually independent of the population level. This could lead to sudden extinction (see Section 3.3.1) at very low fish population levels if harvesting of the fish is not controlled.

3.3.4. The Beverton and Holt fisheries model Beverton and Holt (1957) developed a model that was aimed at maximizing net present value of fishing subject to the biomass of the fish population. Realizing that biomass varies with the age of the fish, the question can be stated at what age, or between what age, the fish should be caught in order to maximize present value. The Beverton and Holt model is less restrictive then the Gordon–Schaefer model since the latter assumes all age groups of the fish population encounter the same amount of effort. Let Nk (t) denote the number of fish, belonging to the kth cohort (also known as year class), alive at time t. The kth cohort is the collection of fish that recruit (become available for the fishery) at year k. Equation (3) for the kth cohort is now replaced by a dynamic process, for all t ≥ k: dNk (t) = − (M + Fkt ) Nk (t) dt

(25)

where M denotes a constant natural mortality rate, and Fkt is the fishing mortality of the kth cohort at time t. Hence, for all t ≥ k: ⎡ ⎤ t Nk (t) = Rk exp ⎣− (M + Fkt ) dt ⎦ (26) k

where Rk denotes initial recruitment to the kth cohort. Note that the age of the kth cohort as measured from the day of recruitment is t − k. If we multiply Nk (t)by a weight function w(t − k), denoting the weight (kg) of a fish at age t − k, we get an expression for the total biomass of the kth cohort at time t.2 Catch in terms of total biomass of the kth cohort alive at time t can then be expressed as: Ck (t) = q E Nk (t)w(t − k)

2

(27)

A form often used for the weight estimation is the Von Bartalanffy curve: w(t − k) = w ∞ (1 − e−k(t−k−a0 ) )3 , where w ∞ , k, k0 are parameters that can be estimated from catch data.

40

PART I: TOOLS AND BASIC INSIGHTS

For the case of a single cohort, let t = 0 be the moment at which the cohort recruits to the fishery. The profit function to be maximized can then be expressed as: Maximize Et

∞

e−δt ( pq Nt w t − c)E t dt

(28)

0

where Nt is the number of fish in the single cohort at age t, and w t is the weight of the fish in that single cohort at time t. This equation is similar to 3.6 in form. However, as opposed to the net revenue function in Equation (6), the age of the single cohort to fish is sought that maximizes the net revenue function of the total cohort biomass (the biomass changes over time as the fish grows). Maximizing the objective function in Equation (6) is subject to the number of fish caught that changes as the size of the fish population changes. Maximizing 28 subject to 25 (for the single cohort case), the singular path Nt∗ for the single cohort now becomes: Nt∗ =

cδ 

pqw t M + δ − dw /dt dw



(29)

To interpret this result, define the net biovalue V (t) of the cohort at age t as being (Clark, 1990): V (t) = pN (t)w(t) − c

(30)

V (t) changes as a result of natural and fishing mortality as well as natural growth by aging. Let V ′ denote the derivative of V (t) with no fishing mortality present. Equation (29) can then be rewritten as: V ′ (t) =δ V (t)

(31)

This means that the optimal cohort size is adjusted to maintain equality between the proportional increase in potential net biovalue and the discount rate. Clark (1990) has shown that the optimal harvest policy follows a bang-bang approach. The age(s) of the cohort to fish is determined by the discount rate. If the discount rate is zero then the cohort should be fished at the age when the cohort’s biomass reaches its maximal height (impulse fishing), down to the level at which profit is reduced to zero. Under an infinite discount rate fishing should start at the age of the cohort where there is zero profit, until the age where biomass is reduced to such a level that further harvesting would incur a loss.

3.3.5. Metapopulation dynamics and age-related dispersal of fisheries Modeling a single fish species population in isolation may often be unrealistic. Local populations of a single species can be dispersed and yet be dependent on each other through migratory patterns. Such a set of local populations of a single species interacting with each other over a larger spatial scale is also known as a metapopulation. Inclusion of metapopulation features adds significant realism when modeling a fishery.

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 41 Quite a few ecological models exists that are devoted to modeling a non-migratory adult fish population tied to relative small areas, where recruitment takes place from a larval pool of the same fish species that is dispersed throughout a much larger area. Population dynamics of such a fish metapopulation is generally modeled as being dependent on physical transportation processes of the sea (Possingham and Roughgarden, 1990; Alexander and Roughgarden, 1996). These processes affect the recruitment process of the larvae to the adult population. Recently, increasing attention has been given to models that combine the age-dependent dispersal characteristics found in ecological models, such as above, with fishery economics. These models have generally been applied to model the dynamics of shellfish and non-migratory fish populations found in estuaries and coral reefs. Pezzey et al. (2000) have studied such a fishery under open-access fishing conditions. Instead of expressing catch as in Equation (2), they express it as: X C = qE (32) K Normalizing K to 1, the resulting catch is thus taken as being proportional to the (nonmigratory) adult fish stock in the area fished, which has a fixed carrying capacity. The optimal bioeconomic equilibrium is the same as found in the standard open-access setting described earlier. Pezzey et al. aimed to study whether imposing a reserve could lead to a higher optimal catch for the fishery. They used an adapted version of the growth Equation (3) found in the Gordon–Schaefer model. To model the population dynamics of the designated fishing ground they used the equation:   XF G F (X F ) = r K F (X F + X R ) 1 − (33) KF For the reserve they are modeled the population dynamics as:   XR G R (X R ) = r K R (X F + X R ) 1 − KR

(34)

where the subscripts F and R refer to the fish population in the fishing ground and reserve respectively. The rationale for Equation (33) is that given their assumption of uniform dispersal of larvae, the stock relevant to juvenile growth is not X F , but K F (X F + X R ) when K F is the proportion of larvae from the total stock X F + X R that will be in the fishing ground. The capacity constraint that limits growth is dependent on the stock in the fishing ground alone. Hence the term [1 − (X F /K F )]. The rationale for the growth equation of the fish in the fishing reserve is similar. Using these growth equations and finding the bioeconomic equilibrium it turns out that imposing a reserve can lead to a higher equilibrium catch given that the cost of fishing is sufficiently low. Brown and Roughgarden (1997) modeled a fishery where a fish population possessing age-dependent dispersal characteristics is fished under sole-owner conditions. Their model possesses population dynamics aimed at describing a shellfish population in a larvae stage and an adult stage separately. For the adult stage the dynamics of the population is given by: F(X, L) = cL(A − a X ) − µX

(35)

42

PART I: TOOLS AND BASIC INSIGHTS

where c is the settlement rate of larvae L, A is the amount of space in the habitat under consideration, a is a the amount of space occupied by an adult and µ the mortality rate of the adult population. For the larvae stage the dynamics of the population are given by: G(X, L) = bX − cL(A − a X ) − vL

(36)

where b is the reproduction rate of adults and v is the mortality rate of larvae. Using the maximum principle, Brown and Roughgarden maximize net present value given by Equation (6) subject to Equations (35) and (36). Taking the assumption that cost of fishing is zero they derive the following two equations to determine the optimal adult and larvae population to be approached following a “bang-bang” control similar to the one described in Section 3.2.1: acL + µ + δ c(A − a X ) − (37) acL + b c(A − a X ) + v + δ and bN − cL(A − a X ) − vL

(38)

Note that the optimal values for X and L are independent of the price of the catch. This is due to their simplifying assumption that cost of fishing is zero. The equations describing the population dynamics used by Brown and Roughgarden are meant to describe a shellfish population. Pezzey et al. have suggested that Brown and Roughgarden’s model could also be used to describe some fish populations that are non-migratory in the adult stage. For instance certain fish species populating coral reefs. It is however important to consider that, adult and larval mortality considerations aside, space is the only capacity constraint in Brown and Roughgarden’s model. There is no catchability coefficient q or capacity constraint K that could capture other possible limiting effects on the size of the fish population, such as due to interdependency characteristics that could be relevant for describing a specific fish population. The model used by Pezzey et al. however, could be appropriate for modeling both fish and shellfish population dynamics. This is due to the fact that they have not made a distinction between larvae and the adult fish population when modeling the specific population dynamics. They have made the distinction on a spatial scale where the population dynamics of different regions affect each other. When considering a nonmigratory adult fish population fished in a specific region, population dynamics of other regions may be important. These population dynamics of regions such as reserves, poorly accessible areas, may contribute a continuing flow of larvae to the region where the fishery is active.

3.3.6. Ecologically related species Assuming the fish studied in the model are ecologically dependent, a straightforward extension is to replace Equation (3) for fish X and Y with: dX = G x (X, Y ) − qx E x X dt (39) dY = G y (X, Y ) − q y E y Y dt

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 43 Table 3.1. Types of species interactions; + or − denote that population size of one species is positive or negative affected by the population size of the other Type of interaction

Species 1

Species 2

Competition Predator-prey Mutualism/symbiosis Commensalism Amensalism

− + + + −

− − + 0 0

When a fisherman targets his efforts at one species specifically, this is expressed through the effort variables: E x = E and E y = 0, or E y = E and E y = 0. The other species that is caught unintentionally is called bycatch. The bycatch species is affected negatively by fishing. But indirectly the effect can be positive when the targeted species is a competitor or predator that is fished away. When E x and E y are independent we speak of multiple use. The growth functions depend on both species. As shown in Table 3.1, one can distinguish five general types of species interaction, which determine G x (X, Y ) and G y (X, Y ). Competition and predator-prey are the most common for fisheries: Predator-prey relations are often modeled using the Lotka-Volterra model, in which G x (X, Y ) = a X − bX Y (40) G y (X, Y ) = −cY + d X Y Here a, b, c, and d are all positive, which accounts for the asymmetry between predator and prey. Other specifications are also widely used. Maynard Smith (1974) gives examples and shows differences in behavior that result from slightly different specifications. The behavior of the predator-prey model is known to be very sensitive to the form and parameters in the growth functions. May (1972), Clark (1976), Hannesson (1982), and Strobele and Wacker (1995) analyze some economic aspects of a predator-prey system. For a system like 39 where both species are harvested independently, two symmetric golden rules can be derived (Clark, 1973; Strobele and Wacker, 1995). ∂G x (X ∗ , Y ∗ ) cx′ (X ∗ )G x (X ∗ , Y ∗ ) ∂G y (X ∗ , Y ∗ ) p y − c y (Y ∗ ) − + =δ ∂X px − cx (X ∗ ) ∂X px − cx (X ∗ ) ∂G y (X ∗ , Y ∗ ) c′y (X ∗ )G y (X ∗ , Y ∗ ) ∂G x (X ∗ , Y ∗ ) px − cx (X ∗ ) − + =δ ∂Y p y − c y (Y ∗ ) ∂Y p y − c y (Y ∗ )

(41)

The equations have an additional term compared with the single species golden rule, due to the predator-prey interaction. The term accounts for the cost decrease in prey harvesting when extra predators are fished. Owing to this less prey are eaten so preystock size increases, which in turn makes fishing less costly. Conversely, fishing extra prey causes predators to starve, resulting in a lower stock size and an increase in the cost of fishing. Wacker (1999) studies harvesting in a mutualistic ecosystem. Mutualistic systems behave completely different from predator-prey systems. When either one species is

44

PART I: TOOLS AND BASIC INSIGHTS

harvested it will negatively affect the other and this in turn will result in an extra decline in the first species. Wacker shows that the stability of unharvested mutualistic systems changes when it is harvested. For certain harvest rates the system becomes unstable and breaks down. When species compete, one may use the Gause model   X G x (X, Y ) = r x X 1 − − ax X Y Kx (42)   Y G y (X, Y ) = r y Y 1 − − ay X Y Ky

The growth function is equal to the Gordon–Schaefer model to which the competition term, ai X Y is added. Both species are explicitly fished upon. Flaaten (1988, 1991) has extensively studied this model. For the single species model, the MSY is a focal point. The corresponding concept can be found in the multispecies model as well. It is called the Maximum Sustainable Frontier (MSF). The MSF is derived by maximizing the sustainable yield of one species for a constant sustainable yield of the other. This results in a curve, hence maximum sustainable frontier. On this curve it is not possible to catch more of one species without catching less of the other. Because of competition both species cannot simultaneously reach their maximum stock level. The maximum stock level for one species is only feasible when the other is fished to extinction. As in the single species case, we are not so much interested in the MSY but we are interested in the maximum sustainable profit. This depends on the price of the fish and the effort costs to catch them. We leave the analysis of the discounted maximum profit to Flaaten (1991). But we note that it can be profitable to harvest one species at a loss when the other is much more valuable. The more valuable species is troubled less by its competitor and can be harvested at a much higher stock level. For more discussion of predator-prey models see Chapter 6.

3.4. MODEL ELEMENTS RELATING TO ECONOMIC BEHAVIOR 3.4.1. Time variations in price and cost parameters To allow for time variation in the sole-owner model, time dependent parameters can be incorporated in the net present value function given in Equation (6): ∞ PV = αt [ pt qt (X t )X t − ct ]E t dt (43) 0

where: ⎛

αt = exp ⎝−

t 0

⎞

δs ds ⎠

(44)

Note that αt allows for a varying discount rate. If the discount rate is taken to be a constant, αt reduces to e−δt . The optimum solution is the time varying fish population

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 45 level X t∗ given by Equation (13): dG c′ (X t∗ )G(X t∗ ) d p/dt + = δt − dX t pt − c(X t∗ ) pt − c(X t∗ )

(45)

It can be shown that when there is a sudden price change from p1 to p2 at time t (independent of the population level X ), a sudden discontinuous leap occurs in the optimal fish population level, say from X 1∗ to X 2∗ (no sudden change, however, can bring about such a discontinuous leap, because obviously the fish population level cannot change discontinuously). The resource owner can perhaps take advantage of any knowledge about an upcoming price change and then adjust his fishing behavior accordingly at a time t1 prior to t. If the price change is such that the price decreases, he will at time t1 apply maximum effort to take advantage of the current, higher price. At time t he will then stop fishing until the optimum population level is brought back to X 2∗ (Clark, 1990). If the price change is an increase in price, he will stop fishing at time t1 to allow the fish population to increase, in order to take advantage of the higher price effective as of time t. He will then apply maximum fishing effort until the fish population is reduced to the optimum population level X 2∗ , corresponding to the new price level. Note that Equation (45) is entirely determined by current values of the parameters, even though the time scope of the PV estimation was infinite. The model is only dependent on current, not future, price changes. It is for this reason that the optimal biomass level X t∗ is sometimes referred to as being a myopic (short-sighted) decision rule.

3.4.2. Supply and demand Supply in a fishery can be thought of as the catch that is delivered to the market. To graph the equilibrium supply curve in an optimally controlled fishery for the soleowner case3 , we use the Golden Rule Equation (7) and the sustained yield condition dX t /dt = G(X ∗ ) − C = 0. We get the following equation for the equilibrium supply curve: S(C) = p = c (X ∗ ) −

c′ (X ∗ ) C δ − C′

(46)

The graphical representation under three different discount levels is shown in Figure 4 (Clark, 1990). We assume that the fish population growth equation is as used in the Gordon–Schaefer model (Equation (4)). It can now easily be shown that the catch output on the market is zero when p < (c/q K ), as the profit that can be obtained falls below zero. This can be seen by noting that the optimum fish population level for the Gordon–Schaefer model (using Equation (7) and a growth function as given by 3

A sole owner is not the same as a monopolist who completely controls the fish market supply and is able to set a price to maximize profits. In the setting discussed here there can be many sole owners who control and manage separate fisheries but act as suppliers of identical fish. This implies that they are price takers on the market for fish.

46

PART I: TOOLS AND BASIC INSIGHTS

Figure 3.4. Supply curves under various discount rates.

Equation (4)) reduces to (Clark, 1990). ⎛ ⎞  2 c cδ K c δ δ ⎝1 + ⎠ 1+ − + − + X∗ = 4 pq K r pq K r pq K r

When p < c/q K this optimum population is negative. For finite δ, the Golden Rule equation assures that as p increases to infinity the supply curve approaches G ′ (X ) = 0, which is exactly the MSY level. Under a finite discount level (under similar reasoning as in the zero discounting case) according to Equation (7), the line G ′ (X ) = δ is approached asymptotically. When the rate of discount is infinitely large, it is optimal to exploit the fishery now and hence the openaccess fishery scenario is approached. It is easy to show that as price increases to infinity, the sustainable catch supplied will approach zero, meaning it is optimal to completely wipe out the fish population. In Figure 3.4, for the finite and infinite discount case, the optimum catch reaches the MSY point as the price level increases, but then subsequently decreases as the price of fish is further increased. As the price is driven upwards, more fishing will take place and the fish population would sustain itself at a lower population level than the MSY. This is known as overfishing: the same or larger catch can be sustainably realized at a higher population level. Given the discounted supply curve, the optimal equilibrium price and catch are set by the intersection of the discounted supply curve and the demand curve. This optimal

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 47 Price

S(C )

p*

C /(q K

p(C )

)

MSY

Catch

Figure 3.5. Optimal equilibrium point: intersection of the discounted supply curve S(C) and the Demand curve p(C).

equilibrium point can be derived from the following adapted version of the Golden Rule Equation (7), where the price is now dependent on the catch. G ′ (X ∗ ) −

c′ (X ∗ )G(X ∗ ) =δ p(C) − c(X ∗ )

(47)

The optimal equilibrium point lies therefore at the intersection of the demand curve and the discounted supply curve, as shown in Figure 3.5. Here p(C) is the demand curve and p ∗ is the equilibrium price.

3.4.3. Investment and fixed cost It seems perfectly reasonable to include fixed cost in the Gordon–Schaefer model to account for capital invested in vessels and the subsequent depreciation of this capital. In the sole-owner model this amounts to adding two extra constraints to the Equations (3) and (4): dK t = It − γ K t dt

(48)

0 ≤ Et ≤ K t

(49)

48

PART I: TOOLS AND BASIC INSIGHTS

Here It is the investment decision at time t and γ is the depreciation rate. The net present value function to be maximized over time, πt , now becomes: [ pq(X t )X t − c] E t − c K It

(50)

4

Here c K denotes the cost of capital. The problem to resolve reduces to the standard model without capital and with variable costs in the case when capital is completely reversible, which is analogous to renting the vessels required by the fishermen. In other words capital can be seen as variable fishing costs. In the situation where the capital invested is irreversible, which is more likely to be the case for fishing vessels, the derivation of a solution of the model becomes very complex (Clark et al., 1979). The optimal strategy includes: if the fish has fallen below a certain minimum population level, do not fish; fish when the fishery has recovered and exceeds the minimum population level; invest a certain amount of capital when the fish population X t exceeds a certain threshold level. The threshold level is determined by the total amount of capital invested and the fish population level.

3.5. MODELS TAKING MANAGEMENT POLICIES INTO CONSIDERATION 3.5.1. Allocated quota’s In order to regulate the total amount of fish caught, fishery regulation offices can predetermine the amount of fish that can be caught by each fisherman. Each fisherman is then given a certain “quota” he is allowed to catch. The quotas discussed here are assumed to be unit quotas, i.e. the size of the quota equals the number of fish he’s allowed to catch, and transferable between fishermen. Assume that fisherman i has a quota Q i . Then the following holds: Ci ≤ Q i

(51)

In other words, his catch cannot exceed his quota. He will try to maximize a profit function:   Ci πi (X i , Ci ) = pCi − ci (52) qX subject to restriction 51. He will not leave any portion of his quota unused. He may fish or sell his quota. His demand function for quotas is specified by the equation (Clark, 1985): dπi (X i , Q i ) = m dQ i

(53)

where m is the price of the quota on the quota market. So Di equals the size of the catch increase that would bring about a net profit increase of m, which is the price of the 4

Compared with the sole-owner model discussed previously in Section 3.2, optimization here is subject to one extra dynamic equation. There are now two control variables, namely harvesting and investment.

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 49 quota. Given Equations (52) and (53), it can be easily shown that the demand function can be determined from:   Di ′ ci = ( p − m)q X (54) qX In the case of a competitive market for quotas the market for quotas will clear and consequently for fisherman i his demand for quotas will equal the amount of fish he will catch. This implies the following equation:   Di ci′ (E i ) = ci′ = ( p − m)q X (55) qX This equation determines the amount of effort that fisherman i will exert in the fishery, given his quota.

3.5.2. Total allowable catch in an open-access fishery In this section, we look at the case where the amount of fish that is allowed to be caught by the fish industry as a whole is regulated by the fishery regulation offices. This implies that during fishing season fish may be caught up to a predetermined level, after which the fishing season is closed. If we look at just one fishing season, the length of which is defined by T , we can as an assumption ignore the growth function term of the fish population in Equation (3). Assume that there are N vessels then the following intraseasonal model to be maximized is: ⎧ ⎫ ⎨ T ⎬ Maximize N [ pq X E t − c (E t )] dt − c f N (56) N ,E t ⎩ ⎭ 0

subject to:

dX t = −q X t N E dt

(57)

X 0 = R, X T = S

(58)

and

In Equation (56) the term c f N denotes the fixed cost of the N vessels. The catchability term q in Equation (57) is assumed to be constant (which in the optimization model leads to E being constant as well), which in a given season is not entirely unreasonable. Equation (58) reflects the quota: R being the initial recruitment level of the fish population at the beginning of the fishing season and S being the predetermined escapement level of the fish population, as set by the authorities prior to the fishing season. Writing the Hamiltonian for this problem and applying the maximum principle will give the optimal effort to apply. The corresponding optimal number of vessels participating in the fishery can then be determined from Equation (57). The Hamiltonian and its necessary conditions for an optimum lead to the following condition that needs

50

PART I: TOOLS AND BASIC INSIGHTS

to be satisfied: c′ (E) = ( p − λ)q X

(59)

The term λ is the shadow price, which reduces to zero in an open-access fishery. Setting λ larger than zero, the open-access fishery can be controlled because it will no longer be profitable to fish at the open-access equilibrium fish population. This leads to the conclusion that the fishery can be controlled by imposing a tax λ on the fishing industry.

3.5.3. Taxes When the government imposes a tax τ on catch so as to regulate the open-access fishery, the after-tax revenue becomes: πτ = ( p − τ ) Ct − c (E)

(60)

Maximization of this equation subject to effort results in: c′ (E) = ( p − τ )q X

(61)

By comparing this with Equation (55), we see that taxes have the same effect on fishing effort as a transferable quota of m. The difference is that the tax is a direct cost to the fisherman, reducing the price received for the catch, the benefits of which go to the government. The transferable quota is an opportunity cost, since it can be sold at the price m to other fishermen. Equation (59) (note the similarity between Equations (55), (59), and (61)) shows that in a competitive fishery where there is a limited total annual catch, the government can force the fishery into the desired mode by imposing a tax τ . This will have an impact on the effort applied by the number of fishermen and subsequently the number of boats active in the fishery. As a final comment it is interesting to note that X will decrease within the fishing season, under the assumption that fishing seasons are set by the regulation authorities in a manner avoiding recruitment and harvesting taking place simultaneously. In an ideal world this would cause taxes to decrease during the season, because the tax imposed depends on the size of the intraseasonal decreasing fish population. Applying such a policy would raise all sorts of practical problems.

3.5.4. Minimum size rules In order to protect local fishery it is possible for the government to impose minimum fish size rules on the catch. For instance, in the American Lobster industry, lobsters below a certain minimum weight cannot be retained. Variations of the Gordon–Schaefer model used by Bell and Smith to predict future catches were inadequate because they predicted that with an increase in effort the lobster population would collapse (Townsend, 1986). This never happened obviously because of the minimum size rule on catch, which their model did not account for. Townsend showed, using the Beverton Holt model and maximizing a similar expression as 12 over a range from the age at which the lobster reaches the minimum

3. DETERMINISTIC ECONOMIC MODELS OF FISHERIES MANAGEMENT AND POLICY 51 weight size to infinity, that unlike in the Gordon–Schaefer model yields can never fall back to zero. The limiting value is the weight that is exogenously recruited into the fishery each year. Lobsters below the minimum weight level will not be part of the lobster population fished.

3.6. CONCLUSION The objective of this chapter was to give an overview of the most basic models that can be used to obtain insight into how a fishery might behave under certain integrated assumptions and conditions. All models, with the exception of those discussed in Sections 3.3.4 and 3.3.5, are based on a growth function as used in the Gordon– Schaefer model (Equation (3)). Although the simplicity of the models might make their immediate practical use doubtful, they serve as a way to gain understanding in how the ecological and economic model elements discussed above interact and could affect a fishery. The existence of a steady-state optimal equilibrium derived by using any of the above models seems unrealistic, considering the forces that affect the ecological and economical elements in a real-life setting. The models should therefore be considered as a tool aiding a fishery management policy instead of a mechanism determining a fishery management policy. Other types of models as mentioned in the introduction can provide complementary information. When considering the equilibrium point under open-access (point B in Figure 3.1, for example), we also see that extinction of a fish population under consideration is heavily influenced by how much fishing costs. Under a growth function as used in the Gordon–Schaefer model the fish population will not go extinct unless the costs of fishing are zero, which is of course unrealistic. The model does, however, draw attention to the elements that need to be added to appropriately address the matter of sustaining the fish population as a whole for the future. The importance of knowledge concerning the growth function of the species fished cannot be underestimated and is vital to sustaining the fishery. The same holds for knowledge on how a species behaves in its habitat. Without this knowledge placing restrictions on fishermen’s quota or taxation is, however useful, more a matter of preventing overfishing than a matter of striving toward optimal control of the fishery. Modeling the interdependency between fish populations is especially relevant toward providing a greater amount of realism. Identifying the major social and economic forces that lead to overfishing can contribute toward sustaining the fishery. Creating economic incentives that internalize the external cost of biodiversity loss brought on by the fishery is often sufficient to motivate the prevention of overfishing (Folke et al., 1996). Modeling a fishery with interaction of multiple fish species becomes extremely complex as the number of fish to be modeled is increased. Furthermore recent studies have made clear that more attention needs to be spent on age structure and fish populations of the same species interacting with each other on a spatial scale by way of migration (Brown and Roughgarden, 1997; Bulte et al., 1998). Regulation of the fishery by way of a system of individual transferable quotas (ITQ) is being applied in some countries. Nevertheless, many questions have been raised regarding its applicability. Matters as enforcing fishermen to stick to their quota

52

PART I: TOOLS AND BASIC INSIGHTS

and how to deal with avoiding bycatch of smaller fish being discarded are some of the problems that can arise (Copes, 1986). Fishery regulation boards controlling the fishery by imposing regulations have also been criticized for a lack of knowledge of the fishermen they are trying to control. Regulations on fishing gear restrictions, for instance, have been supported by politically powerful, but less inefficient fishermen. This has resulted in an increase in their numbers and a decrease in politically less powerful, but efficient, fishermen (Karpoff, 1987). Such suboptimal controls are the result of fishery regulation boards considering the fishermen to be a homogeneous group, when the fishermen are in fact strongly heterogeneous. Evolutionary models might address such issues. Some studies argue in favor of a quota on effort. Nevertheless whether or not a quota on effort is to be preferred over a quota on catch seems to be highly dependent on the nature of the fishery in question. In fact, Hannesson and Steinshamn (1991) argue that neither a constant quota on effort or on catch is preferred. They argue for continuously adjusted catch quotas dictated by economic and other circumstances.

CHAPTER 4

INCORPORATING UNCERTAINTY IN THE ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION

4.1. INTRODUCTION Fisheries management in the past was primarily focused on maximizing sustainable yield from the fishery by regulating fishing effort. As the effects of overfishing became more apparent, fisheries management objectives shifted from maximizing sustainable yield to maintaining a minimum spawning biomass (Weeks and Berkeley, 2001). In order to determine the optimal effort levels that would satisfy fisheries management, a high degree of trust was put into fish population growth models. Fish populations, however, have not proven easy to manage with the result that many fish populations are currently overfished. In large, this was due to the fact that fisheries management tended to rely on oversimplified models that underestimated the complexity of fish population dynamics (Caddy and Cochrane, 2001). The general conclusion that can be drawn from fisheries management in the past is that fish population dynamics are controlled by factors about which information is incomplete. Biological factors that play a role in population dynamics are often unknown, or their role is unclear. The extent to which fish populations are affected by the surrounding marine ecosystem is often very complex and should not be neglected. How environmental fluctuations affect fish populations on a local and global scale is largely unknown. Fishing may also affect targeted species in other ways than just by reducing the fish population due to fishing mortality. Finally, fishing techniques used can alter the marine habitat and affect other non-targeted species as well. Not only should fisheries management recognize that fish population dynamics are complex and influenced by factors that are often poorly understood, but it should also recognize that fishing effort is something that is not easily controlled. Managing fisheries by prescribing a fishing effort requires detailed knowledge regarding factors that influence fishing behavior, which in turn can vary depending on fishermen’s characteristics, fishing technology used, or even nationality (Roughgarden and Smith, 1996). Furthermore, the implications of fishing regulations on fishermen’s behavior, and hence fishing effort, may be uncertain with the possible result that fishing mortality is not reduced. Fishermen may have different ideas about fisheries management than regulators have and may strategically alter their fishing behavior. In order to manage a fishery successfully, much information is required about the population dynamics of the fish population, the interaction of fish populations with their environment, and the factors that influence fishermen’s behavior. Fishery management

53

54

PART I: TOOLS AND BASIC INSIGHTS

authorities increasingly recognize the lack of information they have regarding these issues. As a result, uncertainty in the fisheries is becoming a key concern that management must address and an issue that scientist must try to account for when modeling the fisheries (FAO, 1990; Hilborn and Peterman, 1995). The purpose of this chapter is to look at how uncertainty has been dealt with in managing fisheries. This chapter is organized as follows: Section 4.2 presents the different kinds of uncertainty that may arise when managing a fishery and which should be accounted for. Section 4.3 discusses how uncertainty in the fisheries has been incorporated into ecologically oriented fishery models that mainly focus on fish stock assessments. Section 4.4 discusses how uncertainty has been addressed in bioeconomic models where the focus is predominantly on the profit-maximizing behavior of fishermen. Section 4.5 addresses policy responses to uncertainty. Conclusion are drawn in Section 4.6.

4.2. A TYPOLOGY OF UNCERTAINTY There are several sources of uncertainty that will have an impact on fisheries management. First, there is uncertainty due to the complexity of the fish population being managed. Second, uncertainty will arise as a result of uncertain economic processes that affect fishing behavior. Third, imperfect information regarding ecological and economic systems can be expected to cause errors in modeling, the size of which may be uncertain. These types of uncertainty are generally interrelated and will be discussed below.

4.2.1. Ecological uncertainty In protecting the viability of fish populations, imperfect understanding of population dynamics needs to be accounted for. Shaffer (1981) made a distinction between two environmental (extrinsic) and two demographic (intrinsic) factors that can affect the fish population. According to Shaffer, the two environmental factors are environmental uncertainty and the uncertainty of catastrophic events. Environmental uncertainty results from biotic and abiotic factors. Abiotic factors that have an influence on fish stocks include water temperature, habitat composition, and sea currents. Fluctuations in these factors could have a profound influence on fish population levels, but the size of this influence is often unclear. Without an understanding of environmental uncertainty it may not only be difficult to project fish population sizes, but it may also be difficult to determine whether changes in fish populations are the result of human activities or should be attributed to changes in the environment. Furthermore, if the linkage between the marine habitat and the fish populations is poorly understood it can be difficult to assess what the impact of fishing techniques are on the marine habitat. Dredging, for example, has been associated with long-term changes in the marine environment, but its exact impact on the habitat is unclear or even disputed (Fogarty et al., 1996a; Ens, 2002). Climatological changes on a global scale have also been associated with fish population fluctuations (Klyashtorin, 1998). Biotic factors that influence fish population dynamics include fish species interactions and the availability of vegetation. Not taking account of uncertain predator–prey relationships during the different life stages of fish species can lead to depleted fisheries and seriously alter the composition of fish

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 55 species in the marine environment (Gjøsaeter and Bogstad, 1998; Pauly et al., 2000; Walters and Kitchell, 2001). Fish populations can also be subjected to high environmental variance or catastrophes. Catastrophic events could be naturally induced (e.g. very extreme storms) or the result of human actions (e.g. oil spills). At any rate, catastrophes substantially diminish chances of survival of the fish population when they occur. If fisheries management does not account for the possibility of a catastrophe, extinction of the population can result. The two demographic factors that can affect fish population fluctuations are demographic and genetic uncertainties. Fluctuations in birth and death rates can have serious consequences for fish population levels when numbers are low (Nisbet and Gurney, 1982; Boyce, 1992). At low numbers age composition will also be an important consideration, because generally the number of offspring varies greatly with weight, which varies with age. These are all factors that are subject to uncertainty. The vulnerability of fish stocks to demographic uncertainty diminishes as populations get larger. Uncertainty regarding metapopulation and spatial dispersal characteristics also needs to be addressed by fisheries management. As fish populations are fished down they become more vulnerable to demographic uncertainty: if recolonization from other areas is low local extinction may result (Mode and Jacobson, 1987). Genetic uncertainty can affect fish population as a result of inbreeding and “mutational meltdown.” Fish population persistence can be negatively affected as the genetic variability required to withstand selective natural environments is reduced. This is an issue of particular concern when fish population levels are low (Foley, 2000). The extent to which ecological uncertainty would have an impact on the persistence of the fish populations depends on their population size. It is clear that as the population numbers dwindle, their exposure to the stochastic nature of demographic and genetic uncertainty increases. A catastrophe may lead to extinction of low stocks or reduce the population to levels from which they cannot recover. If fisheries management does not take account of ecological uncertainty it is likely that fish population persistence will be affected. There is an increase in models that incorporate ecological uncertainty, which contribute to understanding the effects of ecological uncertainty on fish populations in the presence of fishing. Nevertheless, knowledge regarding fisheries dynamics is generally incomplete and fisheries management should understand the limitations of fisheries models presented to them.

4.2.2. Economic uncertainty Ecological uncertainty will have a large impact on fishermen’s behavior, as it will affect the price paid for fish, the number of fishermen fishing, the cost of fishing, fishing technology used, discount rates, and preference of future consumers. Yet these economic factors become even more uncertain because they are affected by other sources of uncertainty as well. Uncertainty in the price for fish can be due to changes in supply and changes in demand. Changes in supply are due to ecological uncertainty, which can impact the availability of fish. Changes in demand are due to changes in consumer preference regarding the fish product. Consumer preference will be affected when it concerns

56

PART I: TOOLS AND BASIC INSIGHTS

commercially harvested species that can easily substitute each other. Local fluctuations in demand can also occur if there are alternative sources where similar fish products can be obtained. Furthermore, fishermen may not only compete with other fishermen but also with aquaculture. Aquaculture can compensate for the effect local fluctuations of fish supply have on the price paid for fish. Fishermen will also be affected by uncertainty about the cost of fishing. Ecological uncertainty contributes to uncertainty about variable cost, because the time at sea required to catch the fish varies with the availability of fish. Variable fishing costs will also fluctuate as a result of changes in fuel prices and changes in employment costs. The fixed cost of fishing will increase when there is uncertainty regarding the profitability of the fishing sector. Uncertainty regarding stock levels can lower the value of vessels but at the same time increase interest rates of loans used to purchase the vessels because fishing will be considered a riskier venture. Technological advances have contributed to fishermen being able to fish populations more effectively and efficiently. In the past, technological improvements have primarily increased fish population uncertainty because they allowed fishermen to fish overfished populations more efficiently. Many new technologies are geared at improving fishermen’s ability to fish more selectively in order to reduce bycatch of juvenile fish populations and other non-targeted species. Technological advances will have either a positive or negative impact on the uncertainty about the effects of fishing techniques used on fish population dynamics and on the uncertainty about the cost of fishing. Fishing behavior will also change as a result of changes in fishery management. The fishing industry may be uncertain about the influence of stakeholder and public pressure on fisheries management regimes and thus on future fishing regulations. In the case where interests between the fishery sector and environmentalists have diverged, Hoel (1998) gives examples of how environmentalists have influenced existing fishing regimes. He argues that increased environmentalist’s pressure has turned the International Whaling Commission from being an organization focused on regulating the whaling industry from a conservationists point of view to an organization controlling the industry from an ethical and animal rights point of view. Multilevel governance itself will create complexity and uncertainty, which will influence fishermen’s behavior. Fisheries may be subjected to management regimes that are complex in the sense that they represent small-scale management plans that cannot be seen in isolation from a larger overarching management regime. The complexity of translating large-scale management plans into localized management plans often means that the finer operational details are left at the discretion of local fisheries management regimes, which can mean uncertainty for local fishermen about how large-scale fisheries regimes will impact their future (Young, 1998).

4.2.3. Uncertainty in modeling Analytical modeling of fisheries will inadvertently incorporate uncertainty if it is based on ecological and economic elements discussed above. Caddy and Mahon (1995) list several sources of uncertainty that need to be accounted for when modeling fisheries. They mention measurement, model, estimation, and implementation of uncertainty.

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 57 Measurement uncertainty results from observation errors when sampling data such as catch levels, fishing effort levels, and weight at age. These data introduce uncertainty in the model when it is used to estimate model parameters. Model uncertainty results when the functional form of population dynamics is unclear and the way harvesting has an impact on fish populations and their marine environment in general is not known. The level of uncertainty will be aggravated when fishing behavior is mis-specified in the model. Measurement and modeling errors in turn give rise to estimation uncertainty. Small fluctuations in data can yield entirely unanticipated results (Lines, 1995; Weeks and Berkeley, 2001). Another source of uncertainty is implementation error, which results from the inability to control fishing effort and catch levels in ways as specified by the model. As previously stated, fishermen behavior and fish dynamics can be very uncertain and as a result prescribing what should happen can be very different from what will happen. Therefore, it is important that scientists make the sensitivity of their models very clear. They should be aware not to present the management scenario that is most likely to achieve the desired management goals as the correct management strategy. Ideally, they should present all considered management strategies along with information about their probability of achieving the desired objective (Buhl-Mortensen and Toresen, 2001).

4.3. ECOLOGICAL UNCERTAINTY IN FISHERIES MODELS 4.3.1. Stock assessment Many studies incorporating ecological uncertainty do so in order to determine how this uncertainty along with fishing mortality affects fish population abundance. The term fish population abundance usually refers to the number or the weight of the fish stock under consideration. Although it is clear from the previous section that many factors contribute to fluctuations in fish population abundance, their exact contribution to population fluctuations generally is hard to quantify and many factors are omitted. As an example, consider the influence of climatological fluctuations. Over the last century, commercial fish catches have displayed oscillatory patterns of approximately 30-year intervals consistent with certain key climatological parameters (Klyashtorin, 1998), suggesting that climatological parameters may be useful for predicting long-term trends in stock fluctuations. Nevertheless, the exact relationship between climatological conditions and fish stock abundance is poorly understood. As a result, climatological parameters are, as yet, generally not included in stock assessment models. Most stock assessment models are conducted on the basis of catch data where the relationship between fishing pressure and fish population size is investigated. A common assumption is that catch per unit effort (CPUE) is directly proportional to population abundance (Clark, 1990). Freeman and Kirkwood (1995), for example, present the following model. It uses CPUE data to estimate the population abundance X t at time t of the form CPUEt = q X t + εt

(1)

58

PART I: TOOLS AND BASIC INSIGHTS

where q is a catchability coefficient and εt is a random normally distributed term denoting observation error. The observation error is attributed to imperfectly sampled fish populations. The population is assumed to be subject to discrete population dynamics of the form X t+1 = α X t + Rt − Ct + ξt+1 (2) with recruitment Rt given by Rt+1 = Rt + ηt (3) Here Ct denotes catch, ξt+1 and ηt are process error terms, and α is a survival constant. The error terms are used to account for ecological uncertainty. Although this model obviously is a very simplified version of reality, it emphasized that in the absence of sufficient catch-at-age data it could be a useful approach to model stock dynamics because of its simplicity. The lack of sufficient catch-at-age data, however, is not a valid excuse to assume proportionality between catch and population size as is done in Equation (1). It is a very simplifying assumption, which implicitly assumes that fish populations are uniformly dispersed throughout the area under investigation and it takes no account of spatial fish population dynamics. Based on CPUE and abundance data from the International Council for the Exploration of the Sea, and accounting for observation and estimation error, Harley et al. (2001) find more credibility in a functional relationship between abundance and CPUE data of the form: β

CPUEt = q X t

(4)

where β is a parameter. When β = 1, Equation (4) reduces to the commonly held assumption of proportionality. A situation known as hyperstability occurs when β < 1 (Hilborn and Walters, 1992). In this case, CPUE remains high whilst population abundance decreases. When β > 1 hyperdepletion takes place: CPUE declines at a faster rate than population abundance (Hilborn and Walters, 1992). The parameter β will generally depend on fish population characteristics. Harley et al. (2001) estimated from the data that for most species examined β tended to be smaller than 1. This is the case, for example, for cod and flatfish. Assuming a linear relationship between CPUE and population abundance for such species can lead to overfished fisheries because population abundance is actually lower than it appears to be. Stock dynamics have also been described by an equation of exponential growth of the form X t = X 0 er t

(5)

whereX 0 is the starting population size, X t is the population size at time t and r is the per capita population growth rate. Foley (1997) proposed the following probability density function to estimate r incorporating most of the ecological uncertainty sources discussed in Section 4.2: ⎧   ⎨ N rd , vr + vl with probability (1 − λ) X (6) r (X, t) ∼ ⎩ log(1 − δ) with probability λ

where λ is the probability of a catastrophe destroying a proportion δ of the population, rd is the expected per capita growth rate which is normally distributed, vr is the

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 59 variance due to environmental uncertainty, and vl is the variance due to demographic uncertainty. Rather than estimating the above uncertain parameters by frequentist hypothesis testing (which is an approach many ecologists have followed), Foley (2000) and Ellison (1996) argue that estimation of the above parameters should be done by Bayesian estimation methods. Frequentist hypothesis testing tends to favor the null hypothesis as the amount of data decreases. It is somewhat subjective because tested hypotheses of parameters generally reflect beliefs about what is expected to happen. In contrast, Bayesian estimation methods treat uncertain parameters not as hypothesized beliefs but as random variables whose value is constantly updated as more information about uncertain dynamic features becomes available.

4.3.2. The use of Bayesian methods in stock assessment Bayesian methods consider parameter values to be random variables whose plausibility cannot be accepted or rejected like in hypotheses testing. Bayesian methods use information from two different sources to assign probabilities to a specific parameter under investigation. The first source is observational data that can provide information concerning the parameter. The second source is theoretical insights concerning the probability statements of the parameter. These two sources are brought together in order to provide an updated probability statement about a parameter that takes into account the observational data at hand. Bayes’ theorem states this relationship mathematically as follows: P(ϑ|X ) =

P(X |ϑ)P(ϑ) P(X )

(7)

where ϑ is the parameter (or set of parameters ϑ) under consideration and P(ϑ) is the hypothesized probability distribution concerning ϑ. P(ϑ) is known as the prior probability distribution. X is the observational data and P(X |ϑ) is a conditional probability statement known as the likelihood function. P(ϑ|X ) is called the posterior distribution. The purpose of Bayes’ theorem in decision analysis is to estimate the consequences of a decision (as posterior probabilities) based on uncertainty (prior probability) and events (likelihood functions) (Ellison, 1996). In finding prior probability distributions for important fish population parameters, meta-analysis is an important technique. Its purpose is to synthesize results from independent fish population studies analyzing common uncertain elements in order to reduce the uncertainty inherent to these parameter estimates. It is not hard to imagine that carrying out a meta-analysis in order to estimate population parameters of fish populations is no easy task, as many parameters would appear to be population specific. Nevertheless, Myers and Mertz (1998) provide examples of how meta-analysis has been used in the fisheries. They argue that meta-analysis can greatly reduce uncertainty of the biology of exploited species, but is often of little help in reducing the inherent uncertainty with interannual variability in recruitment (Mertz and Myers, 1995; Myers and Mertz, 1998). This is probably due to the fact that interannual fluctuations in population dynamics are often quite high and poorly understood. It should be noted, however, that drawing inferences by analyzing the results of independent results might lead to many false conclusions. Hedges and

60

PART I: TOOLS AND BASIC INSIGHTS

Olkin (1985) give an overview of the potential pitfalls a na¨ıve meta-analysis approach may lead to. In fisheries, Bayesian methods are becoming important tools for stock assessment. The International Whaling Commission, for example, has used Bayesian methods to draw inferences about uncertain parameters that are part of a population dynamics equation of Bowhead whales (International Whaling Commission, 1995). Nevertheless, it is important to note that the use of Bayesian methods is not undisputed. Two common objections are mentioned here: First, the usage of prior distributions allows subjective elements to come into the analysis, as it is based upon someone’s best judgment as to what the prior distribution is. It has been suggested that in the case that no prior information exists for uncertain parameters, the so-called non-informative priors may be used which minimize the subjective element. Second, as the complexity of the model increases it becomes increasingly difficult and time consuming to use Bayes’ theorem (Punt and Hilborn, 1997; Schweder, 1998). The first criticism of the use of Bayesian methods mentioned above is one of the reasons for the popularity of the likelihood principle as an alternative to Bayesian estimation techniques. The likelihood principle states that the only source of statistical inference concerning uncertain parameters should be the likelihood function (Fisher, 1973). Thus, its premise is that all information, given in the data and the model, can be derived from the observed likelihood function. Maximum likelihood estimation can then be used to provide parameter estimates based on observational data. Besides Bayesian methods, maximum likelihood estimation techniques have been used to estimate uncertain parameters that are key to assess the population dynamics of the Bowhead whale (Punt and Butterworth, 1997). Maximum likelihood estimation techniques for statistical parameter inference are appealing because of their simplicity. Nevertheless, their use as a tool for parameter estimation is not undisputed, because they treat parameters as fixed values, whereas Bayesian inference treats them as variables (Ellison, 1996; Nielsen and Lewy, 2002). In other words, maximum likelihood techniques assume that there is one fixed value that describes parameter uncertainty for the model at hand and that the observational data are sufficient to estimate this parameter. In fisheries, population dynamics assuming that one “true” value of a parameter can be found is particularly debatable, given the fact that observational data are used that has been sampled from an ecological setting that is far from stable and poorly understood. Bayesian inference, on the contrary, considers uncertain parameters to be variables and uses observational data to assess how plausible hypotheses regarding the variables are. Thus, in a setting where it is very difficult to obtain adequate samples from the fish population, as is the case for quite a number of fish species because dynamics are poorly understood, the Bayesian approach is preferable to maximum likelihood estimation methods. Note that the argument favoring Bayesian methods over maximum likelihood methods is based on general reasoning as to why Bayesian methods should be preferred to frequentist hypothesis testing. The second criticism of Bayesian techniques follows from the fact that as the set of parameters becomes large it can be extremely time consuming or impossible to calculate the posterior distribution analytically. Fortunately, simulation techniques such as the Markov Chain Monte Carlo method (MCMC) have reduced the problem of finding the posterior distribution somewhat (Metropolis et al., 1953; Hastings,

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 61 1970; Gilks et al., 1996). These simulation techniques update the current posterior distribution at each step and then either accept or reject it.1 One of the simulation techniques used is the Metropolis algorithm. This algorithm updates ϑ to ϑ ′ at each step according to a “symmetric proposal distribution” and accepts it with probability min(1,P(X |ϑ ′ )P(ϑ ′ )/P(X |ϑ)P(ϑ)). Thus, if the updated parameter set ϑ ′ yields a greater posterior density than before it will be accepted with probability 1. If the process of updating is repeated the updated parameters can eventually converge to the true posterior density. Virtala et al. (1998) and Nielsen and Lewy (2002) give examples of how MCMC methods have been applied in fish stock assessments.

4.3.3. Recruitment Bioeconomic modeling generally assumes that there is a functional relationship between the recruitment and the spawning adult population. In a continuous model this is normally reflected as a population growth equation. In a discrete model there is a functional relationship between the spawning adult population and the recruitment, which either occurs in the next period or with delay. Yet there is a lot of uncertainty concerning the extent to which fishing pressure influences recruitment. Some researchers claim that the effects of fishing pressure on recruitment can be virtually ignored as it is dominated by other factors that are more apparent. These other factors include, among others, fluctuating predation pressure (Gjøsaeter and Bogstad, 1998), and ecosystem and climatic variability (Baltz, 1998; MacKenzie, 2000). The prospects of predicting recruitment based on pre-recruit life stages and environmental influences are generally considered to be poor (Walters, 1986; Mertz and Myers, 1995). Myers and Barrowman (1996) have conducted an extensive review on recruitment and spawner abundance data covering a wide range of different species. They concluded that recruitment and adult fish stocks are almost always related. This is in contrast to often held beliefs that there is no relationship between fish stocks and recruitment and that, therefore, fisheries systems are chaotic (Wilson et al., 1994). The fact that there generally is some relationship between fish stocks and recruitment should provide some legitimacy for modeling recruitment as a stochastically fluctuating function of fish stock levels. The stochastic nature then accounts for the natural variability of the stock. The question remains what specific stock–recruitment relationship is most suitable to describe a particular fishery in question. Various functional forms have been tried out. In the case of the Brazilian Sardine fishery, Vasconcellos (2002) considered three hypotheses regarding the appropriate stock–recruitment relationship of the Beverton– Holt type, which were assigned equal probability of occurrence. One of the hypotheses allowed for depensatory effects at low population levels, another was characterized by periodic recruitment failure due to environmental effects, and a third represented a gradual decline in recruitment to overfishing. Accounting for observation error and variable catchability parameters, the effects of three different management strategies were evaluated in a simulation. It was concluded that the current strategy used in the Brazilian sardine fishery, which is a constant effort strategy, was inadequate. It 1

Virtala et al. (1998) give a clear and brief description of the MCMC methodology.

62

PART I: TOOLS AND BASIC INSIGHTS

was shown that applying a constant harvest rate would be less sensitive to the stock– recruitment relationship and would generally lead to higher catches. There is strong support for stochastic stock–recruitment relationships to be lognormal. Such relationships are characterized by a high probability of fairly low recruitment levels and an occasionally large recruitment level. Furthermore, as the mean value of recruitment increases, so does the variance. Fish populations characterized by a lognormal distribution are characterized by the fact that small changes can lead to large fluctuations in recruitment levels. This makes the use of lognormal distributions very tricky because observational and measurement error as discussed in Section 4.2 can distort the actual stock–recruitment relationship (Walters, 1986). Fogarty et al. (1996b) considered two lognormally distributed stock–recruitment relationships to describe the relationship between the adult stock and the recruitment. One incorporated density independence (a stochastic version of a linear stock–recruitment relationship) and the other incorporated density independence (a stochastic version of the Ricker model). Based on stock assessment of the Atlantic cod fishery on Georges Bank and accounting for errors in fishing mortality and population size estimates, they concluded that the stock–recruitment relationship appeared to be distributed lognormally around a linear line exhibiting no density dependency.

4.4. UNCERTAINTY IN BIOECONOMIC MODELS OF A FISHERY Several bioeconomic models address the consequences of ecological and economic uncertainty on a fishery’s profit function or on a social welfare function. These models try to determine optimal catch and effort levels such that a net profit or welfare function is maximized subject to constraints incorporating uncertain dynamics. This is in stark contrast to stock assessment models as discussed in Section 4.3 where fishing effort is observed and the consequences of fishing on fish population levels examined given the uncertainty of the system. This section discusses how ecological and economic uncertainty can influence profit-maximizing behavior of fishermen.

4.4.1. Harvesting decisions in a single-species fishery Natural fluctuations in fish stocks, as a result of natural variability, will have an impact on harvesting decisions. Pindyck (1984) presented one of the first models addressing a profit optimization problem subject to a stochastic differential equation reflecting simple population dynamics. In his continuous model each fishermen tries to maximize his present value of expected discounted profits subject to the equation dX = [ f (X ) − q(t)] dt + σ (X ) dz

(8)

where f (x) is a population growth equation, q(t) is the extraction rate, and σ (X ) dz is an increment of a Wiener process. The latter term is used to represent ecological uncertainty. Solving this, Pindyck derives a stochastic version of the (deterministic) golden

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 63 rule equation2 (Clark, 1976), which yields the optimum quantity q ∗ to be harvested: E t d( p − c) c′ (X )q ∗ −VX X + f ′ (X ) − = δ + σ ′ (x)σ (X ) ( p − c) dt ( p − c) VX

(9)

where p is the price paid for fish, c is the per unit harvesting cost, and δ is a discount rate. VX and VX X are partial derivatives that can be determined from the present value function of the fishery V (see Pindyck, 1984). Compared with the deterministic golden rule equation, this equation contains one extra term, which is the second term on the right hand side. This term is what Pindyck calls the risk premium that fishermen expect as a compensation for an increase in marginal stock variability. Under the assumption that the cost of fishing is convex (i.e. cost of fishing increases as the stock decreases) and that the population growth function is concave, the risk premium will be positive and will increase as the level of uncertainty increases. Yet, the exact effect of an increase in uncertainty is indeterminate, because an increase in uncertainty will lead on the one hand to higher expected costs of extraction and on the other hand to higher expected profits (see Pindyck, 1984). The former will provide an incentive to increase harvesting levels and the latter will provide an incentive to reduce harvesting levels. Bulte and van Kooten (2001) model a fishery where a resource manager optimizes both the expected benefits of fishing and conservation of a fish population. Maximization of the objective function is subject to natural fluctuations in the population due to demographic and natural variability as well as to catastrophes. They use the following stochastic equations to model natural variability in fish population dynamics: dX = [G(X, f ) − C] dt + σ1 (X ) dw 1 − j(X ) dq

(10)

d f = σ2 ( f ) dw 2

(11)

and

Here G(X, f ) is the population growth equation subject to current population levels X and food availability f , σ1 (X ) dw 1 and σ2 ( f ) dw 2 are terms describing random fluctuations following a Wiener process as a result of demographic and environmental variability, respectively, and j(X ) dq describes the reduction in population levels as a result of a catastrophe following a Poisson process. They derive a golden rule equation from which the optimal fish population level could be deduced. Much like as in the model presented by Pindyck (1984), the effect of demographic and environmental uncertainty on the optimal population is indeterminate, but an increase in the probability of catastrophes provides an incentive to drive population levels down because the profitability of investing in conservation of the stock will be lowered.

4.4.2. Multispecies fisheries It has become increasingly clear that considering a single-species fishery in isolation will generally not lead to good fisheries management. This has led to an increase in focus on multispecies fisheries management (Caddy and Cochrane, 2001). Most multispecies 2

For a complete interpretation of the golden rule see Clark (1990).

64

PART I: TOOLS AND BASIC INSIGHTS

management approaches in fisheries appear empirical rather than analytical. This is probably due to the fact that spatial and dynamic interactions between fish populations are very region specific, as is the way fish populations are targeted. Predator–prey relationships, especially, have a significant effect on fisheries management. Without accounting for mortality as a result of predator pressure, targeted fish population levels may be overestimated. In the case of the Barents Sea capelin fishery, for example, it was found that recruitment failure of the capelin population was associated with high levels of the herring population that prey on capelin larvae (Gjøsaeter and Bogstad, 1998). In order to manage the capelin fisheries and understand the dynamic interactions between the relevant fish species, a simulation model is used that accounts for predator–prey interactions of several species, including cod, herring, capelin, minke whales, and harp seals (Tjelmeland and Bogstad, 1998). This model is a spatial model that accounts for migratory patterns and biological characteristics of the relevant fish species and which is used as input in managing the capelin and cod fisheries. It is rarely the case that fisheries targeting a particular fish species will only extract that species from the marine environment. Bycatch will generally either be discarded or kept as incidental catch. In many fisheries management regimes bycatch is not subject to fisheries regulations. It is highly likely that no, or imperfect, statistical information is recorded for discarded fish. It is also likely that there are many cases where bycatch that is landed as incidental catch is not recorded. This makes it very difficult to assess the impact of fishing pressure on the fish population caught as bycatch. As Milessi and Defeo (2002) point out this impact may be significant. Based on historical data, they estimated catch per unit of effort of a species caught as incidental bycatch in the Uruguayan tuna fishery using a generalized linear model with vessel types and seasons as explanatory variables. They concluded that the fish species caught as incidental catch was at risk of overexploitation. Regulating the bycatch of fish species may be a sensitive issue. Pascoe (2000) studied a multigeared fishing fleet active in a multispecies fishery, which is regulated by a total allowable catch (TAC). In this fishery one fish species generally caught as bycatch is overfished. Based on historical catch and effort data, the costs and benefits associated with conserving the overfished stock were simulated. Pascoe concluded that the costs of conservation of the stock are likely to outweigh the benefits to the fishery associated with stock recovery. This is due to the fact that, as the overfished species is caught as bycatch, conservation would entail reducing the allowable catch of the targeted species.

4.4.3. The timing of adopting policy measures to control the fishery As targeted fish species become increasingly overfished, future fisheries regulations should be expected in order to protect these species. At the same time, current fishing activities may be expected to affect other non-targeted fish populations that are not subjected to regulatory protection. These species could be affected because they are caught as bycatch and therefore are subjected to fishing mortality. They may also be affected indirectly because fishing methods may have an impact on their habitat. Thus as more knowledge becomes available concerning the impact of fishing activities on

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 65 targeted and non-targeted fish species, it may be necessary to adopt new regulations or regulate already regulated fisheries even further. An interesting question is then at what time should regulations be imposed given uncertainty regarding economic conditions that may drive fishermen to deplete fish stocks in the future and given ecological uncertainty regarding fish population dynamics of fish populations affected? This question is not one that is often asked in practice: most fisheries managers start to impose regulations on the fisheries after the resource has become overfished, not before. Perhaps this is the reason why this question is generally not addressed in the fisheries literature, even if it is addressed in other disciplines. Pindyck (2000, 2002) deals with the problem of a firm that pollutes when it extracts resources. He considered whether or not, and if so when, it would be socially optimal to impose regulations given the extent of the environmental damage and the social costs of imposing those regulations. Both ecological and economic uncertainties are considered through stochastic dynamic equations following a Brownian motion. Given this uncertainty he compares the net present value of a social welfare function of a resource extraction policy where no regulations are imposed to a policy where regulations are imposed in order to determine the optimal timing of regulations. If his ideas are carried over to the problem of regulating the fisheries, one could then determine critical fish population levels at which the net present value of the social welfare function where no regulation policy is adopted is equal to one where fish regulations are adopted. Such a procedure would constitute a proactive approach toward fisheries management because critical fish population levels can be identified at which the social benefits of fishing would outweigh the social cost of fishing before these levels are reached. This approach can thus be used to determine if, and if so, at what population levels it is best to adopt fisheries regulations taking into consideration the social costs of imposing regulations on the fishery.

4.4.4. The impact of tradable quotas on fishing behavior An open-access fishery will lead to a situation where fishermen will try to maximize their share of the catch. Such a fishery is characterized by fishing behavior where each fisherman does not take into account the effect of their efforts on the productivity and profits of others. As long as a profit can be made fishermen will enter the fishery until an equilibrium point is reached at which no fisherman makes a profit (Clark, 1990). Clearly this point is economically inefficient, as the industry is not maximizing its profit. Often such a point will also result in biological overfishing, because fish populations are below levels at which maximum sustainable yield is maintained (Clark, 1990). To counter the inefficiency of open-access fisheries, individual transferable quota (ITQ) systems have been introduced. ITQs give fishermen a share of a TAC, which is determined by the government. It should be noted that, although various ITQ arrangements exist and are actively used as management tools, some problems are associated with ITQs (Copes, 1986). In many countries, fishermen are given ITQs or fishing rights for a certain time horizon. All fishermen can then decide whether to fish, sell their ITQ, or lease their ITQ for a certain period. Yet managing a fishery using ITQs may not improve economic

66

PART I: TOOLS AND BASIC INSIGHTS

efficiency if the impact of uncertainty on trading behavior of fishermen is not taken into account. Uncertainty regarding future TAC levels, output price fluctuations, cost fluctuations, and stock dynamics will have an impact on how fishermen value quotas. Grafton (1994) studied the impact of output price uncertainty on quota trading in a fishery where over a certain time horizon each period can be seen in isolation. In his model, output price uncertainty does not fluctuate with time and is characterized as a spherical and symmetric probability distribution. Each period a TAC is set and risk-averse heterogeneous fishermen have to decide what quantity to fish during that period. They do this by maximizing an expected profit function of the form: E(π ji ) = pq ji − ce(q ji ; b) − γ q ji − β j qi2j σ p2

(12)

where π ji is the profit of a fisherman i using technology j, p is the price paid for fish, q ji is the quantity he decides to fish, c is the cost of fishing per unit of effort, e(·) is an effort function, γ is the quota price, β j is a risk-averse parameter, and σ is an uncertainty parameter. The last term on the right is a risk premium associated with random output prices. Given the TAC and taking first order conditions of Equation (12), a solution to this maximization problem can be found for each fisherman. Grafton concludes that increased price uncertainty will lead risk-averse fishermen to value quota’s less and may lead to reduced profits in the fishery. Real options theory, which is the extension of financial option theory to options on real assets, can provide useful insights into the impact of output price uncertainty on investment decisions, such as the purchase of fishing rights. If stochastic price fluctuations follow a geometric Brownian motion, increased price uncertainty will lead to fishing rights being valued higher. At the same time, prospective entrants to the fishery will require higher profits from fishing before they agree to purchase fishing rights. The value of the option of waiting to enter the fishery increases with uncertainty in price. As a result, there will be less investment in purchasing fishing rights even though they will be more highly valued. See Dixit and Pindyck (1994) for an elaborate discussion concerning the impact of price uncertainty on investment decisions. This fishing behavior is illustrated by Weninger and Just (2002) who studied the entry and exit behavior of fishermen in a fishery. In their model, there is uncertainty regarding the cost of fishing instead of prices paid for fish. Fishermen have to make a decision whether or not to sell their permit, which has an indefinite life span. Because expected fishing costs are assumed to grow over time as a result of, for example, wear and tear, fishermen will eventually exit the fishery and sell their quota to fishermen who are willing to invest in the fishery. In their model, cost fluctuations follow a geometric Brownian motion with drift of the form dc (13) = µ dc + σc dz c where c denotes the cost of fishing, µ is the expected growth rate (positive), σc is the cost volatility, and dz is an increment of the Wiener process. Maximizing an active fisherman’s expected present value requires considering at each time t whether to continue fishing, or stop and sell the fishing right. Such a problem is analogous to an optimal stopping problem, which can be solved by writing the net present value

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 67 function as a Bellman equation and using Ito’s lemma to solve (see for example, Dixit and Pindyck, 1994). Under competitive conditions, prospective entrants to the fishery will require that expected payoffs from entering the fishery equal the cost of purchasing fishing rights and their investment costs. Weninger and Just (2002) find that compared with the deterministic case (σc = 0), stochastic cost fluctuations result in higher permit prices and in active fishermen fishing longer, thus delaying the exit of unproductive capital. One of the conclusions that can be drawn from Weninger and Just (2002) is that when governments design an ITQ scenario to improve economic efficiency, an underestimation of the level of uncertainty in the fisheries can lead to suboptimal individual and social outcomes. As a result, quota trading may be sluggish while inefficient fishermen may be active for far too long. This does not mean to say that the general existence of uncertainty in fisheries will cause ITQs to be an inappropriate tool to manage the fisheries. Rather, managing a fishery using ITQ scenarios firstly entails taking uncertainty into account and secondly entails observing how fishermen respond to ITQ scenarios in their fishing and trading behavior (Gudmundsson and Anderson, 2000; Asche, 2001). A lot can be learned from observing the behavior of fishermen, which will improve the design of the ITQ arrangement. For example, if ITQs can be leased on an annual basis and the ratio of the value of an ITQ into perpetuity over a leased ITQ is high, then discount rates are also high (Asche, 2001). This could imply that fishermen consider fishing a risky business because stocks are low. If this is true, the government could lower the TAC, which would reduce the volatility on the market for quotas and improve fish stocks. As such, managing a fishery by way of an ITQ scenario can be an adaptive process. The fact that ITQs are generally based on a TAC, which is determined in each fishing period using stock assessments at the end of the previous fishing period, can undermine the effectiveness of ITQs. Stock assessments will evidently lead to uncertainty about actual escapement levels. The term escapement is generally used to denote the proportion of the fish population that survives the fishing season. Furthermore, uncertainty generally exists regarding the relationship between surviving spawning adults at then end of a fishing period and recruitment in the next period. As a result, setting a TAC at the start of a fishing season can be considered a risky business. Weitzman (2002) illustrates this point in a fishery with ecological uncertainty regarding fish population dynamics. In his model, a choice has to be made between regulating the fishery by imposing landing fees or by issuing quotas. Landing fees regulate the escapement levels of a fishery by controlling marginal fishing effort, whereas a quota system regulates the harvesting levels under great uncertainty about escapement levels. Weitzman shows that a system of landing fees imposed per unit of fish yields a higher expected discounted present value for a profit-maximizing fishery than a quota system does. Weitzman argues that in a fishery where marginal costs of fishing increase as fish populations decrease, a landing fee system does not need to take account of the ecological uncertainty because a landing fee can be set that controls fishing effort in order to achieve any desired escapement level. In contrast, a quota arrangement controls how much will be harvested based on an estimated population size with significant uncertainty about escapement levels.

68

PART I: TOOLS AND BASIC INSIGHTS

4.4.5. Marine reserves and metapopulations Marine reserves can serve to protect fish species from being overfished. Spillover effects from the closed area to adjacent fishing grounds may improve fish stock levels and benefit the fishery in the long run (Roberts et al., 2001). Nevertheless, determining the appropriate size of marine reserves is not straightforward (Boersma and Parrish, 1999). Metapopulation and spawner–recruitment dynamics of spatially distributed fish species are generally not known with certainty (Jennings, 2001). Furthermore, without restricting fishing effort in the fishing grounds, it is likely that the establishment of marine reserves will lead to an increase in fishing effort in the open areas (Sanchirico, 2000). Since the exact response of the fishery and its impact on the fish population are both uncertain, the design of marine reserves is further complicated. Sumaila (1998) studied a model where a fisheries habitat can be divided into an area open to fishing and a marine reserve. The model allows for positive spillover effects from the marine reserve to the fishing ground in the form of migratory fish. Using data on the North-East Atlantic Cod, he shows that in the event of a significant catastrophe due to recruitment failure it may be beneficial to a profit-maximizing fishery to establish a marine reserve, but only if the spillover effects are high. The model shows that the higher the spillover effects are, the smaller the marine reserve has to be in order for it yield higher profits as compared to the case where no marine reserve is established at all. Conrad (1999) and Harford (2000) also studied a simple model of a profit-maximizing fishery where spillover effects between a marine reserve and a fishing ground could occur. In contrast to Sumaila’s model, recruitment was modeled as a stochastic process: a logistic population growth equation was used with a uniformly distributed growth rate. Economic advantages to the establishment of marine reserves were not as apparent as in Sumaila’s simple model. This suggests that different (non-use) benefits of marine reserves should be identified in order to justify and support the establishment of marine reserves. Thus, it seems that simple representations of complex ecosystems and fishing behavior can yield different points of view as to the advantages of marine reserve creation. Bulte and van Kooten (1999) presented a model where migration occurs between two local populations. Migration is modeled in such a way that the smaller of the two stocks will be replenished by the other. The dynamics of both populations are subjected to random fluctuations characterized by a stochastic disturbance term following a Wiener process. In this way account was taken of environmental and demographic uncertainty. The objective was to maximize the value of the harvest and non-use values over an infinite time horizon subject to the two stochastic dynamic equations. In order to find the optimal population levels, stochastic dynamic programming techniques were used. Bulte and van Kooten showed that harvesting decisions for both populations cannot been seen in isolation. This is because due to the migratory effects the size of each of the two populations depends on the other. The optimal situation is such that the population considered to be more valuable should be more heavily harvested. The population that is valued less will then fill the niche left behind. As in the case of the model presented by Pindyck (1984) (discussed above), the effect of stochastic fluctuations on harvesting levels of both fish populations is indeterminate. Under similar assumptions of concavity and convexity, however, Bulte and van Kooten show that

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 69 if stock fluctuations of both populations are negatively correlated, this will lower the impact the risk premium will have on optimal harvesting levels. This is due to the fact that in this case a decrease in one population coincides with an increase in the other, which has a buffering effect on the risk premium as presented by Pindyck (1984).

4.5. POLICY RESPONSES TO UNCERTAINTY Changes in the ecosystem as a result of fishing will bring about changes in fisheries behavior and viceversa as both the ecosystem and the fishing community continuously react and adapt to each other and represent an interdependent dynamic system (Allen and McGlade, 1987; Peterson, 2000; Settle et al., 2002). Uncertainty in what influences the fishing behavior and marine ecosystem dynamics may make it difficult to understand how one will influence the other and to what extent. An important concept here is resilience. Fishing activities can be expected to have an effect on the stability of the fish population, but uncertainty makes it unclear how large this effect may be. If fish population dynamics are such that multiple stable states exist, management strategies can force a fish population from one steady state to another. Holling (1973) defines resilience as the capacity of an ecological system to remain in a stable state following perturbation. Each steady state has a basin of attraction. If a shock to the system is such that the population is shifted from its current level to another population level within this basin of attraction, the population will eventually return to its steady state. If the population is pushed beyond the limits of attraction, it will move to another stable state. Probabilities can be assigned to various management options that will force a fish population from one steady state to another (Perrings, 1998). Thus these probabilities will give an estimate of the resilience of the fish population in its current stable state to outside shocks. Perrings (2001) has given an example of the problem management is faced with when it wants to maximize a net revenue function, but is confronted with discrete fish population dynamics having multiple steady states. When choosing a management strategy, the uncertain impact of that strategy on the resilience, and thus the sustainability of the fish population, needs to be taken into account. The failure of most models used to underpin fisheries management in a sustainable way has led to a call for decision making to take place under “reserved rationality” (Perrings, 1997). This describes the decision-making processes where incomplete information about the probability distribution of potential outcomes of management strategies makes it natural to proceed cautiously and avoid potentially severe outcomes, such as irreversible damage to the fisheries ecosystem. The increase in attention for the precautionary principle is a direct consequence of management making decisions under reserved rationality. It involves a risk-averse attitude toward managing fisheries under uncertainty, where the focus is on maintaining marine ecosystem functions rather than on harvesting fish resources. It also involves a change in the burden of proof for justifying management action, placing the responsibility on the fishing sector to prove that restrictive measures may be too restrictive, rather than on the environmentalists. Under precautionary management, a lack of complete scientific certainty is therefore not used as an excuse to delay actions that prevent degradation of the marine environment (Buhl-Mortensen and Toresen, 2001; Weeks and Berkeley, 2001).

70

PART I: TOOLS AND BASIC INSIGHTS

It is important to note that although a precautionary approach to fisheries management is intuitively appealing, it is often a conceptual idea that is hard to execute in practice because of political pressure. An example is given by the Dutch cockle fishery. In the Netherlands, the shellfish sector became intensively regulated in the early 1990s when shellfish stocks showed signs of depletion, which had negative spillover effects on other species. Yet the regulations imposed have not had the desired effect. Under heavy pressure from the environmentalists the government has funded research to study the impact of shellfishing on the marine ecosystem. Pending the results of this investigation, cockle fishing at present continues at levels deemed unsustainable by the environmentalists. This is in part due to the fact that imposing limits on the fishing sector is a politically sensitive issue. As part of a precautionary approach to managing fisheries adaptive management scenarios can be used as they embrace uncertainty and treat management decisions as hypotheses that need to be tested, evaluated, and adjusted accordingly. Bayesian estimation methods can play an important role since they combine observational data with management strategies. Ellison (1996) concludes that an iterative Bayesian learning and decision-making process is analogous to the concept of adaptive management as both incorporate learning processes. Bayesian methods can help fishery management to consider actions that take the form of feedback-control decision rules (Punt and Hilborn, 1997), since the consequences of alternative management actions can be analyzed given the estimated status of the stock and hypotheses reflecting uncertainty about population dynamics. Iterative Bayesian learning procedures can be set up where the consequences of management actions can be observed (as posterior distributions) and used as new information (new prior probability distributions) for ongoing decision-making processes. Management actions are refined as managers learn how their decisions impact the fisheries. Such an approach to manage the fisheries is attractive because it combines the key features of stock assessment models and economic models generally used to model the fishery: observational data regarding complex ecosystems are coupled with possible management actions that take the economic behavior of fishermen into consideration in order to find feasible fishing strategies in a complex ecosystem about which there is much to learn.

4.6. CONCLUSIONS In comparing the models discussed in Sections 4.3 and 4.4 two distinctions become clear. Firstly, most of the models discussed in Section 4.3 are assessment models where uncertainty is accounted for when trying to gain insight in the relationship between fish stock dynamics and harvesting pressure. Given this uncertain relationship, harvesting levels can be chosen such that their impact on the ecosystem is minimal. Most of the models discussed in Section 4.4 can be seen as bioeconomic models where uncertainty in the fisheries is accounted for when trying to determine profit-maximizing fishing behavior. Secondly, the stock assessment models of Section 4.3 are generally focused on specific localized fisheries recognizing that fisheries take place in complex ecosystems. Most of the bioeconomic models of Section 4.4 are not focused on localized fisheries, and hence tend to be more descriptive in nature when it comes to describing fishery

4. UNCERTAINTY IN ECONOMIC ANALYSIS OF MARINE ECOSYSTEM EXPLOITATION 71 characteristics. As such, stock assessment models rely heavily on observational data, whereas bioeconomic models do not. Where fisheries management is based on stock assessment or bioeconomic models in order to manage a fishery sustainably under uncertainty, it is likely to underestimate the human–ecosystem interactions that occur in a fishery. Stock assessment models place much emphasis on the ecological aspects and underestimate human reactions to changes in the marine ecosystem. These models do not tend to ask questions like “why are fish stocks being fished down?” or “what provides the incentive for fishermen to overfish?” Bioeconomic modeling in the fisheries is likely to do just the opposite. It may focus too much on the economic aspects of the fishery, giving too little attention to ecosystem complexity. Not taking account of minimum viable population levels, for example, may drive certain fish stocks to extinction or to population levels from which they may not recover. Bioeconomic models may not take sufficient account of the fact that low stocks are increasingly sensitive to stochastic fluctuations such as catastrophes, genetic inbreeding, demographic, and environmental fluctuations as they are fished down. An inappropriate representation of fish population dynamics can lead to serious mismanagement. Modeling stock–recruitment relationships as being density dependent when in fact they are not will lead to less conservative fisheries management, which may bring the population close to collapse. Most models that address uncertainty in fishing behavior do so by focusing on the uncertainty in the availability of fish that will impact their catch levels. Few models address uncertainty in economic parameters that will affect fishing behavior. Furthermore, no economic models encountered focused on the impact of changing management regimes on fishermen’s objectives. This is nevertheless a crucial point that will affect fishermen’s behavior. The fact that fisheries models generally are imperfect does not mean to say that it is impossible to manage fisheries successfully. In fact, quite a number of fisheries have been successfully managed based on quota restrictions and effort levels. But many fisheries have also been mismanaged. Therefore, it is important to rely on more than just modeling approaches to manage the fisheries. Policies based on a precautionary approach under which these models are used need to be adopted. This will limit the consequences a lack of knowledge will have on fisheries sustainability. A precautionary approach may not initially lead to the best sustainable profit-maximizing policy. Nevertheless, as understanding regarding fish populations, ecosystem dynamics and fishing behavior improves, the ecological and economic limits under which fisheries are sustainable will be better understood. Thus in the long term a precautionary approach seems to be the best sustainable policy for both the fishing sector and society.

CHAPTER 5

MANAGING THE FISHERIES: A SYNTHESIS OF OLD AND NEW INSIGHTS

5.1. INTRODUCTION After a period of substantial investment of resources in the fishing industry, a significant portion of fish populations worldwide is either overfished or is on the verge of being overfished. Catch levels tend to be unsustainable because economic factors driving the fishing industry are not in balance with ecological factors influencing fish population dynamics. Historically, fishing grounds were considered to be an open-access resource: fish resources were not subject to ownership and access to the fishing grounds was completely unrestricted. Thus anyone wanting to fish was welcome to do so. Since World War II fisheries worldwide have expanded rapidly. The incentives to overfish came from several sources. First, new technological advances occurred during wartime, such as larger and more powerful vessels and synthetic fibers for nets. These made fishing cheaper and more effective and therefore entrance to the fishing grounds more attractive. Second, as the world population and the standard of living in some countries increased, demand for seafood rose. This encouraged the fishing industry to expand rapidly and fill in the niche of supplying consumers worldwide with the fish they wanted. In the 1990s, an increasing demand for fish continued to boost overcapitalization despite lower supplies of fish on the market. Garcia and Newton (1997) argue that, despite diminished catches, expansion of the fishing capacity on a global scale made economic sense as the total revenue of the fishing industry worldwide grew faster than the total amount of capital invested in the business. Third, many governments encouraged the fishing industry to expand for the benefit of the national economy. Encouragement to develop the fishing fleet was offered in the form of subsidies like low interest loans and tax benefits by the government, which drew capital rapidly into the fishing industry. National governments and fishermen alike seemed to be gripped by the misperception of oceans being so vast that the fish resources it contained could not be eroded and hence, need not be managed in order to prevent overexploitation. Fourth, increased competition amongst fishermen in combination with the increased capacity to fish contributed to fish stocks showing signs of depletion. Driven by uncertainty about the fish stocks in the future as well as a mentality that can best be described as “if I don’t catch the fish someone else will,” a race for the fish started that drew even more capital into the fisheries and depleted fish stocks even further. 73

74

PART I: TOOLS AND BASIC INSIGHTS

Recognizing that a major cause of depletion of fish stocks was overfishing, the United Nations Convention on the Law of the Sea (UNCLOS) was established in 1982 (United Nations, 1998). This treaty recognized the sovereign rights of coastal nations over the living marine resources within 200 nautical miles off their shoreline. Based on a sort of property rights notion it turned many coastal open-access fisheries into national property regimes with the intention to encourage coastal nations to manage and conserve fish populations in their waters using the best scientific information available. In many countries where no formal fisheries management structure existed, UNCLOS paved the way for autocratic fisheries management structures to emerge that were characterized by centralized governmental decision-making without participation from the public and the fishing sector (Mc Goodwin, 1990). Nevertheless, it should be noted that some countries, such as Spain and France, have traditionally had longstanding fisheries management structures with strong engagement from the fishing sector (Jentoft and McCay, 1995). The initial impact of UNCLOS was the reduction of foreign competition. For some coastal nations, this simply led to changing the problem of open-access from being an international to a national problem. Although most countries initially benefited from the elimination of foreign fishing within their coastal waters, important fish stocks became depleted as a result of excessive catches, which in many countries were allowed to continue under a national property regime (FAO, 1985; Jentoft et al., 1998; Caddy and Cochrane, 2001). The lack of appropriate governmental action to manage and conserve the fisheries has led to increased pressure from environmental groups, scientists, and fishermen on the government to conserve and restore the depleted fish populations as well as to alleviate the hardship suffered by the fishing community. To revitalize both the fish population and the fishing industry is no easy task because the economy and the environment have requirements for growth that are almost never congruent (Hanna, 1997). Nevertheless, sustainability of the fisheries does require that the government somehow bring the human exploitation systems prevalent in the fisheries more in line with the natural environment of the fisheries and its growth characteristics. If the government wants to restore the balance between the fisheries and the fishing industry, it must remove the incentives for overfishing in the fisheries and replace them with incentives for conservation and sustainable management. This is harder than it seems. Contrary to the way many fisheries have been managed, effective fisheries management will need to take explicit account of uncertainty and human behavior, and will need to be proactive rather than reactive (Ludwig et al., 1993; Folke et al., 1996; Weeks and Berkeley, 2001). Actions curbing fishing activities could be undertaken upon the first signs of depletion and should not be delayed by scientific research. Focusing on the several causes of overfishing mentioned above, the government could either adopt a policy influencing market conditions, change the institutional framework used to manage the fisheries, or employ both options. The purpose of this chapter is to determine the factors that need to be considered to set up a successful fisheries management framework. The organization of the remainder of this chapter is as follows. Section 5.2 examines existing ideas on governmental control in fisheries policy and management, separating between centralized and market based regulations. Section 5.3 addresses

5. MANAGING THE FISHERIES

75

the co-management approach to governmental interference with fisheries. Section 5.4 deals with the question how to correctly frame fishery management, notably in terms of key factors such as temporal and spatial scales, and users and characteristics of exploited marine ecosystem. Section 5.5 presents a case study of the Common Fisheries Policy (CFP) in the EU in which the foregoing issues are examined in more detail, thus clarifying why the CFP has failed. Section 5.6 concludes.

5.2. GOVERNMENT CONTROL IN FISHERIES MANAGEMENT 5.2.1. Centralized government control Traditionally, governmental policy has been aimed at expansion of the fisheries. Now in many fisheries it is aimed at contraction. Reducing catch levels toward a level at which fisheries can be sustained means that in the short run measures need to be implemented that are going to hurt the fishing industry and can be expected to be met with some reluctance. If the government adopts a centralized approach aimed at reducing fishing effort, it can expect to encounter several problems regardless of the measures taken. First, reluctance amongst fishermen to comply with new regulations may be encountered. Centralized governments have a limited capacity to oversee local circumstances and seasonal variations within different regions and sectors of the fishing industry (Noble, 2000). As a result, it is likely for any stakeholder of a specific fishery targeted by some fishing management policy scheme to be skeptical about the government’s ability to set up an efficient fishery management policy. Fishermen may mistrust any fisheries regulation measure imposed by the government without any input from them as actors on the local level. They may feel that any regulation decided upon without their involvement will lack some realism because local experience-gained knowledge is missing without their involvement in the management process (Jentoft et al., 1998). In Denmark, for example, fishery policy decisions are made at a centralized level without any real involvement from the fishermen themselves. Nielsen and Mathiesen (2000) have shown that in the Danish fisheries fishermen’s compliance to fishing regulations is partly influenced by fishermen’s perception of the meaningfulness of regulations regarding stock conservation. As far as fishermen’s safety is concerned, Kaplan and Kite-Powell (2000) argue that conservation regulations may put fishermen under increased pressure as they struggle to comply with regulations and to make a profit, which could result in higher risk-taking and unsafe behavior at sea. Fishermen may have different perceptions regarding safety at sea than the government and therefore may choose not to comply with regulations they feel put them in danger. Clearly, there are situations where it is true that fishermen know more than the government and scientists regarding some aspects of the fisheries. But the reverse must also be true in some cases because otherwise there would be no need for government intervention at all. Nevertheless, irrespective of who is more knowledgeable about the fisheries, imposing fisheries regulations requires a motivational change on behalf of the fishermen because otherwise regulations may not be accepted. Such a motivational change may be hard to bring about when fishermen play no role in the fisheries management process.

76

PART I: TOOLS AND BASIC INSIGHTS

The second problem the government can expect to encounter is closely related to the first problem. When fishermen are solely motivated by economic gains that can be derived from the fishery, fishermen will often consider violating regulations if a significant economic gain can be expected (Sutinen et al., 1990). In fact, Nielsen and Mathiesen (2000) concluded that economic incentives were the most important reason for Danish fishermen to violate regulations. In order for a regulation to be credible therefore, the government needs to impose an enforcement and monitoring structure that will dissuade fishermen from illegal fishing practices. A low probability of detection of illegal fishing activities coupled with high penalties would keep the costs of enforcement and monitoring low for the government. This is generally not feasible due to the fact that courts are often not willing to impose penalties on fishermen that are excessively severe. Therefore, imposing a fisheries regulation mechanism can be expected to yield high enforcement and monitoring costs for the government (Sutinen and Kuperan, 1999) or else be limited in its effectiveness. This also implies that as long as economic gains from violating regulations are high it is reasonable to expect that some form of external governmental control will be necessary, irrespective of the fisheries management approach adopted. A third problem that a centralized approach toward fisheries management may face is related to the execution of governmental decision-making. Strategic long-term decision-making concerning the desired sustainable fishing policy generally occurs at the highest governmental level, usually national. The operational details that are to ensure the success of the strategic fishing policy are usually left at the discretion of regional or local governmental bodies responsible for fisheries management. Regional and local governmental institutions can be far more susceptible to regional and local socio-economic conditions and allow short-term interests of local fisheries to prevail above long-term strategic interests of national fisheries. In the Netherlands, for example, local public authorities were accused in the mid-1980s of being involved in the noncompliance with quota regulations set out by the EU in certain North Sea fisheries (Dubbink and van Vliet, 1996). This can also be seen as a problem of scale: concerns regarding fisheries generally cover a larger spatial and temporal range on the national level than they do on the local or regional level. Strategic fisheries policy-making on a national level ideally should address the aggregate of all concerns shared by the regions and localities constituting that nation. All regions constituting a nation, however, do not necessarily have to express the same level of socio-economic dependency on the fisheries and thus regional governmental institutions responsible for day-to-day fishing operations may disagree with the government as toward what constitutes a good fishing policy. Thus, governments who set out a fisheries policy without taking into consideration regional and local concerns may experience regional or local dissent. The problems mentioned above should make the government reluctant to manage a fishery by itself that has no input from stakeholders and that does not acknowledge the problem of scale that can arise between centralized and local decision-making. Although the government is responsible for ensuring a sustainable fishery that addresses the ecological, economical, and social concerns of all stakeholders involved it is going to be very difficult to persuade all stakeholders in the fishery to adopt a policy that has been decided upon without having involved them in the decision-making process. If the government is to manage the fisheries in order to maintain ecological, economical, and

5. MANAGING THE FISHERIES

77

social health it is going to need the cooperation, expertise, and knowledge of all users as well as biologists and socio-economic scientists to determine what management approach at what scale should be adopted.

5.2.2. Market intervention by the government Advocates of government intervention in the market see the lack of individual property rights as the fundamental source of fisheries problems. They argue that introducing a system that assigns property rights to fishermen could enhance economic efficiency because it allows fishermen with property rights to make a profit that would not be dissipated by others: with property rights fishermen are guaranteed a proportion of the catch that they can harvest in a cost effective way without having to race with others for the fish (Clark, 1990). As long as fishing is profitable, introducing a property right mechanism in the form of individual transferable quotas (ITQs) is also considered as something that will enhance economic efficiency because fishing rights would be transferred to the most efficient fishermen as they purchase the fishing rights of less efficient fishermen (Copes, 1986; Dubbink and van Vliet, 1996). Another argument often brought forward in favor of property right mechanisms is that they can contribute to converging private and public interest by enabling fishermen to act as private owners of a resource having an interest in conservation and sustainable management (Hanna et al., 2000). This generally implies that under certain conditions the behavior of fishermen aimed at maximizing profits is socially efficient. But whether private owners have an interest in conservation and sustainability is a matter of discount rates. Depending on these, private and public interests converge or diverge (Clark, 1973; Asche, 2001). Fishermen’s fishing behavior may also be motivated by other factors than profit maximization, such as the existence of social norms and the appeal of fishing (Saraydar, 1989; Steelman and Wallace, 2001). These can also cause private and public interests to converge or diverge. Moreover, local and family traditions as well as lack of alternative employment may also play a role. Thus, introducing a system of private property rights on the market would involve placing a lot of trust in the market to ensure that private and public interests do not diverge, which as Dubbink and van Vliet (1996, p. 513) point out, is misplaced. They state that the market can only perform “if it is supported by a strong government vested with substantial responsibilities” which continuously has to make sure that public and private interests do not divert. In other words, the government has to keep an eye on the conditions under which the market operates at all times and adjust it in such a way that all private and public interests are met, now and in the future. Such governmental stewardship in the marketplace is cumbersome to say the least as the fisheries are shrouded in uncertainties. Incomplete information concerning temporal and spatial fish population dynamics and how fish species interact with their ecosystem make it difficult to predict future fish stocks at the local, regional, and national levels. Furthermore, the exact effect of human interactions with the marine ecosystem is usually uncertain. Thus, the government cannot expect the market to manage the fisheries in an effective way because the market cannot provide the public and the fishing community with the information on how commercial fish populations change over time. If the government

78

PART I: TOOLS AND BASIC INSIGHTS

would rely on a system of property rights, such as ITQs, to manage the fishery in a sustainable manner it probably would find itself in a position where it would have to intervene in the market place continuously as it cannot rely on the market to signal the effects fishing will have on the fish population. This does not mean that property rights mechanisms must be ruled out as instruments that can be used to create sustainable fisheries. Rather, it means that the government should not rely on a system of property rights alone to manage the fisheries. Furthermore, if the government introduced a system of private property rights into the fishery it would still need the support from the public and the fishing community. Whatever the legal foundation of the introduction of private property rights may be, if it lacks legitimacy because it is not founded on some moral and social beliefs, it is likely to be challenged on grounds of social injustice (Copes, 1986; Jentoft, 2000). In the case of the European Community (EC), for example, ITQs were not considered a management option because many countries in the EC had moral objections against setting up a fishery management structure that would remove the public property nature of the fisheries (Ministerie, 1999).

5.3. MANAGING FISHERIES BY WAY OF CO-MANAGEMENT The appropriate institutional framework to manage the fisheries is seen by many to be a co-management approach in which the government to some extent shares management concerns and responsibilities with other stakeholders in the fishery in order to develop sustainable fisheries over a long-term horizon (Jentoft et al., 1998; Jentoft, 2000; Lane and Stephenson, 2000; Noble, 2000). Jentoft et al. (1998, pp. 423–424) define co-management as “the collaborative and participatory process of regulatory decisionmaking among representatives of user-groups, government agencies and research institutions.” User-groups consist of people that derive value from a fish population. Often user-groups are considered to be people who have a commercial interest in the fishery. Nevertheless, those that place an existence value on the fish population as well as future generations who have an option value on the resource must also be considered as user-groups. On the one hand, co-management involves bringing all parties that have an interest in the fisheries together in order to determine how they are to be managed on a sustainable basis by taking action to enhance stocks, by challenging existing fishing practices, and by examining new fishing techniques that are on the one hand more selective and on the other hand more efficient than current techniques. On the other hand, co-management involves increasing the profitability of the fishing industry as well as improving the viability of the sector as a whole. The institutional framework for co-management can take on different forms with varying degrees of responsibility allocated between user-groups and the government in the decision-making process of how to manage certain aspects of the fisheries at the national, regional, or local level. Not all aspects are necessarily subject to user-group involvement as the economic, social, or political incentives to undertake fisheries management responsibilities are not always evident to user-groups (Nielsen and Vedsmand, 1999). An institutional framework for co-management can be set up when the government is unwilling to deal with the social and political issues that result when it introduces a fisheries management plan that is

5. MANAGING THE FISHERIES

79

scrutinized by user-groups (Sen and Nielsen, 1996). Alternatively, it may be costly to impose and enforce fishing regulations, which can be circumvented by delegating some responsibility to the various user-groups. Generally, however, a co-management scenario is introduced because of a legitimacy issue: user-groups should have a larger say and control in decisions affecting the resources on which they are dependent, in cooperation with government, economic, and administrative functions (Noble, 2000). The literature on voluntary agreements between user-groups and government agencies in an industrial setting (normally in relation to pollution control) displays multiple reasons why government agencies and user-groups may have an interest in engaging in a collaborative arrangement to minimize the consequences of user-groups actions on the environment (for an overview, see Segerson and Li, 1999). Voluntary agreements used in the industrial environment may be seen as a form of co-management because all parties involved in a voluntary agreement generally are endowed with some responsibilities by the government toward minimizing the impact of user-group actions on the environment. Some reasons for establishing a voluntary agreement in the industrial environment can also give insight in why user-groups and the government may have an interest to engage in a co-management approach in managing the fisheries. Usergroups, for example, may be motivated to participate in a voluntary agreement because of the existence of a public centralized regulation threat (Cabugueira, 2001) where usergroups have no capacity to influence the regulation procedures and its implementation. User-groups may also be motivated to participate because it can yield certain benefits such as subsidies. Governments may be motivated to establish voluntary agreements with positive and negative inducements as long as the incentives for overfishing that the market provides remain (Segerson and Li, 1999). Another incentive for user-groups to engage in a voluntary agreement is to get a competitive edge over their competitors (Cabugueira, 2001). In a heterogeneous fishery, where fishermen use different gear in order to catch fish, this may be an important incentive for certain fishermen to influence regulations that may affect the type of fishing gear used to their advantage. In order to achieve a sustainable fishery where so much about the resource and the impact of various fishing techniques used is unknown, it is imperative to have cooperation between user-groups, scientists and the government over long periods of time. Corrective measures to stabilize the fisheries on a short-term horizon are insufficient due to the uncertainties surrounding the fisheries and long-term preventive measures should be sought after (Aggeri, 1999). In order to achieve collaborative arrangements between the user-groups and government agencies, voluntary agreements can exists due to the existence of credible regulation threats from the government on the one hand and the existence of economic incentives on the other. Economic incentives for fishermen to cooperate in a co-management setting can be due to the increased security that reduced uncertainty in the fisheries can bring in the form of more stable fishing populations. Another economic incentive can be provided by the decrease in uncertainty about the consequences of fishing activities on the fisheries resources and the potential discovery of better fishing technologies that can result from research as part of the collaborative management process. Fishermen will be able to invest in fishing gear that is more suitable to the fishery and is less subject to stringent fishing regulations.

80

PART I: TOOLS AND BASIC INSIGHTS

Nevertheless, setting up an adequate co-management system where all participants are expected to sustain fish stocks as well as to respect the social and cultural values of fishing communities is no easy task. The question as to what constitutes an appropriate institutional framework to facilitate co-management has been extensively discussed in the literature (Jentoft and McCay, 1995; Sen and Nielsen, 1996; Noble, 2000). A prerequisite for a potentially successful co-management scenario is that the government is willing to allow user-group involvement of some form in the fisheries management process and that user-groups are willing and able to participate in the decision-making process. The ability of user-groups to participate in the management process is facilitated when user-groups are organized in such a way that they have a common objective, which users in a user-group do not always have. For example, fishermen differ in terms of certain characteristics, such as the fishing gear and type of vessel used in the fishery. When they are to participate in a co-management scheme as one user-group they can be expected to have non-identical and competing views and therefore different and conflicting objectives. Representatives of such user-groups participating in a comanagement arrangement may find it hard to adopt a flexible position in which there is room for compromise with the government on the one hand and other user-groups on the other. The involvement of user-groups in the fisheries management process is also facilitated when they have the organizational skills required to carry out management functions effectively or to participate in the decision-making process directly. Some user-groups are designed for the purpose of handling administrative matters and not necessarily for the purpose of participating in fisheries management (Nielsen and Vedsmand, 1997). Involvement of these groups in the management process may not be constructive as they may slow down the management process or may not be able to carry out management tasks effectively. Involving user-groups in a co-management arrangement who are either unwilling or unable to participate in a co-management framework effectively is likely to yield an ineffective fisheries management scenario. This is supported by the review of case studies by Sen and Nielsen (1996): in situations where the government has singlehandedly imposed a co-management scenario upon unorganized user-groups, the goal of creating sustainable fisheries has not (yet) been achieved. If the government desires involving user-groups in the management process, they could encourage the formation of user-groups in the case where they were previously non-existent, or contribute toward the development of the essential management and administrative skills required in order for user-groups to organize and participate in the co-management process effectively. Encouragement can be provided by positive and negative inducements by the government. The EU, for example, encourages the formation of Producers Organizations possessing the management skills that enable them to coordinate their fishing activities with demand from the market. These Producers Organizations are granted the privilege of withholding fish from the market to influence the price for fish, a privilege that other fishermen do not have. Other ways to encourage participation in a collaborative fisheries management are economic incentives such as subsidies or more stringent fishing regulation threats.

5. MANAGING THE FISHERIES

81

What is often omitted in fisheries management scenarios is a clear objective of what is to be achieved. Seldom specific targets are set that define the desired outcome of fisheries management plans. Having a clear objective is important because it provides a benchmark against which the performance of a management plan can be evaluated. If targets cannot be specified at the outline of the management plan this needs to be clear to all user-groups, while the consequences of perceived failure of the management plan, which could result in further stringent management measures, have to be accepted by all. In a co-management scenario, a clear objective is an indication of a shared vision and a willingness to work together for a common good. This is sometimes not easily achieved given the fact that user-groups can be expected to have different objectives. Objectives of an environmental organization may not coincide with fishermen’s objectives. Due to the difficulty associated with reaching consensus, the objective of a management plan is often kept broad which can lead to its ineffectiveness. For example, the objective of a fishery management plan is often stated as being the creation of a sustainable fishery. The problem here is that the dynamics governing fish populations and the impact fishing has on fish populations is often unclear, which makes the determination of what exactly a sustainable fishery is troublesome, to say the least. Furthermore, what exactly constitutes a sustainable fishery can be something that user-groups differ upon. Fishermen and conservation activists may have different views here. Jentoft (2000) notes that in many cases environmental organizations have long-term ecological objectives that do not necessarily coincide with short-term profit-maximizing objectives of fishing communities. The government can potentially provide a valuable contribution toward establishing a clear and defined common objective by acting as an intermediate between user-groups, ensuring that communication between user-groups proceeds smoothly, and making sure that user-groups understand each other’s concerns (Grafton, 2000).

5.4. FRAMING THE FISHERIES MANAGEMENT PROBLEM CORRECTLY Between the two extremes of having the government being responsible for fisheries management on the one hand and user-groups on the other, lies a whole array of comanagement scenarios where the extent of participation and decision-making in fisheries management plans by user-groups and the government differs. Where fish stocks are overfished, questions have arisen as to what the right approach toward fisheries management should be. As noted earlier, UNCLOS in many cases led to the establishment of centralized governmental decision-making regimes toward managing the fisheries. The failure of many centralized regimes to manage the fisheries successfully has given rise to the premise that the appropriate institutional framework to manage the fisheries is one incorporating some form of co-management arrangement. Whilst this is probably true for many fisheries, it does not have to be true in general. A failure of a management process controlled by the government may lead to overfished fish populations and thus to unsatisfied fishermen who have little faith in the ability of the government to manage the fisheries, but it does not have to mean that co-management is the ultimate solution

82

PART I: TOOLS AND BASIC INSIGHTS

to the management problem. Rather, it suggests that the government has framed the problem of managing the fisheries incorrectly. The appropriate management scenario for a fishery depends on who the users of the resource are and on the characteristics of the resource being exploited. The characteristics of the resource being exploited are influenced by the ecological interactions between the fish population itself and its environment, which also includes the fishermen. A fisheries management scenario is therefore primarily influenced by the geographical scope chosen to manage the fisheries as this determines the users involved and sets the boundaries that define what spatial interactions between the fish population and its ecosystem need to be accounted for. There is no point in setting up a fisheries management framework without having asked these two questions first: which fish species need to be managed and who are the parties concerned? Once it has been decided which fish species are in need of management, the appropriate scale at which they should be managed needs to be determined. Following Gibson et al. (2000), the term “scale” will be used in this chapter to refer to “the spatial, temporal, quantitative, or analytical dimensions used to measure and study objects and processes.” The question of how to determine the scale of the problem of managing a fish species is therefore a strategic one. What is the appropriate timeframe that a fisheries management scenario should cover and what is the size of the fish population to be managed? Determining the appropriate timeframe of a strategic fisheries management plan is no easy task because there generally exists high uncertainty as to what the factors are that determine the fluctuations in the fish population. There is little doubt that fishing pressure is generally a large contributor to fluctuating fish populations, but little is known about what the exact effects of fishing are. Furthermore, fluctuations in the fish population are also the result of natural causes, about which little is known. Thus, determining the appropriate timeframe over which a fish population is to be managed is troublesome, to say the least. Determining the spatial scale of the fish population that needs to be managed should involve consideration of the population dynamics of the fish population and how it interacts with the marine ecosystem. The poor performance of fisheries management in general can be the result of ignorance regarding the fine-scale aspects of the fish populations being managed (Stephenson, 1999; Wilson et al., 1999). For example, ignoring potential metapopulation characteristics of a fish population can have disastrous effects: not accounting for the existence of small localized spawning grounds can result in extinction of the fish population being managed in the area under consideration. The analytical dimensions under which fish populations are managed is also increasingly scrutinized: managing fish species without recognition that they are part of a larger marine ecosystem is considered a fallacy. Yet management policies are bounded by geographical, political, and socio-economic concerns, which makes sound management based on the biological population structure of stocks often very complicated. After having determined the scale of the fish population that needs to be managed, the government needs to focus on who the parties are that will be affected by fisheries management. This is a question of level. Following Gibson et al. (2000), the term “level” is used to refer to a region “along a measurement dimension.” The government needs to recognize who the affected user-groups of the fisheries are at the international, national, regional, and local level. Where it exists, fisheries management at the international

5. MANAGING THE FISHERIES

83

level generally consists of a set of goals agreed upon by multiple nations fishing in international waters in order to maintain a viable fish population. In the Atlantic Ocean, for example, certain migrating species are subject to international management plans. In the EU, the CFP has been in place for the last 20 years. The CFP has amongst its objectives to ensure a fair standard of living of the fishing community within the EU, to assure the availability of fish stocks and to increase productivity by promoting technical progress (Commission of the European Communities, 2001a). Whilst management objectives are generally strategic in nature at the highest level, they need to be refined to tactical plans with measurable and quantifiable objectives as the scope of the fisheries management policy is downscaled from national to regional and local levels. The EU has recognized the necessity of this and has adopted the principle of subsidiarity that states that action should be taken at the lowest level possible in order for fishing policy to respond quickly to local concerns. Downscaling the level at which fisheries management is to take place involves addressing the concerns of the local environment and user-groups in more detail. This includes setting boundaries between localities that need to be managed at a lower level. Setting boundaries involves taking into consideration social, economic, and ecological characteristics, which is generally considered to be the task of the government. Ecological concerns that need to be considered when setting boundaries are the spatial characteristics of the fish population, such as metapopulation and migratory characteristics, as well as the interactions of the fish population with the marine ecosystem. When setting boundaries, the government also needs to take into consideration the fishing patterns of the fishing industry, recognizing the fact that fishermen are not bounded to a single area. Should boundaries be based on where those having an interest in fish resources reside or where they are active? Another socio-economic concern when determining boundaries is the degree of dependency on the fishing industry of certain areas. Some areas are more dependent on the fisheries than others so that fisheries management plans may need to address local concerns differently per region. For example, in the EU there is a policy of contracting the fisheries in well-developed regions where alternative employment outside the fishing industry can be found: subsidies are granted to projects that reduce fishing pressure. In southern European coastal regions, artisan fisheries are still being maintained and kept afloat through subsidies, simply because alternative employment outside the fishing industry in these regions is scarce (Commission of the European Communities, 2001a). Setting boundaries is a complex task when fisheries management is downscaled. Boundaries must be chosen carefully so that they cover relevant ecological and socio-economic concerns at the lower level as much as possible: something that becomes increasingly difficult as the geographical scope of managing fisheries becomes smaller and smaller. Boundaries that are well chosen will facilitate user-group participation at that level in the management process since user-groups can be expected to understand the premises upon which the boundaries are set, see their concerns addressed within these boundaries, and share their experience with usage of that resource for the benefit of the fishery as a whole (Wade, 1987). After having determined the scale of the fisheries to be managed as well as who will be impacted at what level, the government is responsible for setting up an institutional framework for fisheries management at all appropriate levels. It is at this point that the issue of co-management should be brought up whereby the government

84

PART I: TOOLS AND BASIC INSIGHTS

should assume responsibility in bringing together the different user-groups and create arrangements where management problems can be dealt with adequately (Symes, 1997). It is generally agreed upon that in some way or other participants must contribute to, and be represented in, policy setting at the strategic (long-term) as well as the operational (intra-seasonal) level in a participatory co-management approach (Lane and Stephenson, 2000). Participants should be able to come together and set clear objectives for fisheries management that are measurable at all levels. But the role user-groups and government have in the decision-making process are different on each level and context dependent (Steelman and Wallace, 2001). At any rate, the government has an important active role at the national and lower levels because it has to make sure that all consumptive and non-consumptive interests of current users are taken into consideration based upon their geographical location and their usage of the fisheries resource. It has to coordinate and assess the impact interactions of fisheries management plans at a lower level can have and if necessary react and make sure that prompt action is taken to sudden unforeseen detrimental changes to the fishery. It has to make sure that local management decisions do not effectively curtail options open to future generations. At all levels the government can ensure that experience is exchanged, information is shared, and rivalry over the use of fisheries is overcome, all of which will promote better management at all levels.

5.5. A CASE STUDY: THE CFP IN THE EU In managing European fisheries, the EC has as its objective to maintain the fisheries at a sustainable level where fragile fish stocks in European waters are protected and where the fishing industry is assured of an existence. This objective has been formulated in the CFP. The principle of relative stability that the EC has tried to establish in European Fisheries has manifested itself in the allocation of quotas amongst its member states. Each country is allocated a fixed proportion of the TAC for each fishery, which is supposed to be a fixed percentage based upon their fishing history. In this way, the EC tries to maintain the balance of power in the fishing sector that was there before its formation. All fishing countries within the EU, whose fisheries are subject to quotasetting, will proportionally be affected in the same way by any reduction proposed in the TAC and benefit from any increased in quota as a result of recovering fish populations. This section will look at some of the problems the CFP has encountered and see if they could have been prevented had the insights derived in the preceding sections been considered. In managing the fisheries the EC utilizes two tools. The first tool is in the form of setting a TAC based upon advice provided by an independent party, namely the International Council for the Exploration of the Seas (ICES), which itself has delegated the responsibility to the Advisory Committee of Fisheries Management (ACFM). The second tool is reducing the amount of fishing effort exerted in the fisheries. In order to achieve the objected reduction in fishing effort, the EC has periodically introduced fishing capacity reduction plans called Multi Annual Guidance Programs (MAGP’s). Currently, MAGP IV is effective and over its first 3 years enabled EU countries to reduce fishing activity by reducing the number of days spent fishing at sea or by reducing

5. MANAGING THE FISHERIES

85

the fleet capacity. As opposed to earlier MAGP’s, the option for countries to reduce their fishing activity through limiting the number of days spent at sea was a new one brought on by the failure of certain segments to reduce their fleet capacity targets. Nevertheless, the current capacity reduction program does not meet its objectives because several member states are not reducing their fishing capacity, opting to reduce their fishing pressure by reducing the number of days spent at seas alone (Commission of the European Communities, 2000). After a preliminary review of MAGP IV, EU countries are now faced with binding fleet capacity reduction requirements for the remainder of the program. Since its implementation in 1983 the Common Fishing Policy is flawed with errors. At the supra-national level decisions are made concerning the TAC of overfished fish stocks based on input provided by the ICES. Yet, the European Council of fishery ministers structurally sets the TAC higher for most depleted stocks than is advised. On the one hand, this is due to the uncertainty associated with fish stocks data. On the other hand, this appears to be due to social-economic considerations the ministers are faced with in their countries. There seems to be an understanding that fishery ministers set the TACs levels as highly as can be compromised upon in order to limit the short-term effects TAC levels will have on their fishing communities, as well as to reduce costs and problems associated with enforcement of compliance (Daan, 1997). This seems to suggest that the time scale over which the fisheries are managed causes a conflict between the two main elements that need to be managed in order to yield a sustainable fishery, namely the fish population and the community dependent on the fisheries. The short-term effects that TACs have on the community are more clear and tangible than the long-term effects of the TAC on the stability of the fish population. There are also concerns associated with the effectiveness of the TAC as a management tool in European waters due to the fact that current fishing practices in the EU make it very difficult to measure what exactly has been caught. Faced with restrictions on quota high grading may occur where smaller fish are discarded in favor of larger and more valuable fish. Furthermore, some fisheries are confronted with bycatch that is not part of their quota and are therefore subsequently discarded. In the case of the flatfish fishery in the North Sea, for example, fishing techniques used result in the seemingly inevitable bycatch of cod that has to be discarded because the flatfish sector generally has no quota for this species. Thus, the actual fishing mortality for cod, for which a quota exists in European waters, is actually much higher than landing figures might suggest. This obviously undermines the process of performing stock assessments because the data on catch statistics becomes flawed. Setting TAC’s in the EU is based on commercial catch and effort data and research vessel survey data (Karagiannakos, 1996). Thus, landing figures can be expected to yield a higher recommendation for catch than might be appropriate. The ACFM is aware of the fact that due to the incongruity between landings and catch any advice it gives on TAC is flawed and potentially destructive to the fishery and has pointed this out to the EU. Yet, the EU insists that advice is given on TAC’s because it is the only legal tool the EU has in setting restrictions on catch (Karagiannakos, 1996; Daan, 1997). The ACFM also gives information concerning the state of fish stocks in the EU in the form of spawning stock biomass (SSB). If the SSB is above a level at which the fishery is deemed to be sustainable, a range of catch options can be provided for under which the fishery is expected to remain stable. Over the

86

PART I: TOOLS AND BASIC INSIGHTS

period 1980–1994, the degree of association between landings in the EU and SSB was much higher than the degree of association between landings and the recommended TAC, or for that matter landings and actual TAC set by the EU (Karagiannakos, 1996). One may therefore conclude that in managing overfished stocks, the EU would be better off to use information concerning SSB as the basis from which to determine quotas for fish species in their waters. The objective of the CFP is not only to protect overfished fish populations, but also to protect local fish communities dependent on the fishing industry. By setting TAC’s generally higher than the recommended level, EU ministers are essentially trying to limit the destabilizing effect TAC’s might have on fishing communities in their countries. Targets set in MAGP IV for a reduction in fishing capacity have deviated from the proposed targets because they were considered to be too ambitious. Thus, the approach of reducing fishing pressure seems to have led to a conflict on the supra-national level between the objectives of protecting the fish populations on the one hand and the fishing industry on the other. This conflict of objectives is further aggravated on the national level in most countries within the EU (Banks, 1999). In some cases the national government is actively participating in this conflict. The Dutch government, for example, has argued that effort reductions through limiting the number of days spent at sea is enough to counter the excess in fishing activity and that a reduction in fishing capacity is unnecessary (Ministerie, 1999). However, from an EU point of view an excess in fleet capacity is not compatible with the permanent reduction in fishing activity that is sought (Commission of the European Communities, 2000). The EU wants a permanent reduction in fishing activity by permanently reducing the fleet capacity. Reducing fishing activity by limiting fishing effort (in terms of days active at sea) is considered by the EU to be non-permanent because the capacity for fishing at unsustainable catch levels remains. Success of the capacity reduction program can be further undermined on a national level because the fishing industry does not operate as a coherent institution that is transparent and easy to manage. It is in effect made up of different types of fishing vessels and fishing companies that are active in different fisheries that are subjected to different fishing conditions. Although the fishing industry as a whole might agree with the fact that there is too much capacity, individual fishermen and fishing companies might become threatened in their existence when faced with these restrictions as individuals. The decommissioning scheme introduced in the United Kingdom in 1993 (Hatcher, 1997; Banks, 1999), for example, resulted in a reduction in fishing capacity primarily in areas where the fisheries were either in terminal decline or where quotas and days at sea were not fully utilized. It seemed to have virtually no effect on reducing fishing capacity in the sectors were fishing was big business. Thus although there was a reduction in overall fishing capacity, there was no real reduction of fishing capacity that was actively being used. In fact because the UK attached quotas to fishing licenses on the one hand and allowed the aggregation of licenses on the other (subjected to a penalty in total capacity reduction), more productive fishermen could absorb underutilized quota. This in effect resulted in a failure to reduce capacity in sectors where it really mattered. Thus, the objective of the CFP to safeguard fish stocks threatened with collapse through imposing TACs on the one hand and imposing fishing capacity reduction plans

5. MANAGING THE FISHERIES

87

on the other, has encountered problems on the supra-national level as well as on the national level. There is agreement at all levels that something needs to be done to improve the state fish stocks are in. On the supra-national level, there is disagreement on the impact the tools should have. The ACFM as an outside party and the European Commission are not on one line with the European Council of Fisheries ministers as to what the impact of the CFP should be. This seems to be a direct result of the fact that fisheries ministers are players at two different levels in the EU fisheries management process concerned with fisheries management on different scales. Their roles at an EU level can conflict with their roles at the national level because they gain their legitimacy as representatives of national interests and not EU interests. The reduction of fishing capacity strives toward reducing the impact the fishing industry has on overfished fish stocks. By reducing capacity, fish populations will hopefully be sustained at acceptable levels where the fishing industry and fishing communities in the EU are guaranteed a healthy existence with no fear for the future. It is clear that the tools the EU uses to manage the fisheries aim to balance the impact of fishing restrictions on fishing communities based upon traditional fishing patterns in the past, in order to ensure that fishing communities in each country are sustained in the long run. The European Union is responsible for determining the impact these tools should have in each country. The national governments are responsible that capacity reduction targets are met and that their national quota for overfished fish is not exceeded. Yet each country is free to determine how these targets are met and how the national quota is distributed. This, together with the fact that EU fishing fleets have, broadly stated, a tremendous freedom in choosing their place of establishment within the EU (Morin, 2000) has resulted in the fact that stability of local fish communities is no longer ensured. EU fishermen and fishing companies can in effect choose to register their vessel(s) in any EU country and thereby gain access to that country’s fishing quotas. This procedure, commonly known as reflagging, can cause great hardship on local fishing communities as quotas issued for the benefit of local fishing communities are on the one hand fished by fishermen originating from outside this local community and on the other hand potentially processed by the fishing industry elsewhere. Spanish fishermen, for example, have often been criticized for buying licenses (and obtaining the quotas attached to these licenses) from English fishermen who were not making enough money. It has been argued furthermore that the subsequent reflagging of Spanish vessels as British vessels has lead to Spain meeting its capacity reduction targets, whereas Britain has not, resulting in an even higher pressure on the British fishing sector to reduce capacity. As Morin (2000) has argued, the principle of relative stability does not match with the right of EU workers to work anywhere within the EU. The principle of relative stability does not preserve fishing communities because the incentives fishermen have to work elsewhere are not removed. Given the objectives of creating one European economic zone the principle of relative stability seems to be a flawed one. If fishing communities are to be protected than this should be done on principles which would guarantee the exclusive access to certain fishing resources within the EU to certain local communities displaying a high level of dependency on these waters. The strategy to manage fisheries formulated at the EU level does not preserve and protect local fishing communities. Furthermore as responsibilities to manage the

88

PART I: TOOLS AND BASIC INSIGHTS

fisheries are delegated to the national level, the process becomes even more flawed as countries differ in their approach as to how targets determined at the EU level should be met and how to enforce compliance with these targets. Fishermen on the one hand could benefit from this by choosing to fish under the flag where fishing conditions given the national policy are most beneficial to them. On the other hand, fishermen could also be increasingly frustrated by the differences among the EU countries in how regulations are enforced and violations are handled. National EU governments have on numerous occasions tried to protect their fishing industry in vain by introducing legislature that would promote some form of limited exclusiveness to their quota for national fishermen, which turned out to be in conflict with EU law. The CFP thus encounters many operational problems when it comes to providing socio-economic stability within the fishing sector in the EU as it also does when trying to manage overfished stocks. This is for a large part due to the lack of homogeneous operational procedures detailing how the CFP should be carried out at the national level. Again fisheries ministers find themselves operating at the supra-national and the national level, where they are faced with conflicting objectives. The objective of providing economic stability in the fisheries sector becomes flawed because the transboundary nature of the fishing sector at the national level has not been sufficiently recognized. Therefore, managing the fisheries sector at the national level following a supra-national management plan becomes problematic. Set against this background, the principle of relative stability does not meet its objective, namely to continue and protect the existence of national fishing sectors. The principle of relative stability only makes sense if the CFP would promote exclusive access to certain regions heavily dependent on fish resources. If one were to maintain this principle, this would lead one to conclude that either the CFP is using the wrong tools (or is missing a tool in managing the fisheries), or that the objective of providing socio-economic stability in the fisheries sector is ill defined. In the latter case, it seems that the objective of the CFP as far as providing socio-economic stability in the fisheries sector is concerned should be twofold: to provide socio-economic stability in the transboundary fishing industry within the whole EU and to provide socio-economic stability to national regions heavily dependent on the fisheries. The previous sections dealt with critical issues that needed to be addressed in order to frame the fisheries management problem correctly. One can set the CFP against the background of the discussion of the preceding sections. Then, it becomes clear that the CFP has encountering problems that are related to stating a clear objective of the CFP that is commonly agreed upon by all parties concerned. Environmental scientists consistently argue that proposed TAC reductions are not sufficient for creating sustainable fish populations. Some governments argue that the capacity reduction plans set out at the EU level are too ambitious. Fishermen feel that the CFP lacks some realism because their input is very limited. Their limited involvement leads them to having concerns about their future. Thus, the objectives of the CFP can hardly be considered common objectives shared by all user-groups. The CFP is also confronted with problems related to scale. On a temporal scale, national governments are too concerned about the short-term effects the CFP might have on their national fishing industry, inhibiting the long-term recovery path that needs to be followed in order to enable fish populations in EU waters to recover to sustainable levels. On a spatial scale, the EU fails to recognize

5. MANAGING THE FISHERIES

89

the mobility of EU fishermen, requiring national governments to manage their national fisheries in order to achieve goals set out at the EU level. As far as problems related to carrying out the CFP at all levels are concerned, national fisheries ministers seem to have conflicting roles at the EU level and the national level. The CFP at the supra-national level almost resembles a centralized management approach that is encountering the very problems as addressed in Section 5.2: there is little user-group support toward the CFP and fishermen are reluctant to comply with regulations that follow from the CFP. This is partly aggravated by the fact that national fishery management plans in order to meet EU targets differ from country to country. As a result, regulation and monitoring of fishermen is carried out differently, which frustrates fishermen who see that in some areas control is more lenient than in others. This encourages them to question the success of the CFP in managing the fisheries resources and the fishing sector on the one hand and the fairness of the CFP toward them on the other. In some countries, fishermen are included at the national level in co-management arrangements with the government in order to address the operational details in order to meet targets set out in the CFP, and in others they are not. Nevertheless, there seems to be little exchange of information between fishermen and those responsible for the CFP. Given the uncertainty the fisheries are shrouded in and the economic hardship many fishing communities are faced with, this is a something of a missed opportunity.

5.6. CONCLUSION Over the last couple of decades, the discussion on using collaborate arrangements between user-groups and the government in managing the fisheries has increased. Governments are frequently blamed for fisheries mismanagement and new ways to manage overfished fisheries are sought where the balance of power in managing the fisheries is shared with other users. It is not right, however, to say that a centralized management regime will lead to fisheries mismanagement. Nor is it right to say that collaborative approaches will lead to more successful fishery policies. It is true that in many cases centralized governmental control has failed to create sustainable fisheries and healthy socio-economic conditions within the fishing sector. But it is also true that in many cases this was a direct consequence of UNCLOS introducing property rights regimes in places where there was little experience with fisheries management in the first place. If anything, increased experience and frustration with fisheries management regimes worldwide has shown that the uncertainty surrounding the dynamics governing fisheries resources on the one hand and the impact fishermen have on the fish population on the other does not call for short-term regulations in the fishing sector. It has also shown that a long-term management plan having the support of all user-groups involved can generate a common understanding regarding conservation measures in the fisheries and facilitate compliance. This chapter has discussed both the centralized approach and the collaborative approach for fisheries management. It has highlighted why both government agencies and user-groups might have an incentive to manage the fisheries resources together: it is in their mutual interest to create sustainable fisheries and sustainable fishing populations, and to minimize conflicts between all usergroups involved. The degree of power sharing between the government and user-groups

90

PART I: TOOLS AND BASIC INSIGHTS

depends on a number of factors. Most importantly, it depends on the nature of the fishery in question, the amount of uncertainty the fisheries are shrouded in, and the characteristics of the user-groups involved. Fisheries management needs to be formulated addressing the economic and ecological, as well as the temporal and spatial scales of the problem appropriately. Furthermore, as part of the management process, responsibilities of all user-groups and government agencies need to be outlined at all levels. The case of fisheries management in the EU has addressed several pitfalls a badly formulated management plan may face. The objectives of the CFP in terms of capacity reduction are not supported or difficult to execute at lower levels. The objectives of the CFP in terms of desired fish population levels are also disputed. Governments seem to have different objectives at the supra-national level and the national level. Furthermore, the issue of scale is addressed incorrectly. Once targets have been agreed upon centrally at the EU level, the national governments are expected to manage their fishing sector nationally, despite the fact that fishermen have the freedom to fish anywhere they like in European waters. When some form of collaboration is desired, it is imperative that there is a common goal, which all user-groups and the government agree upon and against which the success of a management plan can be measured. The option of excluding “dissenting” user-groups from a fisheries co-management scenario in order to get an agreed upon common objective may invoke public dissent and may mean that interests of excluded users, now and in the future, are not adequately taken at heart. Therefore, if a clear-cut objective is missing, it may be best to adopt a more centralized management approach where the government represents the interest of all user-groups. Both the government and user-groups in question need to realize and accept that successful fisheries management relies on a successful institutional management framework that may be subjected to change as the circumstances the fisheries are in change. Marine ecosystems are generally complex and far from stable. Faced with a large amount of ecological uncertainty, fisheries management plans should adapt and evolve. Increased experience in fisheries management can lead to more stable fisheries management regimes as the thresholds within which exploited fish populations can be said to be sustainable and subject to natural fluctuations are better understood.

PART II

BIOECONOMIC MODELING

91

CHAPTER 6

HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

6.1. INTRODUCTION1 The subject of harvesting in predator–prey systems has been of interest to economists and ecologists for some time now. Most research has focused attention on optimal exploitation, guided entirely by profits from harvesting. The present study emphasizes that the ecosystem offers more than the exploitation of a species. This is expressed through a non-use value, which is indirectly influenced by harvesting. Clark (1976), Hannesson (1982), Ragozin and Brown (1985), and Str¨obele and Wacker (1995) derive golden rules for optimal steady-state harvesting in a multi-species context. In addition, Ragozin and Brown (1985) study the approach path toward the optimal steady state. Semmler and Sieveking (1994) show that an optimal constant harvesting effort can result in a trajectory that does not reach equilibrium but oscillates over time. Except for Tu and Wilman (1992), who consider the stability of an ecosystem for predator control programs in combination with harvesting of prey, no attention has been given to nature protection policies in multi-species systems. Olson and Roy (1996) have studied extinction in a discrete single species model in which the harvest as well as the stock is considered valuable. The present study is the first that combines harvesting and conservation of different but ecologically related species. The present chapter seeks to find optimal exploitation strategies for a predator–prey system, but differs in three respects from the previous studies. First, we investigate the possible extinction of the non-harvested species. Prey is harvested while predators are protected. More specifically, in contrast to other studies, we explicitly value the existence of a species, in this case the predator. Although the predator species is itself not harvested, incorporating its existence value is necessary because it is indirectly affected by the harvesting of its food. Secondly, the ecosystem model is based on a more realistic specification of predation than the familiar Lotka–Volterra type. In particular, it includes search and handling of prey by predators in order to derive the ecosystem dynamics. A critical review of different types of predator–prey relations is found in Berryman (1992) and Yodzis (1994). Thirdly, a very complete analysis of possible optimal harvest regimes is given. The chapter investigates for which economic and ecological parameters the predator will survive.

1

Reprinted from J. Hoekstra and J.C.J.M. van den Bergh (2005). Harvesting and conservation in a predatorprey system. Journal of Economic Dynamics and Control 29(6):1097–1120.

93

94

PART II: BIOECONOMIC MODELING

Although the formal model analyzed will be fairly general, it is motivated by a specific conflict between shellfisheries and the conservation of bird species in many European estuaries such as the Dutch Wadden Sea (Atkinson et al., 2003; Ens, 2003). Therefore, we will mostly refer to birds and shellfish in this chapter. Nevertheless, the model can be used to describe general predator–prey ecosystems and even herbivore– plant systems, as long as their interactions obey what biologists refer to as a Holling type II functional response (Yodzis, 1989 and Section 6.2.1). This chapter is an endeavor to theoretically investigate the balance between exploitation and nature conservation. This balance is reflected in the social welfare function. The social welfare function consists of two terms. One expresses the income generated by harvesting prey. The other represents conservation benefits derived from the presence of a certain amount of predators in the ecosystem. Predators serve as a proxy for the nature values of the ecosystem. We derive the optimal harvest rate that maximizes the social welfare function. An optimal trajectory approaches either a fixed steady state or an optimal cyclic state Ragozin and Brown (1985). The optimal harvest trajectory is divided into two parts: an end state and the approach path. We find the optimal end state(s) and give necessary conditions for its (their) existence. In addition, we present a qualitative analysis of the approach path toward the end state. The structure of the remainder of this chapter is as follows. Section 6.2 describes the ecosystem model and analyzes its behavior. Optimal harvesting is examined in Sections 6.3. Section 6.4 provides a qualitative analysis of the approach paths to the end state. Section 6.5 concludes.

6.2. THE ECOSYSTEM MODEL 6.2.1. Model description The ecosystem model describes the interactive dynamics between predators (birds) and prey (shellfish). The net growth rate of birds is described by Equation (1). The first term expresses growth, which depends on the birth rate, r . The second term expresses decline, which depends on the mortality rate expressed as life expectancy that on turn depends on energy intake.   e0 B B e0 B dB = rB − = rB 1− (1) dt e b r eb where B is the number of birds; r the birth rate of birds; e0 the reference energy intake rate per bird; e the total energy intake rate; and b the reference life expectancy (1/b is the reference mortality rate). Equation (1) is the logistic equation (Yodzis, 1989) and r eb/e0 is called the carrying capacity. The energy intake depends linearly on the amount of shellfish eaten. It is given by: e = cd

(2)

where c is the energy content of a shellfish and d the number of shellfish depleted. The depletion depends on the time a bird needs to search for shellfish and the amount of time it needs to break open the shellfish and swallow its contents. The search time decreases as the number of shellfish becomes larger. This is expressed in the following

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

95

equation: d=

B BS = z/S + h z + hS

(3)

where S is the number of shellfish; z the search time coefficient and h the handling time (per bird per shellfish). This equation is known as Holling’s disc equation or a type II functional response (Yodzis, 1989). Equation (3) shows that the amount of food intake per bird is asymptotically limited. Indirectly, therefore, also the life expectancy of birds has an upper limit. Shellfish die due to being eaten by birds, being harvested by fishers or because of other factors such as lack of food, cold winters, etc. Biologists call the process of larvae turning into new adult shellfish, recruitment. Shellfish recruitment does not depend on stock size. Measurements in the Wadden Sea do not show a strong relationship between recruitment and the adult stock size of shellfish (Beukema, 1993, van der Meer, 1997). Some of the reasons are that larvae arrive in the area by wind driven and tidal currents. Sometimes adult shellfish cannibalize larvae and good settlement spots become scarcer as the stock size increases. Their survival depends on a number of factors such as sea bottom characteristics, food supply, temperature, and preying fish. These factors are, however, too complex to be included in our model. The number of shellfish is given by: dS S =q − −d −y dt a

(4)

where q is the shellfish recruit rate; a the average shellfish age (1/a is the mortality rate), if not eaten by birds or fished, and y the shellfish harvest rate. Substituting Equations (2) and (3) in (1) and (4) and creating the parameter K = r cb/e0 , results in the following predator–prey system: dS BS S = f S (S, B) − y = q − y − − dt z + hS a

(5)

  dB z + hS = f B (S, B) = r B 1 − dt KS

(6)

and

The definition of f S (S, B) and f B (S, B) are introduced for notational simplicity. Besides results expressed in f S (S, B) and f B (S, B) are of general form. They apply to any predator–prey system in which prey is harvested. The parameter K shows the time a bird needs to eat (find and handle) a unit of shellfish in equilibrium. If a bird needs more time, due to a long search time (z/S) because of low stock sizes of shellfish, the numbers of birds will decrease. If it needs less time, due to high shellfish numbers, the numbers of birds will increase. Because we assume shellfish is the only food birds eat, the mass of birds cannot increase more than the mass of eaten shellfish. Hence   dB z + hS BS = rB 1− ≤d= dt KS z + hS

96

PART II: BIOECONOMIC MODELING

For a sufficiently small numbers of shellfish, S ≤ z/(K − h), this condition is true because (1 − (z + h S)/K S) is negative. For larger numbers of shellfish, the condition is true if KS S r ≤ min S z + h S K S − (z + h S) If K − 2h ≤ 0 we find the minimum at S → ∞ and r ≤ K / h(K − h), or if K − 2h > 0 the minimum is at S = 2z/(K − 2h), so that r ≤ 4/K .

6.2.2. Ecosystem behavior under fixed harvest rates The behavior of non-linear systems can be characterized by standard equilibrium, stability, and bifurcation analysis (Strogatz, 1994). Stability analysis in predator–prey models has received widespread attention, in the case of Lotka–Volterra type models, e.g. May (1972). Brauer and Soudack (1979) studied the behavior of a general predator– prey system (which includes Lotka–Volterra models and the model in this chapter), where the predator is harvested at a constant rate. Dai and Tang (1998) examine a more specific model in which predators and prey are harvested. These studies show that, depending on parameter values a stable equilibrium, a limit cycle or a homoclinic2 orbit is possible. The ecosystem described here has two equilibrium points ( S˙ = 0, B˙ = 0): Be1 = (q − y)K −

z Kz , Se1 = (K − h)a K −h

(7)

and Be2 = 0, Se2 = (q − y)a

(8)

Note that, in the first equilibrium, the number of shellfish does not depend on the harvest rate, y. If the harvest rate increases, the number of shellfish in equilibrium does not decrease. Extra harvesting takes food away from birds therefore the ecosystem can sustain fewer birds. This is typical for a predator–prey system with a Holling type II functional response to describe the predator’s dependency of prey. In the second equilibrium, birds have become extinct. As long as (q − y)a > z/(K − h), the existence of the first equilibrium is guaranteed. Fishers can fish and a certain bird population size can be sustained. Figure 6.1 shows the phase diagram of the system. It contains an example trajectory, the S- and B-isoclines, and the direction of the system’s vector field. The B-isocline ( ˙B = 0), is given by S = z/(K − h) = Se1 and by B = 0. The number of shellfish does not change on the S-isocline ( S˙ = 0), which is given by   S (z + h S) B= q−y− a S From the isoclines a vector field is deduced that indicates the direction of the trajectories. 2

In a homoclinic orbit a seperatrix of a saddle point originates from that saddle point.

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

97

B S=0

B=0

B e1 0

B=0

S S e1

S e2

Figure 6.1. Phase diagram of the ecosystem under a fixed harvest rate. Note: An example trajectory of the simple predator–prey system is shown. The isoclines are shown in gray and the equilibrium points as black dots. The arrows denote the direction of the trajectories.

The arrows show that a trajectory will either tend to, or originate from, or circle around the first equilibrium, (Se1 , Be1 ). In Section 6.2.3 we show that every trajectory approaches equilibrium. The second equilibrium (Equation 8) is a saddle point and can only be reached when the number of birds becomes zero. The S-isocline depends on the harvest rate. When fishing increases, the S-isocline moves down and the number of birds in equilibrium (Be1 ) consequently declines. Figure 6.2 shows the case when the harvest rate becomes so large that the isoclines do not intersect in the quadrant where S and B are positive. When y > q − z/((K − h)a), then (Se2 , 0) is the only feasible equilibrium.

6.2.3. Stability analysis From an analysis of the direction of the vector field, it follows immediately that (Se2 , 0) is a saddle point. Equilibrium (Se1 , Be1 ) is Lyapunov stable, when the real part of all the eigenvalues of the Jacobian, D f (Se1 , Be1 ), are negative. ⎤ ⎡ S ∂ f (S, B) ∂ f S (S, B) ⎢ ⎥ ∂S ∂B ⎢ ⎥ D f (S, B) = ⎢ ⎥ ⎣ ∂ f B (S, B) ∂ f B (S, B) ⎦ ∂S

∂B

Lyapunov stability means that a trajectory will stay within a finite distance from the equilibrium whenever it comes close enough to that equilibrium. The eigenvalues

98

PART II: BIOECONOMIC MODELING B B =0

S =0

0

B =0

S S e2

Se1

Figure 6.2. Phase diagram of the ecosystem under a high fixed harvest rate. Note: The harvest rate is so high that birds become extinct. The isoclines are shown in gray and the single equilibrium point as a black dot. The arrows denote the direction of the trajectories.

of are:

with

 1

D ± D 2 + 4a K r z(K z + D) 2a K z

(9)

D = −(q − y)a(K − h)2 − hz < 0

(10)

A negative real part of the eigenvalue means that the equilibrium is stable. Equation (10) shows D is negative. From Equation (9) we can see that the eigenvalue will have a negative real part if the square root term is smaller then D. That is the case when K z + d < 0, which leads to z < (q − y)a (11) (K − h) This condition is satisfied because Se1 < Se2 (Figure 6.1) so that the equilibrium is stable. When the eigenvalue is complex, the equilibrium is a stable focus. If the eigenvalue is real, (Se1 , Be1 ) is a node. The eigenvalue is real if the square root term is positive. The square root term, D 2 + 4a K r z(K z + D), is a parabola in D. For r < 1/a, it has no roots and is positive. Thus, the equilibrium is a node if the birth rate for birds is smaller than the additional mortality rate for shellfish. For r ≥ 1/a, the parabola has two  roots: namely, D1 = 2K z( ar (ar − 1) − ar ) and D2 = −2K z( ar (ar − 1) + ar ). The eigenvalues are real, i.e. the square root term is positive, for  2K (ar − ar (ar − 1)) − h D ≥ D1 ⇒ Se2 ≤ Se1 K −h

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

99

and for D ≤ D2 ⇒ Se2 ≥

2K (ar +

 ar (ar − 1)) − h Se1 K −h

Hence, the system has a node if the two equilibria are either close together or far apart. We have shown that the system is locally stable by looking at the linearized system. We now prove that the system is also globally stable by showing that the system does not tend to a limit cycle. Using the Bendixson–Dulac criterion (Strogatz, 1994), we can prove that the system does not have a limit cycle in phase space. If such a closed trajectory C exists then

   S˙ · n dl = 0 B˙ C

with n the outward normal on C. The dot product must equal zero because the trajectory follows C. Green’s theorem yields:      

 S˙ S˙ · n dl dA = g(S, B) ∇ · g(S, B) ˙ B˙ B C

A

with A, the surface enclosed by C. So, if we can find a function g(S, B) for which the sign of the integrand is always positive or always negative over at least A, then the surface integral must be unequal to zero. Consequently, this means C is not a trajectory. Taking g(S, B) = (z + h S)/(K B S), then   ∂g(S, B) ∂g(S, B) S˙ ∇ · (g(S, B) ˙ = ( f S (S, B) − y) + f B (S, B) B ∂S ∂B h S 2 + (q − y)az 0 and ymax < q. By observing the vector field in Figure 6.2, it is clear that the equilibrium (Se1 , Be1 ) must be globally stable. The system is locally stable, so any trajectory starting at an initial point in a neighborhood  of (Se1 , Be1 ) will tend toward the equilibrium. Suppose that a trajectory starting at a point outside  moved away from the equilibrium. Due to continuity of the system, a trajectory starting on the border of  has to follow that border, which is impossible because it would imply the system has a limit cycle. Therefore (Se1 , Be1 ) must be a globally stable equilibrium. One can get a feel for how the system behaves under dynamic (non-fixed) harvest rates if one considers such a system to approach a moving target: namely, the changing equilibrium because it depends on y. The consequence of global stability is that fishing will not irreversibly change the system. As long as birds are not made extinct by excessive fishing of their food supply, the system is able to recover. Once fishing is stopped, the system will asymptotically approach its natural equilibrium. =−

100

PART II: BIOECONOMIC MODELING

6.3. OPTIMAL HARVESTING AND ECOSYSTEM VALUE 6.3.1. Problem formulation and necessary conditions Suppose a governing body manages the fishery. Its objective is to optimally exploit the resource, while taking into account the social value of the state of the ecosystem. When birds are not socially valued, it could be advantageous to catch shellfish until birds are extinct. The reason is that, in these circumstances either more shellfish are caught or a competitor is eliminated and harvest can therefore proceed at the same rate but, given the higher shellfish stock, at lower costs. The main question here is: what are the implications of a social valuation of birds on optimal harvesting? In order to address this issue, we assume that the governing body strives to maximize the following social welfare function, in which the value of birds in the ecosystem and the profits (revenues minus costs) of fishing are added and discounted over time:  ∞ J = max e−δt (v(B) + py − c(S)y)dt (13) y

0

where δ is the discount rate; v(B) the value assigned to the state of the ecosystem; p the price of a unit shellfish, and c(S) the cost of harvesting. The price of shellfish is set constant, as we assume the amount harvested in this particular area will have a negligible influence on the overall supply of shellfish on the market. The cost of harvesting is assumed to decrease with stock size, c′ (S) ≤ 0, and the value of birds to increase with their stock size, v ′ (B) ≥ 0. We assume that the harvesting costs are linear in y. The interpretation is that labor can be hired at constant cost and is needed proportionally to the harvest rate. Later, we will relax this assumption (Section 6.3.7). In addition, it is assumed that the existence of a fixed number of boats determines the maximum harvest rate (ymax ). The value function of birds is assumed to be concave or S-shaped (Figure 6.3). We apply Pontryagin’s maximum principle (Kamien and Schwartz, 1981). The current value Hamiltonian is: H = v(B) + py − c(S)y + λ S ( f S (S, B) − y) + λ B f B (S, B)

(14)

The equations that form the necessary conditions for a solution are r the maximum condition: max H ⇒ max( p − c(S) − λ S )y, y

y

0 ≤ y ≤ ymax

where ymax is the maximum harvest rate; r the familiar equations of motion for the state (5) and (6), r and the equations of motion for the co-state or shadow prices:   Bz 1 r Bz ˙λ S = δλ S − ∂ H = δλ S + ∂c(S) y + λ S + − λB 2 ∂S ∂S (z + h S) a K S2

(15)

(16)

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

101

Type I S-shape concave

R• B0

B

Type II

Figure 6.3. Graphical representation of possible solutions to Equation (28). Note: The black curves denote the RHS of (28) for two types of hyperbolic functions. The gray curves denote the LHS of (28) for a concave and for an S-shaped bird-value function.

and   z + hS ˙λ B = δλ B − ∂ H = δλ B − ∂v(B) + λ S S − λB r 1 − ∂B ∂B z + hS KS

(17)

r and because S ≥ 0 and B ≥ 0, the transversality condition: lim λ S ≥ 0,

t→∞

lim λ B ≥ 0

t→∞

(18)

The Hamiltonian (10) depends linearly on y with coefficient ( p − c(S) − λs ). Consequently, its maximum value is reached for the extremes of y. Thus, one must fish as much as possible when the shadow price of shellfish is sufficiently low (λs < p − c(S)), and not fish at all when the shadow price is sufficiently high (λs > p − c(S)). Furthermore, when λs = p − c(S), the harvest rate is undetermined. In this case, three solutions for y are possible: namely, 0, ymax , or y˜ (t) the singular control that maintains the condition λs = p − c(S). We assume there is a unique optimal path. After an initial period, the system will arrive at some end state. The optimal trajectory of the system will approach either an equilibrium or a cycle. We call this the end state. The approach path is the beginning of the optimal trajectory until the end state is reached. The remainder of this section is devoted to examining the end state. The following four end states are conceivable: 1) No harvesting: y = 0 and p − c(S) ≤ λs . The end state is in equilibrium (Section 6.3.2). 2) Maximum harvesting: y = ymax and p − c(S) ≥ λs . The end state is in equilibrium (Section 6.3.3).

102

PART II: BIOECONOMIC MODELING

3) A singular state: y = y˜ (t) and p − c(S) = λs (Section 6.3.4). This can result in two sorts of singular equilibria: one in which birds survive (Section 6.3.4.3), and one in which birds are extinct (Section 6.3.4.4), or it can result in a limit cycle (Section 6.3.4.5). 4) A bang-bang cycle, i.e. an oscillation controlled by a harvest rate that flips back and forth between the maximum, zero, and possibly a singular harvest rate: y = ymax , when p − c(S) > λs ; y = 0, when p − c(S) < λs ; and y = y˜ (t), y = 0, or y = ymax when p − c(S) = λs (Section 6.3.5).

6.3.2. Case 1: No harvesting Not harvesting is optimal, when in the equilibrium (Se1 , Be1 ), the total cost of fishing (c(S) + λ S ) exceeds the price of shellfish. This means that, at any harvest rate the loss in social value of birds would be greater than the net gain from fisheries.

6.3.3. Case 2: Maximum harvesting The second possibility is to keep harvesting at the maximum level. In this case, the price of shellfish must exceed the total costs. If the maximum harvest level is relatively small (ymax < q − z/(a(K − h)), see Equation (7)), then the system will asymptotically reach the equilibrium (Se1 , Be1 ). But if ymax ≥ q − z/(a(K − h)) then the system approaches the other equilibrium (Se2 , 0), in which birds are extinct. The condition p − c(S) > λ S and the transversality condition (14) imply that p > c(S). So, obviously fishing must make a monetary profit.

6.3.4. Case 3: The singular state 6.3.4.1. General singular harvesting The third possibility is an end state in which the total system (state and co-state) is kept in a singular state. From the maximum condition (11), it follows that at this end state we have to satisfy: λ S = p − c(S)

(19)

Substituting (15) and its derivative ∂c ∂c λ˙ S = − S˙ = − ( f S − y˜ ) ∂S ∂S in (16) gives the following expression for λ B in the singular state,  B    ∂f ∂c S ∂f S f + λ B = ( p − c) δ − ∂S ∂S ∂S

(20)

provided ∂ f B /∂ S = 0 (i.e. B = 0). We take the time derivative of expression (20) for λ B and substitute it, together with (19) in (17). This eliminates both shadow prices and we get the following expression that implicitly defines y˜ , the harvest rate in the singular

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

103

state:  2    ∂ c S ∂c ∂f S B ∂f S S S ˜ ˜ ( f − y ) f + − δ ( f − y ) + f 2 ∂ S2 ∂S ∂S ∂B  B 2 S 2 S ∂ f ∂f ∂ f − ( p − c) ( f S − y˜ ) + fB 2 ∂S ∂ B∂ S ∂S   2 B    S ∂ f ∂2 f B B ∂f ∂c S S ˜ f × ( f − y ) + f − ( p − c) δ − + ∂S ∂S ∂ S2 ∂ B∂ S   B     S B ∂f ∂f ∂c S ∂ f = δ− f ( p − c) δ − + ∂B ∂S ∂S ∂S     2 ∂v ∂f S ∂f B − + ( p − c) (21) ∂B ∂B ∂S We have ignored a function’s variable list so as not to further complicate the expression. Equation (21) means that we can find the singular harvest rate for every point in the phase diagram. Be aware, however, that 0 ≤ y˜ ≤ ymax , and thus y˜ may not be not feasible for every value of S and B. If a singular harvest rate is employed, an autonomous system results: dS = f S (S, B) − y˜ (S, B) dt

(22)

dB = f B (S, B) dt

(23)

and

with y˜ (S, B) as the harvest rate, implicitly given by Equation (21). The singular system indirectly depends on the cost of fishing and the value of birds. In the next sections (Sections 6.3.4.2–6.3.4.4), we consider various possibilities for a singular end state. These are: an equilibrium with coexistence of birds and shellfish; an equilibrium without birds; or, a limit cycle. In equilibrium, y˜ (Se , Be ) is constant. This means that the equilibria of the singular system must be equal to the equilibria under fixed harvest rates, as studied in Section 6.3 (see Equations (7) and (8)). In Section 6.3.4.5 we investigate the possibility that the singular system has a limit cycle, which means that y˜ (S, B) in the end state is not constant.

6.3.4.2. Singular equilibrium harvesting In this section we restrict ourselves to singular harvesting in equilibrium. From Equations (22) and (23), it follows, that in equilibrium, f S − y˜ = f B = 0. Substitution of this in (21) gives        ∂f B ∂v ∂f S ∂f B ∂f S ∂c S 0= δ− f − + ( p − c) ( p − c) δ − + ∂B ∂S ∂S ∂B ∂B ∂S (24)

104

PART II: BIOECONOMIC MODELING

Equation (20) is the golden rule of an optimal equilibrium. Clark (1976) and Str¨obele and Wacker (1995) find two symmetric golden rules (one for each species) for the optimal harvesting of two interacting species, when both species are harvested. Ragozin and Brown (1985) have generalized those results for harvesting n interacting species (n symmetric equations). The difference with our result in Equation (24) springs from the fact that in our case we attach value to the number of predators remaining in the ecosystem instead of those harvested. Furthermore, the social welfare function depends on a species (birds) that is not harvested and therefore cannot be directly controlled. First, we will interpret (24) in a general way. In Sections 6.3.4.3 and 6.3.4.4 we examine Equation (24) more specifically with regard to the described ecosystem. Equation (24) can be satisfied in many ways. Either the two main terms are non-zero and of equal size, or either (25) or (26), and (27) hold. First, however, we discuss the equilibrium when a combination of the following terms are zero:   ∂f S ∂c S ( p − c) δ − f =0 (25) + ∂S ∂S and δ−

∂f B =0 ∂B

(26)

and 

∂v ∂f S + ( p − c) ∂B ∂B



∂f B =0 ∂S

(27)

Note that ∂ f B /∂ S = 0. Equation (25) is the standard golden rule for the harvesting of a single species (Clark, 1976). It reflects the direct change in profit due to a change in the equilibrium shellfish stock, when birds are kept constant. The terms in (25) express the impact of a change in shellfish stock on social welfare through three channels: r more interest on extra income from catching one unit more fish ( p − c)δ; r less future income through stock effects, i.e. less future harvesting −( p − c)∂ f S /∂ S; r an increased cost of future harvesting due to stock effects, (∂c/∂ S) f S . Note that, only if (25) is satisfied (and thus also (27) holds), can the optimal harvest rate in the present multi-species context equal the optimal harvest rate in the single species context. Expression (27) shows the indirect effect of a change in shellfish stock on social welfare that occurs via changes in the number of birds. A change in the number of shellfish changes the number of birds by a factor ∂ f B /∂ S. This results in a change of the value of birds by ∂v/∂ B. In addition, a change in bird numbers means a change in natural predation, so that the shellfish stock changes (by ∂ f S /∂ B). This is translated into value terms through the net benefit of extra fish to fisheries, ( p − c). When (25) is satisfied, no direct gains can be made by a change in shellfish stock. Then (27) must hold too, so that neither any gains can be made through a change in bird numbers. Secondly, Equation (24) is satisfied when both (27) and an even more elementary golden rule, Equation (26), hold. This means there are no marginal net benefits due to

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

105

a change in bird stock (27), because the marginal growth rate for birds exactly matches the discount rate. In other words, changing the harvest rate does not lead to extra net benefits because benefits (more value of more birds) cancel out against cost (the opportunity cost of waiting, i.e. the discount rate). Third, to better understand (24) as a whole (both main terms are non-zero), one can use (20) to express Equation (24) in terms of the shadow price for birds. This leads to the following recursive expression in λ B :   1 ∂v ∂f S ∂f B λB = + ( p − c) + λB δ ∂B ∂B ∂B The shadow price of birds represents the marginal social benefit (or cost) of a marginal change in the number of birds. In simple terms, this is the price of an extra bird in the ecosystem. In equilibrium, it is equal to the discounted sum of three elements: r a change in the direct value of birds, ∂v/∂ B; r a change in net revenues of harvesting due to a change in the equilibrium shellfish stock, ( p − c)∂ f S /∂ B; r a change in the number of birds due to a change in the number of offspring, valued against the shadow price, (∂ f B /∂ B)λ B .

6.3.4.3. Equilibrium harvesting when S = Se1 Any equilibrium is located on the B-isocline, so that either S = Se1 or B = 0. First, we consider an equilibrium on S = Se1 . In the Section 6.3.4.4, we will analyze an equilibrium on B = 0. From now on, our analysis is less general. We will explicitly use the growth functions from the ecological model, Equations (5) and (6), to determine the location of the singular equilibrium of the harvested ecosystem. To find the optimal number of birds in equilibrium, we take the derivatives of f S (Se1 , Be ) and f B (Se1 , Be ) with respect to S and B, and substitute these together with the equilibrium value of Se1 (Equation (7)), in (24). This gives us the following implicit solution for Be that allows for a graphical interpretation. K zδ (( p − c(Se1 ))(δ + (1/a)) + (B0 /K )(∂c(Se1 )/∂ S)) ∂v(Be ) = R∞ + ∂B (K − h)2r Be

(28)

e1 )/∂ S) where R∞ = ( p − c(SKe1r))(δ + r ) − zδ(∂c(S and B0 = q K − (K Se1 /a) = q K − (K z/ (K −h)2 r (K − h)a) is the number of birds in the natural (non-harvest) equilibrium. Figure 6.3 represents Equation (28). The right hand side (RHS) of Equation (28) is a hyperbolic function shown in black. It can be one of two different forms depending on the sign of the second RHS term. We call the RHS type I when the second RHS term is positive and type II when it is negative. The left hand side (LHS) of Equation (28), the marginal value of birds, is shown in gray. We consider two possible forms of the bird-value function: namely, an S-shaped and a concave function. The intersection point of a black and a gray curve defines a solution for Equation (28). That is the singular equilibrium value of B.

106

PART II: BIOECONOMIC MODELING We call the solution a type I equilibrium when the RHS is of type I, meaning:   1 K ∂c(Se1 ) > −( p − c(Se1 )) δ + (29) ∂S a B0

and a type II equilibrium when the RHS is of type II, meaning:   ∂c(Se1 ) 1 K < −( p − c(Se1 )) δ + ∂S a B0

(30)

A type I (II) system means that if the system is in an optimal equilibrium with birds extinct, S is smaller (larger) than Se1 . For further discussion, see Section 6.3.4.4, where singular equilibria on the S-axis are discussed. From the graph in Figure 6.3 and from the condition 0 < Be < B0 , we can deduce the following conditions for the solutions of Equations (24) or (28). There can be 0, 1, or 2 type I equilibria. The minimum of the LHS (gray) must be smaller than the maximum of the type I RHS (black), or the two curves will not intersect. Therefore, a necessary condition for the existence of a type I equilibrium is:    ∂v(B) 1 r +δ K zδ max δ + > ( p − c(Se1 )) + (31) B ∂B Kr (K − h)2r B0 a The RHS of (31) equals the RHS of (28) for B0 , its minimum value. If v(B) is concave, then max ∂v(B)/∂ B = ∂v(0)/∂ B. There is only one type I equilibrium, when B

   ∂v(B0 ) 1 r +δ K zδ δ+ > ( p − c(Se1 )) + ∂B Kr (K − h)2r B0 a

(32)

This can be seen from Figure 6.3. When the RHS is of type II, up to three equilibria may exist. A necessary condition for a type II equilibrium is:      1 K zδ r +δ ∂v(0) ∂v(B0 ) min δ + , + < ( p − c(Se1 )) (33) ∂B ∂B Kr (K − h)2r B0 a For a concave v(B), no more than one equilibrium exists and (33) is also a sufficient condition. If v(B) is S-shaped, up to three equilibria may exist. Equation (34) gives a necessary condition for the existence of more than one equilibrium:    ∂v(0) 1 r +δ K zδ δ+ < ( p − c(Se1 )) + (34) ∂B Kr (K − h)2r B0 a We are looking for possible end states. Not every singular equilibrium is a possible end state. According to the transversality condition (18), the shadow price of birds needs to be positive for an optimal end state. Let us consider the singular end state on the B-isocline. Substituting S = Se1 and Equation (20) in λ B > 0 leads to: −K 2 z (( p − c(Se1 ))(δ + (1/a)) + (B0 /K )(∂c(Se1 )/∂ S)) B˜ > ( p − c(Se1 ))(K − h)2 − K z(∂c(Se1 )/∂ S)

(35)

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

107

This equation defines the minimum number of birds on a singular trajectory for which the shadow price is positive, i.e. a lower bound to the number of birds in a singular coexistence end state. The denominator is positive. Hence, the number of birds in an equilibrium of type I will satisfy (35). If the equilibrium is of type II, we can combine Equation (24) with (35) to derive the following necessary condition that ensures the type II equilibrium is optimal:   ∂v(0) ∂v(B0 ) ( p − c(Se1 )) ∂v(Be ) min , > (36) > ∂B ∂B ∂B K

6.3.4.4. Equilibrium harvesting when B = 0 Now, we consider the singular end state to be an equilibrium on the S-axis. Of course, the system must be able to reach  this state, so ymax ≥ q − Se1 /a (Equation (8)). Since ∂ f B (S, 0)/∂ S = 0, ∂ f B (S, 0) ∂ S = 0 the condition to derive Equation (17) is not satisfied. For a singular equilibrium on the S-axis, we must have y˜ = f s (Se , 0) (Equation (22)). Given that the system is in equilibrium, it follows from (15) that λ˙ S = 0. Substituting λ˙ S = 0, Equation (15) and B = 0 in Equation (12) eliminates λ S and results in:   f S (Se , 0) ∂c(Se ) ( p − c(Se )) δ − (37) f S (Se , 0) = 0 + ∂S ∂S Again we have the well-known golden rule for harvesting in a single species ecosystem (Clark, 1976). This makes sense because birds are extinct, so they do not influence the amount of shellfish in the end state. Equation (37) implicitly defines the number of shellfish in the singular equilibrium. Substituting f S (Se , 0) = B0 /K and ∂ f S (Se , 0)/∂ S = −1/a in (37), then the LHS of (37) increases monotonically as S increases. Hence, no more than one solution is defined by (37). If, for some S, the LHS of (37) is greater than zero, consequently S > Se . Note that, for a type I equilibrium, the LHS of (25) is greater than zero at Se1 and thus the optimum when birds are extinct is smaller than Se1 . For a type II system, the equilibrium lies to the right of Se1 and is not attainable unless birds are eliminated first. Strictly, following the continuous mathematics of our model, birds can only be eliminated asymptotically. So, theoretically birds cannot be eliminated in finite time. Therefore, only in a type I system, extinction of birds is a possibility. Now we can interpret the two types of equilibria in Section 6.3.4.3. A type I equilibrium exists if, in the absence of birds, fishers would prefer to fish at a shellfish stock level less than Se1 . In contrast, a type II equilibrium exists if fishers would prefer to fish at a stock level higher than Se1 , in the absence of birds. In each case, birds are a nuisance to fishers. In a type I equilibrium, birds decrease the catch because they eat shellfish that fishers would like to have caught. In a type II equilibrium, the birds increase the fishing costs by depleting the shellfish stock. 6.3.4.5. Singular cyclical harvesting Semmler and Sieveking (1994) show that optimal constant harvesting may push a predator–prey system into cyclical behavior, whereas without harvesting the system would reach equilibrium. This does not hold for the system studied here. As shown in Section 6.3 constant harvesting causes the ecosystem to reach an equilibrium. From

108

PART II: BIOECONOMIC MODELING

the dynamic optimization literature, e.g. Brock and Scheinkman, 1977; Benhabib and Nishimura, 1979; Dockner and Feichtinger, 1991; Wirl, 1995 it is known that a nonsteady singular control can cause an optimal path to take the form of a limit cycle. Suppose such a limit cycle exists, then it is necessary to establish whether the associated singular control y˜ (t) is feasible, meaning that 0 ≤ y˜ (t) ≤ ymax . The trace condition (Brock and Scheinkman, 1977) can be applied to rule out limit cycles. Otherwise, the Hopf bifurcation theorem (Benhabib and Nishimura, 1979; Dockner and Feichtinger, 1991; Wirl, 1995) can be used to show that limit cycles are possible for certain combinations of parameter values. Both methods require the Jacobian of the singular system. The Jacobian must have purely imaginary eigenvalues for the Hopf bifurcation. Because the system Equations (22) and (23) consist of second-order partial derivatives of f S (S, B) and f B (S, B), determining eigenvalues is a cumbersome and tedious task. Therefore, we use a less traditional approach to determine the possibility of a limit cycle as the end state. We can exclude limit cycles when ymax ≤ q. This follows from the stability analysis of the original system (Equation (12)). In order to have a limit cycle y(t) ≥ q for some t. Moreover, a limit cycle will follow a closed orbit around an equilibrium. Suppose the equilibrium exists, then we can use the transversality conditions (18) to determine the manifold on which an optimal limit cycle must circle. On this manifold, the shadow prices are strictly positive. From Equation (19), it follows that a profit must be made from the fishery throughout the singular limit cycle, and thus S ≥ S0 , with c(S0 ) = p. Furthermore, an extra condition for the existence of an optimal singular limit cycle can be derived from the knowledge that the shadow price of birds must be strictly positive along the limit cycle. Using (16), λ B > 0 and ∂ f B (S, B)/∂ S > 0, we find that   ∂c(S) ∂ f S (S, B) f S (S, B) > −( p − c(S)) δ − (38) ∂S ∂S At the B-isocline, Equation (38) transforms to (35), which tells us the minimum value of B, for which λ B is positive. Because the limit cycle circles around Be , it will intersect S = Se1 below the equilibrium, Be , and above the minimum for which λ B > 0. Therefore, also Be must lie in the manifold where λ B is positive. So, in order for the system to have an optimal singular limit cycle, it must have a singular equilibrium with positive λ B . This means the equilibrium is either of type I, or that when it is of type II, Equation (36) must hold.

6.3.5. A bang-bang cycle: Case 4 The last conceivable end state is a bang-bang cycle in which y(t) continuously switches between ymax , 0, and y˜ (t). Alternatively, a bang-bang control can move the system closer and closer to some point. We discuss this in Section 6.5 when we examine approach paths. Typically, a cycle will cross the B-isocline because, for every harvest rate, bird numbers increase to the right of it and decrease to the left of it. Figure 6.4 illustrates a typical bang-bang cycle.

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

109

B

B0

S

0

S e2

S e1

Figure 6.4. Phase diagram of a bang-bang cycle. Note: The cycle is caused by alternating between maximum harvesting and not harvesting. The isoclines of the harvested and non-harvested system are shown in gray. The equilibria of both systems are denoted by a black dot. At the intersection of the two trajectories, the harvest rate is switched. Thus, neither the harvest nor the non-harvest equilibrium is ever reached and the bang-bang control is repeated forever.

The phase diagram in Figure 6.3 shows that, when harvesting is stopped, the shellfish stock size increases while the bird population size keeps on decreasing. Only after a while, when S = Se1 , will bird numbers increase again. Also when the fishers start fishing to the right of the B-isocline (S > Se1 ) bird numbers will increase in spite of fishing until the number of shellfish is Se1 . The simplest bang-bang cycle consists of two branches as in Figure 6.3: one on which is harvested and the number of shellfish declines, and another on which no fishing takes place and shellfish and birds recuperate.

6.3.6. The influence of the discount rate and costs The discount rate plays a pivotal role in establishing the singular solutions of the system. Suppose we have a maximum harvest rate at which birds can become extinct: ymax > q − Se1 /a. Then, for a large enough δ, Equation (21) is satisfied, but not 31 so that we can have an end state in which birds are extinct. For a slightly smaller δ, 31 is satisfied, and 32 (just one equilibrium on S = Se1 ) is not satisfied. Then the system has between one and three equilibria, one on the S-axis and none, one or two on the B-isocline. A limit cycle is also possible, provided ymax > q. If δ again becomes slightly smaller, then 32 is satisfied too and no more than two equilibria remain, one at B = 0 and one at S = Se1 . For a sufficiently small enough δ and shellfish mortality 33, the system can have an optimal type II equilibrium, in which birds surely survive. This illustrates the importance of the discount rate. The costs at S = Se1 can also serve as bifurcation parameter. Given the right cost function, every type of singular state is

110

PART II: BIOECONOMIC MODELING

possible. This is relevant because fishers will try to reduce fishing costs, which will lead to new optimal end states. If the reduction leads to an increase in the marginal cost of fishing, the system will lose its type II equilibrium but it will gain a type I equilibrium and an equilibrium in which birds are extinct. For low enough costs at S = Se1 , the singular system has only one type I equilibrium. For even lower costs, birds will become extinct. This illustrates that the incentive of cost reduction can push the system to a different type of equilibrium. Equation (24) shows that through cost reduction the optimal number of birds, and consequently the optimal harvest rate, may go up or down, depending on how marginal costs change. In special circumstances for instance when the harvesting costs decrease in a type I system, but the marginal harvesting costs remain constant, a counterintuitive result of cost reduction is: lower optimal harvest rates, resulting in lower fishing revenues and possibly lower profits. In this case, policy makers would have a very difficult task, provided they realize what is happening. They would have to reward the higher efficiency of fishers by setting higher quota that reduce the income fishers can realize. Of course, the optimum value of the social welfare function increases when costs are reduced.

6.3.7. Costs as a non-linear function of harvesting Until now, we have assumed costs to increase linearly with the harvest rate. Suppose, however, that the cost function is non-linear in y: c(S, y) instead of c(S)y. In this case, the maximum of the Hamiltonian is not necessarily at one of the extremes (y = ymax or y = 0). Instead, an interior solution may exist. The maximum condition becomes ∂ H/∂ y = 0 ⇒ p − ∂c(S, y)/∂ y = λ S , which replaces Equation (19). The analysis of the optimal end state is restricted to analysis of the singular state, leading to the same results for the equilibria and cycles, as derived in the previous sections, but with c(S) substituted by ∂c(S, y)/∂ y. Assuming, furthermore, that the Hamiltonian is concave in S, B, and y, the necessary conditions are also sufficient (Kamien and Schwartz, 1981). The optimal solution, the approach path, and the end state are given by applying the singular harvest rate.

6.4. QUALITATIVE ANALYSIS OF APPROACH PATHS The harvest rate on the approach path (as elsewhere) is either bang-bang or singular. We will qualitatively describe how the system reacts under bang-bang control and it will automatically become clear when and how the singular control must be applied. We make use of the theory of variable structure systems (Costa et al., 2000) to describe the behavior of the ecosystem under bang-bang control. The optimal harvested ecosystem can be considered to consist of three subsystems: one in which y = ymax , one in which y = 0, and one in which y = y˜ (S, B) (singular, Equation (21)). These subsystems correspond to different regions in the phase diagram. A switching line separates the regions of maximum harvesting and no harvesting. The switching line is the projection on the (S, B)-phase plane of the points where the optimal trajectory of the four-dimensional state/co-state system (Equations (5), (6), (16), and (17)) intersects the singular manifold, p − c(S) = λs .

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

111

B

Harvest

B0

No harvest

S

0

S e2

S e1

Figure 6.5. Phase diagram of a system with bang-bang control. Note: The system has one virtual and one real equilibrium. The thick gray curves denote the isoclines of the harvested and non-harvested system. The thin line is the switching line. A trajectory will approach the real equilibrium.

Each subsystem has its own equilibrium points. The position of the switching line will determine if the equilibrium of a subsystem is located in the region where that subsystem is active. That means the equilibrium can be reached. If an equilibrium of a subsystem can be reached, it is called “real.” If it cannot be reached it is called “virtual.” On one side of the switching line, p − c(S) < λs , so that no fishing occurs. On the other side, p − c(S) > λs so that fishing occurs at the maximum rate. When, as in Figure 6.5, the switching line is located such that the equilibrium of the harvested system lies in the region in which no harvest occurs, then the harvest equilibrium cannot be reached. It is a virtual equilibrium, as opposed to the unharvested equilibrium, which is a real equilibrium because it can be reached. This system is an example of Case 1 (Section 6.3.2), where in the end state no shellfish are harvested. Suppose one of the equilibria is real and the other(s) is (are) virtual. Then the real equilibrium will be reached (Figure 6.5). When all equilibria are real, one will be reached depending on the initial state (Figure 6.6). The end state will either be Case 1 or Case 2 (Sections 6.3.2 and 6.3.3). Maximum or no harvesting will take place in the end state. If all equilibria are virtual equilibria, the system will experience bang-bang control (Figures 6.6 and 6.7) until a state is reached where the singular harvest rate must be applied. A bang-bang cycle is a special case where two virtual equilibria exist and the bang-bang control must be applied infinitely. Generally, the system will end in equilibrium with a singular harvest rate. Suppose the trajectory on both sides of the switching line is directed toward the switching line and thus toward the other region. This is best seen in Figure 6.8, but it also happens in Figure 6.7. The system will arrive at the switching line and will not be able to leave; therefore it must slide along it. In effect, this means that the singular harvest rate is applied and the system follows a singular trajectory, because the system moves

112

PART II: BIOECONOMIC MODELING B

No harvest

B0

Harvest

S

0

S e2

S e1

Figure 6.6. Phase diagram of a system with bang-bang control. Note: The system has real two real equilibria. Three trajectories are shown. Depending on the initial condition the trajectory approaches the harvest or non-harvest equilibrium. The thin line is the switching line.

B

Harvest

B0

No harvest

S

0

S e2

S e1

Figure 6.7. Phase diagram of a type II system with bang-bang control. Note: The system has two virtual equilibria and approaches a point (the singular equilibrium) on the B-isocline. The thick gray curves denote the isoclines of the harvested and non-harvested system. The thin line is the switching line.

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

113

B

Harvest

B0

No harvest

0

S S e2

S e1

Figure 6.8. Phase diagram of a type I system with bang-bang control. Note: The system has real two virtual equilibrium and approaches one of the singular equilibria, depending on the initial condition. The thick gray curves denote the isoclines of the harvested and non-harvested system. The thin black line is the switching line. Close to the singular equilibrium, the system cannot cross the switching line and must follow it under singular control.

talong a path on which p − c(S) = λs . Thus, in this case, (part of) the switching line is a singular trajectory. It has turned out to be difficult to prove or disprove whether in fact, the switching line must be a singular trajectory. Figure 6.7 shows an example of a single type II equilibrium. A more complex switching line and resulting system behavior is possible if the system has several singular equilibria. An example is given in Figure 6.8, which shows a type I system. If we assume that the optimal trajectory ends in a singular equilibrium on the B-isocline, then the switching line crosses or at least touches the B-isocline at the equilibrium. The examples in Figures 5–8 show that an end state is an equilibrium at the singular, the maximum or the minimum harvest rate. But in special cases, the approach path never reaches equilibrium. This means the end state is a limit cycle either singular or bang-bang (Figure 6.4). From Figures 6–8, we can conclude that not only the system parameters determine the end state but also the initial conditions play an important role. Intuitively this is clear, for example, if in the initial state the number of birds is very small, the discount rate large and the recovery of bird numbers slow, it would be advantageous to harvest maximally and not wait for the recovery of birds. However, if in the same system bird numbers were high, a much lower harvest rate could be optimal.

6.5. CONCLUSION This chapter has illustrated that several types of optimal harvesting solutions are possible in a predator–prey system when conservation of the predator species is considered

114

PART II: BIOECONOMIC MODELING

valuable. The type of solution depends on economic parameters, e.g. the maximum harvest rate, the discount rate, and the cost of fishing, as well as on ecological parameters such as the predator’s search and handling time of prey. The final optimal harvest rate can be constant, resulting in an equilibrium, either with or without the predator species. After an initial period the system reaches the end state. The end state can be one of three possibilities: not fishing, fishing maximally, or fishing at a singular harvest rate. The necessary conditions for each possible singular end state are derived in Section 6.3. For some combinations of parameters, several end states can exist. In that case, the initial conditions will determine which end state is optimal. This chapter has characterized the approach paths toward the end states. The approach path toward an equilibrium end state will usually consist of some period of bang-bang control followed by a singular harvest rate. Several possibilities are shown graphically. However, when costs do not depend linearly on the harvest rate, the Hamiltonian has an interior maximum. Then the optimal solution is given by applying the singular harvest rate. In that case the conditions stated in Section 6.3.7 formulate when birds can survive. The optimum end state may not be stable over time. Fishers have an incentive to lower costs, because this will seem to increase their income. The regulator will encourage this because it is also beneficial to society. The social welfare function does indeed increase with lower costs. However, what happens to the socially optimal harvest rate depends also on how the marginal cost at the equilibrium (Se1 ) changes. If the optimal harvest rate decreases and lowers the fisher’s income, tension between regulator and fishers will arise. Fishers will feel that increasing efficiency is punished. Evidently, the results provide incomplete information for policy design. For example, the introduction of measures involves the administration of stock levels, harvest rate, and so on. Moreover, then the measures have to be enforced. All this comes with a cost that we have ignored in specifying our welfare function and subsequently in defining the optimal state. For example, it may be much more costly to ensure a cyclical harvest rate than to maintain a constant harvest rate. The most important general conclusion to ensure harvesting and conservation is that birds will surely survive if the system has a type II equilibrium. This means (Equation (30)): 1. 2. 3. 4. 5. 6.

the absolute marginal cost of harvesting at Se1 is large, the cost of harvesting at Se1 is small, the discount rate is small, prey has a low additional mortality rate, the pristine state carries many predators, predators can quickly deplete prey in equilibrium, i.e. predators eat many prey, and (Equation (33)), 7. the marginal value for birds at either B = 0 or B = B0 is small. An example of a value function consistent with the latter condition is an S-shaped function with the sharp increase after 0 or an asymptotic function that is close to its maximum in the pristine state. However, because of (36)

6. HARVESTING AND CONSERVATION IN A PREDATOR–PREY SYSTEM

115

8. the marginal value of birds at B = 0 or B = B0 or the cost of harvesting at Se1 should not be too small. Note that in the case of a type I system, birds are not necessarily doomed. But, assuming that fishers attach no value to birds, it is then critical for society not to wait too long with regulations because if the system deteriorates too far a birdless end state will be optimal.

CHAPTER 7

BIOECONOMIC ANALYSIS OF A SHELLFISHERY WITH HABITAT EFFECTS

7.1. INTRODUCTION1 In accordance with the recognition that fish populations generally are subject to complex dynamic population processes, there has been an increase over the last years in bioeconomic analysis that steps away from simple stock growth modeling techniques. Greater realism has been provided by modeling interdependency of fish species and by recognizing that fish species display different population dynamic characteristics at the adult life and recruitment life stages (Eggert, 1998). In line with these developments, the last decade has seen an increase in the literature focused on modeling sedentary fish population dynamics as metapopulations. The sedentary species considered in the literature generally have a life cycle consisting of a juvenile stage and an adult stage. In the juvenile stage, the species is dispersed throughout a specific predefined region. The species reaches an adult stage once the juvenile has found an appropriate location to settle, after which it is assumed to remain tied to the area. Previous model-based studies have analyzed the spatial population dynamics of the fish based on the ecological linkage to its surroundings (Possingham and Roughgarden, 1990; Alexander and Roughgarden, 1996), the metapopulation movements between a fishing ground and a marine reserve (Man et al., 1995) and the optimal way to exploit the fish species given certain metapopulation characteristics (Brown and Roughgarden, 1997; Bulte and van Kooten, 1999; Sanchirico and Wilen, 1999, 2001; Pezzey et al., 2000). In this chapter, a bioeconomic model is presented of a shellfishery where during each period juveniles are dispersed throughout the sea after adult spawning and, if possible, recruit to the shellfish bed. In each period, the effort exerted in order to catch shellfish is restricted. Fishing is only allowed to occur in the fishing season. The no-fishing season coincides with the adult spawning season. During this season, the newly born juvenile shellfish population is allowed to recruit. In the model discussed here, effort is assumed to have an indirect effect on recruitment, much as discussed by other bioeconomic models covering sedentary fish species, by reducing the adult fish population that can spawn and thus the size of the juvenile population that recruits to the fishery. As opposed to the general literature covering sedentary fish species, however, the model presented here goes one step further by assuming that effort exerted in order to catch shellfish 1

Reprinted from R.J. Imeson and J.C.J.M. van den Bergh (2004). A bioeconomic analysis of a shellfishery: The effects of recruitment and habitat in a metapopulation model. Environmental and Resource Economics 27:65–86.

117

118

PART II: BIOECONOMIC MODELING

also has a direct impact on recruitment. Fishing is expected to affect the recruitment process directly by making space available for juveniles to recruit to in the recruitment season through catching the adult population attached to the shellfish bed in the fishing season. The direct effect effort has on the shellfish habitat can also manifest itself in many other ways: fishing techniques are often criticized for causing habitat destruction and pollution, which obviously have a negative impact on the recruitment process. As an example of a shellfishery were effort has a direct impact on population abundance in different ways consider the cockle fishery in the Dutch Wadden Sea. In this fishery, there is strong evidence that cockle fishing techniques used have a negative impact on shellfish population abundance by altering the habitat. Nevertheless, there is also evidence that fishing has a positive effect on recruitment as well, which is suspected to be the result of competitive adult–juvenile interactions (Piersma et al., 2001). Modeling a shellfishery where fishing has both a direct and an indirect effect on the sedentary fish species is therefore not an unrealistic assumption. Rather, it represents a framework that provides insight into how fishing can affect a shellfishery in multiple, realistic ways. The bioeconomic model formulated hereafter gives rise to a non-linear discrete optimization problem whereby net profit over an infinite horizon is maximized subject to an equation describing seasonal population dynamics. The optimal level of fishing effort that should be exerted and the optimal shellfish population in a steady state can be determined. Intuitively, if fishing frees up space for juveniles to recruit to, fishing under certain conditions can be expected to benefit the growth of the population. The conditions under which fishing should be encouraged will be investigated. It should be noted that the optimization problem has been formulated in discrete time to address multiple generations with different lifecycles (adult shellfish and larval juveniles). The organization of this article is as follows. Section 7.2 presents the general structure of the discrete time model and its optimal steady state solution for the case that net profits are maximized over an infinite time horizon. Section 7.3 considers a particular discrete time recruitment function that captures basic shellfish characteristics, and solves the resulting metapopulation model in the case of instantaneous recruitment. Section 7.4 considers a continuous time recruitment function where recruitment takes place over a certain time period. Section 7.5 compares the general solution of Section 7.2 with the case of fishing effort negatively affecting the shellfish habitat. Conclusions are drawn from Section 7.6.

7.2. GENERAL STRUCTURE OF THE MODEL Consider a fishery that consists of fishermen actively fishing shellfish for commercial purposes. As a result, the adult shellfish population fluctuates each period. Its dynamics from one period to the next are given by: X t+1 = F(X t , E t ),

t = 0, 1, . . . .

(1)

where X t is the stock density of the adult shellfish population at the start of period t and E t is the amount of effort exerted in period t in order to catch shellfish. The stock density of the shellfish population in the next period is a function of the stock density and the effort exerted to catch this stock in the previous period. The stock density is

7. BIOECONOMIC ANALYSIS OF A SHELLFISHERY

119

considered to be equal to the number of shellfish as a proportion of the total carrying capacity of the area under consideration. The total carrying capacity of the area under consideration is the maximum number of shellfish that can be supported in the area. Hence, the stock density must lie between 0 and 1. A stock density of 0 means that there are no shellfish in the area under consideration. A stock density of 1 is assumed to mean that the area is carrying the maximum number of shellfish that it can support. We assume that the fluctuation in the shellfish population is of a discrete nature. Every period t is assumed to consist of two seasons. During the first season, the shellfishery is open and the shellfish population is fished down to a certain level. During the second, the fishery is closed down and recruitment takes place. In this season the fishery is allowed to recover. Fishing and recruitment are assumed to be sequential processes that do not overlap in each period t. For all periods t the dynamics of the model relating the population density at period t + 1 to the population at the previous period t are governed by the following specific version of Equation (1): X t+1 = X t − C(X t , E t ) + R(X t , E t )

(2)

where C(X t , E t ) is the catch in period t and R(X t , E t ) is the larvae population recruited in period t. Note that the size of the catch and the newly recruited shellfish population in a period are considered to be dependent on the shellfish stock and the fishing effort exerted in the same period. Furthermore, we assume that all adult shellfish mortality is due to fishing. There is no natural mortality at the adult life stage. As found in the general literature (Clark, 1990; Neher, 1990), effort exerted has an indirect negative impact on recruitment as it increases catch and therefore reduces the adult population that spawns. But effort exerted in the fishing season can also have a direct effect on recruitment once the fishing season is over. Negative direct effects on recruitment are the result of, for example, pollution or a reduced carrying capacity of the area by destroying the shellfish beds. Positive effects are, for example, the reduction of competition of the adult population facing potential recruits. Such direct effects of effort on recruitment are generally not included in bioeconomic models used in the fisheries, although they do seem to be relevant. Maximizing net profit over an infinite horizon, the direct and indirect effect of effort on the steady state shellfish density will be analyzed later in this section. The abundance of shellfish depends on the availability of shellfish beds in which the shellfish are settled. The shellfish are non-migratory and physically attached to the beds. Since the locations of all beds are presumed known, fishermen will exert effort only in those areas where beds are located. The resulting catch per unit of effort can then be considered to be a function of the areal stock density (Pezzey et al., 2000). A common specification is used: C(X t , E t ) = q E t X t

(3)

where q is a catchability parameter. We consider that when the fishermen jointly decide how much to catch, they do so seeking the optimal effort and catch maximizing the present value of profits over an infinite horizon: ∞  π (X t , E t ) t=0

(1 + i)t

(4)

120

PART II: BIOECONOMIC MODELING

where i is the periodic discount rate and πt (X t , E t ) is the profit in period t. To find the optimal fish population in each period, the discrete maximum principle (Clark, 1990) is applied to the problem given by maximization of Equation (4) subject to Equation (1), i.e. the shellfish population dynamics. This gives the following equation, which determines the optimal population density X t and effort E t (see Appendix A for details): 1+i =

∂ F(X t , E t ) (∂πt (X t , E t )/∂ X t )(∂ F(X t , E t )/∂ E t ) − |||||| ∂ Xt ∂πt (X t , E t )/∂ E t

(5)

This equation is a fundamental result of models of renewable resources (Conrad and Clark, 1987). In order to gain a better understanding of the meaning of this equation in the present setting it is assumed that a steady state optimal solution exists and can be reached after a certain amount of time. Section 7.3 will discuss the conditions for an optimal steady solution and whether it can be reached. Using Equation (2), Equation (5) in a steady state can be rewritten as: ∂C(X, E) ∂ R(X, E) i =− + ∂X ∂X (∂π(X, E)/∂ X )(−(∂C(X, E)/∂ E) + (∂ R(X, E)/∂ E)) − (6) ∂π (X, E)/∂ E At this point, it is useful to specify the profit equation in a general way as: π (X t , E t ) = pC(X t , E t ) − cE t

(7)

where p is the price paid per unit of fish and c is the cost per unit of effort. p and c are assumed to be constant. Equation (7) is simply the difference between total revenue and total cost in period t. Substituting Equation (7) into Equation (6) and rearranging terms allows the optimal steady state solution to be rewritten as: ∂ R(X, E) − ∂X p(∂C(X, E)/∂ X )(−(∂C(X, E)/∂ E) + (∂ R(X, E)/∂ E)) + ( p(∂C(X, E)/∂ E) − c)(∂C(X, E)/∂ X ) p(∂C(X, E)/∂ E) − c

i =

(8) which can be simplified to: ∂ R(X, E) ∂C(X, E) p(∂ R(X, E)/∂ E) − c − (9) i= ∂X ∂X p(∂C(X, E)/∂ E) − c Equation (9) states that when maximizing the net profit over an infinite horizon the periodic discount rate is equal to the marginal impact of the shellfishery, the latter being the sum of two terms. The first term ∂ R(X, E)/∂ X represents the marginal impact the adult shellfish density has on the recruitment density and hence future shellfish generations. The second term represents the marginal impact of the adult shellfish density on catch taking into account that catching shellfish now can be expected to influence future recruitment and hence future profits. To see how the second term draws future profits from shellfishing into consideration, note that p R(X, E) is the total revenue that would be generated from fishing future recruits. The term p R(X, E) − cE can then be interpreted as being the loss or gain of future profits due to shellfishing in the

7. BIOECONOMIC ANALYSIS OF A SHELLFISHERY

121

steady state. It is the opportunity cost of shellfishing in the steady state. The marginal impact of shellfish effort on future profits is then given by p(∂ R(X, E)/∂ E) − c. The expression p(∂C(X, E)/∂ E) − c represents the marginal effect shellfishing has on current profits in a steady state. As the marginal profit ratio p(∂ R(X, E)/∂ E) − c p(∂C(X, E)/∂ E) − c is the ratio of two monetary terms it is an entirely dimensionless factor. Thus in the optimal steady state, the second term of Equation (9) adjusts the marginal effect of the shellfish population on catch by a factor whose value is dependent on the marginal effect of fishing effort on future and current profits derived from shellfishing. The total direct and indirect effect of effort on recruitment is represented by the term (∂ R(X, E)/∂ E). In the extreme case where effort has no effect on recruitment, Equation (9) reduces to the “Golden Rule” found in the general fisheries literature (i.e. Clark, 1990; Neher, 1990) from which the optimal steady state shellfish population density and fishing effort exerted can be derived. In the case where the marginal impact of effort on future profits is equal to the marginal impact of effort on current profits, Equation (9) reduces to i = (∂ R(X, E)/∂ X ) − (∂C(X, E)/∂ X ). In this case, a level of fishing effort exerted should be chosen such that the marginal growth rate of the fish population density is equal to the discount rate. If fishing effort has a negative total direct and indirect effect on recruitment, it can be seen from Equation (9) that the marginal growth of the shellfish population density in an optimal steady state will be negatively affected (assuming of course that current fishing is profitable). If fishing has a sufficiently large positive direct and indirect effect on recruitment it is possible that p(∂ R(X, E)/∂ E) − c will be larger than zero. Fishing will then actually have a positive effect on the marginal increase in the shellfish population in the optimal steady state solution.

7.3. A DISCRETE DENSITY DEPENDENT RECRUITMENT FUNCTION 7.3.1. Recruitment pattern and density function In the no-fishing season, recruitment is influenced by the stock size of the adult population, the average reproductive output per individual parent, and the survival rate of larvae and juveniles (Honkoop and van der Meer, 1998). Juveniles are larvae that are old enough to recruit. Recruitment is considered to be a balance between the availability of substrate that can be acquired and the mortality rate of adults (Petraitis, 1995). This suggests that intraspecific competition can limit successful recruitment by juveniles to the adult population when the occupancy rate of the shellfish bed is high. To model recruitment, we make the assumption that all individuals in the adult population spawn simultaneously and only once during each period. After spawning, the larvae are dispersed throughout the sea. It can take them a number of days to reach the beds and recruit to the adult shellfish population as juveniles, given there is enough space available in the beds. Not all juveniles recruit to the shellfish bed simultaneously, however, because some juveniles reach the shellfish bed faster than others do (de Vooys, 1999). As more juveniles reach and recruit to the adult population over time, the amount of

122

PART II: BIOECONOMIC MODELING

free space to settle will become scarcer. In other words, previous recruits will have already taken up space for latecomers to recruit to. Once there is no more substrate to acquire juveniles that have not yet settled will die. The minimum number of days it takes to recruit to the beds is assumed to be 1. Against this background a formal model of the total recruitment density in period t can be formulated starting with: Rt =

n max 

(10)

Rtn

n=1

where Rt is the total recruitment density in period t, Rtn is the density of juveniles that recruit in period t on n days after the start of the no-fishing season, and n max is the last day on which recruitment can take place. The total recruitment density Rt cannot exceed 1, which is the maximum carrying capacity of the area. Rtn is given by (11)

Rtn = stn L tn

where in period t, stn is the proportion of space available for recruitment in the beds on day n and L tn is the density of the potential recruits who will recruit given there is enough space available in the beds. The proportion of space available on the beds, stn , lies between 0 and 1 (see below). Equation (11) thus expresses a recruitment profile where the density of the potential recruits, L tn , that will actually recruit is equal to the proportion of space available. As space gets scarcer it will be harder for the potential recruits to find space. The proportion of space (stn ) to recruit to depends on the size of the shellfish catch during the fishing season. On each day n, when recruitment takes place, availability of space to recruit to is furthermore dependent on space taken up by recruits on previous days. The total proportion of space available on day n, stn can then be specified as follows:   n−1  stn = 1 − β Rtm , n = 1, . . . , n max (12) X t − Ct + m=1

where

β=

⎧ ⎪ ⎪ ⎪ ⎪ ⎨1 ⎪ ⎪ ⎪ ⎪ ⎩

1  X t − Ct + n−1 m=1 Rtm

if X t − Ct + if X t − Ct +

n−1 

m=1 n−1 

Rtm ≤ 1 (13) Rtm > 1

m=1

As shown by Equation (12), fishing reduces the adult population and hence frees up space for juveniles to recruit to in the no-fishing season. stn has to be larger than 0 (stn < 0 would imply that recruitment up until day n has exceeded the space available for recruitment and that the proportion of space left to recruit to is negative). The term β is used to ensure that the proportion of space to recruit to is non-negative for all days  n. As long as the total population density on day n, X t − Ct + n−1 m=1 Rtm , does not exceed the carrying capacity of the area, β reduces to 1 and the proportion of space to recruit to will be non-negative. If the total population density exceeds the maximum carrying capacity, β ensures that the proportion of space to recruit to becomes 0.

7. BIOECONOMIC ANALYSIS OF A SHELLFISHERY

123

The potential density of the recruits L tn in period t on day n is assumed to be dependent on two factors: on the mortality rate during the larvae stage, which can be as high as 99% (Sprung, 1984); and whether or not larvae reach the shellfish bed after their dispersal in the seas. An arrival rate will be used in the model to reflect the density of juveniles that actually reach the shellfish beds and subsequently are able to recruit given enough space available. The density of juveniles aged n days that reach the shellfish bed in period t is given by: L tn = α(1 − α)n−1 γ (X t − Ct )(1 − µ)n ,

n = 1, . . . , n max

(14)

where α is the daily arrival rate of juveniles at the shellfish bed, µ is the daily mortality rate of the larvae population while they are dispersed throughout the sea, and γ is the reproduction rate of adults. α is assumed to be a value between 0 and 1. If α were equal to 1, Equation (14) would make no sense as the group of juveniles to reach the shellfish bed would be reduced to zero for all ages. To understand the rationale behind Equation (14) note that in period t after spawning of the adult population the density of juveniles dispersed throughout the sea is γ (X t − Ct ). On each day n in this period a proportion α(1 − µ) of the juvenile density will reach the shellfish bed. A proportion of (1 − α)(1 − µ) will not reach the shellfish bed, but survives and may recruit on later days in this period. Thus on day 1 the density of potential recruits is L t1 = αγ (X t − Ct )(1 − µ), while the density of juveniles that do not reach the beds but is still alive after n − 1 days is given by (1 − α)n−1 γ (X t − Ct )(1 − µ)n−1 . Of the juvenile population density that has not recruited after n − 1 days, a proportion of α(1 − µ) will reach the shellfish bed on day n, giving Equation (14). Taking Equations (12)–(14) together we can rewrite recruitment in period t as:    n max n−1   Rt (X t , E t ) = Rtm 1 − β X t − Ct (X t , E t ) + m=1

n=1

n−1

× (α(1 − α)

γ (X t − Ct (X t , E t ))(1 − µ)n )

(15)

7.3.2. Solution to the model in the case of immediate recruitment Let us first consider the simplest case and assume that the juvenile shellfish population recruits in 1 day to the shellfish bed without exceeding the proportion of space available for recruitment. In this case β in Equation (13) reduces to 1. The recruitment function 15 then becomes: Rt = αγ (1 − (X t − Ct ))(X t − Ct )(1 − µ)

(16)

Using Equations (2), (3), and (16), the population density in period t + 1 can be determined from: X t+1 = X t − q E t X t + (1 − (X t − q E t X ))α(1 − µ)γ (X t − q E t X t )

(17)

In a steady state equilibrium, the shellfish population density remains unchanged from one period to the next. Hence: X t+1 = X t . This is equivalent to stating that in each period t recruitment has to be equal to catch. Omitting time subscripts, Equation (17)

124

PART II: BIOECONOMIC MODELING

can be rewritten as: (1 − X + q E X )ξ (X − q E X ) − q E X = 0

(18)

where ξ = α(1 − µ)γ . This equation can be rewritten to yield an expression where the steady state fish population is a function of effort in the steady state: X=

ξ − q E(1 + ξ ) ξ (1 − q E)2

(19)

ξ can be interpreted as being the periodic increase in adult shellfish population density due to recruitment given there is enough space available. Note that the shellfish population is also in a steady state in the rather trivial case that the adult density is equal to zero. If (ξ − 1)/qξ ≤ E ≤ ξ/(q(1 + ξ )), the adult population density in a steady state will be a feasible solution between 0 and 1. In the extreme case where there is no mortality due to shellfishing (i.e. E = 0), the steady state solution is X = 1. In this case there is no room for new recruits on the bed. This is due to the assumption that there is no natural adult shellfish mortality. In Section 7.2, the optimal shellfish population density and fishing effort in each period t that maximizes net profit over an infinite horizon was determined, given that the adult shellfish population and effort exerted were both in a steady state. Having derived the equation that yields the steady state shellfish population density, Equations (19) and (5) can be combined to yield a set of equations from which the optimal steady state solution(s) for the shellfish population density (X ∗ ) and effort to be exerted (E ∗ ) can be determined. With the population dynamics described by Equation (17), Equation (5) becomes:2 1 + i = (1 + ξ )(1 − q E) − 2ξ X (1 − q E)2 −

pq E(−q X (1 + ξ ) + 2qξ X 2 (1 − q E)) pq X − c

(20)

Both Equations (19) and (20) are quadratic expressions in E and Equation (20) is also quadratic in X. Nevertheless, for every possible steady state adult shellfish population density level, Equation (19) yields only one feasible solution of fishing effort.3 Substituting this solution for E in Equation (20) gives an equation yielding multiple solutions, which is analogous to the Golden Rule equation found in Clark (1990). Whether or not they are feasible depends on whether (ξ − 1)/qξ ≤ E ∗ ≤ ξ/(q(1 + ξ )) (and E ∗ ≥ 0). 2

Note that we have: (∂ F/∂ X ) = (1 + ξ )(1 − q E) − 2ξ X (1 − q E)2 and (∂ F/∂ E) = −q X (1 + ξ ) + 2qξ X 2 (1 − q E). 3 Solving Equation (19) to find the optimal level of effort gives solutions 2ξ X − 1 − ξ + (1 + ξ )2 + 4ξ X E= 2ξ Xq and 2ξ X − 1 − ξ − (1 + ξ )2 + 4ξ X E= 2ξ Xq In order for the first solution to be feasible, X > 0 should hold. Feasibility of the second solution requires that X > (2/ξ ) + 1. Since X has to be lie between 0 and 1, the second solution for E is not feasible.

7. BIOECONOMIC ANALYSIS OF A SHELLFISHERY

125

The optimal approach path to the steady state solution (X ∗ , E ∗ ) will not follow a “bang-bang” approach. To see this note that optimal control theory states that the optimal path to the steady state solution will be one where effort E t maximizes a Hamiltonian expression of the form: H (X t , E t ) = δ t−1 π (X t , E t ) + λF(X t , E t ) (21) where λ is the adjoint variable. Substituting Equations (2), (3), and (7) into this expression, we get: H (X t , E t ) = δ t−1 (( pq X t − c) − λq X t )E t + λX t + λR(X t , E t ) (22) Maximizing the Hamiltonian with respect to effort will result in a “bang-bang” approach to the optimal steady state solution only if the Hamiltonian is linear in effort. This means applying either maximum fishing effort or no-fishing effort in order to drive the shellfish population toward the optimal steady state shellfish population density as fast as possible. Nevertheless, expression 22 is linear in effort only if the recruitment function is linear in effort. In our setting, this is not the case as can be seen after rewriting Equation (16): R(X t , E t ) = ξ (1 − X t + q E t X t )(X t − q E t X t ) (23) which is a quadratic expression in E t . Recruitment will only be a linear function of effort if fishing effort exerted in the fishing season has no effect on the recruitment process of juveniles to the shellfish beds. To see this more clearly, note that although fishing in period 1 has an indirect impact on recruitment by reducing the number of adult shellfish population that can spawn, it has a direct effect as well. The direct effect of effort manifests itself in the fact that freeing up space on the shellfish beds for juveniles to recruit to (through the term 1 − X t + q E t X t ) makes recruitment more successful. Only in the ideal world where there are no direct effects due to shellfishing will recruitment be linear in E t and will an optimal approach path exist that follows a “bang-bang” pattern. Since this shellfish model possesses a non-linear recruitment function in effort, the approach path to the optimal steady state solution, if it exists, will be asymptotic (see also Spence and Starret, 1975). The total direct and indirect marginal effect of effort on recruitment (∂ R(X t , E t )/∂ E t ) is given by 2qξ X t2 (1 − q E t ) − qξ X t . The marginal effect of effort on recruitment will always be negative in the case that effort only has an indirect effect on recruitment since in this case the marginal indirect effect of effort on recruitment is equal to −qξ X t . The direct effect of effort on recruitment is equal to the term 2qξ X t2 (1 − q E t ). The total marginal effect effort has on recruitment is only positive if the direct marginal effect of effort is larger than the indirect marginal effect of effort on recruitment, i.e. 2qξ X t2 (1 − q E t ) > qξ X t . This expression can be rewritten to 1 (24) 2 Thus in order for effort to have a positive total marginal effect on the recruitment process, the adult fish population must not be fished down beyond half the carrying capacity of the area. Fishing should be encouraged if the adult population density at the start of the fishing season exceeds half the carrying capacity of the shellfish bed. Substituting Equation (20) into Equation (24) one finds that if an optimal steady state solution exists where effort has a positive total direct and indirect marginal effect (X t − q E t X t ) >

126

PART II: BIOECONOMIC MODELING

on effort one must have i < −q E − ( pq E(−q X ))/( pq X − c), or rather i (1/2). Thus effort will not have a positive marginal effect on recruitment if the adult population density after the fishing season is not larger than half the carrying capacity of the area. Substituting Y for the left-hand side of Equation (35), Figure 7.3 shows that as ϑ increases, the minimal parent population density required to ensure that fishing has a positive effect on recruitment decreases. Thus if in an optimal steady state solution 6

Note that ∂F = (1 − q E)(1 − ϑ(1 − (X − q E X )))eϑ(X −q E X ) ∂X and that ∂F = −q X (1 − ϑ(1 − (X − q E X )))eϑ(X −q E X ) ∂E

7. BIOECONOMIC ANALYSIS OF A SHELLFISHERY Y

131

4 3 2 1

0

0.2

0.4

0.6

ϑ = −10

0.8

1 P ϑ = −1

−1

ϑ = −0.1

Figure 7.3. Parent population densities at which there is a positive effect of effort on recruitment. Note: The curves are given by the left-hand side of inequality 35, where X − q E X is the spawning adult population density that survived the fishing season. The curves are depicted over the range of surviving adult population densities satisfying inequality 35. As ϑ decreases, fishing will have a marginal positive impact over a wider range of surviving adult population densities.

the parent population is larger than 1/2, the total direct and indirect marginal effect of effort is always positive. The optimal steady state solution (X ∗ , E ∗ ) has to satisfy Equation (32). Using Equation (32), expression 35 can be rewritten as i < −q E −

pq E(−q X ) pq X − c

In other words, in order for effort to have a positive indirect and direct marginal effect on recruitment we must have the condition that ∂C (∂π/∂ X )(∂C/∂ E) i VX 1 , we have σ (Z )

∂σ (Z ) VZ Z ∂σ (Z ) VZ Z > σ (Z ) ∂ X 1 VX 1 ∂ X 2 VX 2

Quite unsurprisingly, this means that based on this inequality one can determine that one would attach a higher value to conserving shellfish in the source than in the sink: one would pay more to reduce the uncertainty regarding future conservation requirements in the source. Equation (26) states that in a steady state solution the expected immediate gains of establishing a marine reserve (left hand side) have to equal expected future gains from establishing a marine reserve (right hand side). The second and third term on the left are the lost profits associated with immediate conservation of a unit of shellfish stock. Equations (24) and (25) together suggest that the shellfish at the source should be conserved. This intuitively makes sense as shellfish in the sink yields higher profits and the source serves to replenish the sink. The effect of uncertainty about future conservation requirements is that fishermen will be willing to conserve a larger part of the stock than in the case where there is no uncertainty present. However, Equations (24) and (25) do not consider at the margin what establishing a marine reserve in the sink or the source does for population dynamics in both the sink and the source. Equation (26) does do that. It may be true that establishing a marine reserve in the sink will increase the influx of shellfish from the sink to the source, where it is considered to be more valuable. If, however, the population dynamics are such that the marginal growth of the shellfish stock in the source is larger than in the sink and the influx of new shellfish from the source is small, establishing a marine reserve in the source may not be beneficial.

8.4. CONCLUSIONS One of the main concerns with marine reserve design is the human response. How does the fishing community respond to the fact that an area previously open to fishing is now closed? It is generally assumed in the literature on marine reserves that fishing pressure on the fishing grounds will increase as a result of marine reserve creation. The impact that uncertainty about future conservation requirements has on fishing behavior is a neglected theme in the literature. It is likely that if this uncertainty is large, the fishing community will engage in some form of self-regulation and adjust its behavior in order to preserve future profits.

148

PART II: BIOECONOMIC MODELING

This chapter presented a model of a shellfishery with metapopulation dynamics corresponding to a sink-source model. In the model future conservation requirements were uncertain because the effect of fluctuations of the targeted fish population on the marine ecosystem was poorly understood. It was shown that if there was uncertainty about future conservation requirements, the fishery would be willing to forfeit a proportion of the possible catch in order to reduce uncertainty regarding future conservation requirements. Whether the fishing community considers it more desirable to preserve additional stock in the source or in the sink, depends on which stock is considered to be more valuable and on the population dynamics of the stock in the source and in the sink. At any rate the possibility of additional closure prompts the fishing community to reduce fishing efforts. It is therefore not straightforward to assume that fishing pressure will increase as a result of establishing a marine reserve.

CHAPTER 9

A SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS IN A MARINE ECOSYSTEM

9.1. INTRODUCTION The study of management of renewable resources has a long-standing tradition. Fisheries for example have been studied extensively since the 1950s. Most of the early studies focus on single-species and single-user systems with just one management goal, usually profit maximization. For these types of systems, analytical solutions can sometimes be found (Clark, 1976). Obviously, the importance of species interactions such as predator–prey relations or competition (see e.g. May et al., 1979) are important for the dynamic behavior of the system and make it difficult to find the optimal control. Therefore, numerical modeling and simulation have become necessary to study management strategies in complex ecosystems. A natural resource that is complex and difficult to manage is a coastal zone or an estuary with shellfisheries in which protected birds prey on shellfish. The Dutch Wadden Sea is such an ecosystem, where fishermen harvest cockles Cerastoderma edule and mussels Mytilus edulis (Smit et al., 1998; Kamermans and Smaal, 2002; Ens, 2003). Furthermore, these shellfish serve as prey for bird species such as oystercatchers Haematopus ostralegus and common eiders Somateria mollissima. Both of which need to be conserved due to Dutch and European environmental law. Several other European estuaries such as the Bury Inlet and the Wash in the UK, the Oosterschelde and Westerschelde in the Netherlands, the Baie de Somme in France and the German and Danish parts of the Wadden Sea, are similar areas with birds and shellfish where the balance between exploitation and conservation is sought (Bell et al., 2001; Stillmann et al., 2001; Atkinson et al., 2003; Ens et al., 2004; Goss-Custard et al., 2004). This chapter describes a conceptual model of a bird/shellfish ecosystem with spatial dynamics. The model incorporates multiple species [two predator species (birds) and two prey species (shellfish) divided into two age classes] and multiple users (two fisheries and a public authority). Stillmann et al. (2001) examine a similar system of a shellfishery and its effect on the oystercatcher population. They developed an individual based model of foraging behavior of the oystercatcher, for the Exe estuary and the Bury inlet in the UK. The model presented here is not as detailed when it comes to oystercatcher behavior and calibration to a particular estuary, but incorporates an extra bird and shellfish species and can investigate the effect of mussel and cockle fisheries simultaneously.

149

150

PART II: BIOECONOMIC MODELING

The chapter concentrates on the behavior and sensitivities of the ecosystem model and does not investigate fishery policies yet. Fishery policies can only be analyzed if the underlying ecosystem model is well understood. This chapter is about understanding the ecosystem model and its behavior. For that purpose, we simulate the “pristine” ecosystem (no fishing) and undertake a sensitivity study. Future work will build on the insights gained here and focus on fishing policies. The model is built with the aid of a cellular automata program called Cormas1 . Programs like Cormas allow the easy introduction of spatial relationships between variables. We use Cormas to solve a system of differential equations that describe the feeding and migrating behavior of birds. Alonso and Sol´e (2000) have used a similar approach to study rainforest dynamics. Olson and Sequeira (1995) give a review of similar types of ecosystem modeling. The remainder of this chapter is organized as follows. In Section 9.2, an extended description of the model is given. Section 9.3 motivates the choice of parameter values. Section 9.4 describes the dynamic behavior of the model. Section 9.5 provides a sensitivity analysis. Section 9.6 draws attention to the effect of stochastic recruitment. Finally, conclusions are drawn in Section 9.7.

9.2. MODEL DESCRIPTION The model simulates the population dynamics and spatial distribution of birds and shellfish species. Birds prey on shellfish. If the shellfish stock is too low, birds will starve and die. Each year new birds are born and new shellfish are recruited. Shellfish are eaten by birds (and fished by humans). Adult shellfish are immobile and stay at the location where they settled as larvae. Birds move around searching for food. The model simulates the stock size of each species in the ecosystem in space and time. Time is divided into discrete time steps. Space is divided in grid cells. At each time step, the stock size of each species is computed in each grid cell. Figure 9.1 shows schematically the simulation area. The area is divided into cells. Each cell has certain habitat characteristics and contains birds and shellfish. Birds feed in a cell, depending amongst others on the shellfish density in that cell. Birds can move freely from one cell to another. Shellfish are immobile. Due to the tide, each grid cell, depending on its depth, will be submerged during part of the day. Depending on the characteristics of a cell, it can be a more or less suitable environment for shellfish. The area is bounded, meaning that birds do not leave the grid. Birds remain in the area, but they do move from one cell to another. Time is divided into two periods: winter (210 days, approximately September to April) and summer (the rest of the year). In winter, the birds stay in the modeled area (the grid). In summer, most birds migrate to other areas to breed. Parents and offspring return the next winter. Because birds are situated in the modeled area only during winter, the winter period is computed for daily time steps for which food intake, starvation, and migration between grid cells are simulated whereas the summer period is computed in one time step only. In this single

1

Information about Cormas can be found on the internet at: http://cormas.cirad.fr/indexeng.htm.

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 151 0

1

movement of birds cockle suitability

mussel suitability

1

0 0

1 depth (part of the day above water)

Figure 9.1. A simulation grid. Note: Stock levels of each species are simulated in each grid cell. Each cell has certain characteristics such as depth and suitability for shellfish recruitment, birds can move freely from one cell to another.

time step, recruitment of shellfish and birth of new birds are modeled for the entire summer period. The model consists of two types of birds (parameterized so as to resemble oystercatchers and common eiders) and two types of shellfish (parameterized so as to resemble mussels and cockles). The shellfish are divided into two cohorts, juveniles and adults. Five processes control the change in stock size of shellfish: r r r r r

food intake by birds, additional mortality, recruitment of juveniles and aging (and growth) of adults, fishing. The change in stock size of birds in a grid cell is controlled by three processes:

r birth, r death (starvation), r emigration from and immigration to a grid cell. Each of these processes depends on the stock size of shellfish and birds. Starvation and migration depend on food intake by birds. Birds can feed on alternative prey and the mass of shellfish declines during winter. When birds do not eat enough they either die or go away. A schematic overview of the model is shown in Figure 9.2.

152

PART II: BIOECONOMIC MODELING

mass decrease starvation fishing new-borns

add. mortality Shellfish

Birds food intake

ageing

recruitment

migration

Figure 9.2. Schematic outline of the model processes within one grid cell. Note: Processes in gray take place during the summer period in one time step. Other processes take place during all other time steps in the winter period. Arrows pointing toward a box denote an increase of the value of that state variable, arrows pointing outward denote a decrease. Dashed arrows show that starvation and migration are influenced by food intake.

Mathematically the system is described by the following differential equations: dn im (x, t) = Rim (x, t, n im (x, t)) − Mim (n im (x, t)) dt  ˜ im j (x, t, n lm (x, t), pk (x, t)) − Fim (x, t) − W

(1)

j

d p j (x, t) = B j (t, p j (x, t)) − S j (x, t, n lm (x, t), pk (x, t)) dt − E j (x, t, n lm (x, t), pk (x, t)) + I j (x, t, n lm (y, t), pk (y, t))

(2)

for every predator j and every prey i, of every age class m, in every grid cell x, where, n im (x, t) [#]: the number of prey i of age class min cell x, at time t, p j (x, t) [#]: the number of predator j in cell x, at time t, Rim (x, t, n im ) [#/summer]: the recruitment (juveniles) or aging (adults) of prey i, age class m, in cell x, during summer, Mim (n im ) [#/s]: additional mortality rate of prey i, age class m, in cell x, ˜ im j (x, t, n l , pk )[#/s]: the rate at which predator j eats prey i, age class m, in cell W x, with prey density n l and predator density pk Fim (x, t) [#/summer]: the number of prey i, age class m, fished in cell x, during summer, B j (t, p j ) [#/summer]: the number of newborn predators j in cell x, during summer, S j (x, t, n l , pk ) [#/s]: the starvation rate of predators j in cell x, E j (x, t, n l , pk ), [#/s]: the emigration rate of predator j from cell x, I j (x, t, n l , pk ), [#/s]: the immigration rate of predator j to cell x. Here the notation n lm and pk indicate that a function depends on every prey, including each age class, and every predator, respectively. Further y indicates that the

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 153 function depends on every grid cell but x. The unit in which each variable is expressed is given between square brackets. For notational convenience and clarity, the functions’ variable lists are partly omitted in the list of descriptions and in the following. For computational convenience, aging is modeled as a form of recruiting from one age class into the next. The function Ri1 (x, t, n i1 ) is the net effect of the gains of juveniles through recruitment and the loss of juveniles through aging into adults. Consequently Ri2 (x, t, n i1 ) are those juveniles that become adults. The recruitment of adults per definition equals the survival of the juveniles. The model solves numerically the coupled differential Equations (1) and (2) by using a fourth order Runge–Kutta scheme. The following paragraphs describe in detail the processes denoted by capitals. These control the dynamics of the ecosystem.

9.2.1. Food intake by birds Feeding by birds directly links the stock levels of the predator and prey through depletion of shellfish on one hand and starvation and movement of birds on the other. If a bird cannot find enough food it will either starve or move away to another grid cell. Feeding depends on two factors: (1) whether a shellfish species is a suitable prey and (2) the time it takes a bird to find and eat a shellfish. The diet, meaning the prey species that will be eaten, depend on prey densities that change over time. A bird’s diet is determined by Charnov’s model (Charnov, 1976). According to this model a predator will add a certain prey to its diet if this increases the predator’s intake rate, measured in energy per time. This means that prey with a relatively high energy content and relatively low handling time are eaten. The average intake rate per bird as a function of both the densities of prey and predators, known as the generalized functional response (van der Meer and Ens, 1997) is given by Beddington’s equation (Beddington, 1975). It is adjusted for multiple prey species as in (Charnov, 1976). Two extra factors are included. One to take into account that birds may feed only during daylight hours and a second one to account for the tide. Some species feed during the time the grid cell is under water while other species feed while the cell is above water. The following equation expresses the intake rate:  a j m lm (t)(n lm (x, t)/A)  lm W j (x, t, n lm , p j ) = L j d j (x) , (3) h lm j (n lm (x, t)/A) + q j p j (x, t)/A 1 + aj lm

where

W j (x, n lm , p j ) (kg/s): intake rate of predators of species j in cell x, m lm (t) (kg): mass of prey species l, age class m, at time t a j (m2 /s): search coefficient of predator of species j, h lm j (s): handling time of predator j of prey l, age class m, q j (m2 ): interference area of predator j, A (m2 ): surface of a grid cell that is covered with shellfish beds d j (x)[1]: part of the tidal period that predator j can catch prey in cell x, L j [1]: part of the day during which predator j feeds (e.g. daylight hours). The intake rate is expressed in ash-free dry mass (AFDM) per second. We assume mass is proportional to energy content. Here the notation lm indicates that only those

154

PART II: BIOECONOMIC MODELING

prey species, l, and age classes, m, are incorporated in Equation (3) that are part of a predator’s diet. Prey are ranked according to their mass/handling time ratio and subsequently added to the diet as long as the intake rate increases. The result is that those prey are incorporated in the diet that maximize the predator’s intake rate. The equation gives the rate of potential food intake for a bird species. This sets the maximum amount a bird can eat in 1 day. It is assumed that birds eat during daylight hours only, hence the factor L j . Furthermore only part of the tidal cycle (day) can be used to feed because some birds dive for their food while others feed on dry mudflats. The factor d j expresses this. Birds eat only so much per day. So, their daily intake has a maximum that is the minimum of Wnec, j and W j , expressed in kg per second. Wnec, j is the daily intake a bird needs in order not to starve. Furthermore, birds will not feed on shellfish when the intake rate is below some minimum, Wmin . Whenever the intake rate becomes less than Wmin , birds will feed on alternative prey. In the case of oystercatchers, these prey are found upshore or on inland fields. The total daily intake is allocated over the prey species in the predator’s diet, proportional to the mass of each prey species present in the grid cell. This defines ˜ im j (x, t, n lm , p j ), the rate at which predator j eats prey i of age class m. W

9.2.2. Seasonal decline in shellfish condition The intake rate of birds depends on the mass of shellfish. During winter the mass of shellfish decreases. This decrease is more or less linear for both cockles (Klepper, 1989) and mussels (Goss-Custard et al., 2001). Therefore, we also assume that the mass of shellfish decreases linearly over winter. m lm (t) = m 0im − bim t where m 0im : the mass of prey species i, age class m, at the start of winter bim : the mass decline coefficient of prey species i, age class m

9.2.3. Additional mortality of shellfish Apart from being eaten by birds or fished by humans, other factors cause shellfish to die, e.g. diseases, food shortage, storms, ice winters, or being preyed upon by animals not explicitly modeled. This additional mortality is assumed linearly dependent on stock size for adult shellfish. It is expressed as: Mim (n im ) = kim n im (x, t),

(4)

where kim (1/s): additional mortality coefficient. Bell et al. (2001) suggest that additional mortality may not be proportional to stock levels and in fact may compensate for mortality from predation. We choose to keep the additional mortality coefficient constant so that unknown varying additional mortality does not obscure the effect of predation. A fixed additional mortality rate for cockles of 1% per month was used by Rappoldt et al. (2004) in a model of oystercatchers preying

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 155 on cockles. This model predicted quite well the cockle stocks remaining at the end of winter for both the Dutch Wadden Sea and the Oosterschelde (Rappoldt et al., 2003a,b; Ens et al., 2004). The additional mortality of juvenile shellfish is taken care of in the aging process. Not all juveniles become adults.

9.2.4. Recruitment of shellfish Recruitment of shellfish occurs during the summer period (one time step). The number of juvenile recruits that settle in a grid cell depends on the suitability (e.g. bottom characteristics) of that particular grid cell. Recruitment also depends on a random factor that expresses exogenous effects such as harsh winters in the year preceding recruitment, which may greatly influence subsequent recruitment (Beukema, 1982; Beukema and Dekker, 2005). Here, these random effects are mostly ignored. Hence, X is set at 1. The following equation denotes recruitment. Ri1 (x, t, n i1 ) = si (x)ci1 (t)X,

(5)

where Ri1 (x, t, n i1 ) [#/s]: recruitment rate of juvenile prey of species i in cell x, si (x) [1]: suitability of grid cell x, for species i [a number in the interval (0,1)], ci1 (t) [#/summer/m2 ] juvenile recruitment coefficient of species i, X [1]: exogenous effect coefficient, stochast with average 1 and a lognormal probability distribution function. In the winter period shellfish do not recruit.

9.2.5. Aging of shellfish The existing juvenile stock turns into adults at the end of each year in the summer period. Not all juveniles achieve adulthood. This mortality in the summer period is expressed through the recruitment rate of adults. The following equation expresses ageing of shellfish. Ri2 (t, n i2 ) = ci2 (t)n i1 (x, t),

(6)

where Ri2 (t, n i2 ) [#/s]: recruitment of adults of species i in cell x, per unit of time, ci2 (t) [1]: adult recruitment coefficient of species i (i.e. juvenile survival). In the winter period shellfish do not age.

9.2.6. Fishing Fishing takes place during the summer period. During this time step, shellfish is withdrawn from the stock, depending on, e.g. shellfish densities but foremost on the fishing scenario that is employed. Several fishing scenarios can be implemented. For this study, only the Nature scenario is used. In this scenario, fishing is not allowed anywhere at any time.

156

PART II: BIOECONOMIC MODELING

9.2.7. Recruitment of birds Every summer period new birds are born. The survival of newborns is expressed through a recruitment coefficient. The birds return to the grid cell they left last winter. The surviving newborns return to the same grid cell as their parents. There is no evidence that offspring of migratory oystercatchers and common eiders actually do return to the same spot as their parents, but because birds quickly disperse over the grid, we see no harm in modeling it in this way. The birth rate includes additional mortality. It is a net growth rate that includes everything except starvation and migration. The birth rate of birds is expressed by the following equation. B j (t, p j ) = r j (t) p j (x, t),

(7)

where r j (t) (1/s): the recruitment coefficient of predator species j. For both oystercatchers (Goss-Custard et al., 1996) and common eiders (Swennen, 2002) there is evidence that production of fledglings is density dependent at the scale of the local breeding population or the breeding colony, meaning that parameter r j is a function of p j . However, populations of both oystercatchers and common eiders have increased, because they occupied new breeding space: oystercatchers moved inland (Goss-Custard, 1996) and common eiders founded new colonies on each of the Wadden Sea Islands (Swennen, 1991). In the model we effectively describe the overall recruitment of birds belonging to different colonies or local breeding areas that overwinter in the same area instead of the recruitment within one colony or breeding area. Because it is not clear how we should describe r j dependence of p j in this case, we choose to keep the recruitment function (Equation 7) without this effect. It would involve at least one extra parameter for which we do not have good estimates. Furthermore, in Section 9.5 of this chapter we show that the overall population dynamics are not very sensitive to the value of r j . Finally, in the winter period no birds are born.

9.2.8. Starvation of birds Birds have a chance of dying when during a time step their intake rate of shellfish is too low. A bird has a maximum probability of dying ( f j ) when its intake rate is 0. The probability of starvation is supposed to decreases proportionally until it is 0 when the intake rate is Wnec, j , the necessary and the maximum daily intake. Birds need Wnec, j to survive. If a grid cell cannot provide every bird with the food it needs, some birds will die. The number of starving birds is proportional to the lack of available food. The following equation expresses the starvation rate:   W j (x, n l , pk ) + Walt, j S j (x, n l , pk ) = f j 1 − p j (x, t), W j (x, n l , pk ) Wnec, j + Walt, j < Wnec, j S j (x, n l , pk ) = 0, W j (x, n l , pk ) where

+ Walt, j ≥ Wnec, j

(8)

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 157

starvation probability

1

f

0

W nec - W alt

W nec

intake rate

Figure 9.3. Starvation, the probability that a bird starves to death during a time unit increases linearly when its food supply is less than its daily need. Note: The gray line depicts the starvation probability when there is an alternative source of food. The black line depicts the starvation probability when only shellfish are available as food source.

f j (1/s): the probability that predator j will starve during a unit of time when no food is available, Wnec, j (kg/s): the minimum daily intake rate necessary for predator j, to survive Walt, j (kg/s): the daily intake rate of predator j that can be achieved by foraging on alternative prey Walt denotes the potential intake of alternative prey and is considered constant. Birds will first feed on shellfish. If this is not enough for survival they can then feed on alternative prey for at most Walt . Figure 9.3 shows the effect of the alternative prey on the starvation probability graphically.

9.2.9. Emigration of birds Birds prefer grid cells where they can find their daily food rather quickly. They remain in cells where food is abundant and emigrate with higher probability if their intake rate becomes lower. When a bird cannot find any food, it will surely emigrate. If the intake rate is higher than 0, the emigration probability decreases proportionally with the potential intake rate. At some point, which we indicate as the emigration threshold intake rate, a bird considers the intake rate high enough and it remains at that site. The emigration rule is a variant of the delayed departure rule (Blaine and DeAngelis, 1997).

158

PART II: BIOECONOMIC MODELING

emigration probability

1

0 Wemi

intake rate

Figure 9.4. Emigration, the probability that a bird emigrates during a time unit increases linearly when its intake rate is less than satisfying.

The following equation expresses the emigration rate: E j (x, n lm , pk ) = e j



W j (x, n lm , pk ) 1− Wemi, j

E j (x, n lm , pk ) = 0,



p j (x, t), W j (x, n lm , pk ) < Wemi, j

W j (x, n lm , pk ) ≥ Wemi, j

(9)

where e j ,(1/s): the probability that predator j will emigrate during a unit of time in which no food is available, Wemi, j (kg/s): the minimum daily intake rate necessary for predator j, to remain at its location The graph in Figure 9.4 depicts Equation (9). The emigration rule works out slightly different for common eiders compared to oystercatchers, because we assume that common eiders do not suffer from interference, whereas oystercatchers do (see later). When the shellfish density is decreased to a certain level, common eiders will leave and will continue to emigrate from that cell during the rest of the winter because the shellfish density and thus the intake rate, will only decrease further. The intake rate of oystercatchers does not only depend on shellfish density, but is also negatively influenced by the density of conspecifics through interference. When the shellfish density has decreased to the emigration threshold level, some birds will leave. Therefore, in the next period, the remaining birds are bothered

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 159 less by interference. So, although the shellfish density has decreased, the intake rate may increase and the remaining birds may stay for one or more periods.

9.2.10. Immigration We assume that birds can move to each grid cell within a day, i.e. a time step. A bird that emigrates from a grid cell will move randomly to one of the other grid cells. According to Lande (1987) and, King and With (2002), only the amount of habitat matters and not the spatial structure in case of random dispersal. Here, that means that it does not matter where a cell is located and who its neighbors are as long as its habitat parameters (depth and recruitment suitability) are maintained. No other nearby feeding grounds is considered. In winter, all birds remain in the modeled area. So, the total number of birds (summed over all cells) that immigrate is equal to the total number of birds that emigrate. Moreover, the immigration to a grid cell is the sum of the emigration to that cell from all other cells. 1  I j (x, n lm , pk ) = E j (y, n lm , pk ), (10) N y=x where N [#]: the number of grid cells.

9.3. PARAMETER VALUES The time step of the simulation is 2 days. The spatial grid has a 10 × 10 resolution, each grid cell measures 1 km2 (1 × 1 km). That is, it represents a surface of which 2 1 km is covered with shellfish beds. In the real world only a few percent of the tidal and subtidal areas is covered with shellfish beds (Ens et al., 2004), but it is only this area that concerns us here. Depth increases from east to west. It is represented by the part of the tidal cycle that a grid cell is above water. This value is set at 0 for the first 2 km (deep water) and from then on increases linearly from 0 to 1 at the east side. The suitability for mussels increases linearly from 0 at the north side to 1 at the south side, and the suitability for cockles decreases linearly from 1 at the north to 0 in the south. Due to the random dispersal of birds it does not matter where in the grid each cell is located. For reasons of easier interpretation, the habitat parameters are increasing linearly along each side of the grid. Recruitment per cell is kept constant over time for analytical purposes. Physically, this is interpreted as the absence of external effects. The initial conditions are chosen as follows. At the start of the simulation, birds and adult shellfish are distributed homogeneously over the grid; birds with a density of 0.001/m2 and adult shellfish with a density of 37.5/m2 . Initial juvenile shellfish distribution is governed by the suitability of the grid cell, according to si (x) × 100/m2 . The parameter values in the model are shown in Tables 9.1 and 9.2, one for the predators (birds) and one for the prey (shellfish). In the following paragraphs, we motivate the choice of parameter values.

Table 9.1. Parameter values for the birds r (year−1 ) Oystercatcher Juvenile cockle Adult cockle Juvenile mussel Adult mussel Common eider Juvenile cockle Adult cockle Juvenile mussel Adult mussel

a (m2 /s)

q (m2 )

0.0015 0.003 0.002 0.005

500 500 500 2000 0

0.1

0.1 0.0015 0.003 0.002 0.005

L (%)

Wnec (kg/s)

Walt (kg/s)

Wmin (kg/s)

Wemi (kg/s)

f (day−1 )

e (day−1 )

100

0.5 × 10−6

0.25 × 10−6

0.6 × 10−6

2 × 10−6

0.1

1

100

2 × 10−6

1 × 10−6

2.4 × 10−6

8 × 10−6

0.1

1

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 161 Table 9.2. Parameter values for the shellfish and shellfish-dependent bird parameters m 0 (kg) 50 × 10−6

b (kg/s) 1.38 × 10−12

Juvenile mussel Adult mussel 600 × 10−6 16.52 × 10−12 Juvenile cockle 35 × 10−6 0.96 × 10−12 Adult cockle 400 × 10−6 11.02 × 10−12

c1 (year−1 m−2 ) c2 (%) k (s−1 )

h OC (s) h eider (s)

100 – 100 –

25 75 15 50

– 75 – 75

– 5 × 10−9 – 5 × 10−9

5 15 5 10

9.3.1. Oystercatcher mortality and production of chicks (r) Goss-Custard et al. (1995a) assembled all available information on production of chicks and mortality of oystercatchers in Europe. On the basis of this information Goss-Custard et al. (1995b) constructed parameters for their modeling of the population dynamics of the oystercatcher. Annual mortality of adults is estimated at 0.09 for Atlantic birds and 0.04 for Continental birds. About every seven winters the Continental birds experience a severe winter and mortality is much higher, between 0.1 and 0.2. Mortality of juveniles and immatures is higher than mortality of adults, so 0.1 seems a reasonable estimate of annual mortality rate in our model, where we do not distinguish between age classes and populations. In a stable population, the mortality is compensated by the production of chicks. From accumulated data on clutch size, egg mortality, chick mortality, and juvenile survival, it is possible to calculate that each breeding pair produces between 0.21 and 0.44 chicks that survive the first winter. This is between 0.11 and 0.22 chicks per adult. If 80% of the oystercatchers breed (a high estimate), this amounts to 0.08–0.17 surviving chicks per oystercatcher. If 50% of the oystercatchers breed (a low estimate), this amounts to 0.05–0.11 surviving chicks per oystercatcher. Thus, both from mortality estimates and from chick production figures a recruitment rate of 0.1 seems reasonable.

9.3.2. Common eider mortality and production of chicks (r) Swennen (1991) provides information on the population dynamics of common eiders for the Vlieland colony, which is the oldest and the largest colony in the Dutch Wadden Sea. Annual survival of female breeders is 0.95. The age at first breeding lies between 3 and 4 years, as 37% of the females first breed at 3 years of age and the remainder at the age of 4 years. Survival of ducklings to recruitment age is 0.5. From this, Swennen calculates that each female should fledge a brood of 0.25 per year for population stability. This is 0.125 fledgling per adult common eider when the sex ratio is 1. At the time of his calculation, the population seemed stable, but the population of adult nonbreeders was assumed 0. Christensen and Noer (2001) provide data for a Danish colony. Clutch size is 4.33 eggs/nest, 91.5% hatching success, and adult annual survival of 0.87. The cumulative percentage of females that breeds increases with age: 14%, 45%, 77% and 100% at age of 2, 3, 4, and 5 years, respectively. Based on these data we choose a value of 0.1 for r

162

PART II: BIOECONOMIC MODELING

9.3.3. Oystercatcher food consumption and feeding time (Wnec , Walt , Wmin , L) According to the review by Zwarts et al. (1996c) the daily energy needs of free-living oystercatchers (E in kJ per day) depend on body mass (M in g) in the following way: E = 0.061M 1.489 This is only true for birds which do not change weight and which live under thermoneutral conditions. Kersten and Piersma (1987) state that extra costs for thermoregulation are incurred when the ambient temperature (T in ◦ C) drops below 10 ◦ C. For each degree Celsius below this point the bird needs an extra 30 kJ per day. Thus: E = 0.061M 1.489 + 30(10 − T ) when T < 10 ◦ C When birds increase in weight, they require extra energy, whilst they require less energy when they are losing weight. According to Kersten and Piersma (1987) 1 g increase in body mass requires an extra food consumption of 1.2 g AFDM of food. Similarly, losing 1 g body mass means that the birds have to eat 1.2 g AFDM less food. The digestive efficiency of oystercatchers is estimated at 85% (Kersten and Piersma, 1987; Speakman, 1987; Kersten and Visser, 1996; Zwarts and Blomert, 1996). According to a review by Zwarts et al. (1996b), the energy content of 1 g AFDM corresponds to between 22 and 22.5 kJ for marine invertebrates (and the pellets used by Kersten and Piersma, 1987). This means that a gross intake of 1 g AFDM corresponds to 19 kJ metabolizable energy. Thus, a change in body mass of 1 g requires or yields 22.8 kJ, depending on the direction of the change. Thus, when temperatures are below 10 ◦ C and the birds change weight d in g per day: E = 0.061M 1.489 + 30(10 − T ) + 22.8d when T < 10 ◦ C and d = 0 The gross daily food intake (G) is measured in gram AFDM per day and can be calculated from E as follows: G = E/19 This leads to an estimation of daily food over the year as indicated in Table 9.3. Thus, the total daily food consumption varies between 36 and 59 g AFDM in the course of the year. For Wnec , this is expressed in kg/s: between 4.2 × 10−7 and 6.9 × 10−7 . According to (Zwarts et al. 1996a) oystercatchers need 5 h feeding per day to obtain their maximum daily energy needs at the average year-round rate of 2 mg AFDM/s feeding. This results in Wnec,oy = 4.2 × 10−7 kg/s. We choose 5 × 10−7 kg/s for the winter period. Oystercatchers can find alternative food sources on the mud flats during low tide and in the fields during high tide. Stillmann et al. (1996) report intake rates of 0.67 mg/s AFDM upshore and 0.53 mg/s in the fields. Therefore we take Wmin , the minimum intake rate at which oystercatchers start feeding on alternative prey at 0.6 × 10−6 kg/s. Studies show (e.g. Goss-Custard et al., 2004) that when shellfish are scarce oystercatchers feed for some 50% on alternative prey. Therefore we choose Walt,oy = 2.5 × 10−7 , which is half of Wnec,oy .

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 163 Table 9.3. Daily food consumption of an oystercatcher in the Dutch Wadden Sea Jan

Feb Mar

Apr May Jun

Jul

Aug Sep Oct Nov Dec

◦

Temperature ( C) 2 2 5 8 12 15 17 17 Body weight (g) 607 621 575 566 539 530 535 549 Basic consumption 45 46 41 40 37 37 37 39 (g AFDM/day) Thermoregulation 13 13 8 3 0 0 0 0 (g AFDM/day) Weight change 0 1 −2 0 −1 0 0 1 (g AFDM/day) Total daily 57 59 47 43 36 36 37 39 consumption (g AFDM)

14 8 5 546 555 593 38 39 43

3 610 45

0

3

8

11

0

0

1

1

38

43

53

57

Oystercatchers feed regularly during the night. Hulscher (1996) has reviewed the available data on the feeding behavior during the night and found some indication that intake rates are lower during the night than during the day. Some observations suggest that intake rates are lower during the night, but field observations covering the entire low water period indicate little difference between the intake rates during the day and during the night. In the most comprehensive and sophisticated study to date, Sitters (2000) shows differences between birds employing different feeding techniques. However, the general conclusion is that intake rates between day and night are roughly similar. In winter, energy needs are high and a digestive bottleneck prevents the birds meeting their daily energy needs during daylight only (Zwarts et al., 1996c). Thus, it seems reasonable to assume that oystercatchers feed during both day and night, i.e. L oy = 100%.

9.3.4. Common eider food consumption and feeding time (Wnec , Walt , Wmin , L) The best estimate of the food consumption of free-living common eider ducks has been made by (Nehls, 1995). In a subsequent paper, Nehls et al. (1997) provide an estimate of the daily food consumption (in terms of AFDM) of the common eiders in the Wadden Sea throughout the year (Table 9.4). Thus, the total daily food consumption varies between 130 g AFDM and 180 g AFDM. For Wnec , this is expressed in kg/s: between 1.5 × 10−6 and 2.1 × 10−6 . We Table 9.4. Daily food consumption of a common eider

Total daily consumption (g AFDM)

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

180

180

180

170

170

150

130

130

130

150

170

180

Source: Nehls et al. (1997).

164

PART II: BIOECONOMIC MODELING

assume that their daily need is at the high end of their average daily consumption. So, we choose 2 × 10−7 kg/s for the winter period. Common eiders switch to less preferred alternative food sources when cockles and mussels are scarce. Common eiders can feed on crabs, with a high risk of parasitism, or move to deeper water, where they can feed on the shellfish Spisula subtruncata (Swennen, 1976; Camphuysen et al., 2002). We assume that like oystercatchers, common eiders feed for some 50% on alternative prey when shellfish are scarce. Therefore we choose Walt,ei = 2.5 × 10−7 . Furthermore, we assume that common eiders switch to alternative prey at a value for Wmin, ei = 2.4 × 10−6 . Nehls (1995) found that common eiders are partly active during the night. From January to March, 17% of all dives was made during the night, but on occasion 50% of all dives occurred during the night. Therefore we take L ei = 100%.

9.3.5. Handling time of oystercatchers (h) Zwarts et al. (1996b) reviewed all published studies on feeding behavior of oystercatchers, and provided relationships between handling time and prey size for all prey species. Handling time depends on prey type, prey size, and the way the prey is opened (Table 9.5). Norris and Johnstone (1998) estimate handling times of 50 s for large cockles and 15–20 s for the smallest size classes. Triplet et al. (1999) measure a handling time of 14 s for cockles 23 mm in size. Given this variability, handling times of 15, 25, 50, and 75 s for juvenile, respectively, adult cockles and mussels fall within the range of observations.

9.3.6. Handling time of common eiders (h) Common eiders swallow the prey whole, so handling times are shorter compared to oystercatchers, which spend a lot of time opening the prey and cutting the flesh loose from the shell. Nehls (1995) has published information on the handling times of common eiders feeding on large mussels in the wild (Table 9.6). Table 9.5. Handling times (s) for different species and sizes of prey Mussel ventral a b Size (mm) 10 20 30 40 50 60

0.712 1.313 15 36 62 90 121 154

Mussel dorsal 0.443 1.432 12 32 58 87 120 156

Mussel stab 0.975 1.081 12 25 39 53 67 82

Cockle hammer 0.054 1.945 5 18 40 71 109

Note: Handling time h = aLb (handling time h in s and prey length L in mm). Source: Zwarts et al., 1996b.

Cockle stab 0.053 1.846 4 13 28 48 73

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 165 Table 9.6. Handling (s) of common eiders feeding on large mussels

“Swimming” Diving

June/July

August/September

November/March

7.8 12.8

14.4 12.8

22.0 16.1

Source: Tables 5.1 and 5.2 in Nehls, 1995.

Unpublished personal field observations of common eiders feeding on cockles indicate that swallowing of cockles takes only a few seconds. We roughly estimate the handling times for juveniles cockles and mussels, to be 5 s and those of adult cockles to be 10 s and 15 s for adult mussels.

9.3.7. Functional response of oystercatchers (a, q) Van der Meer and Ens (1997) estimated the parameters of the Beddington functional response equation for cockle and mussel feeding oystercatchers based on the literature They arrive at a = 0.0007 m2 /s, q = 1000 m2 . Rappoldt et al. (2004) estimate a value of 0.00,086 m2 /s for a in the Holling disc equation which ignores interference. Stillmann et al. (1996) find much larger values. They observed a search speed and a search path width. If we multiply these values for search speed and search path width, we arrive at a value for a of about 0.009 m2 /s. Norris and Johnstone (1998) using an equation given by Zwarts and Blomert (1996) find values varying between 0.001 and 0.005 depending on the size of the shellfish. We base the value for a on Norris and Johnstone (1998) because using those values mean that oystercatchers experience little interference when the predator density is less than 100, which coincides with the observations shown in Stillmann et al. (2001). Interference among mussel feeding oystercatchers is stronger than interference among cockle feeding oystercatchers (Ens and Cayford, 1996). Here, we assume that the interference of mussel feeding oystercatchers is twice the value estimated by Van der Meer and Ens (1997) and interference of cockle feeding oystercatchers is half of this. These values are such that the intake rate of oystercatchers matches the observations of Stillmann et al. (2001) more or less. The intake rates are shown in Figure 9.5.

9.3.8. Functional response of common eiders (a, q) No published information exists on the functional response of common eiders or on interference. To date, the only evidence for interference comes from a study of Nehls and Ketzenberg (2002) who show that fewer birds fed on a preferred mussel bed when the total population increased. However, since common eiders often feed in large tight flocks, it is assumed that interference is so small that it can be ignored. Hence, q = 0 m2 . Furthermore, we assume that common eiders search as quickly and efficiently as oystercatchers. Therefore the same values for a apply as for oystercatchers. This is a rough guess, but a better estimate is not available yet and by taking the same values as for oystercatchers we are better able to study the effect of those parameters, which we do know to be different.

166

PART II: BIOECONOMIC MODELING 10

W [mg AFDW/s]

8 6 4 2 0 1

10

100

1000

--1

oystercatcher density [ha ] a

b

c

d

e

Figure 9.5. Graphical representation of the generalized functional response of the oystercatcher. Note: The intake rate is plotted against the density of conspecifics for several densities of adult prey; (a) cockle density:0/m2 and mussel density: 500/m2 , (b) cockle density: 0/m2 and mussel density: 100/m2 , (c) cockle density: 100/m2 and mussel density: 100/m2 , (d) cockle density: 100/m2 and mussel density: 0/m2 , (e) cockle density: 500/m2 and mussel density: 0/m2 .

9.3.9. Starvation (f) Oystercatchers and common eiders have energy reserves in winter, which carry them through periods during which they cannot feed. Hulscher (1989) calculated the expected survival of oystercatchers during a frost period based on the measured weight distribution and the known energetics of the birds. The first birds would have died after 3 days and virtually none would have survived more than 10 days without food. Such a mortality curve cannot be achieved with a constant daily mortality rate for birds that do not feed. However, a mortality rate of 0.1 is a reasonable approximation. After 3 days without food, 27% would have died (which is too high) and after 10 days, 65% (which is too low). Common eiders have a lower mass-specific metabolic rate. So, they can probably survive a little longer without food for a same proportion of extra reserves. But for the common eider, a mortality rate of 0.1 also seems a reasonable estimate.

9.3.10. Shellfish mass decrease (b) According to Goss-Custard et al. (2001), the overwinter decrease in flesh mass (AFDM) of mussels 45 mm long is 62.5 mg/month, i.e. a decreases by about 50% per winter. Norris and Johnstone (1998) measure a 49.3% decrease in flesh mass of cockles during winter in the Burry Inlet. Similar declines are reported by Klepper (1989) and Zwarts (1991). In line with the literature, we choose a 50% decrease of shellfish mass during winter. This leads to the values in Table 9.2.

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 167

9.3.11. Additional mortality (k) Rappoldt et al. (2004) estimate additional mortality of shellfish to be 1% per month. Thus k is set at 5 × 10−9 /s, assuming exponential decrease (Equation 4) for cockles as well as mussels. To account for the fact that in the model all additional mortality takes place during 210 days of winter instead of 365 days per year, k is rounded above, i.e. the ceiling of ln(0.99)/30 × 24 × 3600. According to these parameter values, an oystercatcher needs 43.2 g AFDM of shellfish meat per day, which are 108 adult cockles at the start of winter. An oystercatcher can find and eat this amount in less than 2 h, when 10 oystercatchers occupy an area of 104 m2 with a cockle density of 50/m2 (Equation 3). The intake rate for oystercatchers is shown in Figure 9.5, for several prey densities at the start of winter. This is the intake rate while the bird is feeding, when the tide is out. Triplet et al. (1999) measured an average feeding rate of 5.7 cockles per 5 min in the Baie de Somme. The average oystercatcher density was 165/ha, the average cockle density was 1200/ha, and their average size 23 mm. Assuming those cockles had an AFDM of 150 mg, the intake rate of the oystercatchers was 2.85 mg/s. The Beddington Equation (3) combined with the parameter values in Tables 9.1 and 9.2 estimates a intake rate of 2.85 mg/s as well for these values for densities and mass. The interference area for common eiders is 0, thus in effect common eiders have a Holling type II functional response (Yodzis, 1989). It does not matter how many common eiders occupy a grid cell, their intake rate will remain the same. Table 9.7 shows intake rates for the same prey densities as used in Figure 9.5. A common eider needs 173 g AFDM shellfish meat per day, which equals 288 adult mussels at the start of winter. In an area with a mussel density of 50/m2 , a common eider can find and eat those in less than 1 h and a half. Common eiders dive for prey, so they only eat on intertidal flats when these are covered with water. Oystercatchers feed on exposed mudflats. Thus, the two birds compete for food in those areas that are dry during a certain period of the tide and are sufficiently submerged in another period. In our model, adult cockles and mussels have the same mass/handling time ratio for common eiders as well as oystercatchers. As the mussels are quicker to find and heavier, common eiders prefer mussels over cockles. When the cockle density is sufficiently high (about 60%) compared to the mussel density cockles are added to the common eider’s diet. For oystercatchers it is more complicated because of interference. With few competing birds, oystercatchers prefer mussels because they are quicker to find. However, with many competitors, oystercatchers will prefer cockles because interference is lower. Table 9.7. Intake rates of common eiders for different densities of adult cockles and adult mussels. (The same shellfish densities are used as in Figure 9.5)

2

Cockles (/m ) Mussels (/m2 ) W (mg/s)

a

b

c

d

e

0 500 38.3

0 100 35.3

100 100 36.2

100 0 30

500 0 36

168

PART II: BIOECONOMIC MODELING

9.4. SYSTEM BEHAVIOR The model is simulated on a hypothetical 10 × 10 grid. The geography and spatial heterogeneity of the model world do not resemble a real world ecosystem. However, differences in depth and suitability for recruitment are characteristics of any real world system. The geography, a square along which sides, depth, and suitability for shellfish recruitment increase linearly, is chosen to interpret the population dynamics of the ecosystem model easily. Annual recruitment of shellfish is taken constant. This is not realistic. Recruitment varies randomly over the years. For the sake of studying the internal dynamics of the system, we choose to exclude this stochastic external driving force. In Section 9.5, the stochastic behavior of the model is illustrated when random recruitment is applied. The ecosystem produces a certain amount of adult shellfish each year. This amount is described by variables such as recruitment, mortality, and the mass of shellfish. With this annual production of shellfish, an average number of birds can be sustained. The feeding strategy of the two bird species represented by parameters such as a, h, and Wemi , create the dynamics and the composition of the equilibrium. Depending on the parameter values, birds either efficiently eat the new supply each year, resulting in a constant number of birds and shellfish at the start of winter, or the ecosystem experiences an ongoing over- and undershooting of the sustainable stock levels. When shellfish stock levels are high, bird numbers will grow. When they are low bird numbers decrease. When the size of the bird population has grown above the sustainable number, birds eventually will add (more) juvenile shellfish to their diet. Therefore, less adult shellfish are available the next year, resulting in a fast collapse of shellfish and bird numbers. Bird species can switch to less preferred alternative food sources when cockles and mussels are scarce. The availability of alternative prey dampens the extreme collapse of both bird species. Although one could say that birds live on annually produced shellfish, a sufficiently high stock level of shellfish is necessary because birds need a high enough density to find their daily meal. In this model, there is no stock–recruitment relation for shellfish. Thus, the shellfish stock level is not needed to produce offspring that can mature to become bird food. In the model, common eiders have a refuge for competition. On the west side of the grid, it is too deep for the oystercatcher to feed. Fishermen can fish in the deep cells, so when fishing is allowed common eiders lose their exclusive use. During the winter period, the population sizes of all animals decrease. Shellfish are eaten and some birds die of starvation. The following summer, stock sizes change suddenly due to newborns and recruitment. This is seen in Figure 9.6. The saw-tooth like time series occurs because of the different length of the time step for summer and winter. The condensed time in summer (one time step) when shellfish recruit and birds get chicks causes the upward jump in stock levels. In the following, the output data are annually sampled at the beginning of winter (the highest values) to smoothen the yearly decreases and sudden summer recoveries. The simulation shows that out of the irregular local dynamics a regular global behavior emerges. Global stock levels vary almost periodically with periods of 20–25 years. Birds lag some 5 years behind compared to shellfish. Figure 9.7 shows the 300-year

3.0E+10

150000

2.0E+10

100000

1.0E+10

50000

Birds

Shellfish

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 169

0.0E+00

0 35

40

45

Year juvenile mussels adult cockles

adult mussels common eider

juvenile cockles oystercatchers

Figure 9.6. 10-year time series of global stock levels of shellfish and birds.

time series of the number of birds and shellfish. The oystercatcher population collapses a few years before the eider population. The system is characterized by periods of collapse and recovery. Locally, at the grid cell level, no semi-periodic behavior can be seen (Figures 9.8 and 9.9). The local behavior is characterized by two periods, one of steady growth, 150000

2.0E+10

100000

1.0E+10

50000

Birds

Shellfish

3.0E+10

0.0E+00 0

50

100

150

200

250

0 300

Year adult mussels

adult cockles

common eiders

oystercatchers

Figure 9.7. 300-year time series of stock levels in numbers of shellfish and birds. Note: Stock levels are sampled at the start of winter.

170

PART II: BIOECONOMIC MODELING 4000

60000

Common eider

3000 40000 2000 20000 1000

0 250

260

270

280

290

0 300

Year

Oystercatchers

Figure 9.8. The last 50-year time series of the local eider population in each cell and the global population. Note: The global population is denoted by the fat gray line. The values for the local population can be read from the primary y-axis, values for the global population can be read from the secondary y-axis. 10000

160000

7500

120000

5000

80000

2500

40000

0 200

225

250

275

0 300

Year Figure 9.9. The last 100-year time series of the local oystercatcher population in each cell and the global population. Note: The global population is denoted by the thick black line. The values for the local population can be read from the primary y-axis, values for the global population can be read from the secondary y-axis.

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 171

Cockle mortality

100%

75%

50%

25%

0% 0

50

100

150

200

250

300

Year additional

oystercatcher

common eider

Figure 9.10. The mortality of adult cockles through predation by oystercatchers, common eiders, or through additional causes.

which coincides for a large part with the global recovery period and one of chaotic behavior, which starts before the collapse of the global system and ends when the global recovery begins. In cells with large bird populations, the chaotic behavior starts earlier. In cells with small bird populations, the chaotic period may be absent. The period of growth is followed by a period of decline. Globally, the annually sampled system shows a semi-periodic behavior with the period and amplitude depending on parameter values, as we will show later (Figure 9.14). As can be seen from Figures 9.8 and 9.9, the oystercatcher population behaves chaotically much earlier than the eider population, for some grid cells almost during the entire cycle. According to findings of (Bascompte and Sol´e, 1995; Jansen and de Roos, 2000), space can stabilize the dynamics, meaning that although the dynamics can be chaotic, the overall deviation of a long term mean is relatively small and shifts toward other global attractors are more difficult in large areas. We see the same mechanism in our model. Local surpluses are compensated by local shortages elsewhere and potential fluctuations are smoothed by emigration of birds from other areas. Local behavior is more irregular than global behavior. The global predation pressure and additional shellfish mortality are shown in Figures 9.10 and 9.11. On average some 20% of both the cockle and mussel population is eaten by oystercatchers. These values are within the range reported in field studies on oystercatchers (Zwarts and Drent, 1981; Goss-Custard et al., 2001; Ens et al., 2004). Common eiders eat another 20% of both shellfish species and some 13% dies of additional causes. In years of collapse up to 50% of the shellfish stock is eaten by common eiders. Oystercatchers can eat up to 35% of the shellfish stock. At times of collapse when the common eider population is at its peak, eiders eat more of the shellfish stock than oystercatchers. In the beginning and most of the recovery period oystercatchers eat more shellfish than the common eiders.

172

PART II: BIOECONOMIC MODELING

Mussel mortality

100%

75%

50%

25%

0% 0

50

100

150

200

250

300

Year additional

oystercatcher

common eider

Figure 9.11. The mortality of adult mussels through predation by oystercatchers, common eiders, or through additional causes.

The mortality of the oystercatcher is larger than that of the common eider in most years, but not as severe in the years of collapse (Figure 9.12), which explains the much faster collapse of the common eider population. Interference causes the difference between the two bird species. The common eider population can grow until there is not enough food to support them and the population almost at once collapses. The 100%

Bird mortality

75%

50%

25%

0% 0

50

100

150

200

250

Year eider mortality

oystercatcher mortality

Figure 9.12. The mortality of birds through starvation.

300

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 173 100%

Bird mortality

75%

50%

25%

0% 0

2

4

6

8

surplus fraction oystercatcher mortality

common eider mortality

Figure 9.13. A scatterplot of bird mortality versus the surplus fraction of food in the ecosystem.

oystercatchers on the other hand start to hinder each other at high densities causing many of them to starve before there is an actual food shortage. This dampens the collapse. The amount of food that needs to be present in the ecosystem per individual so that it can survive, which Ens et al. (2004) have called the ecological food requirement, is higher than the amount the birds actually eat in the course of winter, which is called the physiological food requirement. Reviewing studies in British and French estuaries, Goss-Custard et al. (2004) conclude that oystercatchers need 2.5–7.7 times as much shellfish mass at the start of winter than they actually eat. Rappoldt et al. (2003a,b) found values of the ratio between ecological and physiological food requirement of 3 and 2.5 for Wadden Sea and Oosterschelde, respectively. Our finding fits that range. Figure 9.13 shows the surplus fraction, i.e. the amount of shellfish in mass the ecosystem provides divided by what birds physiologically need, in relation to bird mortality. Bird mortality is almost always between 0% and 10% except when the surplus fraction is lower than 2. For a higher surplus fraction than about 5 there is a decreasing mortality trend. When the data are examined in a time series it shows that the oystercatcher population start their collapse at about a surplus fraction of 2.5 and the common eider population collapses at a surplus fraction of about 2. The choice of parameter values is important. Slightly different values can have a great effect on the behavior of the model. Figure 9.14 shows five common eider time series in which 12 parameters are increased with 10% and another 11 parameters are decreased with 10%. These five simulation runs are the first five shown in Table 9.8. They illustrate that a slightly different set of parameter values (10% change), e.g. due to measurement errors, can make a significant difference. Slightly different parameter values can change the period and amplitude of the predator–prey cycle. Other

174

PART II: BIOECONOMIC MODELING

Common Eider

60000

40000

20000

0 0

50

100

150

200

Year Figure 9.14. Five 200-year time series of the global common eider population for parameter values that differ plus or minus 10% from the standard values (Section 9.3).

parameter values can change the composition of the system. In the first run for instance, the common eider becomes extinct. It is also possible that the system dynamics change from being cyclic to approaching equilibrium. In the next section, we will examine for which parameters the system is most sensitive.

9.5. SENSITIVITY ANALYSIS The model has many parameters. Some parameters are more important than others. Some have a large influence on the system’s behavior and some hardly matter. It is not intuitively clear how model behavior changes when one of the parameters is slightly altered. Therefore, a sensitivity analysis is applied to better understand the role different parameters play in the model. We consider a sensitivity analysis good modeling practise and important to understand the behavior of a complicated model. Apart from the physical parameters described in Section 9.2, the model has two computational parameters, namely the time step and grid size. These parameters are needed for the numerical solution of the model. The resolution in time and space of the model is of great importance to the numerical error in the simulation. The time step and grid size must be small and in balance with each other to avoid large errors. Otherwise, numerical noise drives the dynamics of the model. We used trial and error to find the time step and grid size that were accurate enough. The difference of the bird population levels between simulations on a 10 × 10 and a 20 × 20 grid were negligible. However, the simulation on the 5 × 5 grid results in bird populations that fluctuate faster, with higher and lower extremes and a lower average than on the 10 × 10 grid. We obtained similar results for the time step. A time step of 10 days instead of 2 or 1 day(s) similarly

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 175

Wemi,ei

Wmin,ei

aoy

hoy,m2

hoy,c2

Wmin,oy

roy

mm2

mc2

cm1

cc1

cm2

cc2

km2

kc2

Wnec,oy

Commoneider

hei,c2

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

0

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

22953 86998

3

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

19906 93520

4

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

22842 75205

Wemi,oy

hei,m2

+

2

qoy

aei

1

rei

Wnec,ei

Oyster-catcher

Table 9.8. Plackett–Burman design matrix of the sensitivity analysis RUN

126185

5

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

27440 94214

6

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

28851 76544

7

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

22412 109354

8

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

44538 73430

9

+

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

31656 156159

10

+

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

28824 139040

11

-

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

30452 110972

12

-

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

37761 154531

13

+

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

26620 145105

14

+

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

31062 117373

15

-

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

38212 77915

16

-

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

30431 107881

17

+

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

28936 102402

18

-

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

31303 102376

19

+

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

26159 125983

20

-

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

32036 83706

21

-

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

40952 92215

22

-

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

35519 138936

23

-

+

+

+

+

+

-

+

-

+

+

-

-

+

+

-

-

+

-

+

-

-

-

34905 129423

24

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

22537 89658

25

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

35318 109125

26

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

31096 147863

27

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

38012 136685

28

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

34221 160481

29

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

38273 122130

30

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

27424 152705

31

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

34211 110689

32

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

26174 125694

33

-

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

27534 77752 28563 79839

34

-

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

35

+

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

27839 61435

36

+

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

0

37

-

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

33353 81198 28820 106529

67115

38

-

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

39

+

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

24966 135506

40

+

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

31981 102454

41

-

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

35780 104646

42

+

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

0

43

-

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

32786 99639

44

+

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

27161 139160

45

+

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

24449 111660

46

+

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

20768 87314

47

+

-

-

-

-

-

+

-

+

-

-

+

+

-

-

+

+

-

+

-

+

+

+

24640 92393

48

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

35276 138485

109136

Note: Rows represent a simulation run, columns represent a parameter. ‘+’ denotes the maximum value is used, ‘−’ (also marked in gray) that the minimum value is used. The last two columns give the results (over last 150 years).

176

PART II: BIOECONOMIC MODELING

causes instabilities. A course resolution can cause instabilities. This is a well-known result from numerical mathematics but nonetheless important to note here. Numerical instability can easily be attributed to chaotic non-linear behavior of the system under study. Following Beres and Hawkins (2001), a foldover Plackett–Burman design is used to find the sensitivities of the physical parameters. Cryer and Havens (1999) give another example of applying the Plackett–Burman design for a sensitivity analysis of an environmental model. A Plackett–Burman design is a fractional factorial design, which allows detecting very efficiently the main effects of each parameter (Box et al., 1978). The main or first order effect is an estimate of ∂ X/∂ p where X is the variable of interest and p is the parameter for which the sensitivity of X is tested. Simulations are executed with maximum and minimum values of each parameter that is included in the sensitivity analysis. Compared with a traditional one-at-a-time approach, in which only one parameter is varied per run, the effect of changing a parameter is estimated as the average over variations of the other parameters. Also, interaction between parameters (second and higher order effects) can be shown although they may be confounded with other two- or multiple-parameter interactions (Box et al., 1978). The design matrix (Table 9.8) shows 48 runs that have been simulated for a period of 200 years. Each row of the matrix represents a simulation run. The signs indicate whether the high or the low extreme of a variable is used. The nature scenario is applied in which there is no fishing. The model is simulated on the before mentioned 10 × 10 grid. Section 9.3 gives more details on the simulation grid and the choice of parameter values. The maximum and minimum values deviate plus and minus 10% from those given in Section 9.3. For each run the average bird population is computed over the last 150 years. After 50 years the system is expected to be stable (Figure 9.7). The first order effect of a parameter on a variable is computed by subtracting the average of the variable when the parameter has its minimum value, from the average of the variable when it has its maximum value. This results in the average change when the parameter changes from its minimum to its maximum value. For example the main effect of adult mussel mass (m m2 ) on the common eider population is calculated as follows: −0 − 22,953 + 19,906 . . . /24 = 3459. The pluses and minuses are found in the adult mussel mass (m m2 ) column of Table 9.8. When we divide this number by the average common eider population over all 48 runs, we get 0.12, the number in Table 9.9. This number reflects the average relative change in the common eider population when the adult mussel mass changes from −10% to +10% of the standard value. Tables 9.9 and 9.10 show the first order effect of the parameters for which the common eider and the oystercatcher population are the most sensitive respectively. The numbers give the average rate of change for the stock levels and variances when the parameter is changed from the minimum to the maximum. As expected the first order effects show that the bird populations are predominately sensitive to the presence of food in relation to how much they need. This is reflected through parameters such as the mass and mortality of shellfish and Wnec . However, especially the common eider is not overly sensitive to recruitment, cm1 and cc1 . Adult shellfish are the main source of food. Presumably, it is more important how many recruits survive into adulthood reflected by the parameters cm2 and cc2 , than are actually recruited. The oystercatcher is also sensitive to Wemi , which reflects how easily birds

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 177 Table 9.9. The results of the Plackett–Burman sensitivity analysis for the eider population Common eider Wnec,ei m c2 Wmin,ei cm2 m m2 qoy cc2 h ei,c2 aoy Wemi,ei Wmin, oy h oy,c2 rei km2 Wnec,oy cm1 roy aei h ei,m2 cc1 Wemi,oy h oy,m2 kc2

Necessary daily intake rate below which a common eider will starve Mass of an adult cockle at the start of winter Minimum intake rate below which a common eider will feed on alternative prey Summer survival of juvenile mussels Mass of an adult mussel at the start of winter Interference area of oystercatchers Summer survival of juvenile cockles Handling time of an adult cockle by a common eider Search coefficient of oystercatchers Emigration intake rate below which a common eider will emigrate Minimum intake rate below which an oystercatcher will feed on alternative prey Handling time of an adult cockle by an oystercatcher Recruitment coefficient of common eiders Additional mortality of adult mussels Necessary daily intake rate below which an oystercatcher will starve Recruitment of (juvenile) mussels Recruitment coefficient of oystercatchers Search coefficient of common eiders Handling time of an adult mussel by a common eider Recruitment of (juvenile) cockles Emigration intake rate below which an oystercatcher will emigrate Handling time of an adult mussel by an oystercatcher Additional mortality of adult cockles

−0.36 0.22 0.17 0.16 0.12 0.11 0.11 −0.10 −0.07 −0.06 0.05 0.05 −0.05 0.04 0.04 0.03 −0.02 0.02 −0.02 0.01 0.01 0.01 0.00

Note: Each column shows the main effect of each parameter on the common eider population. The parameters are ranked according to their absolute importance. The values show the relative change when the parameter is changed by 10%. Negative sensitivities are marked in gray.

disperse over the area to find food. The oystercatcher population would decrease if birds would search for a better feeding area sooner, i.e. a greater Wemi . There must be a trade-off, however, between leaving a spot too soon (without being able to find a better feeding place) and staying too long (when elsewhere higher intake rates could be realized). The relatively high sensitivity of Wemi is rather unfortunate because it is a parameter that has not been measured in the field yet. Consequently, the errors in our estimates are probably large and may have serious influence on the simulation results. The high sensitivity of Wemi,oy in comparison with the sensitivities of search and handling time (a and h) shows that the macrofeeding behavior of a bird, i.e. how to find the best musselbank, is more important than the micro behavior, i.e how quickly a mussel is processed. We have investigated the sensitivity for migration further. A simulation is conducted in which a bird emigrates to neighboring grid cells within a 2 km radius instead of to a random grid cell if its intake rate is less than Wemi . The global differences are very small. The cycles for the oystercatcher are somewhat smoother and the minimum population level is about 10,000 higher.

178

PART II: BIOECONOMIC MODELING Table 9.10 The results of the Plackett=Burmansensitivity analysis for the oystercatcher population

Oystercatcher m m2 Wnec,oy m c2 cm2 cm1 Wemi,oy cc1 h oy,m2 cc2 Wnec,ei Wemi,ei h oy,c2 roy kc2 qoy rei aoy h ei,m2 Wmin,oy Wmin,ei hei,c2 aei k m2

Mass of an adult mussel at the start of winter Necessary daily intake rate below which an oystercatcher will starve Mass of an adult cockle at the start of winter Summer survival of juvenile mussels Recruitment of (juvenile) mussels Emigration intake rate below which an oystercatcher will emigrate Recruitment of (juvenile) cockles Handling time of an adult mussel by an oystercatcher Summer survival of juvenile cockles Necessary daily intake rate below which a common eider will starve Emigration intake rate below which a common eider will emigrate Handling time of an adult cockle by an oystercatcher Recruitment coeffcient of oystercatchers Additional mortality of adult cockles Interference area of oystercatchers Recruitment coeffcient of common eiders Search coeffcient of oystercatchers Handling time of an adult mussel by an oystercatcher Minimum intake rate below which an oystercatcher will feed on alternative prey Minimum intake rate below which a common eider will feed on alternative prey Handling time of an adult mussel by a common eider Search coeffcient of common eiders Additional mortality of adult mussels

0.26 0.14 0.12 0.12 0.09 0.06 0.05 0.04 0.03

0.02 0.02

0.01 0.01

Note: Each column shows the main effect of each parameter on the oystercatcher population. The parameters are ranked according to their absolute importance. The values show the relative change when the parameter is changed by 10%. Negative sensitivities are marked in gray.

Finally, Table 9.9 on common eiders contains eight oystercatcher parameters, whereas Table 9.10 on oystercatchers contains only seven common eider parameters, whose sensitivities are also generally smaller than the sensitivities of eiders to oystercatcher parameters. Thus, eiders seem more sensitive to oystercatchers than oystercatchers are to eiders. Probably, this is related to the fact that oystercatcher populations decrease as a result of food shortage before eider populations do.

9.6. STOCHASTIC RECRUITMENT The system is not very sensitive to recruitment. Nevertheless, recruitment does influence the behavior of the ecosystem. For the sake of the sensitivity analysis, they were taken constant but these parameters can vary greatly over the years. The model will be used to develop and analyze management scenarios. Because of the sensitivity for

4.0E+10

200000

3.0E+10

150000

2.0E+10

100000

1.0E+10

50000

0.0E+00 0

50

100

150

Birds

Shellfish

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 179

0 200

Year juvenile mussels adult cockles

adult mussels eiderducks

juvenile cockles oystercatchers

Figure 9.15. 200-year time series of stock levels of shellfish and birds. Recruitment of shellfish is stochastic. Note: The probability density of X is lognormal distributed with an average of 100 m−2 and a standard deviation of 100 m−2 .

shellfish recruitment and the stochastic nature of recruitment, management scenarios should be analyzed for robustness using stochastic recruitment, long time series and ideally evaluate several values for mean and variance. To illustrate the importance of stochastic recruitment, a simulation with random recruitment of shellfish is undertaken. Figure 9.15 shows the dynamics of the model when stochastic recruitment is used. Recruitment of shellfish often has a skewed probability density function. Many years will show relatively little recruitment and only in very few years recruitment is high. Therefore the value of the stochastic recruitment parameter X is taken from a lognormal distribution. Thus, there will be relatively many years with recruitment below average. Figure 9.15 shows that random recruitment causes the bird populations also to vary randomly but as expected at a much slower pace than the shellfish populations. The number of birds is strongly correlated to the number of birds in the year before. The good years do not add much to the food supply of birds because most of the extra shellfish die before they can be eaten. The bad years however hamper growth of the bird population by either causing extra starvation directly though lack of food or indirectly by causing searching inefficiencies. In a test run with higher variability (standard deviation, 200 m−2 ) the common eider population becomes extinct at some time. A few bad years in a time when the bird population is relatively high can cause a collapse from which recovery is not possible.

180

PART II: BIOECONOMIC MODELING

The model obviously ignores that common eiders from other geographical areas may recolonize the area and also that birds may have fled to these areas before they become extinct. Nevertheless, it is shown that severe randomness can have a devastating effect on the bird population.

9.7. CONCLUSIONS The chapter describes and analyzes a model of the interaction between two types of birds and two types of shellfish. The model simulates the movement of birds in space and the dynamics of the populations. It is not our aim to contribute with this model to our understanding of the population dynamics of this type of system. Our aim is to contribute to an understanding how different shellfish fishery policies will affect the ecosystem and the economic profits of the fishermen. However, we can only use the model for scenario calculations once it has gained sufficient credibility. We also need a sufficient understanding of the dynamics of the model. Before we discuss the behavior of the model, we will deal with the question to what extent the model succeeds in capturing the important features of the empirical study system. For the most part, we have characterized the model system via basic processes whose parameters can and have been measured. Consequently, the model can be considered as a bookkeeping exercise of shellfish biomass and bird numbers. When there are no mistakes in the bookkeeping, the results should follow. However, we also deliberately ignored several well-known features of the ecology of both shellfish and birds. We will discuss the most prominent of these in turn. (1) We did not include a digestive bottleneck for the birds (Kersten and Visser, 1996). When food is abundant, birds can collect food at a faster rate than they can digest, so they are forced to interrupt their food searching with digestive pauses. This process was not included as it operates at the scale of hours, whereas 1 day was a convenient time step for our simulations. The digestive bottleneck is a problem only when the birds have a limited feeding time. Thus, the possibility for the birds to maintain a positive energy balance is overestimated in grid cells with short feeding times. We suspect that this may affect the details, but not the overall working of the model. (2) We did not include emigration to different geographical areas in our model. Other estuaries or coastal zones could offer the birds a refuge when the mussel and cockle populations have crashed. In that case we expect an effect on the dynamics of the system, because birds would move to the refuge, instead of starve to death. This situation may apply to common eiders that have moved from the Wadden Sea to the North Sea coastal zone to feed on S. subtruncata, apparently in response to a shortage of their primary food (Camphuysen et al., 2002). (3) We did not include differences between individual birds in our model. For oystercatchers it is well known that individuals differ in prey selection (Sutherland et al., 1996), the rate at which they feed in the absence of interference (Goss-Custard and Durell, 1988) and dominance (Ens and Cayford, 1996). Explicit inclusion of individual differences requires that the model keeps track of each individual, which it currently does not do. Keeping track of individuals would greatly increase both the complexity of the model and the duration of the simulations. Implicitly, individual differences are incorporated by modeling emigration and starvation as probabilities that depend on the

9. SPATIAL–TEMPORAL MODEL OF THE INTERACTION OF SHELLFISH AND BIRDS 181 average intake rate of the individuals in the grid cell. Our feeling, which remains to be tested, is that this suffices for our purposes. (4) We did not include seasonal changes in climatic conditions in the model. The birds have higher energy needs during periods of cold weather and oystercatchers may find it difficult to feed when the mud flats are covered with ice. Our feeling, which remains to be tested, is that inclusion of seasonal changes in climate and climatic variability would increase the noise in the outcome, but would not dramatically alter the dynamics of the system. We only included stochasticity in recruitment of shellfish as an option in the model, yet it is known that recruitment of shellfish is highly erratic between years (Beukema, 1982). Our rationale for this choice was that inclusion of stochasticity in recruitment from the start would have made it difficult, if not impossible, to comprehend the dynamics of the system and to carry out a sensitivity analysis. The constant recruitment case provided a baseline for studying the model. Nevertheless, an illustration of the effect of adding noise to the recruitment is shown. The behavior of the system is characterized by predator–prey cycles with a period of 20 to 25 years. In order to calibrate and/or validate the model accurately, one would need a data set of at least 40 years. Stochastic recruitment obscures the systems behavior and prolongs the cycle (Figure 9.15). Consequently, one would probably need more than 40 years of data. To date, the only data set on bivalve stocks that covers such a long period is the study of the benthic animals on the Balgzand area, which was initiated in 1970, e.g. (Beukema 1982, 1993). However, the Balgzand area is only a small part of the entire Wadden Sea. Besides, our grid is a hypothetical area that does not exactly match a real world situation. Therefore, a validation can only be tentative. Nevertheless, predation pressure is of the correct order of magnitude and so is the ecological food requirement. A striking feature of the model is that it predicts occasional mass mortalities of the birds as a result of food shortage. In the real world, there is good evidence for mass mortality as a result of food shortage for both common eiders (Camphuysen et al., 2002; Ens et al., 2002a,b; Ens et al., 2004) and oystercatchers (Camphuysen et al., 1996; Atkinson et al., 2003). A sensitivity analysis shows that average population levels of birds are most sensitive to parameters that directly govern the amount of shellfish in the ecosystem in relation to what they need to survive. These are parameters such as the average mass and survival into adulthood of shellfish and Wnec , the daily intake a bird needs to survive. The efficiency with which birds choose to emigrate to better feeding grounds represented by Wemi , happens to be important. Unfortunately this is a parameter that is not measured accurately in the field. The sensitivity for Wemi shows that macrofeeding behavior is more important than micro feeding behavior. A bird population can gain more by efficiently detecting the best mussel bed than by more efficiently depleting a mussel bed. Finally, when this model is used to analyze management options for the fisheries we obviously recommend to model recruitment of shellfish as a stochastic process. Recruitment of shellfish varies greatly over the years and since bird populations are very sensitive to the presence of shellfish in the ecosystem a great deal of variability in fishery profits and stock levels of birds is not accounted for when one uses constant average recruitment.

PART III

MONETARY VALUATION AND STAKEHOLDER ANALYSIS

183

CHAPTER 10

POLICY FAILURE AND STAKEHOLDER DISSATISFACTION IN THE DUTCH WADDEN SEA SHELLFISHERY

10.1. INTRODUCTION Shellfishing has been an important economic activity in Dutch coastal areas for centuries. The shellfishery has been active predominantly in the Dutch province of Zeeland and in the Wadden Sea (Dijkema, 1997). The Wadden Sea is a shallow inshore body of water, which is characterized by large expanses of intertidal mud flats. It extends from the Netherlands to Denmark and is bordered by a row of islands. The total area of the Wadden Sea that is part of the Netherlands is 2500 km2 (Dijkema, 1997). Fishing activities in the Wadden Sea have largely focused on shellfish and shrimps. The shellfish fishery has predominantly been targeted at oyster (Ostrea edulis), cockle (Cerastoderma edule), and mussel (Mytilus edulis) populations. While the oyster population disappeared from the Wadden Sea between 1940 and 1950 due to overfishing (Dijkema, 1997), shellfishing for mussels, and cockles continues at present. Over the second half of the last century, fishing pressure has dramatically increased in the cockle sector. The increase in landings in the cockle fishing sector has resulted from the advent of mechanized fishing techniques, whereby suction dredges were implemented to harvest cockles. Until recently, the shellfishery in the Wadden Sea was in a precarious state. Fishermen faced great uncertainty about the impact potential future regulations may have on their activities. Furthermore, the future of the mechanized cockle fishery as a whole was very uncertain, since it was threatened with closure. Until recently, the fisheries management policy in the Wadden Sea was aimed at stock conservation, but now nature protection is also being taken into account (Ministerie LNV, 1993). In fact, it has been decided that the primary function of the Wadden Sea is to be a nature area; human activities are allowed as long as they do not cause significant harm to the natural values of the Wadden Sea. The crash of the mussel and cockle populations in 1990–1991 gave rise to a conflict between environmental organizations and the fishing sector. This conflict concerns the impact of fishing on the ability of the Wadden Sea and other marine environments (the Oosterschelde region in the province of Zeeland) in the Netherlands to sustain important ecological values relating to bird species, sea grass, and mussel beds. Since then, the management policy, as formulated in the Sea and Coastal Fisheries Policy (SCFP) (Ministerie LNV, 1993), demands that fishing activities take into consideration the aforementioned ecological values. The SCFP requires that the food requirements of bird species wintering in the Wadden Sea be taken into account. The

185

186

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

SCFP also demands that fishing activities not interfere with important ecological values, such as intertidal mussel banks and sea grass beds. With the introduction of the SCFP, the government has laid down a framework allowing themselves more control over the impact of shellfishing on ecological values of the Waddenzee. In terms of the terminology introduced by Ens (2002), shellfish management can be seen to be shifting from a focus on shellfish stock sustainability to a focus on ecological sustainability. The SCFP has transformed the shellfishery from a virtually unrestricted fishery to a fishery that has become more regulated in line with a comanagement arrangement. The initial effect of the regulations was that 25% of the intertidal flats of the Wadden Sea were permanently closed for fishing and that in years of low shellfish stocks the shellfish fishery would be closed altogether to reserve food for the bird populations. The official aim of this policy was that the remaining shellfish stocks would cover 60% of the birds’ food needs, but this goal was not met because the calculations underlying the policy were deficient (Ens, 2000). Furthermore, the mussel fishery is restricted to the sublittoral areas of the Wadden Sea, and constrained to a total allowable catch (TAC) of 65 million kg of mussel seed. This mussel seed is sown on culture plots in the Wadden Sea and the southern Dutch province of Zeeland and should be sufficient to create a target production of 100 million kg of mussels for consumption (Smaal et al., 2001). Originally, the SCFP did not specify a maximum catch for the cockle fishery. Later on, following a decision in court, the cockle fishery was effectively constrained by a TAC of 10 million kg of cockle flesh from the Wadden Sea. This sector has also been subjected to fishing capacity constraints. These fisheries management measures have not prevented fluctuations in the status of the mussel and cockle populations from occurring, which is not surprising, since large fluctuations are natural for the populations of these species, especially for cockles (Kamermans and Smaal, 2002). However, there are several indications that in its current form, the SCFP cannot guarantee that important ecological values in the Wadden Sea will be sustained. A mid-term evaluation of the SCFP in 1998 resulted in further closure of fishing grounds due to the fact that the intertidal mussel banks had not recovered. Moreover, fisheries management so far has not prevented a decline in bird populations: both the Common Eider (Somateria mollissima) and the Oystercatcher (Haematopus ostralegus), which have continued to decrease in numbers, winter in the Wadden Sea (Smit et al., 2000; Camphuysen et al., 2002). There is strong disagreement among the various stakeholders about the causes of this decline, and about how much of the decline is related to fishing pressure. To improve the effectiveness of a fisheries management plan that attempts to achieve the sustainability of important ecological values in the Waddenzee, various future scenarios can be considered. This requires that the management plan has clear objectives that are accepted by all the stakeholders involved. The present management plan has not achieved its objectives and has succeeded in frustrating all stakeholders, as evidenced by the fact that both fishermen and conservationists regularly go to court to fight government decisions. In other words, the co-management arrangement has not worked well. This frustration is not only based upon the SCFP’s failure to meet many of its objectives, but also due to disagreement about what the objectives actually are. The purpose of this chapter is to understand what lies at the heart of the shellfishery management conflict. Stakeholders’ objectives are taken as the starting point of the

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

187

analysis. It is determined why the current management plan has failed as well as what alternative management scenarios might achieve. The remainder of this chapter is organized as follows. Section 10.2 examines the objectives of the various stakeholders of the current Wadden Sea fisheries management plan, and whether these have been met. Section 10.3 considers factors that are considered essential in promoting marine ecosystem sustainability, as well as possible scenarios that have been suggested for the future of the SCFP. Conclusions are drawn in Section 10.4.

10.2. HAVE THE OBJECTIVES OF ALL PARTIES INVOLVED IN THE SCFP BEEN MET? 10.2.1. The stakeholders and their objectives The most important, and influential, stakeholders in the case of the Wadden Sea are the government, conservationists, the fishing sector, and researchers. The latter group includes both consultants and scientists. Table 10.1 lists the most important objectives the stakeholders are expected to have and whether or not they have been achieved under the current shellfish policy. Each objective listed in this table will be discussed in further detail in this chapter and the reason for whether they are to be considered met or not will be outlined. These objectives have been selected for analysis because they represent the most often raised issues in stakeholder-sponsored publications, most of which will be discussed later on in this chapter. Objectives 1–4 of the government are consistent with the goals of the SCFP: the government is the party that is ultimately responsible for the SCFP and thus the objectives of the SCFP must also be government objectives. The outcomes of the SCFP must satisfy objectives 5 and 6, although these objectives also include issues beyond the scope of the Wadden Sea. With regard to objective 7, when the government handles monitoring and enforcement, costs associated with these tasks tend to be higher than when some of these tasks are handled by stakeholders themselves. In the SCFP some of the tasks involving monitoring and enforcement of fishing activities are delegated to shellfisheries’ producers organizations (POs). There are several conservation organizations that focus attention on the ecological and environmental values of the Wadden Sea and the threat posed to the region by economic interests. These organizations include the national union for the conservation of the Wadden Sea (over 50,000 members), The Dutch Bird Conservation Society (108,000 members), and Nature Monuments (approximately 1 million members). The first conservationists’ objective is assumed to hold for all conservation organizations. Objective 9 does not necessarily hold for all conservationists. There are organizations that feel that there is no room for any economic activity whatsoever in the Wadden Sea. Other organizations are less extreme and recognize that economic activities have taken place in the Wadden Sea for centuries. The most lenient point of view found amongst conservationists is one in which only the mussel seed fishery would be tolerated and regulated. This chapter takes the most lenient conservationists’ view into consideration when it comes to regulating economic activities in the Wadden Sea. The scope of the environmental and ecological concerns covered by objectives 8 and 9 goes beyond the

188

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 10.1. Objectives of the stakeholders

Stakeholder

Objectives

Government

1. Restore wintering bird population numbers to average population levels over the period 1980–1990 2. Restore sea grass fields 3. Restore area of stable intertidal mussel banks to former levels of 20–40 km2 4. Maintain a healthy and competitive national mussel and cockle fishery. 5. Comply with national and international regulations concerning the protection of ecological values in the Wadden Sea 6. Be consistent with national and international management plans affecting the Wadden Sea 7. Reduce enforcement and monitoring costs 8. Maintain the Wadden Sea as an area of ecological and environmental importance 9. Restrict the mussel seed fishery, close the cockle fishery 10. Maintain a profitable fishery 11. Prevent a reduction in fishing capacity and maintain robustness against fluctuations in supply 12. Guarantee the future of the fishing sector 13. Participate in the SCFP discussion in order to have scientific concerns addressed 14. Establish marine reserves to research ecological processes in a “pristine” environment

Conservationists

Fishing Sector

Researchers

Objective realized No

No Yes Yes Currently being researched Currently being researched Yes No No Yes Yes No Yes Partly

concerns that the SCFP addresses. In addition to shellfishing, conservation organizations consider all human activities that may have a detrimental impact on the Wadden Sea. As such, these organizations have exerted particularly strong pressure to prevent the exploitation of natural gas reserves and, to a lesser extent, the placement of wind turbines in the region. These issues of concern are, however, not part of the SCFP.

10.2.2. Objectives of the government 10.2.2.1. Restoration of bird population levels Objective 1, bringing the winter bird population back to the average population size it had during the period 1980–1990, has not been met. Compared with the their size during the benchmark period, 1980–1990, both Oystercatcher and Common Eider populations have declined in the Wadden Sea. The cause of the decline in the Oystercatcher population is most likely a shortage of shellfish available for consumption (Smit et al., 1998; Ens, 2000). The total number of Common Eiders has declined in the Wadden Sea. Uncertainty about their diet and foraging behavior makes it difficult to state

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

189

90 Data: Dijkema (1997)

80

Data: Produktschap Vis (1999) Million kg fresh weight

70 60 50 40 30 20 10 0 1951

1956

1961

1966

1971

1976

1981

1986

1991

1996

Year

Figure 10.1. Cockle landings in Dutch coastal waters in million kg fresh weight (note that data from the periods 1968–1970 and 1974–1975 is missing). Source: Ens (2002).

unequivocally that the decrease in the number of Common Eiders in the Wadden Sea can be attributed to a shortage of suitable food (Higler et al., 1998) and not to other factors. Studies by Berk et al. (2000), Ens (2000), and Camphuysen et al. (2002) seem to indicate that the Common Eider population’s shift to the North Sea since 1990, as well as the decline in their numbers in the Wadden Sea, is due to a shortage of food in the Wadden Sea. These studies have ruled out many other possible reasons for the decline in Common Eider population levels, such as parasite infections and diseases. Instead they hypothesize that the decline in the number of Common Eiders can be attributed to a combination of the heavy fishing pressure on Spisula in the North Sea and cockles in the Wadden Sea, a lack of available mussels, and seasons in which shellfish were not abundant due to unfavorable weather conditions. Until the late 1980s catches of cockles increased, but since then they have been fluctuating according to natural fluctuations in the cockle population (Figure 10.1). This suggests that increased fishing efficiency caused the large catches (Ens, 2002), but since then the ability to expand has been saturated. Mussels in the sublittoral regions are an important source of food for Common Eiders as well as other species that are present in smaller numbers, such as diving ducks. Since the creation of the SCFP, the mussel population on the intertidal flats has generally been so small (the fall of 1994 is an exception) that fishing activities on the intertidal flats were restricted as part of the 60% food reservation rule for birds (Higler et al., 1998). Currently, fishing for mussel seed occurs in the subtidal regions. With the exception of 1994, less than 100 million kg of mussels have been landed since the introduction of the SCFP in 1993 (Figure 10.2). Between 1973 and the late 1980s landings of around 100 million kg of fresh weight were the norm. Thus one can conclude that the mussel population is not as large as it used to be.

190

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

180

Million kg fresh weight

160 140 120 100 80 60 40 20 0 1951-1952

1959-1960

1967-1968

1975-1976

1983-1984

1991-1992

1999-2000

Season

Figure 10.2. Mussel landings in Dutch coastal waters in million kg fresh weight (note that mussel culture was introduced in the Dutch Wadden Sea in 1951). Source data: Mosselkantoor.

10.2.2.2. Restoration of sea grass and mussel beds Objective 2, the restoration of sea grass fields has not been met. Fields of sea grass in the Dutch Wadden Sea virtually disappeared in the 1930s as a result of a disease, although other factors may have contributed. Since the end of the 1980s areas with sea grass beds have been closed to the fishery. Additional areas where sea grass has been discovered since the start of the management plan are voluntarily being avoided by the fishing sector (Higler et al., 1998). The recovery of sea grass fields has been very slow and it has been hypothesized that the failure to recover can be attributed to a lack of sea grass fields that could serve as a donor population for new fields (De Jonge et al., 2000). In any case, there is no clear evidence that current shellfishing practices are responsible for the failure of sea grass fields to recover in the Wadden Sea, but at the same time there is no evidence to prove that current shellfishing practices are not responsible for this failure. Objective 3 concerns the recovery of mussel banks in the intertidal area. Overfishing, natural mortality, and weather conditions had virtually wiped out the mussel beds before the implementation of the SCFP. In 1991 there were only 0.10 km2 of intertidal mussel beds in the Wadden Sea. Partly due to good spatfalls in the years of 1999 and 2001, the total area covered by mussel banks in the Dutch Wadden Sea is around 30 km2 at present, according to a recent expert judgement made by van Stralen (2002). The goal of the SCFP was to create an area of 20–40 km2 of stable intertidal mussel beds. Since fishing practices are not allowed where mussel beds are located, it would appear that the SCFP has had some success in the restoration of mussel banks in the intertidal area. Nevertheless, it should be noted that mussel beds are vulnerable to weather conditions and mussel bank densities are thus subject to natural fluctuations. The mechanized cockle fishery could be a factor that hampers mussel bank recovery since its fishermen have been active on the intertidal flats since the start of the SCFP.

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

191

Table 10.2. Profit of the mussel fishing sector, average price of mussels, and supply of mussels over the period 1991–2001

Profit of the mussel fishing sector (in million euros) Average price (euros/100 kg)b Supply of Mussels (in mil. kg)b a data b data

91/92

92/93

93/94

18.7a

11.5a

14.2a

94/95

95/96

96/97

97/98

98/99

99/00

24.5b

32.7b

21.8b

22.7b

16.4b

24.1b

00/01

45

68

52

61

47

55

125

104

84

92

88

97

96

58

40.9b

from Agricultural Economic Institute (LEI, 1998). from Agricultural Economic Institute (LEI, 2001).

Nevertheless, as is laid out in fishing plans, the fishermen are to avoid areas where there are signs of mussel bank recovery or where mussel seed spatfall is generally high.

10.2.2.3. Sustaining the mussel fishing sector Objective 4 concerns the government’s aim of maintaining a healthy and profitable shellfishing sector and shellfish processing industry over the long term. Under the SCFP the mussel fishing sector has been profitable (see Table 10.2), although fewer mussels were supplied to the processing industry over the period 1987–1997, with an average of 79 million kg/year, than over the period 1977–1986, with an average of 90 million kg/year. The decrease in mussel supply has resulted in an increase in the price of mussels over recent years. Figure 10.3 depicts the dynamics of the ratio between prices paid and mussels supplied, as well as the ratio between profits and mussels supplied, since the introduction of the SCFP. The figure shows that in the year 2000 prices increased beyond what could reasonably have been expected based on data from the preceding years. The price paid for mussels may have been the result of an increase in demand (LEI, 1998). Figure 10.3 shows no real evidence of this for the period 1993–1998, but it could be the case for the years after 1998. The mussel fishing sector has not yet had negative effects in terms of a loss in profits of the fishery, despite reduced landings. Figure 10.3 also shows that profits were significantly higher in 2000 than could reasonably have been expected. If we look at the relationship between mussel supply and prices paid since 1973, Figure 10.4 confirms that, corrected for inflation, there is a clear tendency for prices to go up when supply goes down. The linear trend line in the figure, developed by a least squares procedure, can be interpreted as being the expected relationship between changes in supply and changes in price. A comparison between the trend line and the actual data lead to the conclusion that since 1989 (with the exception of 1993) prices paid for mussels have been higher than would generally have been expected. An explanation for this could be that the conflict between the shellfishery and the conservationists began around 1990. The conflict may have led to a situation in which mussels were finally recognized as a scarce resource. The availability of mussel seed in the Wadden Sea is considered to be in a more precarious state than previously. One can therefore expect the prices paid for mussels

192

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS 1.8

6

price/supply 1.6

profit/supply

5

1.2

4

1 3 0.8 0.6

2

Profit/supply ratio

Price/supply ratio

1.4

0.4 1 0.2 0

0 1993

1994

1995

1996

1997

1998

1999

2000

Year

Figure 10.3. The ratio between prices paid (in euros per 100 kg) and mussels supplied (in million kg), as well as the ratio between profits (in million euros) and mussels supplied (in million kg), over the period 1993–2000. Note: Own calculations based on data from Mosselkantoor and Agricultural Economic Institute (LEI, 1998; LEI, 2001). Prices have been corrected for inflation.

to remain high. In 2000, approximately 85% of fresh mussels caught in the Netherlands were exported, in particular to Belgium and France (LEI, 2002a). Approximately 55% of the fresh mussels went to Belgium, where Dutch mussels constitute 95% of the market (LEI, 2002a). Figure 10.5 shows the development of the Belgian and French mussel import markets and Dutch and Spanish export over the last decades. The figure shows that until 1994 fluctuations in Belgian imports of mussels have largely followed 140

Price (euros) per 100 kgs

120 1991

2000

100 1992 80

1995 1988

1993 1986

60

1990

1997 1999

1980

1996 1979

1994

1998

1987

40

1989

1976 1978

1975

1984

1973

20

1985 1974

1983 1977 1982 1981

0 41.01

66.72

86.71

91.87

103.55

125.56

Mussels supply (million kgs)

Figure 10.4. The relationship between mussels supplied and prices paid since 1973. Prices have been corrected for inflation. Source data: Mosselkantoor.

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

193

Dutch export

Export/import of fresh and chilled mussels (in tons)

60000

French import

40000

Belgium import Spanish export

20000

0

1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1999

Year

Figure 10.5. Export and import data for fresh and chilled mussels (in tons) for the period 1976–1999. Source data: FAO Fishstat database.

fluctuations in the Dutch export market, even if not all the Dutch mussels are exported to Belgium. Belgium imports have started to increase since 1995, despite the fact that Dutch export levels in 1998 were the lowest in decades. This suggests that Belgium must have imported a lot of mussels in 1998 from countries other than the Netherlands. Dutch mussel producers may feel increased competition and pressure to supply enough mussels as soon as the season starts. They may not be able to supply enough mussels, however, since the SCFP is expected to limit the mussel seed available for fishing, and thus the supply of mussels on the market. Other factors that may negatively affect the availability of mussel seed for future fishing activities include changes in the natural habitat of the mussel population. These are caused by, among other reasons, global warming, a decrease in eutrophication leading to a decrease in the availability of algae as a food resource for mussels, and fishery-induced deterioration of the sublittoral habitat (Ens, 2002).

10.2.2.4. Sustaining the cockle fishing sector The consequences of the fishery management plan for the mechanized cockle fishing sector are significant. Had the cockle fishing sector been able to continue its activities

194

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 10.3. Data on supply of cockles and prices paid for cockles

Supply of cockle flesh (tons) Price paid per kilo of cockle flesh (euros)

97/98

98/99

99/00

00/01

1697 3.80

10,220 3.00

7800 2.68

2590 2.18

Source: Data fromAgricultural Economic Institute (LEI, 2001).

without any restrictions, their catch would be an estimated 75% higher, according to the LEI (1998). The restrictions on fishing activities cost the mechanized cockle fishing sector roughly 4.5 million euros a year. The profits derived from the fishery were negative for the years 1991, 1996, and 2001. The fishing sector’s average profit over the period from 1993 to 1997 was an estimated 3.2 million euros. The average net profit over the period from 1997 to 2001 was an estimated 8.6 million euros (LEI, 2002a). Compared with prices paid for mussels, prices paid for cockles do not seem to respond as strongly to changes in supply. This is partly due to the fact that cockles are also supplied from abroad and by the manual cockle fishing sector. Another reason is that cockles are generally processed into products with a long shelf life, so that fluctuations in the yearly supply of cockles can easily be accounted for (LEI, 1998). The profitability of the cockle sector may be further threatened by a decrease in demand. Data on the supply of cockles and the prices paid for cockles, which are given in Table 10.3, show that over the last 3 years both the supply of Dutch cockles and the prices paid for Dutch cockles have decreased. A possible explanation is that 80% of these cockles are exported as conserves to Spain. On the Spanish market, Dutch cockles are expensive and can easily be replaced by other conserved shellfish products. Figure 10.6 shows export and import data for canned mollusc products. It can be seen that in recent years both Dutch exports and Spanish imports of canned mollusc products have decreased and Spanish production of canned mollusc products has increased. This would lead one to conclude that the Dutch cockle sector is losing its market share in Spain to local shellfish producers, which may explain the lower cockle prices seen in recent years. It should be noted that 1985 marked the start of the Spisula fishery, targeting Spisula subtruncata, which occurs just outside the Wadden Sea in the North Sea coastal zone. Eighty percent of the vessels that engage in this fishery are also active in the cockle fishery (Craeymeersch et al., 2001). Fishing techniques used in the cockle and Spisula fishery are fairly simple; minor adoptions need to be made for a cockle vessel to engage in Spisula fishing. Cockle fishing remains the prime business of these vessels. Approximately 10% of the revenues derived by firms in the cockle fishing sector come from the Spisula fishery (LEI, 2002a). The fact that the cockle fishing sector has become active in the Spisula fishery is a direct result of the SCFP (Craeymeersch et al., 2001) and the limitations imposed on the cockle fishery. Without the profits derived from the Spisula fishery, the profitability of the cockle fishing sector would be lower. Increased competition from foreign suppliers and the expansion of fishing activities to the Spisula fishery are signs that cockle fishing as a stand alone business is not as profitable as has

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

195

25000

Spanish Import

Tons

20000

15000

Dutch Export 10000

5000

0

Spanish Production

1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999

Year

Figure 10.6. Dutch export of canned mollusc products and Spanish production and general import of canned mollusc products. Source data: FAO Fishstat database.

been suggested by LEI (2002a). In addition, the mechanical cockle fishery is threatened because it is suspected to have a long-term negative effect on the cockle population and other species endemic to the Wadden Sea (Piersma et al., 2001).

10.2.2.5. Sustainability of the mussel and cockle processing sector The government is interested in maintaining not only a healthy national shellfishing sector, but also a healthy shellfish-processing sector. The mussel processing sector is concerned with either the preparation of fresh mussels for consumption, or the processing of mussels as conserves, or both. Since natural factors and a TAC limit the mussel seed fishery, the supply of mussels to the processing industry is also limited and hence there is little room for expansion of this industry. As a result the revenues of the mussel processing sector have fluctuated around 91 million euros. The sector is not very profitable due to heavy competition and high auction prices for mussels. A reduction in mussel landings can be expected to have an effect on employment within this sector. According to LEI (1998), the SCFP has led to a decrease in employment of roughly 10% in the sector, compared to a situation with an unregulated fishery. Fluctuations in landings from the Wadden Sea can be expected to affect the cockle processing industry, but will not necessarily affect it to the same extent as they will affect the cockle fishing sector. According to LEI (1998), the SCFP costs the processing sector an average of 4.5 million euros a year in lost profits.

196

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

10.2.2.6. Synchronization with other national and international responsibilities Objective 5 captures the government’s responsibility to ensure that the SCFP does not violate international and national legal agreements. On an international level, the management plan has to comply with EU regulations dealing with the protection of bird species, as laid out in the EC Bird Directive, which has designated the Dutch Wadden Sea as a special protection area. The Wadden Sea is also designated as a special area of conservation under the EC Habitat Directive. Both directives are non-voluntary. They require that the selected habitats and flora and fauna of special concern under either of the above-mentioned directives be protected and conserved. National laws that need to be considered in the SCFP are the fisheries law and the law on nature protection. The fisheries law provides the legal basis upon which fishing activities can be restricted by issuing fishing licenses, which govern the conditions for entry into Dutch waters. The nature protection law influences fishing activities in the Wadden Sea because it lays out the conditions according to which access to the Wadden Sea can be granted. These conditions are based upon the environmental considerations that the protectoral status of the area demands. Objective 6 reflects the government’s responsibility for making sure that the SCFP fits well within existing international, national, and regional management plans aiming to protect the ecological values that are generated by the Wadden Sea. Objective 6 differs from objective 5 in that there is no legal requirement for the SCFP to comply with these management plans. On an international level, for example, the current management plan needs to fit in with the Trilateral Wadden Sea Plan, which aims to manage the Dutch, German, and Danish section of the Wadden Sea as a whole based on seven principles (de Jong et al., 1999). These principles include the restoration of environmental values to optimal conditions, the avoidance of activities that are potentially harmful to the area, and the precautionary principle. On a national level the SCFP needs to fit in with the aforementioned Integrated Management Plan for the Wadden Sea, which aims to protect and develop the Wadden Sea as a nature preserve. This forms the basis upon which fishing activities are restricted. In 1984, the Dutch Wadden Sea was also declared a wetland of international importance under the Ramsar convention on Wetlands (1971). The aim of the current Wadden Sea management plan is consistent with the national and international management plans insofar as it incorporates ecological considerations deemed valuable by national and international management plans that might be affected by the shellfishery. Whether or not the SCFP has managed to protect and sustain these ecological values successfully is another concern. The evaluation program of the SCFP, EVA II, has concluded that this is not the case (EVA II, 2003). This has contributed to the recent change in policy as explained at the end of the conclusions to this chapter. 10.2.2.7. Reducing monitoring and enforcement costs Objective 7 concerns the government’s interest in minimizing the costs associated with the monitoring and enforcement of fisheries regulations. It should be noted that prior to the implementation of the SCFP, shellfishing in the Wadden Sea was basically unregulated by the government. The government restricted its efforts primarily to issuing fishing licenses (Steins, 1999). Thus, prior to the introduction of the SCFP, the only regulatory costs related to shellfishing in the Wadden Sea were derived from regulating and monitoring access to the fishery in the Wadden Sea. The government has reduced

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

197

the costs associated with monitoring and enforcement that it would otherwise have incurred through the formation of the PO mussels and PO cockles. These POs have a legal foundation that allow them to monitor compliance with fisheries regulations as determined by the SCFP and to issue fines if necessary. Both the internal controls among fishermen within the POs as well as the introduction of black boxes tracking fishermen’s movements have proven to be successful. By transferring the responsibilities associated with enforcement of the regulations and the monitoring of compliance to the POs, the government has avoided costs it would have incurred in the absence of the POs. Furthermore, the POs are responsible for allocating quotas and fishing days among their members, thus reducing costs even further. The SCFP also necessitates the monitoring of shellfish stocks and natural values. Monitoring of shellfish stocks is partly paid for by the government and partly by the shellfish industry. The monitoring of bird numbers and other natural values is partly sponsored by the government and partly by conservation organizations.

10.2.3. Objectives of the conservationists Objective 8, of the conservationists, is primarily focused on ecological sustainability. Fishing activities can be defined as ecologically sustainable if they do not cause major and irreversible damage to the natural habitat (Ens, 2002). As such, emphasis should be placed upon the restoration of important habitats as well as natural fluctuations in the shellfish populations and the prevention of structural reductions in average shellfish populations as a result of fishing activities. This is in line with the common conservationist’s perspective: conservation of biodiversity should take precedence over maximizing profits from the fisheries (Agardi, 2000). The shellfishing methods utilized must minimize their effects on habitats and non-targeted species. Important habitats include sea grass and mussel beds. In addition to pressures produced by national and international obligations (see objectives 5 and 6), pressure on the government by the conservationists has contributed to the incorporation of environmental and ecological concerns into the SCFP as policy objectives. These policy objectives are listed in Table 10.1 as government objectives 1–3. They are not necessarily far-reaching enough for conservationists. In any case, because the SCFP has failed in its ecological objectives to restore bird populations and sea grass fields (see Section 10.2.2), it can be concluded that objective 8 of the conservationists has not been met. Objective 9 reflects a prime source of concern for conservationists: the effects of fishing methods on the habitat. Cockle fishing methods that employ suction dredges have been heavily criticized and are the subject of ongoing debate. There is no doubt that suction dredging causes considerable mortality among non-target fauna, including undersized cockles and other benthic animals. However, estimates of the time it takes the ecosystem to recover differ wildly, ranging from 56 days to over a year (de Vlas, 1982), to more than ten years (Piersma et al., 2001). The latter study proposes that suction dredging leads to the loss of fine sediments, reducing the seabed’s suitability for the recruitment of cockles and other shellfish. This conclusion has been questioned by Duiker et al. (1998) and is at odds with what the cockle sector believes (Stichting ODUS, 2001). Nevertheless, as the situation stands today, conservationists share the

198

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 10.4. Indicators of size for the mussel fishing sector over the period 1994–2001

Size indicators

1994

1995

1996

1997

1998

1999

2000

2001

Number of companies Number of boats Total engine power (1000 PK) of sector Investments (million euros)

68 77 39

64 76 40

64 75 42

66 75 43

65 76 47

63 77 49

58 73 52

55 71 51

4

7

6

8

6

3

7

3

Source: Data fromAgricultural Economic Institute (LEI, 2001; LEI, 2002b).

opinion that the mechanized cockle fishing sector should be closed. Because the cockle fishery at present still continues, objective 9 has not been met.

10.2.4. Objectives of the fishing sector Objective 10 concerns the fishing sector’s desire to be a profitable business. When it comes to shellfishing in the Waddenzee, fishermen generally believe fishing has an impact on the habitat, but they tend to disagree with conservationists about how large this impact exactly is. Their prime concern is shellfish stock sustainability and not ecological sustainability. Section 10.2.2 has shown that despite the catch restrictions imposed as a part of the SCFP, the mussel sector still has been profitable. The mussel fishing sector experienced record profits in 2000/2001. Profits derived in the cockle fishing sector were negative in 1991 (before the SCFP came into effect), 1996, and 2001. Nevertheless, the average profit in this sector over the last decade has been positive. While it is clear that cockle fishing activities have been limited by the SCFP, they have recovered some of the revenues lost in the cockle fishery by becoming active in the Spisula fishery. Objective 11 concerns the shellfishery’s desire to be a healthy business with a future. Ongoing investment in the fishery is a sign that the industry is healthy and expects to be profitable over a long horizon. Fishing capacity has increased, as shown by the increase in engine power in the sector (see Table 10.4). Investment in the mussel sector over the period 1994–2000 has reduced the proportion of boats older than 20 years from 56% in 1990 to 39% in 2001 (LEI, 2001; LEI, 2002b). The number of people employed in the mussel fishery in 1990 was 250 (LEI, 1998). There has been a slight decrease compared to 1990, but in recent years employment has been fairly stable, fluctuating between 214 and 230 over the period 1995–2001 (LEI; 2002b). LEI (1998) expects that a ten percent decrease in profits will have little effect on employment. Table 10.5 presents data regarding the Dutch cockle fishery. It concerns cockle fishing activities that have taken place in areas in the Netherlands other than besides the Wadden Sea, namely the Oosterschelde and the Westerschelde in the province of Zeeland. Nevertheless the Wadden Sea is by far the most important area for the cockle fishery. Furthermore the fishery was closed in the Oosterschelde between 1997 and 2000. Table 10.5 suggests that employment in the cockle fishery between 1995 and 2000 has been fairly stable. This is in contrast with employment data for the first half

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

199

Table 10.5. Indicators of size of the mechanized cockle fishing sector in the Netherlands over the period 1995–2001 Indicators

1995

1996

1997

1998

1999

2000

2001

Employment Total engine power (X 1000 PK) of the sector

66 15

66 7

66 11

70 18

66 17

64 10

64 10

Source: Data from Agricultural Economic Institute (LEI, 2002b).

of the 1990s (RLG, 1998), which suggests that employment was correlated with the profitability of the sector on a yearly basis. With the introduction of the SCFP in 1993, a total of 37 licenses have been issued to the mechanized cockle fishery, each giving the right to use one vessel. These licenses are in the hands of 12 companies or concerns. Thus only a few players control the cockle fishery (Ernst and Young Accountants, 2001). Nevertheless, the number of active vessels is far smaller than the number of licenses for two reasons. First, the cockle fishing sector requires fishermen to use one license per suction dredge (Steins, 1999). Since many cockle vessels have two suction dredges on board they now require two licenses each. Second, the number of vessels active in the fishery is limited by an annually fluctuating TAC. It should be noted that the reduction in vessels does not necessarily mean that there has been a reduction in fishing capacity as it is not clear whether the number of suction dredges used has been reduced. A strong level of vertical integration characterizes the cockle business (LEI, 2002a; LEI, 2002b). Most companies have integrated fishing activities with processing activities. Four processing businesses control the market. In recent years they have had a share of approximately 75%. In the season 1999/2000 their share of the market was practically 100% (LEI, 2002a). There is also a high level of horizontal integration in the cockle processing sector since it also engages in mussel and spisula processing activities (LEI, 2002a). As discussed in Section 10.2.2, over the last decade profit levels in the cockle fishery have fluctuated enormously and in some years even were negative. Section 10.2.2 also shows that profits derived in the cockle processing industry were less volatile and higher than those derived in the cockle fishery. The high level of vertical integration witnessed in the cockle business makes it more robust to environmental fluctuations than it would be if different companies conducted fishing and processing activities separately. As far as the mussel business is concerned, vertical integration is likely to benefit the processing sector, whose profits are much lower than those of the fishing sector. One of the factors that could have a negative impact on the future profitability of the Dutch shellfishery is the increased competition from Spain and Italy in the Western European market for shellfish (LEI, 2002a). As long as those active in the shellfishery cannot change the general perception that their activities in the Wadden Sea conflict with ecological objectives, their future is threatened. So far, they have failed to do so and as a result objective 12 has not been met.

200

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

10.2.5. Objectives of the researchers Objective 13 has been met: researchers are actively involved in the management process to assess the impact of the SCFP on the ecological values of the Waddenzee and on the shellfish fishing and processing sector. Scientists can be considered a stakeholder group because their research can benefit from the establishment of areas permanently closed to fishing. They may have roles as consultants or scientists. Research institutes that fulfil a consultative role become involved in the debate concerning the SCFP at moment they are hired to conduct research. These institutes generally conduct research on behalf of just one of the stakeholder groups (note that in this chapter the government is also considered a stakeholder). Other stakeholder groups may suspect that the outcome of such research is biased, especially when it presents conclusions that are favorable for those who have asked for the research to be conducted. Scientific researchers may feel the need to criticize the SCFP’s effectiveness because it addresses concerns they feel very strongly about. Notably, some ecological scientists adopt a more militant stance than conservationists do when it comes to the impact the SCFP may have on ecological systems (van der Have, 2002). In order to study the impact of fishing activities on marine ecosystems, a benchmark in the form of a similar ecosystem not impacted by the fishery is required, but in practice this is often unavailable. Whether the closed areas are representative of areas free from the effects of fishing activities depends on the size of the areas and the extent to which fishing activities can have spillover effects on regions outside of the fishing grounds. Because some of the areas permanently closed to fishing in the Wadden Sea are relatively small and because the area open to fishing is much larger than the closed area, fishing activities may well have an impact on the protected areas. Thus objective 14 has partly been met: although areas closed to fishing have been established, they are relatively small.

10.3. ESTABLISHING SUSTAINABLE POLICIES 10.3.1. Are current fishing practices sustainable? Pauly et al. (2000) suggest that most fisheries have historically not been sustainable. They argue that overfishing may have caused fish populations to fluctuate according to environmental fluctuations. Their argument is that fishing practices can cause ecosystem changes, which in turn can lead to increased sensitivity of already reduced fish resources to environmental fluctuations. Figure 10.1 and 10.2 in Section 10.2 show that increased fishing intensity in the Wadden Sea has led to an increase in landings over the latter half of the last century and that fluctuations in landings seem to be fairly normal. Nevertheless, during the 1990s there have been several years in which there have were fewer landings than there have been in any year between 1960 and 1990. The years in which there were extremely few landings in the 1990s, have also shown an abnormally high bird mortality rate. In short, what Pauly et al. attempt to point out for fisheries in general probably also holds true for the mussel and cockle population in the Wadden Sea: increased fishing pressure has led to an increased sensitivity of the mussel and cockle stock to environmental fluctuations. As a result bird population mortality has also experienced an increased sensitivity to environmental fluctuations.

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

201

Historically, fisheries science has not helped much to sustain fisheries. Bioeconomic models with various degrees of complexity have been used to perform stock assessments that could be used as input for fisheries management (Clark, 1990; see also Chapter 3). In practice these models have failed to improve fisheries management for several reasons, such as their lack of complete information about fish population and ecosystem dynamics (Caddy and Cochrane, 2001). The uncertainty fisheries are shrouded in, as well fisheries management’s failure to make use of bioeconomic models, has lead scientists to suggest that the key to creating sustainable fisheries may be the introduction of marine protected areas, the reduction of overcapacity, or a combination of these initiatives.

10.3.2. Reducing overcapacity Reducing overcapacity in a fishery has two main goals. The first is to reduce fishing pressure in order to protect fish stocks. The second is to create a sustainable and profitable fishery through the reduction of competition within the fishing sector and the revitalization of fish resources. In order to reduce overcapacity, fisheries management approaches have either adopted decommissioning schemes or have reduced the fishery’s permitted aggregate capacity (Jensen, 2002). Decommissioning schemes generally involve programs in which the government purchases and retires fishing vessels or fishing licenses (or both) from the fishing sector. The effectiveness of strategies for reducing overcapacity is not undisputed. As Jensen (2002) points out, they do not remove the economic incentive to create overcapacity. Capacity reductions have to be well defined and permanent and must target inputs into the fishery that cannot be compensated for by other inputs. Decommissioning programs can be mandatory or voluntary. Voluntary schemes that compensate fishermen based on their catch history have to make sure that they do not encourage fishermen to overfish in an attempt to maximize their buyout grant. Decommissioning schemes may compensate fishermen for the loss of their vessel and future profits, but they do not necessarily provide programs to assist fishermen with finding a job elsewhere. People dependent on the fishery for making a living, such as the crew and those working in processing industries, are also affected by decommissioning schemes, which may leave them without work and with few job skills to use in seeking employment elsewhere (Holland et al., 1999). Thus, decommissioning schemes may prove very expensive if they offer adjustment programs to all the people affected by buyouts. The question of whether there is overcapacity in the Wadden Sea shellfishery is difficult to answer. For the mussel sector, the SCFP has established a TAC of 65 million kg of mussel seed. Table 10.6 shows that since the introduction of the SCFP in 1993 the mussel seed catch has always been lower than the TAC. Thus one would expect that either there is not enough mussel seed in the Wadden Sea to fish profitably, or that there is undercapacity in the shellfish sector. Since fishing capacity in terms of engine power has increased over the last decade (see Table 10.4) and catches of mussel seed have exceeded the TAC in years prior to 1993, one would not expect undercapacity, but simply a shortage of mussel seed. Further evidence for this can be gathered by

202

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 10.6. Mussel seed catch

Year

1988

1993

1994

1995

1996

1997

1998

Mussel seed catch (in million kg fresh weight)

70

50

32

38

44

60

46

Source: Data from Mosselkantoor.

comparing actual catches to the available stock of seed mussels. Between 1991 and 1997, seed fishing took place on 11 occasions (either in spring or in autumn) and, on average, 73% of the fishable sublittoral stock was fished (Van Stralen, 1998). Each year, fishing takes only a few weeks. In terms of the cockle fishing sector, there are currently 22 cockle vessels that have 37 licenses. Table 10.7 lists the number of participating vessels in the cockle fishery over the last 7 years. From Table 10.7 we can conclude that all the vessels have been active in the fishery, except during years when the quota available to the fishing sector was very small as a result of the SCFP. These vessels were able to take their TAC of 10 million kg of cockle flesh in 1998, showing that there is no undercapacity in the sector. Whether or not there is overcapacity in the sector depends on how many boats are required to fish all the cockles in the period when the cockle fishery is open. This is not an easy question to answer, since several issues need to be taken into consideration. First, the amount of cockle fishing depends on the TAC, which is determined annually. Second, not all areas where cockles can be found are accessible to the vessels. An area’s inaccessibility can be permanent or temporary: suction dredging occurs during high tide and its use can be limited by weather conditions. Third, the number of boats required depends on the cockle vessels’ fishing capacity per unit of time. Through the introduction of fishing plans in 1992 (Steins, 1999), the PO has reduced the fleet size from 36 to 22 vessels and has restricted fishing activity by allocating days at sea between the vessels. Whether or not the reduction in fleet size has resulted in a significant reduction in fishing capacity is questionable: the number of licenses for the use of suction dredges has not been reduced; the remaining vessels in the cockle fleet are capable of fishing with 2 suction dredges each. The number of days at sea determined by the PO cockle fishery varies annually in accordance with fluctuations in the total cockle stock and the catch (see Figure 10.7). One may therefore conclude that controlling annual fishing activity by determining the total number of fishing days before the season starts has worked successfully. Nevertheless, a reduction in fishing activity is not the same as a reduction in fishing capacity. The former is a temporary reduction and the latter is a permanent one. Because the number of licenses issued has Table 10.7. Number of cockle vessels active in the fishery in the Wadden Sea Year

1994

1995

1996

1997

1998

1999

2000

Vessels

22

22

12

9–11

22

22

14

Source: Data from Ens (from personal communication with Holstein).

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION 1400

203 12.0

fishing days catch

1200

1000

Fishing days

8.0 800 6.0 600 4.0 400

Catch (million kg flesh weight)

10.0

2.0

200

0 1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

0.0 2000

Year

Figure 10.7. Annual fluctuations of fishing days allowed with the total catch. Source data: assembled by Ens (2002).

not been reduced, the potential fishing capacity has remained unchanged and the policy measures implemented by the cockle fishery create a temporary capacity reduction only. This situation is notably similar to that of the Dutch fishery targeting demersal and flatfish species. These fisheries were forced to restrict their fishing efforts as part of a European Union (EU) program to reduce fishing effort (Commission of the European Communities, 2001a and b). The so-called Multi Annual Guidance Programs (MAGP) failed to address the problem of overcapacity in the Dutch demersal and flatfish fisheries. As in the case of the Dutch cockle fishery, the fisheries’ response was to reduce fishing activity rather than fishing capacity. The Dutch fleet’s fishing activity reduction was the largest in the EU, whereas their capacity reduction was one of the smallest in the EU (Commission of the European Communities, 2000). The European community has specifically stated that any policy dealing with the environment should be based on the precautionary principle. As part of this principle, the European Community argues that fishing capacity should be permanently reduced to levels at which fishing mortality will not threaten the sustainability of fish populations. Thus the EU has shifted the focus of its MAGPs from fishing effort reduction to fishing capacity reduction. The Dutch government, as a member state of the EU, is responsible for upholding the precautionary principle when managing its shellfishery. The Dutch government should actively engage in research that could indicate the size of cockle fishery’s fishing capacity in the Wadden Sea, and should limit the fishing capacity to levels at which

204

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

the sustainability of the Wadden Sea ecosystem is not threatened, given natural environmental fluctuations. So far, such research has not been conducted. This could be due to the fact that the Dutch government may believe that the shellfishery should be managed in the same way as the demersal and flatfish fishery. The Dutch viewpoint here is, firstly, that decisions about the way fishing effort be restricted should be left to the POs, and secondly, that a system of TACs is at present the best way to manage fish stocks (Directie Visserij, 1999).

10.3.3. Establishing marine reserves The creation of marine reserves to manage fisheries is in line with the precautionary approach because it contains explicit recognition of the uncertainty of fish population dynamics and the uncertainty about fishing activities’ impact on the marine ecosystem. One of the main advantages of marine protected areas is that they provide a hedge against the economic bias of the policy makers who determine quotas. The fact that policy makers cannot deal well with scientific uncertainty strengthens the fishing sector’s position when they demand more lenient regulations (Clark, 1996). Experience suggests that well-defined marine protected areas can protect against the overexploitation of fishing grounds (Roberts et al., 2001). Gu´enette et al. (1998) provide an overview of how marine protected areas may benefit overfished fishing grounds elsewhere. These benefits include the dispersion of new recruits to overfished stocks beyond the marine reserve and the maintenance of ecosystem integrity. The latter is an important point because fishing techniques may have detrimental impacts on the habitats of important fish populations. Marine reserves may benefit fisheries managers since they are easier to enforce than gear restrictions and quota regulations (Gu´enette et al., 1998). Furthermore, although the short-term impact of establishing marine reserves on the fishing sector is probably negative, in the long-term the fishing sector may benefit from the closed areas: stocks in marine reserves may recover, which could lead to a significant influx of new recruits to the fishing grounds. Even in the case in which the benefits to the fishing sector are difficult to determine, managers may conclude that establishing a marine reserve is justifiable on the grounds that external benefits may arise through, for example, increased recreational use of the area (Farrow, 1996). Marine Protected Areas have often been associated with sedentary species that disperse larvae over a wide region (Gu´enette et al., 1998). For this reason one would expect the Dutch Wadden Sea to be a suitable candidate area for the establishment of marine reserves to protect ecosystem integrity. Designing a marine reserve is not necessarily straightforward. Boersma and Parrish (1999) argue that on a local scale marine reserves can be useful tools if their design take the biological characteristics of the population into account along with the physics of the local environment. The Dutch government has stated that more use of marine reserves should be made for certain fisheries (although it is not clear whether this entails permanent or temporary closure), instead of instruments targeted at reducing overcapacity (Directie Visserij, 1999). As part of the SCFP, 31% of the intertidal area in the Wadden Sea is currently closed to fishing. This is not a single large area, but comprises several separate areas. Whether the closure of these areas is effective depends on whether or

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

205

not the areas are large enough to provide a hedge against the influence of fisheries. This requires knowledge of the ecological population dynamics of shellfish, sea grass, and mussel banks. When the closed areas were established, insufficient knowledge of these ecological values existed. The evaluation programs of the SCFP are currently trying to correct this deficit. The fact that not all the ecological objectives of the SCFP have been met indicates that the closed areas have been poorly chosen.

10.3.4. Alternatives to the SCFP The Dutch government has recognized that changes in its shellfish management policy may be necessary in the future. It has initiated research on the impact of alternative SCFP management strategies on the fishery and the Wadden Sea ecosystem. The options that have been investigated so far consist of alternative designs for marine protected areas in the Wadden Sea or methods of reducing current fishing capacity, or both. One of the alternative strategies researched by the government is a voluntary buyout of the mechanized cockle fishery. Ernst and Young accountants (2001) concluded in their advisory report to the government that such a buyout should be based on a payment of eight times the average net profit on top of a compensation fee for liquidation costs. They concluded that this compensation would be a realistic alternative to fishing for the companies active in the cockle fishery. Based on their estimates of fisheries’ net profit over the period 1993–1997, they concluded that the liquidation of the whole cockle fishery, that is the fishing and processing sector, would cost the Dutch government an estimated 55.4 million euros. The government has also asked several agencies to conduct joint research in order to explore changes in the shellfisheries policy that might yield a “win–win” situation for both the shellfishery and the conservationists. The following options were explored (Wolff et al., 2001): 1. An exchange of areas currently open to the fishery for areas closed to the fishery; 2. Changing the way the shellfishery is managed in such a way that areas open to the fishery are significantly reduced in size; 3. Making intertidal areas with a high probability of having good mussel seed spatfall but a low probability of developing stable mussel beds available to the mussel seed fishery, in exchange for no, or limited, mussel seed fishing in parts of the subtidal Wadden Sea. 4. Gathering small cockles in areas with unfavorable conditions for cockle growth and scattering them in places with favorable conditions for cockle growth. These options have been explored taking the objectives of the SCFP into account. The degree to which these options would affect areas that have already been closed to shellfishing has also been considered. Wolff et al. (2001) have estimated that all options could yield a win–win situation, albeit with various degrees of uncertainty. They also stated that if one of the options were chosen it should have a time horizon long enough for the ecological benefits to manifest themselves.

206

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

Option 1 could yield a win–win situation if an exchange of areas would entail closing several areas now open to the fishery in exchange for a specific area with favorable prospects for the shellfishery (Wolff et al., 2001). The area to be closed to the fishery should be substantially larger than the area opened to the fishery. The benefits of such an exchange would be small for both parties. This option does not reduce overcapacity and thus does not necessarily take away the incentive to fish. Reducing the area open to fishing without restricting fishing pressure may lead to severe overfishing, which could have negative long-term consequences for both the marine ecosystem and the fishery. Option 2 could yield a win–win situation if the areas opened to fishing changed annually (Wolff et al., 2001). If option 2 entails that all areas in the Wadden Sea be fished at some time, then this could mean that areas previously fished may not be given time to recover. This option also does not reduce the incentive to fish. Given the fact that this option limits the effect of marine reserves, it seems almost impossible that this option could benefit conservation without the adoption of fishing capacity reduction programs. Option 3 does address limiting the mussel seed fishing capacity, but it does not address the cockle fishery. Closing intertidal areas may have a severe impact on the cockle sector. Thus this scenario might not be a win–win situation for that sector. Option 4 may be beneficial to both the cockle fishermen and the bird population, since it makes more cockles available to both of them. However, this assumes that it is possible to improve the growth and survival of cockles by relocating them, but this is not known. Furthermore, the sediment will be disturbed twice and this option does not address the ecological objectives of restoring sea grass and mussel beds. Therefore, sustainability concerns may still be an issue, which can imply that fishing capacity has to be reduced. Another change in policy that has been suggested (although not by the government) is granting conservation groups ownership rights in the intertidal areas of the Wadden Sea (Van der Have, 2002). Conservation groups could then lease fishing rights to fishermen and could use the payments to finance their activities. Although it is likely that a sustainable fishery would be established in this way, this policy would obviously meet with tremendous disapproval from the fishery sector. Conservationists and fishermen have very different ideas of what constitutes a sustainable fishery and do not have the same interests at heart. The common factor in all these suggestions for policy changes is the large degree of uncertainty about Wadden Sea ecosystem dynamics and about the extent to which current shellfishing methods impact the ecosystem. Although the 2003 reevaluation of the SCFP may reduce these uncertainties, it is hard to imagine that they will be completely eliminated for such a complex ecosystem. The application of some operational version of the precautionary principle will be unavoidable and the government should make sure that any change in management policy respects this.

10.4. CONCLUSION Managing the shellfishery in the Wadden Sea clearly illustrates the problems associated with managing a complex ecosystem with multiple stakeholders. The SCFP has encountered problems that are representative of those encountered by many fisheries

10. POLICY FAILURE AND STAKEHOLDER DISSATISFACTION

207

management regimes in complex ecosystems worldwide. Broadly stated, the problems the SCFP has come across fall in two categories: those that relate to the co-management arrangement in place and those that relate to choosing the right management policy for the future if the current one should fail. The main problems encountered within the co-management arrangement are the lack of communication between stakeholders and the lack of a common objective shared by all stakeholders. The latter circumstance has produced an additional problem for the SCFP: disagreement about which policies are appropriate. This has led to a situation in which the government, after the first review of the SCFP in 1998, chose a policy option that could be seen as offering a compromise between the objectives of the fisheries sector and the conservationists, rather than an adaptive policy that would change the fisheries policy according to the objectives set out by the SCFP. It is difficult to create a shellfisheries’ management plan that can meet all stakeholders’ objectives and such a management plan has not yet been found. It is clear that the fisheries sector has understood the need for sacrifice. Nevertheless, the shellfisheries sector currently perceives that its existence is threatened. Previously it did not see this. Although it is clear that the primary goal of the SCFP was to restore and preserve the ecological values of the Wadden Sea, fishermen thought that the measures taken would be temporary. The fisheries sector should not be held responsible for this misperception on their part. The government should have ensured that all parties understood what the SCFP meant and what the adaptive responses would be if its objectives were not met. If it had done so, stakeholders would probably have been more accepting of appropriate adaptive responses. Furthermore it is questionable whether a management plan could be devised that would satisfy all stakeholders involved. The government should make clear that it is constrained by EU regulations to protect the Wadden Sea, irrespective of what happens to the fishery. Until recently, the government has taken an approach in which the current fishing policy is maintained until further research concludes that precautionary steps are necessary. The fact that further research needs to be conducted is in itself justification enough to engage in adopting a precautionary approach right now. The question really is: how precautionary should a precautionary approach be? An idea that is gaining increasing acceptance is that in order to justify fishing practices it should be proved that they do not threaten the sustainability of fisheries. Currently, it is the other way around; fishing practices are tolerated unless it is proved that they threaten the sustainability of marine ecosystems. In this sense, the Wadden Sea shellfishery management policy has been no different from management regimes in many other complex ecosystems worldwide where all stakeholders have seen their objectives fail. However, it strongly contrasts with the way the Dutch government treated the Dutch Natural Gas Company (NAM), which wants to extract additional gas from areas below the Wadden Sea. Because it was impossible to prove with absolute certainty that gas extraction would not harm the natural values of the Wadden Sea, the NAM was not allowed to extract the gas. Thus, the Dutch government is inconsistent in its application of the precautionary principle. Whereas the NAM’s case may have been treated with too much precaution, the shellfish fishery’s problem has been treated without sufficient precaution. While this chapter was being completed, the situation surrounding the Wadden Sea shellfisheries has changed considerably. Based on a report by the “Advisory Group Wadden Sea Policy” (better known as the “Meijer committee”), in June 2004 the Dutch

208

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

government proposed to no longer allow mechanical cockle fishing in the Wadden Sea as of 1st of January 2005. This decision was approved by parliament. In addition, environmental organizations went to court and succeeded in preventing cockle fishing already in the second half of 2004, in spite of obtained permits for that year. This decision was upheld by the highest level of legal ruling, which is the “Raad van State” (“Council of State”, which is the supreme administrative court of the Netherlands). This means that mechanical cockle fisheries in the Wadden Sea has come to a definite end.

CHAPTER 11

STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS IN THE DUTCH WADDEN SEA

11.1. INTRODUCTION A particular type of shellfishery in the Netherlands, namely cockle fishery in the Wadden Sea, has received much attention from both policy makers and environmentalists. Until the 1950s, most of the catch resulted from fishing by hand. In the beginning of the 1960s the sector introduced mechanical techniques of fishing. According to environmental organizations (notably “De Waddenvereniging”1 ), the process of mechanical shellfishing altered the sediment structure of the seabed in an irreversible way (Stichting Odus, 20012 ). An additional negative impact of cockle fishery is that it withdraws a great amount of cockles from the food web in the Wadden Sea. Cockles constitute an important element of the diet of the bird population in the Wadden Sea (EVA II, 2003). This does not imply, however, that the relationship between cockle fishery and the environmental quality of the Wadden Sea is fully clear and understood. For a long time, it has been debated which policy measures should be taken. Some believed it is possible to make the cockle fishery more sustainable, while others pled in favor of forbidding the cockle fishery entirely. Recently, the government has decided in favor of the latter, based on the advice of a special committee for assessing policy in the Wadden Sea (Meijer, 2004). Since 1993, the main function of the Wadden Sea is to be a (national) natural area. This implies that human activities are allowed as long as they do not cause significant harm to natural values of the Wadden Sea. A relevant question that has never been answered is whether different stakeholders prefer similar or different levels of ecosystem quality. In this chapter, we want to address this question in a systematic way. In particular, it is assessed which policy measures and associated levels of environmental quality the different stakeholder groups prefer. Furthermore, the question who will benefit and lose from specific policies is addressed by measuring the welfare gains and welfare losses associated with them. The stakeholders we wanted to include in the analysis covered ‘ermen, Dutch citizens, local residents, tourists, policy makers, and natural scientists. The cockle fisherman, however, were (the only group) unwilling to participate in the survey. Due to the fact that the Wadden Sea is a national natural area, Dutch citizens can be identified as a relevant stakeholder group. Local residents are the people living in the coastal 1 2

Source: http://www.waddenvereniging.nl. Source: http://www.wildekokkels.nl.

209

210

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

area of the Wadden Sea. They include the inhabitants of the Dutch Wadden islands, and people who live less than 5 km from the coast in the northern provinces of the Netherlands. The policy makers are local politicians, local and national civil servants, and members of lobby groups who are involved in preparing formal policy.3 Natural scientists include biologists at universities or other (scientific) research institutes who have shown a special interest in the Wadden Sea. In order to assess and identify differences between individual preferences of the various stakeholders, we make use of the attribute-based stated choice (SC) methodology (Louviere et al., 2000). This non-market, economic valuation method uses a questionnaire in which respondents choose between alternative goods, in the present context interpreted as different policy options. This study further attempts to measure consumer motivations for the act of giving, sometimes referred to as “warm glow” (Nunes and Schokkaert, 2003). This is operationalized by including in the questionnaire a list of 26 attitudinal questions. The remainder of this chapter is organized as follows. Section 11.2 describes the current situation concerning the management of the Wadden Sea and identifies the alternative policy scenarios in the survey. Section 11.3 describes the SC valuation method used, and discusses the valuation question as well as the econometric models on which the data analysis will be based. In addition, individual motivational profiles will be explained, where special attention is given to the “warm-glow effect.” Sections 11.4 and 11.5 present the results. Section 11.6 concludes.

11.2. STATEMENT OF THE MARINE RESOURCE PROBLEM It is widely recognized that cockle fishery causes potentially significant damage to the environmental quality of the Wadden Sea. This has two dimensions. First, cockles live on water-soils, which makes them very sensitive to soil movements and accumulation of water sediments. Since the harvest of shellfish implies soil movements, due to the use of mechanical and vibrating fishing equipment, negative impacts on the morphology, as well as on marine life functions of the Wadden Sea are inevitable. A second dimension relates to the fact that shellfishing reduces food available to birds. The demand for cockles has increased rapidly over the last decades, notably due to increasing demand on the Spanish market, which has stimulated an increase in cockle fishing efforts in the Wadden Sea (Dijkema, 1997). This has resulted in fewer cockles in the food chain, which in turn threatens the life of many birds in the Wadden Sea. For example, the population of the Oystercatchers, a migratory bird that chooses the Wadden Sea for breading, decreased from 260,000 in the 1980s to about 150,000 in 1998. The last few years have shown a small recovering trend in the number of birds. Today, the population of the Oystercatchers is about 170,000. Another bird that has been very much affected by the reduction of cockles in the Wadden Sea is the Eider duck. In fact, the Eider duck population has decreased from 130,000 to 100,000 over the last decade (EVA II, 2003). 3

This is an example of the “polder-model,” a consultation framework that underlies much of the public decision making in the Netherlands.

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

211

In order to minimize the negative environmental impact of the shellfishery, the current fishery is regulated by the central government. The overall objective is to define precise food requirements, on the advice of biological experts, for the bird species wintering in the Wadden Sea. These can be translated into harvest standards for the cockle fishery. A policy proposal on the government’s agenda is identification of sensitive areas in the Wadden Sea, in which fishing is not allowed. Many individuals and organizations (including politicians and environmental groups such as de Waddenvereniging and Wilde Kokkels) propose a total ban of the cockle fishery in the Wadden Sea. Nevertheless, the perception and evaluation of the impacts of the cockle fishery is not the same for all stakeholder groups. In fact, the recent public debate about cockle fishery in the Wadden Sea makes clear that opinions about whether current cockle fishery is ecologically sustainable differ. So the current fishery policy involves two main restrictions on fishermen, namely fixed areas where fishing is not allowed; and quota to save food for birds. There is no consensus about whether these measures are sufficient in realizing ecological sustainability of the Wadden Sea ecosystem. In this study we propose additional, hypothetical but feasible, fishery management options for regulation of the cockle fishery in the Wadden Sea. We choose to include the following three groups of policy measures: (1) area policy measures (‘area’) – these change the surface area where fishing is allowed; (2) quota policy measures (‘quota’) – these reduce the maximum harvest of cockles the cockle fishery sector is allowed to fish; (3) rotation policy measures (‘rotation’) – these rotate the areas where fishing is not allowed. The third-policy option, “rotation,” has been proposed by the cockle fisherman. They see it as an opportunity to increase the harvest as well as improve the sustainability of fishing (Stichting Odus, 2001). Natural scientists, however, have argued that rotation implies a disturbance of the ecosystem that is more widely spread in the Wadden Sea. We study individual preferences for alternative fishing management scenarios in the Wadden Sea, wherein we make a distinction between relevant stakeholder groups. The latter include the cockle fishery, Dutch residents, local residents, tourists, policy makers, and natural scientists. Despite various efforts on our side to stimulate the cockle fishery sector to participate in this study they turned out to be unwilling to do so.

11.3. STATED CHOICE METHOD AND MODEL FOR DATA ANALYSIS For assessing individual preferences and identifying which measures are the most preferred by the different stakeholders, we make use of the attribute-based SC methodology (Louviere et al., 2000). This SC valuation method asks respondents to make a choice between alternative goods, defined in terms of their attributes. The SC method belongs to a family of attribute-based methods (ABMs). These are a special case of conjoint analysis (CA). By incorporating price as an attribute, ABMs can be used for the purpose of applied welfare analysis of changes in the attributes, thus providing information

212

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 11.1. Illustration of a SC question

Assuming that the following two fishing management practices were the only alternatives available, which one would you consider more attractive, if any?

Policy measures Surface area where fishing on cockles is allowed Allowed harvest of cockles Rotation Ecosystem effect Number of birds Costs Costs per household

Current situation

Policy proposal

Current area

Half of the current level

Current level No rotation Current level

Lower level No rotation More than the current level

0 Euro A

50 Euro B No preference

about willingness to pay (WTP) (Holmes and Adamowicz, 2003). ABMs assume that a respondent’s WTP consistently relates to his of her underlying preferences. The SC method was originally developed by Louviere and Hensher (1982) and Louviere and Woodworth (1983). Hanley et al. (2001) give an overview of SC studies carried out in the field of environmental economics. The use of this method allows us to estimate the value of an improvement in the environmental quality in the Wadden Sea, a non-market benefit, in monetary terms. In addition, as opposed to the older and much applied contingent valuation method, this valuation method offers a wide range of information concerning benefit tradeoffs between attributes (Adamowicz et al., 1998). The questionnaire used in the SC valuation experiment consists of four parts. In the first part, respondents are asked to report their behavior with respect to, among others, visiting the Wadden Sea area and food habitats in relation to the consumption of fish and shellfish. The second part of the questionnaire presents the context of the SC questions, after which a series of SC questions are posed. Each respondent faces nine SC questions, proposing all different tradeoffs among the attributes under consideration. More precisely, respondents are asked to compare the current fishing management practice with nine alternative policy options. In each question, the respondent is asked to choose one of both. In the third part of the questionnaire, respondents are asked to state their opinion about alternative marine management policies in the Wadden Sea. Finally, the fourth part of the survey assesses the socio-economic characteristics of the respondents. Table 11.1 gives an example of a SC question used. The attributes included in the SC question are the alternative cockle fishery policy options, the level of birds and the price level. The number of birds are regarded as an indicator of the environmental benefits of the policy measures, that is, as an indication of the quality of the ecosystem.

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

213

In each SC question, three different policy measures for regulation of cockles fishery are presented to the respondent. These policy measures refer to the three different variables presented in Section 11.2: “area,” the surface area where fishing is allowed; “quota,” the number of cockles that is allowed to fish; and “rotation,” whether the areas where fishing is allowed are fixed or rotated (see Table 11.1). A fourth variable gives the changes in the number of birds, which are interpreted as a proxy for ecosystem quality. The last variable in the choice experiment is a monetary variable that refers to a one-time lump-sum amount for an associated policy option. This gives a total of five distinct variables. In Table 11.2, the attributes and their levels as used in the analysis are presented. To set the level of the attributes for the policy attributes and the level of birds attribute, advice by a marine biologist with expertise on the study area was used. This conformed that we did indeed capture the most important policy measures and that the number of birds is a good indicator for the ecosystem quality of the Wadden Sea. For the current situation, the price is always equal to zero (i.e. the price is normalized on the status quo). For the alternative situation, the price is always larger than zero. Including the current situation in each SC question allows us to identify whether stakeholders prefer to keep the status quo policy scenario. The combination of all SC attributes, and respective attribute levels, allowed us to create a full factorial design. The outcome of such a procedure results in a very large choice set. We resized this by (1) eliminating all the dominant alternatives, and (2) eliminating all policy combinations that were internally inconsistent. We only constructed choice questions consisting of the current situation (status quo) combined with an alternative scenario. This results in 168 different choice sets. Presenting each respondent with 168 choice questions is not a feasible option. We choose to present eight different

Table 11.2. Attributes and their levels as used in the survey Attribute

Levels

Policy measure: “area” – surface area where it is allowed to fish on cockles

The whole Wadden Sea Current level Half of the current level No take (ban) Current level Lower level Rotation No rotation Lower level Current level More than in current situation Much more than in current situation. Nine different bid amounts between 0 and 250 Euro

Policy measure: “quota” – allowed number of cockles harvest Policy measure 3: “rotation” – rotation (or fixed) areas Indicator of the quality of the ecosystem: – number of birds

Financial cost

214

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

choice questions to each respondent. The 168 sets are then divided over 21 versions of eight questions, which were distributed randomly among respondents. In addition, we consider an “extreme policy scenario” (a ninth question). This reflects entirely banning the mechanical cockle fishery sector from the Wadden Sea area. To make this policy scenario realistic, the survey design always portrays to the respondents a higher level of birds associated with this scenario, indicating a better ecosystem quality. We worked with 16 different choice sets for “the banning option” (1 policy measure × 2 levels of birds × 8 price levels). For each choice question there are three possible answers, namely the current policy option, the alternative, and no choice (i.e. no preference). It is necessary to indicate the third no choice option, for two reasons: to deal with sample selection bias, and to develop demand models consistent with economic theory. Omission of this option can yield biased and misleading WTP estimates (Louviere and Street, 2000). Before executing the survey, a series of pretests were done on students in order to check the overall level of comprehension. This showed that the proposed alternative scenarios were accepted and understood. The data retrieved by carrying out a choice experiment can be analyzed by using as the theoretical framework the random utility (RU) model (McFadden, 1973). RU models estimate the probability that respondents will select an alternative based on its attributes. The probability that an individual, q, chooses alternative i is the same as the probability that the indirect utility of that alternative (Uiq ) is greater than the utility of the other choice alternative, j(U jq ). Uiq > U jq ⇒ (Viq + εiq ) > (V jq + ε jq )

for all

j = i, j ∈ Cn

(1)

The error terms are unobserved. Rearranging Equation (1) leads to the RU model given in Equation (2). The aim of this model is to estimate the value and statistical significance of the determinants of the utility function. Piq = P[{ε jq − εiq } < {Viq − V jq }]

for all

j = i.

(2)

Piq is the probability that respondent q choose option 1, given the indirect utility functions. To make this model tractable, the probability that an alternative is chosen needs to be between 0 and 1. This is the case when the error terms associated with each alternative are independently and identically distributed (IID) according to a type I extreme value distribution. This means that we should assume that the independence from irrelevant alternatives (IIA) condition holds. The IIA condition means that the probability ratio of two options should be unaffected by the inclusion or omission of other alternatives. Under these assumptions, the solution of the choice model becomes,

Pi = exp(Viq )



J  j=1

exp(V jq )

(3)

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

215

The utility function of the Wadden Sea is defined as an additive function of the attributes, as shown in Equation (4). Vi = Ci +

K 

βiq xikq

(4)

k=1

For all respondents we know the values of the attributes of the different alternatives (the values of the attributes vary over the respondents) as well as their choice with respect to the attributes. This model is estimated using the maximum likelihood procedure. The coefficient for an attribute is the marginal utility of that attribute. The (Hicksian) compensating variation (Hanemann, 1984) can be written as:    1 0 CV = − (5) V jq α V jq j

j

This model can be applied to predict individual choice behavior within the various stakeholder groups. Furthermore, it can be examined whether and to what extent preferences regarding the Wadden Sea differ among the stakeholder groups. A particular innovation of the current economic valuation exercise is that it combines socio-economic differences and (estimated) differences in warm glow motivation. The purpose of this is to test the significance of the warm glow effect. Warm glow is measured by using an attitudinal scale. In order to obtain internally coherent measures of warm glow motivation we identify, using factor analysis, on the basis of the answers to the attitudinal questions a set of latent underlying motivations. In addition, we can estimate for each respondent his or her individual motivational profile. This, in turn will allow the derivation of a consumer warm glow motivational profile. Together with the respondents’ socio-economic characteristics, this provides a firm basis for understanding and predicting WTP responses.

11.4. ANALYSIS OF THE DATA 11.4.1. Some basic statistics of the questionnaire Interviewers for the tourists and local residents were recruited among students. The tourists were interviewed in the summer of 2003, June till August, while the local residents were interviewed in autumn 2003, October till November. The sample-sizes for the tourists and the locals are 332 and 420, respectively. Most of the tourists were interviewed on the Islands Texel and Vlieland, and on the ferries to these Islands. The local residents were interviewed on the Islands Schiermonnikoog and Terschelling and in the northern provinces Friesland, Groningen, and Noord-Holland which border to the Wadden Sea. The sample of policy makers was constructed by collecting relevant addresses via the Ministries of Agriculture and the Environment. To construct the sample of natural scientists, addresses were collected via Dr. B. Ens, chairman of a large scientific research group responsible for an authoritative report on the impact of shellfishery,

216

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 11.3. Information about the stakeholder groups

Sample size Response rate Income Low income group (3000 Euro) Consume fish Consume cockles Know fisherman Member of environmental association Average age

Dutch citizens

Local residents

Tourists

Policy makers

Natural scientists

1558 70%1

420 49%

332 n.a.

39 42%

29 54%

28% 54% 18% 90% 11% 11% 34% 40.2

25% 46% 12% 85% 6% 51% 45% 43.4

14% 39% 33% 90% 4% 13% 56% 42.7

0% 33% 64% 97% 21% 62% 82% 50.1

3% 31% 66% 90% 48% 52% 90% 48.2

Note: The Dutch citizen sample was reached via the Internet. A contract with an Internet research company promised at least 1500 respondents. After 1558 respondents filled in the questionnaire, access to the questionnaire was closed.

commissioned by the Ministry of Agriculture. Both policy makers and natural scientists were sent the questionnaire by mail or email. After having sent out the questionnaires, we were confronted with the “politicalsensitivity factor” of our research topic. Several policy makers gave their judgement about the questions (instead of answering them). One recurrent comment was that they were not aware of this research project (as if they should have been). Others could not discover in the questionnaire what was the purpose of our research, and neither what it adds to the current debate (so at least our aim to minimize strategic behavior was realized). Some of them appeared to be afraid of the influence of any results. And some asked who financed this research (which is the Netherlands Organization of Scientific Research), because they worried about partiality of the research. Nevertheless, 39 policy makers, who deal with Wadden Sea policy sent back their questionnaire. At this stage, it became clear that fishermen and policy makers are in close (institutional) contact. Representatives of the shellfishery sector received the questionnaire that was originally distributed among policy makers. They read it and even discussed it at a formal meeting of their organization. Before this meeting, we were in contact with a representative of this organization, and tried to make clear what we were doing. The fishermen organization interpreted this type of socio-economic analysis with a certain degree of hostility, as they felt that results could be used against them. Therefore, the fishermen decided that they did not want to participate in our research. A few natural scientists made some comments on the questionnaire. They indicated that the survey described scenarios that were not completely “correct.” Our response was that the scenarios are hypothetical and not meant to be “correct.” but just realistic. In total 29 natural scientists sent back a filled in questionnaire. Table 11.3 summarizes some characteristics of the respondents representing the five stakeholder groups. The characteristics of the different stakeholder groups differ. The policy makers and the natural scientists have on average a larger income than the other two groups. The local residents have on average the lowest income. It is striking

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

217

to see that the natural scientists stakeholder group consumes most cockles. More than 50% of the local residents know a fisherman. This is also the case for the policy makers and natural scientists, but we guess that they know them from a work relation, while the local residents have private contacts. The average age of the policy makers and the natural scientists is higher than that of the tourists and locals.

11.4.2. Valuation of the stakeholder preferences for alternative policy instruments The RU model is estimated with a conditional logit model. Instead of operationalizing the use of the qualitative attribute levels with the use of the “status quo” as the omitted variable for each attribute, we made use of effect codes. In this way the omitted levels of each attribute can be estimated (Holmes and Adamowicz, 2003). Table 11.4 presents the estimation results for Equation (4) for the five stakeholder groups. Preferences regarding policy measures differ among the stakeholders. Because a monetary variable is included in our valuation exercise, it is possible to estimate the welfare changes due to the different policy measures. Table 11.5 shows the welfare changes, measured in monetary terms and according to Equation (4), associated with the potential adoption of the alternative policy measures. In this section, we will discuss the results with respect to the constant term and the fishery policies. In Section 11.4.3, the results with respect to the ecosystem quality will be discussed. The model specification allows us to capture consumer preferences with respect the status quo situation. A positive estimate indicates that choosing the status quo situation, independently of the proposed policy alternatives, increases respondents’ indirect utility. In other words, respondents prefer the current cockle fishery situation. According to Table 11.4, Dutch citizens, tourists, local residents, and policy makers have a positive estimate for the constant term. However, such an estimate only reveals to be statistically significant from zero for Dutch citizens, tourists, and local residents. This estimation result can be interpreted as a signal indicating that these stakeholder groups are averse to a policy change, or that they prefer the way things are. On the contrary, natural scientists show a clear preference for a policy change since the constant term estimate is negative. Nonetheless, this parameter estimate is not significantly different from zero. These estimation results are consistent with the descriptive statistics provided in Section 11.3 where it was shown that 29% of the local residents are characterized by always choosing the current fishery policy scenario. In contrast, this kind of behavioral pattern is not found for any of the respondents in the natural scientist group. As far as the “area” attribute is concerned, the preferences of the five groups are comparable. In fact, all stakeholder groups prefer the policy option characterized by cockle fishing “half of the current area” over the “current situation.” In addition, all stakeholder groups show consistent preferences for the “current situation” over a fishery policy that brings along with it the possibility to fish in the “whole area” of the Wadden Sea. In short, the ranking of stakeholder preferences with respect to cockle fishing is “half of the current area” over the “current situation” over the “whole area.” Furthermore, according to the monetary valuation results (see Table 11.5), natural scientists show the strongest (average) magnitude with respect to this type of preference order.

Table 11.4. Estimation results for different stakeholder groups Dutch citizens ∗∗∗

Constant term 0.56 Area where is it allowed to fish Whole area −0.35∗∗∗ Half of the current area 0.20∗∗∗ Ban cockles fishery −0.22∗∗∗ Quota Half of the current quota 0.01 Rotation Rotation present 0.11∗∗∗ Level of birds Less birds −0.955∗∗∗ More Birds 0.684∗∗∗ Much more birds 0.697∗∗∗ Price −0.014∗∗∗ Log likelihood −7493 Number. of observations 12981 0.17 Adjusted ρ 2

(7.94) (−7.41) (6.53) (−6.77)

Tourists 0.24

∗

Local residents (1.93)

−0.28∗∗∗ (−3.35) 0.20∗∗∗ (3.15) −0.44∗∗∗ (−6.74)

(0.49)

0.04

(0.83)

(4.40)

0.10∗

(1.91)

(−8.57) (15.13) (13.86) (−33.17)

−1.48∗∗∗ 0.80∗∗∗ 0.98∗∗∗ −0.010∗∗∗ −1712 2932 0.15

(−6.17) (8.35) (9.06) (−16.19)

∗∗∗

0.711

(6.19)

−0.32∗∗∗ (−4.08) 0.15∗∗∗ (2.52) −0.36∗∗∗ (−5.55) −0.05 0.08∗ −0.59∗∗∗ 0.50∗∗∗ 0.362∗∗∗ −0.007∗∗∗ −2147 3578 0.13

Policy makers 0.371

(0.944)

−0.363 (−1.479) 0.359∗∗ (1.953) n.a.

Natural scientists −0.455

( −1.043)

−0.777∗∗∗ (−2.908) 0.674∗∗∗ (3.025) n.a.

(−1.30)

−0.0033 (−0.25)

(1.88)

−0.091

(−0.712)

−0.296∗

(−1.900)

−0.953 0.0053 0.870∗∗∗ −0.006∗∗∗ −180 293 0.08

(−1.514) (0.020) (2.930) (−3.085)

−0.334 0.012 0.558∗ −0.008∗∗∗ −129 218 0.11

(−0.580) (0.044) (1.768) (−3.789)

(−3.01) (6.13) (3.99) (−11.83)

The significance of the estimates is indicated by ∗∗∗ , ∗∗ and ∗ , referring to the 1%, 5% and 10% level, respectively; t-values are between brackets.

0.062

(0.394)

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

219

Table 11.5. Marginal WTP estimates (in Euro) Dutch citizens Constant term Area where is it allowed to fish Whole area Current area Half of the current area Ban cockles fishery Quota Current quota Half of the current quota Rotation of the fishing area No rotation (current situation) Rotation Level of birds Less birds Current level of birds More Birds Much more birds

Tourists

41∗∗∗

Local residents

Policy makers

Natural scientists

24∗

101∗∗∗

65

−58

−25∗∗∗ −26 15∗∗∗ −16∗∗∗

−28∗∗∗ −51 20∗∗∗ −44∗∗∗

−45∗∗∗ −75 21∗∗∗ −51∗∗∗

−51 −1 50∗ n.a.

−109∗∗ −13 94∗∗ n.a.

−1 1

−4 4

8 −8

−8 8∗∗∗

−10 10∗

−12 12∗

−70∗∗∗ −31 50∗∗∗ 51∗∗∗

−146∗∗∗ −30 80∗∗∗ 97∗∗∗

The significance of the estimates is indicated by respectively; t-values are given between brackets.

∗∗∗ , ∗∗ ,

−84∗∗∗ −38 71∗∗∗ 51∗∗∗

0 0

−9 9

13 −13

41∗ −41

−133 −14 1 122∗∗∗

−47 30 2 78∗

and ∗ , referring to the 1%, 5%, and 10% level,

In fact, the marginal WTP for adoption of “half of the current area” policy is about 94 Euro, whereas the marginal willingness to accept (WTA) the no take (“ban”) fishing option is 109 Euro. One SC question includes the scenario regarding the ban of the totality of the cockle fishery activities of the Wadden Sea. This scenario is interpreted as an extreme policy option. It appears that Dutch residents, tourists, and locals prefer the current situation overbanning cockles fishery. The sample for policy makers and natural scientists is too small, which implies that the responses to the banning policy option by these stakeholders could not be included in the analysis. Due to this, we do not have (welfare) estimates for these groups. Estimation results show that none of the stakeholders reveal a clear preference with respect to the “quota” policy measure. In fact, Table 11.4 shows that the respective parameter estimates are not statistically significant from zero. We can conclude that the reduction of the current quota by half is an unattractive policy measure for all stakeholder groups. As mentioned, including rotation as a policy attribute in the management of the cockle fishery in the Wadden Sea has been originally proposed by the cockle-fishing sector. Table 11.4 further shows that the introduction of a rotation principle is welcomed among Dutch citizens, tourists, and local residents. On the contrary, this policy is strongly rejected by the natural scientists. According to the welfare estimates, Dutch citizens, tourists, and local residents present a marginal WTP for the adoption of the rotation principle that ranges from 10 to 12 Euro. In contrast, a natural scientist has on average a WTP of 41 Euro to prevent the introduction of rotation. This result reiterates the validity of the information provided by Table 11.3 that 45% of the

220

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 11.6. Different valuation of “more birds” and “much more birds”

Likelihood ratio test

Dutch citizens

Tourists

Local residents

Policy makers

Natural scientists

0

3.2∗

2.8∗

8.7∗∗∗

3.1∗

The significance is indicated by ∗∗∗ , ∗∗ , and ∗ , referring to the 1%, 5%, and 10% level.

natural scientists rejected rotation completely. A possible explanation for this answering pattern can be the fact that lay-people perceive the rotation principle as positive. In contrast, expert individuals, such as natural scientists and policy makers, clearly see the disadvantages of such a procedure. These relate to a wider spread of the human activity in the natural marine environment, creating an additional threat to many of the sensitive marine areas.

11.4.3. Valuation of the stakeholder preferences for environmental quality Ecosystem quality changes in the Wadden Sea area are tackled by changes in the population of birds registered in that same area. For all stakeholders groups, “more birds” and “much more birds” are preferred to the current situation. This result can be interpreted as a clear signal that all stakeholders are in favor of an improved ecosystem quality of the Wadden Sea. A closer look at the estimation results shows that the intensity of stakeholder’s preferences is not the same. In other words, the desired level of ecosystem quality of the Wadden Sea, measured in terms of the number of birds in the Wadden Sea, varies across the different stakeholders. In fact, with the exception of the local residents, all the stakeholders have preferences such that “much more birds” are preferred over “more birds.” For local residents the marginal WTP for “more birds” is 72 Euro and thus higher than the marginal WTP for “much more birds,” which equals 51 Euro. These results can be interpreted as signalling a conservative propensity towards an improvement in the ecosystem quality of the Wadden Sea area. In other words, “more birds” is their preferred level. Among the other stakeholders, “much more birds” is the preferred level. Furthermore, policy makers is the stakeholder group that presents the highest marginal WTP to improve the environmental quality. According to Table 11.5, they are willing to pay 122 Euro to have “much more birds” in the Wadden Sea area. Tourists and natural scientists value the “much more birds” scenario as 96 Euro and 51 Euro, respectively. For the Dutch citizens, the difference in WTP to have “more birds” and “much more birds” is only 1 Euro. A question that emerges from this analysis is to test formally whether stakeholder groups make a distinction between the different proposed changes in the environmental quality of the Wadden Sea. Bearing in mind Equation (3), we want to test the following hypotheses: i i H0 : βmore birds = βmuch more birds

i i Ha : βmore birds = βmuch more birds

with i denoting the range of stakeholders in the analysis. We run a likelihood ratio-test (χ 2 -test). Test results are presented in Table 11.6.

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

221

Table 11.7. Individual entrance fee (in Euro) Household WTP = 81.33 Euro

Discount rate (%)

15 years

30 years

Category = adults

3 4 5 3 4 5 3 4 5

1.85 1.70 1.57 0.92 0.85 0.78 4.54 4.18 3.84

0.72 0.61 0.51 0.36 0.30 0.25 1.77 1.49 1.24

Category = children

Per household

Likelihood ratio test results confirm that four stakeholder groups make a distinction between “more birds” and “much more birds,” at a 10% significance-level. Dutch citizens do not make a significant distinction between the two proposed changes in environmental quality. As an example, we will give a practical application of our estimation results. To do this, we will make use of the tourists data for which we estimated the model with only one variable for a higher level of birds, including “more birds” and “much more birds.” The results of this model indicate a marginal WTP for a higher level of birds of 87 Euro per household. An estimate of the total welfare gain for the tourist’s amounts to 31 million Euro. This value is calculated by dividing the value per household by the average number of persons in a household, and subsequently by multiplying this amount with the total number of individual visitors to the Wadden Sea area, estimated to be around 1 million (990.000 in 2002; CBS, 2003).4 The remaining value can form the basis for compensating cockle fishermen so as to reduce their fishing effort, or adopting environmental sustainable fishing methods that are characterized by a lower productivity. In this context, one could explore the idea of introducing an entrance fee to the tourists visiting the Wadden Sea area (for example, as an additional cost of the boat ticket or a tourist tax). The amount of such a fee would depend on both (i) discount rates and (ii) number of years that such a tax would be collected. The official Dutch discount rate used to account the net present value of public financed projects is 4%. The European Union advises a discount rate of 5%. In other Western European countries the discount rates varies between 3% (Germany) and 8% (France) (Ministerie van Financien, 1995: Eijgenraam et al., 2000). Table 11.7 presents some results, assuming that the entrance fee is set differently among the visitors, in particular that adults pay twice as much as children. Based on these assumptions the entrance fee for adults ranges between 0.51 and 1.85 Euro. The entrance fee per household (without extra assumptions) ranges between 1.24 and 4.54 Euro. 4

According to our survey responses, each household is composed of 2.75 persons (2.14 adults and 0.61 children).

222

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

11.5. TESTING FOR A WARM GLOW EFFECT IN A SC SETTING 11.5.1. Retrieving motivational profiles of warm glow Following the methodological guidelines proposed by Nunes and Schokkaert (2003) to identify and measure warm glow in economic valuation surveys, the questionnaire of this study included a list of 12 attitudinal questions to be answered by the respondents on a five-point Likert-scale (Likert, 1967), with values ranging from 1 for “I disagree completely” to 5 for “I agree completely.” The application of factor analytical methods allows the identification of consumer motivations regarding cockle fishing in the Dutch Wadden Sea in general, and of consumer warm glow motivation in particular. Formally, we use the following model: av = f +  where av is the matrix representing the answers by the sample respondents to the 12 attitudinal questions; f captures the matrix of factor scores giving the position of the sample respondents on four latent factors Fi (i = 1, . . . ,4) interpreted here as general motivations;  captures the matrix of factor loadings showing the correlations between the answers on the 12 items and the respondents’ factor scores; and  captures the matrix of the residual terms. The varimax rotation estimation procedure is used to perform the factor analysis. This method allows an easy interpretation of the results, and attempts to construct “factor loadings,” which are interpreted as the correlations between the attitudinal questions and latent (unobserved) motivational profile (see Harman, 1976; and Nunes, 2002 for technical details). The factor loadings after varimax rotation are shown in Table 11.8. The asterisks denote values above 0.45 (or 45 in the table, as all numbers are multiplied by 100).5 The main items loading on a given factor share the same conceptual motivations and items that load on different factors are associated with different conceptual motivations. The items loading on factor 1 (F1 ) relate to the human, non-recreational use of the Wadden Sea. Therefore, this latent variable is interpreted as the “non-recreational use” motivation. Factor 2 (F2 ) is associated with items that indicate respondents’ perception with respect to the ecosystem’s quality of the Wadden Sea and the current bird population size. Therefore, this latent variable is interpreted as the “nature quality of the Wadden Sea” motivation. Factor 3 (F3 ) is associated with items that underpin a respondent’s feeling of well being or satisfaction generated by the act of giving. We interpret it as the “warm glow” motivation. Finally, factor 4 (F4 ) is associated with items indicating the respondent’s moral considerations with respect to the protection of the existing cockle fishery community operating in the Wadden Sea. We interpret it as the “community” motivation After having defined the content of the factors, the next step is to determine the position of the individuals on these latent motivational factors. These are given by the standardized factor scores (again with mean zero and variance equal to 1; see Harman, 1976). For example, a higher value for Factor 3 indicates that the respondent has a stronger propensity for warm glow of giving. Then we are in a condition to introduce 5

The factor loadings are analogous to standardized regression coefficients. 0.45 is set as a minimum correlation. Items that have a low correlation with the common factors are not taken into consideration.

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

223

Table 11.8. Loading factors after varimax rotation (multiplied by 100) Items 1 6 7 8 9

10 11 2

4

12

3 5

Statement It is important that the Wadden Sea is intact It should be possible to use the Wadden Sea for commercial purposes Tapping Wadden Sea gas should be allowed. Military training in the Wadden Sea should be possible The current number of birds in the Wadden Sea can be considered as a good indicator for the quality of the ecosystem. The Wadden Sea is the most important natural area of the Netherlands To keep the birds population in tact, it is necessary that there are enough cockles. There are some funding campaigns to which my family and I show much sympathy and therefore do not hesitate to support by contributing with a donation. My family and I like admires individuals who, on a voluntary basis, participate in collecting donations for national programs for social aid and solidarity I am happy with myself whenever I give a financial contribution to a charity organization There are ecological risks related to cockle fishery in the Wadden Sea Cockle fishermen should have the possibility to earn enough for a living.

F1

F2

F3

F4

∗

−47 66∗

41 −8

20 −9

−20 18

82∗ 78∗

−13 −15

−9 0

3 6

−15

72∗

5

−5

−18

67∗

16

12

−3

73∗

5

−27

−17

11

66∗

−21

−7

3

72∗

11

−2

16

76∗

2

−12

35

28

−58∗

14

14

−8

78∗

these latent variables in the valuation function and check whether they play a significant role in explaining reported WTP responses. Therefore, we estimate a full model with all attributes of the SC question as explanatory variables, including interaction effects with the motivational factor scores.

11.5.2. Estimation results The estimation results of the warm glow effect in the multivariate setting are presented in Table 11.9. Most of the results speak for themselves and are quite similar to the ones already reported in Table 11.4. We, therefore, focus the attention here on the effects of the motivational factors scores.6 6

Since the valuation of environmental quality is portrayed in terms of the number of birds, we only include the explanatory factor score variables for the **bird-levels and for price. Due to small sample sizes for policy makers and natural scientists, we restrict ourselves to three stakeholder groups, namely Duch citizens, local residents and tourists.

224

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 11.9. Full estimation model with individual latent motivational factors

Variable Constant term Area where is it allowed to fish Whole area Half of the current area Ban cockles fishery Quota Half of the current quota Rotation of the fishing area Rotation Level of birds Less birds More Birds Much more birds Price F1 × more birds F1 × much more birds F1 × price F2 × more birds F2 × much more birds F2 × price F3 × more birds F3 × much more birds F3 × price F4 × more birds F4 × much more birds F4 × price Log likelihood Number. of observations Adjusted ρ 2

Dutch citizens 0.48∗∗∗ (6.43) −0.38∗∗∗ (−7.47) 0.23∗∗∗ (6.75) −0.26∗∗∗ (−7.51) 0.02

(0.59)

0.12∗∗∗ (4.74) −1.06∗∗∗ 0.73∗∗∗ 0.76∗∗∗ −0.018∗∗∗ 0.10∗∗∗ 0.15∗∗∗ 0.003∗∗∗ −0.12∗∗∗ −0.15∗∗∗ −0.003∗∗∗ 0.23∗∗∗ 0.13∗∗∗ 0.002∗∗∗ −0.20∗∗∗ −0.14∗∗∗ −0.002∗∗∗ −6549 12981 0.27

(−9.22) (15.30) (14.29) (−31.20) (3.95) (5.49) (6.11) (−4.70) (−5.49) (−5.28) (8.58) (4.48) (4.04) (−7.10) (−4.60) (−4.44) 2932 0.22

Tourists 0.17

Local residents (1.33)

−0.30∗∗∗ (−3.41) 0.21∗∗∗ (3.11) −0.49∗∗∗ (−7.12) 0.05

(0.91)

0.11∗∗

(2.15)

−1.49∗∗∗ 0.75∗∗∗ 0.90∗∗∗ −0.015∗∗∗ 0.21∗∗∗ −0.02 0.003∗∗∗ −0.13∗∗∗ −0.04 −0.003∗∗∗ 0.07 0.16∗∗∗ 0.001∗∗ −0.04 −0.20∗∗∗ −0.002∗∗ −1574 3578 0.20

(−6.12) (7.35) (7.86) (−13.98) (3.55) (−0.34) (3.34) (−2.76) (−0.82) (−3.35) (1.42) (2.99) (2.08) (−0.83) (−3.54) (−2.54)

0.68∗∗∗ (5.72) −0.31∗∗∗ (−3.90) 0.14∗∗∗ (2.33) −0.41∗∗∗ (−5.96) −0.06 0.09∗∗ −0.64∗∗∗ 0.56∗∗∗ 0.37∗∗∗ −0.008∗∗∗ 0.05 0.17∗∗∗ 0.001 0.02 −0.13∗∗∗ −0.000 0.09∗∗ 0.02 −0.000 −0.09∗∗ −0.25∗∗∗ −0.001∗ −1975

(−1.29) (2.02) (−3.17) (6.46) (3.80) (−11.07) (1.03) (3.23) (1.57) (0.37) (−2.67) (−0.69) (2.15) (0.51) (−1.37) (−2.22) (−5.80) (−1.77)

The significance of the preference weights is indicated by ∗∗∗ , ∗∗ and ∗ , referring, respectively, to the 1%, 5%, and 10% level, with t-value *****between brackets.

The overall pattern presents a remarkable econometric robustness, which suggests an important contribution to explaining the individual choice. In particular, we can see that evaluation of the environmental quality depends on the importance attached by the stakeholders to the different motivations. The estimated coefficients relating to the warm glow motivational factor F3 are statistically significant in all stakeholder groups. We therefore can conclude that warm glow is present in the consumer responses to the SC valuation questions. In addition, the stronger is the individual’s moral consideration regarding the right of the cockle fishery community to operate in the Wadden Sea, factor 4, the lower is the respondent’s propensity to endorse a governmental measure to promote an increase in the bird population – see cross effect “F4 × more birds” and “F4 × much more birds.” This effect is valid for all the three stakeholders under consideration. The highest estimation magnitude is obtained for “local residents” valuing a governmental program that ensures the provision of “much more birds” in the Wadden Sea.

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

225

A similar result, however of a smaller magnitude, is present for both “Dutch residents” and “tourists.” The effects of the “community” motivational profile, factor 4, can be interpreted as indicative of the support to the local fishermen community. It can be seen as an important cultural element of the Wadden Sea area that needs to be protected. All in all, the statistical significance of the influence of this motivational profile in choice behavior indicates that respondents are interested in a diversity of functions and uses of the Wadden Sea, covering both the presence of a protected marine ecosystem and human economic activities such a fishing. Table 11.9 shows that the warm glow has two different effects. A first effect concerns the direct effect of the warm glow motivational profile on the choice of the specific environmental quality program, captured by “F3 × more birds” and “F3 × much more birds.” The respective parameter estimates capture the warm glow or moral satisfaction provided by contributing to a specific project, in this case to a specific environmental quality protection program. For this reason let us name this effect as “project specific warm glow.” According to the estimation results, the empirical magnitude of this “warm glow mechanism” is particular significant for the “Dutch citizens” sample. In fact, respondents belonging to this stakeholder type who are, ceteris paribus, relatively sensitive to warm glow, reveal a higher choice propensity for environmental quality. At a first sight the estimates in Table 11.9 suggest that the marginal effect of differences in the warm glow motivation profile, when measured in terms of the “project specific valuation mechanism,” is different across the two environmental quality protection programs under consideration. In other words, for the “Dutch citizens” sample, the magnitude of the “project specific warm glow” is weaker in the scenario with “much more birds” than in the scenario with “more birds.” Indeed, formal testing confirms this idea. The likelihood ratio test statistic for the restriction of equal warm glow effects is 4.503, well above the 95% critical level of the chi-square distribution with one degree of freedom. This result suggests that the marginal effect of differences in warm glow motivation on individual choices is not the same for the different environmental quality protection programs under consideration. In fact, it shows a marginal decreasing warm glow mechanism, i.e. warm glow effect increases with the proposed environmental quality protection program, however it increases at a decreasing rate. A second effect of the warm glow motivational profile is the impact on the cost and thus likelihood of contributing. This is captured in Table 11.9 by “F3 × price.” The respective parameter estimates capture the general feeling of the warm glow of contributing, independently of the environmental policy program. For this reason, this effect can be referred to as “global warm glow.” The empirical magnitude of the effect of global warm glow is significant for both “Dutch citizens” and “Tourists” sub-samples. According to the estimation results for these stakeholders, respondents who are relatively sensitive to warm glow show a relatively low responsiveness to the financial cost by choosing independently of whether the environmental program refers to “more birds” or “much more birds.” In other words, for the same financial cost (price), these respondents are more likely to choose for environmental protection rather than for the current situation, independently of the specific protection scenario under consideration. Table 11.9 suggests that the magnitude of this warm glow mechanism is comparable for the two stakeholders “Dutch citizens” and “Tourists.” This result is confirmed by formal testing of the “F3 × price” parameter estimate across the two sub-samples.

226

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS 0 ,0 0 3 0 0 0 ,0 0 2 8 3 0 ,0 0 2 7 6

0 ,0 0 2 5 0

Estimate for F3*pice

0 ,0 0 2 3 9

0 ,0 0 2 0 0 0 ,0 0 1 9 0

0 ,0 0 1 6 3 0 ,0 0 1 5 0 0 ,0 0 1 4 2

0 ,0 0 1 0 0

0 ,0 0 0 9 8 0 ,0 0 0 8 7

0 ,0 0 0 5 0

0 ,0 0 0 0 9 0 ,0 0 0 0 0 D u tc h

To u r is t s

Po o le d m o d e l

Figure 11.1. Monetary estimates for the warm glow-price cross effect

The estimate for “F3 × price” is not statistically significantly different between the two sub-samples so that a single estimate makes sense. This idea is confirmed by Figure 11.1, which depicts the monetary estimates expressed for a 90% level of confidence. As we can see, this figure indicates a large overlap between the confidence intervals regarding the two interval estimates. It confirms that the magnitudes of the warm glow mechanism are comparable for the two stakeholders under consideration, namely “Dutch citizens” and “Tourists.” This result corresponds to the general feeling that respondents get moral satisfaction from the act of giving per se, and that this feeling is present whenever the project refers to good causes, such as environmental protection of the Wadden Sea. Moreover, it is not very sensitive to the specific type of regulatory intervention that is being proposed.

11.6. CONCLUSION The aim of this chapter was to assess the preferences of five stakeholder groups, namely Dutch citizens, local residents, tourists, policy makers, and scientists, for management of cockles fishery in the Wadden Sea. An economic SC approach was used for this purpose. The estimation results show that the stakeholder groups clearly differ in terms of average preferences over the policy measures. While all groups prefer the policy measure that fishing is allowed in half the area, the tourists, the Dutch citizens, and the local residents do not like the “extreme scenario” in which fishing is banned. The tourists the Dutch citizens and the local residents like the “rotation” policy measure, whereas the policy makers, and more so the natural scientists, dislike this measure. The latter believe that this fishery policy may destroy the ecosystem in the Wadden Sea. In addition, the monetary values assigned by the different stakeholder groups to the different policy measures were estimated. All stakeholder groups prefer half the current area for fishing. The WTP for a smaller quota is zero or almost zero for all

11. STATED CHOICE VALUATION OF MULTIPLE STAKEHOLDERS

227

Table 11.10. Ranking of policies Stakeholder

Ranking

Dutch Tourists Locals Policy makers Natural scientists

Much more birds ≈ more birds φ current situation Much more birds φ more birds φ current situation More birds φ much more birds φ current situation Much more birds φ more birds ≈ current situation Much more birds φ more birds ≈ current situation

φ means “preferred to”, and ≈ means “indifferent to.”

stakeholder groups. Tourists and local residents prefer the “rotation” policy measure, while natural scientists prefer no rotation. The latter group has a negative willingness to pay (WTP) for rotation. The qualitative results of the analysis are summarized in Table 11.10. Three stakeholder groups prefer more birds than in the current situation, while two are indifferent. Local residents prefer the option “more birds” above “much more birds.” The other stakeholder groups prefer the option “much more birds” over “more birds” and the latter over the current situation. The WTP of local residents is the highest for “more birds,” while the WTP of the other three stakeholder groups is the highest for “much more birds.” Finally, the motivational factors derived from the questionnaire have enabled us to retrieve individual motivational profiles for warm glow, which represents a general feeling of well-being or satisfaction generated by the act of giving. Interpersonal differences in warm glow motivation were estimated with factor analysis, performed on a list of attitudinal items. The results confirm the presence of a warm glow motivational factor and the respective parameter estimate reveals a robust effect on the SC answers. The influence of warm glow on the economic valuation of the environmental protection programs for the Wadden Sea turns out to vary across the stakeholders under consideration. Taking into account the various survey characteristics of individuals, we were able to test formally the nature of warm glow across different types of stakeholders as well as alternative protection programs. Formal testing confirms that the overall magnitude of the warm glow valuation mechanism depends both on the public good that is the object of it and on the individual stakeholder under consideration. Therefore, we can conclude that the use of direct attitudinal information may play a crucial role in obtaining a better understanding of the real content of SC answers, and respective policy options.

CHAPTER 12

THE COST OF EXOTIC MARINE SPECIES: A JOINT TRAVEL COST – CONTINGENT VALUATION SURVEY

12.1. INTRODUCTION1 This study performs an economic valuation of a marine protection program targeted at the prevention of harmful algal blooms (HABs) species along the North Holland open sea coastline of the Netherlands. The term “harmful” refers to a set of algal species that have significant damages to human health, to beach recreation, and to the marine ecosystem, including “red tides” (Anderson, 1994; Perrings et al., 2000; van den Bergh et al., 2002). Algal are primarily introduced in North European waters through ballast water of ships, i.e. water transported by ships across the oceans so as to keep a vessel in balance. It therefore makes perfect sense for port authorities to impose standards on ballast water treatment. Evidently, costs are associated with this activity. This, in turn, leads to the question which benefits are associated with marine protection programs. The present valuation exercise is targeted at a quantitative assessment of the non-market benefits associated with the introduction of ballast water standards. The valuation exercise consists of the three elements. First, it uses travel cost (TC) method so as to measure recreation benefits derived from the prevention of HABs. Second, it uses of contingent valuation (CV) so as to measure bio-ecological benefits derived from the prevention of red tides in the marine ecosystem. Finally, it combines CV and TC valuation results so as to arrive at a complete picture of the benefits provided by the marine protection program, which can serve as a basis for regulation of ballast water. The organization of the remained of this chapter is as follows: Section 12.2 discusses why HABs can be considered a problem for marine resources in the North European waters. Section 12.3 discusses the design and execution of the questionnaire. Section 12.4 performs the TC analysis. Section 12.5 presents the CV exercise. Section 12.6 integrates the economic value estimates and discusses policy implications.

12.2. STATEMENT OF THE NATURAL RESOURCE PROBLEM 12.2.1. Harmful algal species blooms in North European waters There is a long history of accidental introductions of HABs in North European waters (Ostenfeld, 1908). A number of HABs, collectively known as “flagellates,” has caused 1

Reprinted from P.A.L.D. Nunes, and J.C.J.M. van den Bergh (2004). Can people value protection against invasive marine species? Evidence from a joint TC-CV survey in the Netherlands. Environmental and Resource Economics 28(4):517–532.

229

230

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

considerable ecological and economic damage in the past 30 years. For example the flagellate, Gymnodinium mikimotoi, a native Japanese species, has been responsible for recent massive kills of fish and bottom-living animals in North European waters (Boalch, 1987). Another toxic micro-alga, Gymnodinium catenatum, produces potent neurotoxins that accumulate through the food webs, which has led to the death of marine life in the North Sea (Nehring, 1995, 1998; Hallegraeff and Fraga, 1998; Scholin et al., 2000). In addition, HABs can cause a variety of risks to human health, ranging from gastrointestinal disorders, through the consumption of contaminated seafood, to skin irritations, due to direct contact with the algal in the water (e.g. at the beach). Moreover, certain HABs cause important constraints on marine tourism and recreation due to the production of thick foams with repellent odors and repulsive coloration of the beach water. Therefore, port authorities have recently implemented monitoring water quality programs and proposed the introduction of standards on ballast water treatment. Both types of measures are associated with financial costs. This raises the question which benefits are associated with the implementation of such policies. This monetary information is crucial for both resource damage assessment and for benefit-cost analysis. The latter creates an important basis for policy guidance since it consents the ranking of alternative marine policy options (for a list of these, see van den Bergh et al., 2002).

12.2.2. HABs as a source of economic damage In a conceptual framework, one can define the total value of the damages caused by HABs in terms of use and nonuse values. Table 12.1 offers a detailed classification of the damages caused by HABs. For example, direct use values of damages caused by HABs include the loss of marine tourism and coastal recreation benefits; the loss of natural and cultured marine species with commercial value; and risks to human health. On the contrary, indirect use values of damages caused by HABs relate to damages to the functioning of the marine ecosystem and marine living resources, even if these have no direct commercial value. Table 12.1 also shows that the monetary assessment of the damages of HABs requires the application of specific monetary valuation tools, which depends on the type of damage that is under consideration. As one can see, CV can assess the marine ecosystem damages caused by HABs. In addition, TC method is the most suitable valuation method for monetary value assessment of damages caused by HABs that relate to beach recreation. In the present study, both CV and TC methods are used. These valuation techniques have not yet been applied to value HABs damages.

12.3. THE SURVEY INSTRUMENT 12.3.1. Structure of the questionnaire The survey questionnaire consists of two parts. The first is designed so as to gather information on TCs, travel time and on-site expenditures. This information is used to shed light on the recreational use benefits. The second part includes the CV exercise. This is designed to shed light on marine ecosystem non-market benefits. The welfare measure adopted reflects the consumer’s maximum willingness to pay (WTP). This implies that

12. THE COST OF EXOTIC MARINE SPECIES

231

Table 12.1. Classification of economic damages caused by HABs

Value component

Example of damages

Use value (UV)

Loss of tourism and recreational benefits, e.g. visits to the beach, swimming, and sailing. Effects on marine resources with commercial value, e.g. destruction fish, shellfish, and mollusk Effects on human health, e.g. skin allergies and gastrointestinal disorders Effects on marine ecological system, e.g. unbalancing local chemical composition of the water, loss of local marine living resources diversity Risk of loss of legacy benefits, e.g. no legacy of marine living resources for future generations Risk of loss of existence benefits, e.g. no knowledge guarantee that some marine living resources are locally extinct

Direct use value (DUV)

Indirect use value (IUV)

Nonuse value (NUV)

Bequest value (BV) Existence value (EV)

∗ Market

Most suitable valuation technique Travel cost method Aggregate price analysis∗ Contingent valuation Contingent valuation

Contingent valuation Contingent valuation

price valuation technique.

the present distribution of property rights is respected, so that individuals have to pay for the ballast water program. Moreover, such a choice has the advantage that it forces respondents to look forward, which closely mimics a market situation (NOAA, 1993). As respondents are expected to be unfamiliar with the proposed CV market, the ballast water program is carefully described. This covers: (a) the construction of a ballast water disposal treatment complex in the Rotterdam harbor, an internal circuit where ballast water will be transferred to and submitted to an appropriate physicalchemical treatment, which eliminates all the HABs; and (b) the implementation of an algal monitoring program of the water quality in the open North Holland sea. In selecting a payment mechanism, we closely followed the NOAA guidelines to convince respondents that the payment mechanism is appropriate to address the problem and reflects a fair method of payment. Furthermore, respondents are informed that the current budgets of shipping companies are insufficient to allow the implementation of marine prevention and monitoring schemes. In the view of these conditions, we adopted a national tax scheme, rejecting alternatives, such as beach entrance fees. Before the execution of survey a first draft of the questionnaire was tested in a number of pilot interviews. We opted for an in-person survey because it generally leads to the highest survey response, as well as allows the use of the double bounded dichotomous WTP elicitation question format.

232

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

Table 12.2. Opinion of the respondents with respect to alternative marine management policies Ranking N. 1 N. 2 N. 3 N. 4 N. 5 N. 6

Index* Ensure that the beach is clean from rubbish such as cans and cigarettes. Ensure that the marine ecosystem is not threatened by water pollution. Ensure that the quality of the seawater is such that it does not provoke skin allergies. Ensure that the dunes continue to be recreational areas and not for housing. Ensure that some dunes are protected as areas for nature and closed to the public. Ensure that the sand along the waterfront of the beach is clean from algal and foams.

18.6 18.4 17.6 17.5 15.2 12.5

∗ The

ranking index was computed by weighting response percentages to each possible response, assigning a weight of 20 to “very important”; 15 to “important”; 10 to “somewhat important;” and finally, 0 to “of little importance.”

12.3.2. Descriptive statistics The administration of the questionnaire took place during August 2001 in Zandvoort. This is the most popular beach resort in the Netherlands and it is located in the province of North Holland. Zandvoort offers beach recreation activities, ranging from swimming, biking, kiting, windsurfing to sailing. Using a simple random sample selection mechanism, the interviewer contacted 352 groups of visitors, 242 of which completed the questionnaire. The participation rate is therefore 69%. Unlike many other survey applications the most often mentioned reason for refusal is “not speaking Dutch.” Tourists from abroad accounted for one third of all refusals.2 The questionnaire’s demographics and socio-economic characteristics indicate that the median respondent is 41 years old and has a household monthly net income in the range of 1,800–2,300˚. The majority of the respondents were women. When confronting the data of our survey with sociodemographic statistics for the Netherlands we are unable to find major differences. In other words, the sample is rather representative for the Netherlands. Survey responses on the travel data show that more than 50% of the respondents took a car to travel to Zandvoort, 20% arrived by public transport (i.e. train), 10.7% came by bicycle, and 15.6% of the respondent arrived to the beach on foot. We used public domain software of the Dutch Railways to computed travel time for public transport. The sample results show that beach visitors that traveled by train spent on average two hours and twenty-five minutes on the two-way journey.3 Each respondent traveling by car was asked for the postal zip code of their address, the brand and model 2

3

The relative high overall nonresponse rate needs to be examined carefully. In fact, from the policy perspective, the relevant population is the domestic one, i.e. the ones who are in conditions to be taxed by the Dutch national government. Having said that, the nonresponse ranges the nineteen%. The train station of Zandvoort is less than five minutes walking from the beach. As the origin we used the closest train-station to the zip code reported by the respondent. For respondents living far from a railway station we included an extra twenty minutes for transport from home to the railway station.

12. THE COST OF EXOTIC MARINE SPECIES

233

of their car, the size of the engine, the type of fuel used, and the parking fees at the beach. These respondents traveled on average 104 kilometers and spent on average one hour and twenty-three minutes on the two-way journey. On the basis of this information we were able to compute individual car TCs. Finally, respondents were asked to state their opinion about alternative marine management policies. As we can see from Table 12.2 coastal management activities related to “protecting the marine ecosystem from water pollution” and “ensuring seawater quality that does not provoke skin allergies” are associated to a high ranking index. Such a ranking pattern confirms the relevance of the present valuation study of damages caused by HABs’.

12.4. ESTIMATION OF NON-MARKET BENEFITS OF MARINE RECREATION 12.4.1. Travel cost method: transportation costs and monetized travel time The generalized TCs of a visit are defined as the sum of two components: transportation costs and travel time. The transportation cost is calculated as a function of the respondent’s means of transport; in this case (i) car, (ii) train, (iii) bicycle, and (iv) walking. For the latter two categories the transportation costs are assumed to be zero. The train costs are estimated for a two-way train ticket. For respondents traveling by car, the transportation cost is calculated according to the fuel used per kilometer. The time cost is estimated by multiplying the amount of time that a respondent spends on the two-way trip by the value of time. As we can see from Table 12.3, the value of time varies according to respondent’s monthly income and selected means of transport. The different values reflect differences in consumers’ preferences with respect to the choice of the means of travel, incorporating expected congestion time for car users.

12.4.2. Recreation demand function In addition to the individual TC data, Xt , socio-economic-demographic variables, Xs , and site attributes, Xa , are included in the recreation demand function: V = f (Xt , Xs , Xa ; β, ) Here V denotes the number of visits to the Zandvoort beaches during the past twelve months, β the vector of parameters to be estimated, and  the error-term vector. In order to achieve robust, non-biased estimates of β the econometric model specification Table 12.3. Value of time by income and travel mode (˚ ) Travel mode

Car Train

Monthly income categories 3400

4.236 3.645

5.090 4.200

6.268 4.718

11.541 7.627

Source: Gunn et al., 1998.

234

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

and estimation method need to respect the intrinsic characteristics of the TC data. In particular, it needs to take into account the fact that the number of visits is nonnegative, i.e. the dependent variable is truncated at zero (Hellerstein, 1992). An additional characteristic of the TC data is that since it draws from on-site surveys, meaning that frequent visitors are more likely to be interviewed. As a consequence, the econometric model specification and estimation method needs to be corrected for self-selection – see Heckman (1979) and Maddala (1983) for further discussion. This gives rise to the following model specification:    β0′ + βs′ xs + βa′ xa +ε j s a Prob (V = j) = F p ( j) = e(−λ)λ j! with λ = e Here j denotes the possible values for the annual visits to the Zandvoort beaches ( j = 1, 2 . . .), F p (·) the cumulative distribution function of the standard Poisson probability model, and λ (non-negative) Poisson parameter to be estimated (see Greene, 2000). β i are the parameters to be estimated, with β s and β a denoting the parameters of the socio-economic-demographic characteristics of the respondents and the attributes of the Zandvoort beaches, respectively. ε denotes the error term. Furthermore, in order to test for the presence of overdispersion a Binominal Negative model specification was estimated as well. This model specification gave similar estimation results and the estimate of the overdispersion parameter revealed not to be statistically significant from zero. Estimation results for the visitation to the Zandvoort beaches are presented in Table 12.4. This shows the partial derivatives with respect to the vector of explanatory Table 12.4. Estimation results with respect to visitation behavior Variable Intercept Travel costs Transport Travel time Parking Site characteristics Bloemendaal beach Sunny weather Week-end Personal characteristics Male Age Net income (by the hundreds of ˚ ) Field of studies: Economics Living with a partner Stay at the beach all day Log-likelihood Annual gross recreation benefit per individual

Estimate

p-value

4.8966*

0.00

−0.0139* −0.0101 −0.0015

0.04 0.34 0.91

0.8720** 0.0724 0.4847

0.11 0.85 0.25

−0.3192 −0.0118 0.0082** −0.3865 −0.5590 0.7079** −280.65 55˚

0.37 0.42 0.06 0.28 0.20 0.09

Marginal effects model. Full information maximum likelihood estimator with correction for sample selection is estimated with LIMDEP. ∗ Statistically significant at 5%. ∗∗ Statistically significant at 15%.

12. THE COST OF EXOTIC MARINE SPECIES

235

variables, i.e. the marginal effects. By adopting this specification we are able to provide key elasticities, which are more telling for policy purposes. In this context, we can see that all the three variables that represent the generalized TC have a negative sign, reflecting theoretical expectations. Second, transport expenditure estimates are statistically significant different from zero. Third, transport expenditure parameter magnitude when compared to the remaining TC variables reveals to be highest estimate. This shows that the price elasticity of demand of trips, measured in terms of transport expenditures, is particularly relevant in explaining consumer recreational behavior, i.e. in determining the number of annual visits to the beach. Such information reveals to be of crucial information for policy guidance since any measure that targets directly the transport expenditures, including taxes on fuel, will influence the annual trips to Zandvoort. Fourth, any policy option that is characterized by changing the pricing rates of the parking space available at the beach will not change consumer recreational behavior. In addition, Table 12.4 shows that an increase in the travel time would reduce the number of visitors to the beach of Zandvoort. However, the respective price elasticity is not statistical significant. Therefore, one can interpret this result as signal that the value of travel time savings is not particularly relevant to explain visitation behavior. Table 12.4 also shows that the demand of visits to the beach reveals not to be particularly sensitive to income, although the income elasticity reveals to be statistically different from zero. A 1% increase of the net household monthly income corresponds to an increase of 0.82% in the total annual visits. With regard to personal characteristics, the estimation results show that the number of visits is expected to be lower for male, older respondents living with a partner than for other respondents. Furthermore, visitors who plan to stay all day long at the beach have a higher annual visit frequency than respondents who visit the beach for half-a-day or a couple of hours. Finally, estimation results indicate that respondents who visit the beach of Bloemendaal, which corresponds to the location of the Kennemer dunes natural park, have a higher annual visit frequency than respondents that visit the beaches of Zandvoort.

12.4.3. Assessment of the individual recreation benefits The assessment of the individual recreation benefits is performed by deriving a standard Marshallian demand curve for yearly visits to the Zandvoort beach area. We evaluate all the explanatory variables of the demand function at their sample mean, with the exception of the individual TCs variable – see Loomis and. Walsh (1997) for further discussion. As mentioned, the generalized TCs variable is defined as the sum of transport, travel time and parking costs, which is denoted by P in the following reduced form of the demand curve,4 P = 67.7153 − 41.6777 × LogN where P denotes the generalized TC and N the yearly number of trips. The underlying consumer surplus of recreation corresponds to area underneath the resulting demand curve. By integrating this equation we obtain a gross recreational benefits per individual equal to 55˚ per year. In other words, we can predict that if the beach area of Zandvoort 4

See Nunes and van den Bergh (2002) for additional econometric details.

236

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

is closed to visitors for an entire year the total recreational welfare loss would equal, on average, 55˚ per individual.

12.5. ESTIMATION OF MARINE-ECOSYSTEM BENEFITS USING CV 12.5.1. The “no-no”-zero willingness to pay responses To better mimic market behavior, the CV uses a dichotomous-choice elicitation question format (Cameron and James, 1987). Following Hanemann et al. (1991), a follow-up valuation question was included so as to improve the statistical efficiency of the WTP estimates. All respondents that answered “no-no” faced a follow-up, open-ended WTP question. If the response to this was a zero willingness-to-pay, then the respondent is asked to indicate her major motivation for this choice. The most important motivation for not willing to pay anything is, by far with 74.5%, that the financial costs should be entirely met by shipping companies. The second most important motivation is that the government should cover the costs. These two arguments – as opposed to the remaining ones – reflect protest responses, i.e. do not reflect a zero valuation of the program that rather reflect a disapproval of the proposed payment mechanism. For this same reason these respondents will not be handled in the valuation exercise, i.e. they are excluded from the CV – see Strazzera et al. (2003a) and Strazzera et al. (2003b) for further discussion. Finally, some respondents simply did not believe that the program would work. The latter is supported by explicit statements by respondents like: “Do not believe in the ship monitoring activities,” “Do not believe the risk will go down with 90%,” or “The program also needs to be planned for Amsterdam and Antwerp harbors,” or “The program needs to be to be arranged on a European level, otherwise it will not work.”

12.5.2. WTP estimation results According to the double bounded response model, for each respondent j four possible response outcomes are possible: “no/no,” “no/yes,” “yes/no,” and “yes/yes,” coded as j j j j rnn , rny , r yn , and r yy binary indicators variables, respectively. As one would expect from economic theory, the proportion of “yes-yes” responses falls sharply with the amount the respondent is asked to pay – see Table 12.5. Only 3.8% of respondents Table 12.5. Bid cards Monetary amounts∗ Bid card N. 1 N. 2 N. 3 N. 4 ∗ Originally

Initial 6.5˚ 14˚ 20˚ 40˚

Distribution of the WTP responses (%) High 20˚ 34˚ 52˚ 123˚

formulated in Dutch Guilders.

Low 2.5˚ 7˚ 11˚ 16˚

Yes-yes 15.5 11.7 10.5 3.8

Yes-no 4.6 6.7 10.9 9.2

No-yes 0.4 0.8 0.8 5.9

No-no 4.2 5.9 2.5 6.7

12. THE COST OF EXOTIC MARINE SPECIES

237

Table 12.6. Lognormal mean and median WTP estimates

Location (β) Scale (σ ) Log-likelihood

Mean Median

Estimate

Standard error

4.6221 0.9997 −205.430

0.08856 0.08081

Point estimate

90% Confidence interval estimate

76.2˚ 46.2˚

[58.2–101.5˚ ] [39.9–53.5˚ ]

R . Calculations are performed using the PROC LIFEREG procedure in SAS

states a WTP above 123˚ . Such a low sample proportion at the highest bid indicates that the bid card has captured the range of the WTP distribution well. In addition, the proportion of “no-no” responses increases the bid amount. The remaining answering patterns, “yes-no” (“no-yes”) responses, indicate that the respondent’s maximum WTP lies between the initial bid amount and the increased (decreased) bid amounts. Bearing in mind the four possible response outcomes, the sum of contributions to the likelihood function L(θ) over the sample is maximized (Cameron and Quiggin, 1994): L (θ ) =

N   j  j    j   j  j rnn ln F bl ; θ + rny ln F bi ; θ − F bl ; θ j=1

  j   j    j 

j j + r yn ln F bh ; θ − F bi ; θ + r yy ln 1 − F bh ; θ

where F(·) is a statistical distribution function with parameter vector θ ≡ (β, σ ), where β and σ denote the location and scale parameters of the distribution. The ML numerical estimator for the double-bounded model, θˆ , is the solution to the equation ˆ ∂ L(θ)/∂θ = 0. For a univariate model with a lognormal distribution, the mean WTP is 2 given by WTP = eβ+0.5×σ , where β and σ represent the location and scale parameters (Johnson and Kotz, 1970). As we can see in Table 12.6, the β and σ standard errors indicate that the parameters are estimated precisely. The annual mean WTP estimate is 76.2˚. In order to improve the statistical quality of the WTP estimates we explore the use of introducing information into the double bounded dichotomous choice (DBDC) model. Two types of information are considered. On one hand, we consider the use a follow-up open-ended (OE) elicitation question. For example, all the respondents who answered “yes” to the DBDC question erer asked “What is the most that you are willing to pay per year, for the next two years?” The answers to this mixed model provide more accurate information, allowing a more efficient WTP estimation. In order to control for any concerns from introducing the OE question after the DBDC question, we consider introducing a second type of information. It refers the use of the individual income information. This assumption, which is frequently made by economists, states that reported WTP is income constrained. In particular, we make use of the mixed model in which WTP responses are truncated whenever it exceeds a fixed proportion of the net monthly income the respondent. Table 12.7 offers the respective WTP estimates.

238

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS Table 12.7. WTP estimates for the Mixed Lognormal model

Mixed model Mixed model and WTP truncation at 5% Mixed model and WTP truncation at 2.5%

Mean

90% CI

Median

90% CI

50.1˚ 45.4˚ 38.8˚

[44.7–58.8˚ ] [41.0–50.5˚ ] [35.5–43.6˚ ]

43.6˚ 40.6˚ 36.0˚

[40.1–47.4˚ ] [37.6–43.9˚ ] [32.4–38.9˚ ]

A comparison of the original DBCD WTP estimates with the mixed model estimates shows two important differences. First, when we consider the income information we get more conservative estimates: the mean and median WTP point estimates generated by the mixed model are lower than the point mean estimate resulting from the DBDC model. Second, when we consider the income information the degree of precision of the WTP estimates is higher than in the original DBDC model. As a consequence, the mixed model estimates’ 90% confidence intervals are tighter. In addition, the income constraint on the individual WTP responses provides another piece of information, i.e. the direction of the weak monotonicity (see, Carson et al. 1997, 1999; for additional details). The WTP estimator that incorporates this restriction is the Turnbull estimator (Turnbull, 1976). According to this non-parametric estimator the lower bound mean WTP amounts to 39.4˚ . Such an estimate does not only respect the weak monotonicity restriction but also lies within the interval range of the parametric mixed model with truncation at 2.5%, and thus reiterates the validity of the proposed mixed model estimation procedure.

12.5.3. WTP valuation function A large number of possible predictors are available to be integrated in the valuation function. Formal testing procedures, based on the used of the log-likelihood test statistic, show that the model specification as presented in Table 12.8 fits the data the best. The estimation results show that: r Respondents who visit the Zandvoort beaches two to three times per year – which corresponds to the second lowest visiting frequency presented – and plan to stay at the beach the whole day are willing to pay more than the average respondent. r Respondents who visited the beach during the weekends and choose to stay at the Bloemendaal beach are willing to pay more than the average respondent. r Respondents who visit the site during the winter are willing to pay, on average, less for the described protection program. This may be due to the fact that the marine biological pollution, as described in the questionnaire, is less likely in the winter than in the summer, because of the lower temperature of the water. r Respondents who ranked the protection of coastal reserve areas “the most important” priority for beach management are willing to pay more for the marine protection program than the average respondent. r Respondents who spent a longer time traveling or who incurred higher parking costs have lower WTP. This suggests that the values obtained with the TC and CV methods are largely complementary. In other words, the TC exercise has captured other value categories than the CV exercise.

12. THE COST OF EXOTIC MARINE SPECIES

239

Table 12.8. WTP function Beach attributes

Estimate

Visited area: Bloemendaal Day of the visit: week-end Recreational profile of the visitor Number of adults in the group Presence of a child in the group Two or three visits per year Visits in the winter season Time planned to stay on the beach: all day Marine management policy options Nature reserves closed to the public Infrastructure support Socio-economic characteristics Age Income University degree Education in economics or business management Travel costs and on-site expenditures Transport Travel time Parking fee Expenditures on beach materials Expenditures on food and drinks at the beach-house Lognormal parameters Intercept Scale Log-likelihood

Std. error

p-value

33 24

0.53 0.08

0.093 0.086 0.633* −0.534* 0.427**

0.09 0.30 0.31 0.23 0.24

0.30 0.77 0.04 0.02 0.08

0.271** 0.358*

0.16 0.13

0.09 0.01

−0.005 0.008 0.169 −0.288**

0.08 0.01 0.24 0.22

0.52 0.61 0.48 0.20

0.003 −0.014** −0.056* 0.021** 0.005**

0.00 0.01 0.02 0.01 0.00

0.55 0.20 0.01 0.06 0.18

0.206 0.418**

3.323 0.870 –116.40

R . Calculations are performed using the PROC LIFEREG procedure in SAS  *Significant at 5%, **Significant at 10%.

These results show that preferences for a marine protection program are not independent of the respondent “profile.” The consequence of this fact, from the policy design perspective, is that the marine protection program is not equally desired among the Dutch population. For example, the stronger supporters of the program are the taxpayers who live nearby and visit the beach on regular basis (e.g. on the weekends). On the contrary, the taxpayers who incur in relatively TCs, e.g. who spent a longer time traveling or who incurred higher parking, are willing to pay less for the ballast water treatment program.

12.6. CONCLUSION This chapter has offered an economic value assessment of the non-market benefits of a marine protection program. This program focuses on the prevention of HABs. This invasive exotic species are responsible for causing important damages to the marine ecosystem, as well as beach recreation. The valuation is based on a questionnaire

240

PART III: MONETARY VALUATION AND STAKEHOLDER ANALYSIS

undertaken at Zandvoort, a famous beach resort situated in the North Holland coastline. According to the TC model estimates, if the beach area of Zandvoort is closed to visitors for an entire year the total recreational welfare loss equals 55˚ per individual per year. In addition, the time component of the TC reveals not to be particularly influent in explaining individual visitation behavior even if the region of the study in the Netherlands is affected by traffic congestion. The CV estimates indicate that the annual WTP amounts to 76˚ per respondent, per year. The comparison of the TC and CV estimates indicate the importance of marine ecosystem non-market benefits, even if we admit the presence of double counting in the CV estimates. However, we are rather confident that CV estimates mainly refer to welfare impacts that are not directly with recreational use since the marine protection program, as accurately described in the CV survey, strictly conveys marine ecosystem non-market impacts caused by HABs. Furthermore, WTP multivariate regression estimates reiterate the achievement of the CV survey in identifying and measuring consumer preferences with respect to marine ecosystem benefits and nonuse impacts caused by HABs. The combination of TC and CV estimation results shows that respondents with relatively high TCs have a relatively low WTP for the marine protection program. This result suggests that some budget constraint is active, and that TC and CV estimates are, to some extent, complementary in terms of achieving a complete picture of the overall monetary value. Finally, TC and CV results show that if no policy action is undertaken so as to prevent a HABs marine pollution event in the coast of the Netherlands a significant welfare loss may result. An estimate of the total welfare loss amounts to 326,190,000˚ , which corresponds to 0.08% of the Dutch GDP measured at market prices for the year 2000. This value is obtained by: (1) assuming that the respondents that participated in the survey are representative for the entire visitor population of the beaches along the North Holland coastline; and (2) multiplying the sum of the use recreational benefits, which is derived from the TC model and amounts to 55˚ , with the marine ecosystem benefits, which is assessed with the CV exercise and amount to 76˚ , by the total number of visitors to the North Holland coastline, estimated at 2,400,000 per year (CBS, 2000). In addition, if we assume that the respondents that refused to participate in the survey have a zero WTP, then the total welfare loss amounts to 225,071,100˚. The two estimates can be interpreted upper and lower bounds to the financial cost of implementation of an efficient marine protection program, respectively. In other words, the marine protection program can be defended from a cost–benefit perspective if its cost is in the range 225–326 million euro. For example, if we assume that the costs of cleaning sea water and coastal beaches have a similar magnitude to the clear up program of polluted water soils in the Netherlands, which annual costs amount to 0.03% over the period 1994–1998 (CBS, 1999), then we can conclude that the benefits from such a marine protection program far exceeds the costs and thus recommended from a cost benefit perspective.

REFERENCES

Adamowicz, W.L., P.C. Boxall, M. Williams, and J.J. Louviere (1998). Stated preference approaches for measuring passive use values: Choice experiments and contingent valuation. American Journal of Agricultural Economics 80(1):64–75. Agardi, T. (2000). Effects of fisheries on marine ecosystems: A conservationists perspective. ICES Journal of Marine Science 57:765–765. Aggeri, F. (1999). Environmental policies and innovation. A knowledge-based perspective on cooperative approaches. Research Policy 28:699–717. Alexander, S.E., and J. Roughgarden (1996). Larval transport and population dynamics of intertidal barnacles: A coupled benthic/oceanic model. Ecological Monographs 66(3):259–275. Allen, P.M., and J.M. McGlade (1987). Modelling complex human systems: A fisheries example. European Journal of Operational Research 30:147–167. Allison, G.W., J. Lubchenco, and M.H. Carr (1998). Marine reserves are necessary but not sufficient for marine conservation. Ecological Applications 8:S79–S92. Alonso, D., and R.V. Sol´e (2000). The DivGame simulator: A stochastic cellular automata model of rainforest dynamics. Ecological Modelling 133:131–141. Amundsen, E.S., and T. Bjorndal (1999). Optimal exploitation of a biomass confronted with the threat of collapse. Land Economics 75:185–202. Anderson, D.M. (1994). Red tides. Scientific American 271:52–58. Asche, F. (2001). Fishermens discount rates in ITQ systems. Environmental and Resource Economics 19:403– 410. Atkinson, P.W., N.A. Clark, M.C. Bell, P.J. Dare, J.A. Clark, and P.L. Ireland (2003). Changes in commercially fished shellfish stocks and shorebird populations in the Wash, England. Biological Conservation 114:127–141. Bailey, P.D. (1997). IEA: A new methodology for environmental policy. Environmental Impact Assessment Review 17:221–226. Ballantine, B. (1997). No-take marine reserve networks support fisheries. Developing and sustaining world fisheries resources. Proceedings of the 2nd World Fisheries, Brisbane, Australia, pp. 702–706. Baltz, D.M. (1998). Food, density, and microhabitat: Factors affecting growth and recruitment potential of juvenile saltmarsh fishes. Environmental Biology of Fishes 53:89–103. Banks, R. (1999). Subsidizing EU fleets: Capacity reduction or capital subsidization. In: A. Hatcher and K. Robinson (eds.), Overcapacity, Overcapitalisation and Subsidies in European Fisheries. Proceedings of the first workshop held in Portsmouth, CEMARE, 28–30 October, UK. Bascompte, J., and R.V. Sol´e (1995). Rethinking complexity: Modelling spatiotemporal dynamics in ecology. Trends in Ecology and Evolution 10:361–366. Beddington, J.R. (1975). Mutual interference between parasites and predators and its effect on searching efficiency. Journal of Animal Ecology 44:331–340. Bell, M.C., R.C.A. Bannister, P. Walker, P.W. Atkinson, and N.A. Clark (2001). Fisheries and bird predation as components of cockle (Cerastoderma edule) mortality in the Burry inlet, South Wales. ICES Journal CM 2001/J:02. Benhabib, J., and K. Nishimura (1979). The Hopf bifurcation and the existence and stability of closed orbits in multisector models of optimal economic growth. Journal of Economic Theory 21:421–444.

241

242

REFERENCES

Beres, D.L., and D.M. Hawkins (2001). Plackett–Burman technique for sensitivity analysis of many parameter models. Ecological Modelling 141:171–183. Berk, V.M. van de, S. Dirksen, and M.J.M. Poot (2000). Sterfte onder eidereenden in de Waddenzee 1999– 2000, Een zoektocht naar de oorzaak van de massale sterfte van eidereenden in de Waddenzee, Werkdocument EC-LNV nr. 186, Wageningen. Berryman, A.A. (1992). The origins and evolution of predator–prey theory. Ecology 73:1530–1535. Beukema, J.J. (1982). Annual variation in reproductive success and biomass of the major macrozoobenthic species living in a tidal flat area of the Wadden Sea. Netherlands Journal of Sea Research 16:37–45. Beukema, J.J. (1993). Annual variation in reproductive success and biomass of the major macrozoobenthic species living in a tidal flat area of the Wadden Sea. Netherlands Journal of Sea Research 31:395–406. Beukema, J.J., and G.C. Cad´ee (1996). Consequences of the sudden removal of nearly all mussels and cockels from the Dutch Wadden Sea. Marine Ecology 17:279–289. Beukema, J.J., and R. Dekker (2005). Decline of recruitment success of cockles and other bivalves in the Dutch Wadden Sea: Possible role of climate change, predation on postlarvae and fisheries. Marine Ecology Progress Series 287:149–167. Beverton, R.J., and S.J. Holt (1957). On the Dynamics of Exploited Fish Populations. Chapman and Hall, London (also Fisheries Investigation Series 2(19). Ministry of Agriculture, Fisheries and Food, London). Birkeland, C. (2004). Ratcheting down the coral reefs. Bioscience 54(11):1021–1027. Bjorndal, T. (1988). The optimal management of North Sea Herring. Journal of Environmental Economics and Management 15:9–29. Blaine, T.W., and D.L. DeAngelis (1997). The interaction of spatial scale and predator–prey functional response. Ecological Modelling 95:319–328. Blamey, R.K., and M.S. Common (1999). Valuation and ethics in environmental economics. In: J.C.J.M. van den Bergh (ed.), Handbook of Environmental and Resource Economics. Edward Elgar, Cheltenham. Boalch, G.T. (1987). Changes in the phytoplankton of the western English Channel in recent years. British Phycological Journal 22:225–235. Boersma, P.D., and J.K. Parrish (1999). Limiting abuse: Marine protected areas, a limited solution. Ecological Economics 31:287–304. Bohnsack, J.A. (1993). Marine reserves. They enhance fisheries, reduce conflicts, and protect resources. Oceanus 36:63–71. Box, G.E.P., W.G. Hunter, and J.S. Hunter (1978). Statistics for Experimenters: An Introduction to Design, Data, Analysis and Model Building. Wiley, New York. Boyce, M.S. (1992). Population viability analysis. Annual Review of Ecology and Systematics 23:481–506. Brauer, F., and A.C. Soudack (1979). Stability regions and transition phenomena for harvested predator–prey systems. Journal of Mathematical Biology 7:319–337. Brock, W.A., and J.A. Scheinkman (1977). The global asymptotic stability of optimal control with applications to dynamic economic systems. In: J.D. Pitchford and S.J. Turnovsky (eds.), Applications of Control Theory to Economic Analysis. North-Holland Publishing Company, Amsterdam. Brown, G. (1974). An optimal program for managing common property resources with congestion externalities. The Journal of Political Economy 82:163–173. Brown G., and J. Roughgarden (1997). A metapopulation model with private property and a common pool. Ecological Economics 22:65–71. Buhl-Mortensen, L., and R. Toresen (2001). Fisheries management in a sea of uncertainty: The role and responsibility of scientists in attaining a precautionary approach. International Journal of Sustainable Development 4(3):245–264. Bulte, E. (1997). Essays in Economics of Renewable Resources. Ph.D. Dissertation, Wageningen University, Wageningen. Bulte, E.H., H. Folmer, and W. Heijman (1998). Dynamic and static approaches to mixed good management: The case of mink whales in the Northeast Atlantic. European Review of Agricultural Economics 25(1):73–92. Bulte, E.H., and G.C. van Kooten (1999). Metapopulation dynamics and stochastic bioeconomic modelling. Ecological Economics 30:293–299. Bulte, E.H., and G.C. van Kooten (2001). Harvesting and conserving a species when numbers are low: Population viability and gamblers ruin in bioeconomic models. Ecological Economics 37:87–100.

REFERENCES

243

Cabugueira, M.F.M. (2001). Voluntary agreements as an environmental policy instrument – Evaluation criteria. Journal of Cleaner Production 9:121–133. Caddy, J.F., and K.L. Cochrane (2001). A review of fisheries management past and present and some future perspectives for the third millennium. Ocean and Coastal Management 44:653–682. Caddy, J.F., and R. Mahon (1995). Reference points for fisheries management. FAO Fisheries Technical Paper No. 347, FAO, Rome. Cameron, T.A., and M.D. James (1987). Efficient estimation methods for closed-ended contingent valuation surveys. The Review of Economics and Statistic 69:269–276. Cameron, T.A., and J. Quiggin (1994). Estimation using contingent valuation data from a dichotomous choice with follow-up questionnaire. Journal of Environmental Economics and Management 27:218–234. Camphuysen, C.J., C.M. Berrevoets, H.J.W.M. Cremers, A. Dekinga, R. Dekker, B.J. Ens, T.M. van der Have, R.K.H. Katz, T. Kuiken, M.F. Leopold, J. van der Meer, and T. Piersma (2002). Mass mortality of Common Eiders (Somateria mollisima) in the Dutch Wadden Sea, winter 1999/2000: Starvation in a commercially exploited wetland of international importance. Biological Conservation 106:303–317. Camphuysen, C.J., B.J. Ens, D. Heg, J.B. Hulscher, J. van der Meer, and C.J. Smit (1996). Oystercatcher Haematopus ostralegus winter mortality in The Netherlands: The effect of severe weather and food supply. Ardea 84A:469–492. Carpenter, S., W. Brock, and P. Hanson (1999). Ecological and social dynamics in simple models of ecosystem management. Conservation Ecology 3:1–30. Carson, R.T., J. Yongil, and D.R. McCubbin (1997). The Relationship Between Air Pollution Emissions and Income: U.S. Data. Environment and Development Economics, 2: 433–450. CBS. (1999). Environmental Compendium 1999 Netherlands. Centraal Bureau voor de Statistiek (Statistics Netherlands), Voorburg/Bilthoven, The Netherlands. CBS. (2003). Tourisme en Recreatie in Cijfers 2003. Centraal Bureau voor de Statistiek (Statistics Netherlands), Voorburg/Heerlen. CBS. (2000). Statistical Year Book. Statistics Netherlands, Voorburg/Heerlen, The Netherlands. Cesar H.S.J. (ed.). (2000). Collected Essays on the Economics of Coral Reefs. CORDIO, Kalmar University, Kalmar, Sweden. Charnov, E.L. (1976). Optimal foraging: Attack strategy of a mantid. American Naturalist 110:141–151. Christensen, T.K., and H. Noer (2001). Monitoring of Breeding Eiders at Saltholm. NERI, Copenhagen. Clark, C.W. (1973). Profit maximization and the extinction of animal species. Journal of Political Economy 81:950–961. Clark, C.W. (1976). Mathematical Bioeconomics: The Optimal Management of Renewable Resources. WileyInterscience, New York. Clark, C.W. (1985). Bioeconomic Modelling and Fisheries Management. Wiley-Interscience, New York. Clark, C.W. (1990). Mathematical Bioeconomics: The Optimal Management of Renewable Resources, 2nd ed. Wiley-Interscience, New York (first edition 1976). Clark, C.W. (1996). Marine reserves and the precautionary management of fisheries. Ecological Applications 6:369–370. Clark, C.W. (1999). Renewable resources: Fisheries. In: J.C.J.M. van den Bergh (ed.), Handbook of Environmental and Resource Economics. Edward Elgar, Cheltenham. Clark, C.W., F.H. Clarke, and G.R. Munro (1979). The optimal exploitation of renewable resource stocks: Problems of irreversible investment. Econometrica 47:25–47. Commission of the European Communities. (2000). Report from the Commission to the Council: Preparation for a mid term review of the Multi-annual Guidance Programmes (MAGP). COM(2000) 272 final, Brussels. Commission of the European Communities. (2001a). Green paper on the future of the Common Fisheries Policy. COM(2001) 135, Brussels. Commission of the European Communities. (2001b). Elements of a strategy for the integration of environmental protection requirements into the Common Fisheries Policy, Communication from the Commission to the Council and the European Parliament. COM(2001) 143, Brussels. Conrad, J.M. (1999). The bioeconomics of marine sanctuaries. Journal of Bioeconomics 1:205–217. Conrad, J.M., and C.W. Clark (1987). Natural Resource Economics: Notes and Problems. Cambridge University Press, Cambridge.

244

REFERENCES

Copes, P. (1986). A critical review of the individual quota as a device in fisheries management. Land Economics 62:278–291. Costa, M.I.S., E. Kaszkurewicz, A. Bhaya, and L. Hsu (2000). Achieving global convergence to an equilibrium population in predator–prey systems by the use of a discontinuous harvesting policy. Ecological Modelling 128:89–99. Costanza, R., R. dArge, R. de Groot, S. Farbar, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R. O’Neill, R.V. Paruelo, R.G. Raskin, P. Sutton, and M. van der Belt. (1997). The value of the worlds ecosystem services and natural capital. Nature 387:253–260. Costanza, R., L. Wainger, C. Folke, and K.-G. Maler (1993). Modeling complex ecological economic systems: Towards an evolutionary, dynamic understanding of people and nature. Bioscience 43:545–555. Craeymeersch, J.A., M.F. Leopold, and M.O. van Wijk (2001). Halfgeknotte strandschelp en Amerikaanse zwaardschede: Een overzicht van bestaande kennis over visserij, economische betekenis, regelgeving, ecologie van de beviste soorten en effecten op het ecosysteem. RIVO Report C033/01, 1–34. RIVO, Yerseke. Cryer, S.A., and P.L. Havens (1999). Regional sensitivity analysis using a fractional factorial method for the USDA model GLEAMS. Environmental Modelling and Software 14:613–624. Daan, R. (1997). TAC Management in North Sea flatfish fisheries. Journal of Sea Research 37:321–341. Dai, G., and M. Tang (1998). Coexistence region and global dynamics of a harvested predator–prey system. Siam Journal of Applied Mathematics 58:193–210. de Jong, F., J. Bakker, C. van Berkel, K. Dahl, N. Dankers, C. Gatje, H. Marencic, and P. Potel (1999). Wadden Sea quality status report 1999. Quality Status Report Group, Wilhelmshaven. De Jonge, V.N., D.J. de Jong, and M.M. van Katwijk (2000). Policy plans and management measures to restore eelgrass (Zostera marina L.) in the Dutch Wadden Sea. Helgoland Marine Research 54:51–158. de Vlas, J. (1982). De effecten van kokkelvisserij op de bodemfauna van waddenzee en oosterschelde. RIN-Rapport 82/19, 2–99, Texel, RIN. de Vooys, C.G.N. (1999). Numbers of larvae and primary plantigrades of the mussel Mytilus edulis in the western Dutch Wadden Sea. Journal of Sea Research 41:189–201. Dijkema, R. (1997). Molluscan Fisheries and Culture in the Netherlands. Technical Report 129, U.S. Department of Commerce, NOAA, Washington D.C. pp. 115–135. Directie Visserij. (1999). Ministerie van Landbouw, Natuur en Visserij, Gemeenschappelijk Visserij beleid 2002. http: //www.minlnv.nl/infomart/parlemnt/1999/par99187.pdf. Dixit, A.K., and R.S. Pindyck (1994). Investment under Uncertainty. Princeton University Press, Princeton. Dockner, E.J., and G. Feichtinger (1991). On the optimality of limit cycles in dynamic economic systems. Journal of Economics 53:31–50. Dolmer, P.P., S. Kristensen, and E. Hoffman (1999). Dredging of Blue Mussels in a Danish Sound: Stock sizes and fishery-effects on mussel population dynamic. Fisheries Research 40:73–80. Dowlatabadi, H. (1995). Integrated assessment models of climate change. Energy Policy 23:289–296. Dubbink, W., and M. van Vliet (1996). Market regulation versus co-management? Two perspectives on regulating fisheries compared. Marine Policy 20:499–516. Duiker, J.M.C., E. Nomden, and A.C. Smaal (1998). De invloed van mechanische schelpdiervisserij op de sedimenthuishouding van de waddenzee, R 98-01, 2-46. Instituut voor Marien en Atmosferisch onderzoek, Utrecht. Eggert, H. (1998). Bioeconomic analysis and management: The case of fisheries. Environmental and Resource Economics 11:399–411. Eijgenraam, C.J.J., C.C. Koopmans, P.J.G. Tang, and A.C.P. Verster (2000). Evaluatie van Grote Infrastructuurprojecten. Leidraad voor Kosten-baten Analyse. Onderzoeksprogramma Economische Effecten Infrastructuur. Ministerie van Verkeer en Waterstaat and Ministerie van Economische Zaken, The Hague. Ellison, A.M. (1996). An introduction to Bayesian inference for ecological research and environmental decision making. Ecological Applications 6(4):1036–1046. Ens, B.J. (2000). Berekeningsmethodiek voedselreservering Wadden Zee. Alterra Rapport 136, Alterra, Wageningen. Ens, B.J. (2002). What we know and what we should know about mollusk fisheries and aquacultures in the Wadden Sea. In: W.J. Wolff, K. Essink, A. Kellerman, and M.A. van Leeuwe (eds.), Proceedings of the 10th International Scientific Wadden Sea Symposium. 2002. Den Haag, Ministerie van LNV.

REFERENCES

245

Ens, B.J. (2003). What we know and what we should know about mollusc fisheries and aquacultures in the Wadden Sea. In: W.J. Wolff, K. Essink, A. Kellerman, and M.A. van Leeuwe (eds.), Proceedings of the 10th International Scientific Wadden Sea Symposium, 2000. Ministry of Agriculture, Nature Management and Fisheries/University of Groningen, Department of Marine Biology, Groningen. Ens, B.J., F.H.M. Borgsteede, C.J. Camphuysen, G.M. Dorrestein, R.K.H. Kats, and M.F. Leopold (2002a). Eidereendensterfte in de winter 2001/2002. Alterra-report 521, Alterra, Wageningen. Ens, B.J., and J.T. Cayford (1996). Feeding with other oystercatchers. In: J.D. Goss-Custard (ed.), The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford, pp. 77–104. Ens, B.J., R. Lanters, and A. Smaal (2002b). EVA II: Evaluating the Dutch policy of shellfish fishing in 2003. Wadden Sea Newsletter 2:22–24. Ens, B.J., A.C. Smaal, and J. de Vlas (2004). The effects of shellfish fishery on the ecosystems of the Dutch Wadden Sea and Oosterschelde. Final report on the second phase of the scientific evaluation of the Dutch shellfish fishery policy (EVA II). Alterra-report 1011. RIVO-Report C056/04. RIKZ-Report RKZ/2004.031, Alterra, Wageningen. Ernst and Young Accountants. (2001). Adviesrapport inzake vrijwillige saneringsregeling mechanische kokkelvisserij in de Waddenzee. http: //www.minlnv.nl/infomart/parlemnt/2001/par01283a.pdf. EVA II. (2003). Resultaten Wetenschappelijk Onderzoek EVA II. Publieksversie. Ministerie van Landbouw, Natuur en Voedselkwaliteit, Den Haag. FAO. (1985). Adjustment to changes in fisheries law and economics. Fisheries Technical Paper (FAO) No. 269, Rome, Italy. FAO (1990). The State of World Fisheries and Aquaculture. FAO, Rome. Farrow, S. (1996). Marine protected areas: Emerging economics. Marine Policy 20:439–446. Finley, J.R., J.D. Pint´er, and M.G. Satish (1998). Automatic model calibration applying global optimization techniques. Environmental Modeling and Assessment 3:117–126. Fisher, R.A. (1973). Statistical Methods and Scientific Inference. Hafner, New York. Flaaten, O. (1988). The Economics of Multispecies Harvesting: Theory and Application to the Barents Sea Fisheries. Springer-Verlag, Berlin. Flaaten, O. (1991). Bioeconomics of sustainable harvest of competing species. Journal of Environmental Economics and Management 20:163–180. Fogarty, M.J., R. Hilborn, and D. Gunderson (1996b). Rejoinder: Chaos and parametric management. Marine Policy 21:187–194. Fogarty, M.J., R.K. Mayo, L. OBrien, F.M. Serchuk, and A.A. Rosenberg (1996a). Assessing uncertainty and risk in exploited marine populations. Reliability Engineering and System Safety 54:183–195. Foley, P. (1997). Extinction models for local populations. In: I. Hanski and M. Gilpin (eds.), Metapopulation Dynamics. Academic Press, London, pp. 215–246. Foley, P. (2000). Problems in extinction model selection and parameter estimation. Environmental Management 26:S55–S73. Folke, C., C.S. Holling, and C. Perrings (1996). Biological diversity, ecosystems, and the human scale. Ecological Applications 6(4):1018–1024. Freeman, A.M., III (1993). The Measurement of Environmental and Resource Values: Theory and Methods. Resources for the Future, Baltimore. Freeman, S.N., and G.P. Kirkwood (1995). On a structural time series method for estimating stock biomass and recruitment from catch and effort data. Fisheries Research 22:77–98. Garcia, S.M., and C. Newton (1997). Current situations, trends and prospects in world capture fisheries. Global Trends: Fisheries Management. American Fisheries Society Symposium No. 20, American Fisheries Society, Bethesda, Maryland. Gibson, C.C., E. Ostrom, and T.K. Ahn (2000). The concept of scale and the human dimension of global change: A survey. Ecological Economics 32:217–239. Gilks, W.R., S. Richardson, and D.J. Spiegelhalter (1996). Markov Chain Monte Carlo in Practice. Chapman and Hall, London. Gjøsaeter, H., and B. Bogstad (1998). Effects of the presence of herring (Clupea harengus) on the stock–recruitment relationship of Barents Sea capelin (Mallotus villosus). Fisheries Research 38:57– 71. Gordon, H.S. (1954). The economic theory of a common property resource: The fishery. Journal of Political Economy 62:124–142.

246

REFERENCES

Goss-Custard, J.D. (ed.). (1996). The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford. Goss-Custard, J.D., R.T. Clarke, K.B. Briggs, B.J. Ens, K.M. Exo, C.J. Smit, A.J. Beintema, R.W.G. Caldow, D.C. Catt, N.C. Clark, S.E.A.L.V.D. Durell, M.P. Harris, J.B. Hulscher, P.L. Meininger, N. Picozzi, R. Prys-Jones, U.N. Safriel, and A.D. West (1995a). Population consequences of winter habitat loss in a migratory shorebird: I. Estimating parameters. Journal of Applied Ecology 32:320– 336. Goss-Custard, J.D., R.T. Clarke, S.E.A.L.V.D. Durell, R.W.G. Caldow, and B.J. Ens (1995b). Population consequences of winter habitat loss in a migratory shorebird: II. Model predictions. Journal of Applied Ecology 32:337–351. Goss-Custard, J.D., and S.E.A.L.V.D. Durell (1988). The effect of dominance and feeding method on the intake rate of oystercatchers, Haematopus ostralegus, feeding on mussels. Journal of Animal Ecology 57:827–844. Goss-Custard, J.D., S.E.A.L.V.D. Durell, R.T. Clarke, A.J. Beintema, R.W.G. Caldow, P.L. Meininger, and C.J. Smit (1996). Population dynamics: Predicting the consequences of habitat changes at the continental scale. In: J.D. Goss-Custard (ed.), The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford, pp. 352–383. Goss-Custard, J.D., R.A. Stillmann, A.D. West, R.W.G. Caldow, J. Triplet, S.E.A.L.V.D. Durell, and S. McGrogerty (2004). When enough is not enough: Shorebirds and shellfishing. The Royal Society of London B-Biological Sciences 271:233–237. Goss-Custard, J.D., A.D. West, R.A. Stillmann, S.E.A.L.V.D. Durell, R.W.G. Caldow, S. McGrorty, and R. Nagarajan (2001). Density-dependent starvation in a vertebrate without significant depletion. Journal of Animal Ecology 70:955–965. Gough, C., N. Castells, and S. Functowicz (1997). Integrated assessment: An emerging methodology for complex issues. Environmental Modeling and Assessment 3:19–29. Gowdy, J. (1997). The value of biodiversity: Markets, society and ecosystems. Land Economics 73:25–41. Grafton, R.Q. (1994). A note on uncertainty and rent capture in an ITQ fishery. Journal of Environmental Economics and Management 27:286–294. Grafton, R.Q. (2000). Governing the commons: A role for the state? Land Economics 76:504–517. Greene, W.H. (2000). Econometric Analysis. MacMillan, New Jersey, USA. Gudmundsson, E., and J.L. Anderson (2000). Asset Pricing and Option Trading for Harvesting Rights to Renewable Resources. IIFET Conference 2000, Oregon. Gu´enette, S., T. Lauck, and C. Clark (1998). Marine reserves: From Beverton and Holt to the present. Reviews in Fish Biology and Fisheries 8:251–272. Gu´enette, S., and T.J. Pitcher (1999). An age-structured model showing the benefits of marine reserves in controlling overexploitation. Fisheries Research 39:295–303. Guichard, F., S.A. Levin, A. Hastings, and D. Siegel (2004). Toward a dynamic metacommunity approach to marine reserve theory. Bioscience 54(11):1003–1011. Gunderson, L.H., and C.S. Holling (eds.). (2002). Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington, D.C. Gunn, H.F., J.G. Tuinenga, Y.H.F. Cheung, and H.J. Kleijn (1998). Value of Dutch travel time savings in 1997. 8th WCTR Proceedings 3:513–526. Hallegraeff, G.M., and S. Fraga (1998). Bloom dynamics of the toxic dinoflagellate Gymnodinium catenatum, with emphasis on Tasmanian and Spanish coastal waters. In: D.M. Anderson, A.D. Cembella, and G.M. Hallegraeff (eds.), Physiological Ecology of Harmful Algal Blooms. Springer-Verlag, Berlin, pp. 59– 80. Hanemann, M. W., J. Loomis, and B. Kanninen (1991). Statistical efficiency of double-bounded dichotomous choice contingent valuation. American Journal of Agricultural Economics 73(4): 1255–1263. Hanemann, W.M. (1984). Welfare evaluations in contingent valuation experiments with discrete responses. American Journal of Agricultural Economics 66:332–341. Hanley, N., S. Mourato, and R.E. Wright (2001). Choice modelling approaches: A superior alternative for environmental valuation? Journal of Economic Surveys 15(3):435–462. Hanley, N., and C.L. Spash (1993). Cost-Benefit Analysis and the Environment. Edward Elgar, Aldershot. Hanna, S. (1997). The new frontier of American Fisheries governance. Ecological Economics 28:221– 233.

REFERENCES

247

Hanna, S., H. Blough, R. Allen, S. Iudicello, G. Matlock, and B. McCay (2000). Fishing Grounds: Defining a New Era for American Fisheries Management. Island Press, Washington, D.C. Hannesson, R. (1982). Optimal harvesting of ecologically interdependent fish species. Journal of Environmental Economics and Management 10:329–345. Hannesson, R., and I.S. Steinshamn (1991). How to set catch quotas: Constant effort or constant catch? Journal of Environmental Economics and Management 20:71–91. Harford, J.D. (2000). A reconsideration of the bioeconomics of marine sanctuaries: Commentary on Jon Conrads paper. Journal of Bioeconomics 2:133–138. Harley, S.J., R.A. Myers, and A. Dunn (2001). Is catch-per-unit effort proportional to abundance? Canadian Journal of Fisheries and Aquatic Science 58:1760–1772. Harman, H.H. (1976). Modern Factor Analysis. Chicago University Press, Chicago. Harris, G. (2002). Integrated assessment and modelling: An essential way of doing science. Environmental Modelling and Software 17:201–207. Hastings, W.K. (1970). Monte Carlo sampling methods using Markov chains and their applications. Biometrika 57:97–109. Hatcher, A.C. (1997). Producers organizations and devolved fisheries management in the United Kingdom: Collective and individual quota systems. Marine Policy 21:519–533. Heckman, J. (1979). Sample selection bias as a specification error. Econometrica 47:153–161. Hedges, L.V., and I. Olkin (1985). Statistical Methods for Meta-Analysis. Academic Press, New York. Heemink, A. (1991). Rekenen met Onzekerheid. Delft University Press, Delft. Hellerstein, D. (1992). The treatment of nonparticipants in travel cost analysis and other demand models. Water Resources Research 28:1999–2004. Higler, B., N. Dankers, A. Smaal, and V. de Jonge (1998). Evaluatie van de ecologische effecten van het reguleren van schelpdiervisserij in de Wadden Zee en Delta op bodemorganismen en vogels, Structuurnota Zee-en Kustvisserij Evaluatie Kustvisserij-beleid 1993–1997. Ministerie van landbouw, Natuurbeheer en Visserij, The Hague. Hilborn, R., and R.M. Peterman (1995). The development of scientific advice with incomplete information in the context of the precautionary approach. Technical consultation on the precautionary approach to capture fisheries, Lysekil, Sweden. Hilborn, R., and C.J. Walters (1992). Quantitative Fisheries Stock Assessment: Choice, Dynamics and Uncertainty. Chapman and Hall, New York. Hoekstra, J., and J.C.J.M. van den Bergh (2005). Harvesting and conservation in a predator–prey system. Journal of Economic Dynamics and Control 29(6):1097–1120. Hoel, A.H. (1998). Political uncertainty in international fisheries management. Fisheries Research 37:239– 250. Holland, D., E. Gudmunsson, and J. Gates (1999). Do fishing vessel buyback programs work: A survey of the evidence. Marine Policy 23:47–69. Holling, C.S. (1973). Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4:1–23. Holling, C.S. (1978). Adaptive Environmental Assessment and Management. John Wiley & Sons, New York. Holling, C.S. (1986). The resilience of terrestrial ecosystems: Local surprise and global change. In: W.C. Clark and R.E. Munn (eds.), Sustainable Development of the Biosphere. Cambridge University Press, Cambridge. Holmes, T.P., and W.L. Adamowicz (2003). Attribute-based methods. In: P.A. Champ, K.J. Boyle, and T.C. Brown (eds.), A Primer on Nonmarket Valuation. Kluwer Academic Publishers, Dordrecht, pp. 171–219. Honkoop, P.J.C., and J. van der Meer (1998). Experimentally induced effects of water temperature and immersion time on reproductive output of bivalves in the Wadden Sea. Journal of Experimental Marine Biology and Ecology 220:227–246. Horan, R.D., and J.S. Shortle (1999). Optimal management of multiple renewable resource stocks: An application to Minke Whales. Environmental and Resource Economics 13:435–458. Hotelling, H. (1931). The economics of exhaustible resources. Journal of Political Economy 39:137–175. Hulscher, J.B. (1989). Sterfte en overleving van scholeksters, Haematopus ostralegus, bij strenge vorst. Limosa 62:177–181. Hulscher, J.B. (1996). Food and feeding behaviour. In: J.D. Goss-Custard (ed.), The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford, pp. 7–29.

248

REFERENCES

ICES. (1994). Report of the study group on the North Sea Plaice Box. ICES CM 1994/Assess, Copenhagen. Imeson, R., and J.C.J.M. van den Bergh (2004). A bioeconomic analysis of a shellfishery: The effects of recruitment and habitat in a metapopulation model. Environmental and Resource Economics 27:65–86. Imeson, R., J.C.J.M. van den Bergh, and J. Hoekstra (2002). Integrated models of fisheries management and policy. Environmental Modelling and Assessment 7(4):259–271. International Whaling Commission. (1995). Report of the scientific committee, Annex F. Report of the subcommittee on aboriginal subsistence whaling, Report of the International Whaling Commission, Vol. 45, Cambridge, UK. Jansen, V.A.A., and A.M. de Roos (2000). The role of space in reducing predator–prey cycles. In: U. Dieckmann, R. Law, and J.A.J. Metz (eds.), The Geometry of Ecological Interactions: Simplifying Spatial Complexity. Cambridge University Press, Cambridge. Janssen, M.A., J.M. Anderies, and B.H. Walker (2004). Robust strategies for managing rangelands with multiple stable attractors. Journal of Environmental Economics and Management 47(1):140–162. Janssen, P.H.M., and P.S.C. Heuberger (1995). Calibration of process-oriented models. Ecological Modelling 83:55–66. Jennings, S. (2001). Patterns and prediction of population recovery in marine reserves. Reviews in Fish Biology and Fisheries 10:209–231. Jensen, C.L. (2002). Reduction of the fishing capacity in common pool fisheries. Marine Policy 26:155–158. Jentoft, S. (2000). Legitimacy and disappointment in fisheries management. Marine Policy 24:141– 148. Jentoft, S., and B.J. McCay (1995). User participation in fisheries management: Lessons drawn from international experiences. Marine Policy 19:227–246. Jentoft, S., B.J. McCay, and D.C. Wilson (1998). Social theory and fisheries co-management. Marine Policy 22:423–436. Johnson, N.L., and S. Kotz (1970). Continuous Univariate Distributions, Vol. 2. Wiley, New York. Kamermans, P., and A.C. Smaal (2002). Mussel culture and cockle fisheries in the Netherlands: Finding a balance between economy and ecology. Journal of Shellfish Research 21:509–517. Kamien, M.I., and N.L. Schwartz (1981). Dynamic Optimization: The Calculus of Variations and Optimal Control. Elsevier Science Publishers, Amsterdam. Kaplan, I.M., and H.L. Kite-Powell (2000). Safety at sea and fisheries management: Fishermens attitudes and the need for co-management. Marine Policy 34:493–497. Karagiannakos, A. (1996). Total allowable catch (TAC) and quota management system in the European Union. Marine Policy 20:235–248. Karpoff, J.M. (1987). Suboptimal controls in common resource management: The case of the fishery. Journal of Political Economy 95:179–194. Kelly, D.L., and C.D. Kolstad (1999). Integrated assessment models for climate change control. In: H. Folmer and T. Tietenberg (eds.), The International Yearbook of Environmental and Resource Economics 1999/2000. Edward Elgar, Cheltenham. Kersten, M., and T. Piersma (1987). High levels of energy expenditure in shorebirds: Metabolic adaptations to an energetically expansive way of life. Ardea 75:175–187. Kersten, M., and W. Visser (1996). The rate of food processing in the oystercatcher: Food intake and energy expenditure constrained by digestive bottlenecks. Functional Ecology 10:440–448. King, A.W., and K.A. With (2002). Dispersal success on spatially structured landscapes: When do spatial pattern and dispersal behavior really matter? Ecological Modelling 147:23–39. Klepper, O. (1989). A model of carbon flows in relation to macrobenthic food supply in the Oosterschelde estuary (S.W. Netherlands). Agricultural University Wageningen, Wageningen. Klyashtorin, L.B. (1998). Long-term climate change and main commercial fish production in the Atlantic and Pacific. Fisheries Research 37:115–125. Lande, R. (1987). Extinction thresholds in demographic models of territorial populations. American Naturalist 130:624–635. Lane, D.E., and R.L. Stephenson (2000). Institutional arrangements for fisheries: Alternate structures and impediments to change. Marine Policy 24:385–393. Lauck, T., C.W. Clark, M. Mangel, and G.R. Munro (1998). Implementing the precautionary principle in fisheries management through marine reserves. Ecological Applications 8:S72–S78.

REFERENCES

249

LEI (Smit, W., J. de Jager, and M.O. van Wijk). (1998). Structuurnota Zee-en Kustvisserij, Evaluatie Beleid Kustwateren: Evaluatie Economische Effect op de Mossel-en Kokkelsector. Agricultural Economic Institute, The Hague. LEI (van Wijk, M.O., C. de Ruijter, M.H. Smit, and C. Taal). (2001). Visserij in Cijfers. Agricultural Economic Institute, The Hague. LEI (Salz, P., M.O. van Wijk, J.G.P. Smit, and C. Taal). (2002a). De Nederlandse Schelpdiersector: Korte Termijn Studie. Agricultural Economic Institute, The Hague. LEI (van Wijk, M.O., C. Taal, J.W. de Wilde, and M.H. Smit). (2002b). Visserij in Cijfers. Agricultural Economic Institute, The Hague. Likert, R. (1967). The method of constructing an attitude scale. In: M. Fishbein (ed.), Readings in Attitudes Measurement. Columbia University Press, New York, pp 90–95, originally in Archives of Psychology 140:44–53. Lines, M. (1995). Dynamics and uncertainty. In: H. Folmer, H. Landis Gabel, and H. Opschoor (eds.), Principles of Environmental and Resource Economics. Edward Elgar, Aldershot, UK, pp. 67–105. Loomis, J.B., and R.G. Walsh (1997). Recreation Economic Decisions: Comparing Benefits and Costs. Venture Publishing, Pennsylvania. Louviere, J.J., and D. Hensher (1982). On the design and analysis of simulated choice or allocation experiments in travel choice modelling. Transportation Research Record 890:11–17. Louviere, J.J., D.A. Hensher, and J.D. Swait (2000). Stated Choice Methods: Analysis and Applications. Cambridge University Press, Cambridge. Louviere, J.J., and D.J. Street (2000). Stated-preference methods. In: D.A. Hensher and K.J. Button (eds.), Handbook of Transport Modelling, Elsevier Science, Amsterdem. Louviere, J.J., and G. Woodworth (1983). Design and analysis of simulated consumer choice or allocation experiments: An approach based on aggregate data. Journal of Marketing Research 20:350–367. Ludwig, D., R. Hilborn, and C. Walters (1993). Uncertainty, resource exploitation, and conservation: Lessons from history. Science 260:7–36. MacKenzie, B. (2000). Turbulence, larval fish ecology and fisheries recruitment: A review of field studies. Oceanologica Acta 23(4):357–375. Maddala, G.S. (1983). Limited Dependent and Qualitative Variables in Econometrics. Cambridge University Press, New York. Makowski, M. (2000). Modeling paradigms applied to the analysis of European air quality. European Journal of Operational Research 122:219–241. Man, A., R. Law, and N.V.C. Polunin (1995). Role of marine reserves in recruitment to reef fisheries: A metapopulation model. Biological Conservation 71:197–204. Mangel, M. (2000). On the fraction of habitat allocated to marine reserves. Ecology Letters 3:15–22. Manne, A.S., and R.G. Richels (1992). Buying Greenhouse Insurance: The Economic Costs of CO2 Emission Limits. The MIT Press, Cambridge, MA. May, R.M. (1972). Limit cycles in predator–prey communities. Science 177:900–902. May, R.M., J.R. Beddington, C.W. Clark, S.J. Holt, and R.M. Laws (1979). Management of multispecies fisheries. Science 205(4403):267–277. Maynard Smith, J. (1974). Models in Ecology. Cambridge University Press, Cambridge. McFadden, D. (1973). Conditional logit analysis of qualitative choice behaviour. In: P. Zarembka (ed.), Frontiers in Econometrics. Academic Press, New York. McGlade, J., and P.M. Allen (1987). Modelling complex human systems: A fisheries example. European Journal of Operational Research 30:147–167. McGoodwin, J.R. (1990). Crisis in the Worlds Fisheries: People, Problems and Policy. Stanford University Press, Stanford. Meijer, W., P.C. Lodders-Elfferich, and L. Hermans (2004). Ruimte voor de Wadden – Eindrapport Adviesgroep Waddenzee. Ministry of Economics and Ministry of Housing, Spatial Planning and the Environment, The Hague (www.waddenzee.nl). Mensink, P.G.C. (2000). On the Relativity of Quantitative Model Results for Policy Purposes in the Social Sciences. Wageningen Economic Papers 12, Department of Social Sciences, Wageningen University, Wageningen. Mertz, G., and R.A. Myers (1995). Estimating the predictability of recruitment. Fisheries Bulletin 93:439– 453.

250

REFERENCES

Metropolis, N., A. Rosenbluth, M. Rosenbluth, A. Teller, and E. Teller (1953). Equations of state calculations by fast computing machines. Journal of Chemical Physics 21:1087–1091. Milessi, A.C., and O. Defeo (2002). Long-term impact of incidental catches by tuna longlines: The black escolar (Lepidocybium flavobrunneum). Fisheries Research 58:203–213. Ministerie LNV. (1993). Vissen naar evenwicht: Regeringsbeslissing Structuurnota Zee-en kustvisserij. Ministerie van Landbouw, Natuurbeheer en Visserij, The Hague. Ministerie LNV. (1998). Structuurnota Zee-en Kustvisserij. Evaluatie Kustvisserijbeleid 1993–1997 (Evaluation of Coastal Fishery Policy). Ministerie van Landbouw, Natuurbeheer en Visserij, maart 1998, Den Haag. Ministerie LNV. (1999). Gemeenschappelijk Visserijbeleid 2002. Ministerie van Landbouw, Natuur en Visserij, Directie Visserij. http: //www.minlnv.nl/thema/visserij/nottv05.pdf. Ministerie van Financi¨en. (1995). Kabinetsstandpunt Heroverweging Disconteringsvoet. Ministerie van Financi¨en (Ministry of Finance), Den Haag. Mode, C.J., and M.E. Jacobson (1987). Demographic stochasticity and uncertainty in population projections – A study by computer simulation. Mathematical Bioscience 85:185–209. Morin, M. (2000). The fisheries resources in the European Union. The distribution of TACs: Principle of relative stability and quota-hopping. Marine Policy 4:265–273. Munro, A. (1997). Economics and biological evolution. Environmental and Resource Economics 12:429– 449. Myers, R.A., and N.J. Barrowman (1996). Is fish population related to spawner abundance? Fishery Bulletin 94:707–724. Myers, R.A., and G. Mertz (1998). Reducing uncertainty in the biological basis of fisheries management by meta-analysis of data from many populations. Fisheries Research 37:51–60. Neher, P.A. (1990). Natural Resource Economics: Conservation and Exploitation. Cambridge University Press, Cambridge. ¨ Nehls, G. (1995). Strategien der Ern¨ahrung und ihre Bedeutung f¨ur Energiehaushalt und Okologie der Eiderente. Ph.D. Dissertation, FTZ-Report nr.10, University of Kiel. Nehls, G., I. Hertzler, C. Ketzenberg, and G. Scheiffarth (1997). Stable mussel beds as a resource for birds. In: C. G¨atje and K. Reise (eds.), The Wadden Sea Ecosystem: Exchange, Transport and Transformation Processes. Springer Verlag, Berlin, pp. 421–435. Nehls, G., and C. Ketzenberg (2002). Do Common Eiders Somateria mollissima exhaust their food resources? A study on natural Mussel Mytilus edulis beds in the Wadden Sea. Danish Review of Game Biology 16:47–61. Nehring, S. (1995). Gymnodinium catenatum Graham (Dinophyceae) in Europe: A growing problem? Journal of Plankton Research 17:85–102. Nehring, S. (1998). Establishment of thermophilic phytoplankton species in the North Sea: Biological indicators of climatic changes? ICES Journal of Marine Science 55:818–823. Nielsen, A., and P. Lewy (2002). Comparison of the frequentist properties of Bayes and the maximum likelihood estimators in an age-structured fish stock assessment model. Canadian Journal of Fisheries and Aquatic Science 59:136–143. Nielsen, J.R., and C. Mathiesen (2000). Important factors influencing rule compliance in fisheries: Lessons from Danish fisheries. Paper presented at the IIFET 2000 Conference in Oregon USA, 10–14 July 2000, Working Paper No. 4-2000, Institute for Fisheries Management and Coastal Community Development. Nielsen, J.R., and T. Vedsmand (1997). Fishermens organisations in fisheries management. Marine Policy 21:277–299. Nielsen, J.R., and T. Vedsmand (1999). User participation and institutional change in fisheries management: A viable alternative to the failures of top-down driven control? Ocean and Coastal Management 42:19– 37. Nijkamp, P., and J.C.J.M. van den Bergh (1998). Economic aspects of global environmental change impacts and response strategies in the Netherlands. Journal of Coastal Conservation 4:161–168. Nisbet, R.M., and W.S.C. Gurney (1982). Modeling Fluctuating Populations. Wiley, Chichester, UK. NOAA (National Oceanic and Atmospheric Administration). (1993). Report of the NOAA panel on contingent valuation. Federal Register (US) 58:4601–4614. Noailly, J., J.C.J.M. van den Bergh, and C.A. Withagen (2003). Evolution of harvesting strategies: Replicator and resource dynamics. Journal of Evolutionary Economics 13(2):183–200.

REFERENCES

251

Noble, B.F. (2000). Institutional criteria for co-management. Marine Policy 24:69–77. Nordhaus, W.D. (1992). An optimal transition path for controlling greenhouse gases. Science 258:1315– 1319. Nordhaus, W.D., and Z. Yang (1996). A regional dynamic general equilibrium model of alternative climate change strategies. American Economic Review 86(4):741–765. Norris, K., and Johnstone, I.G. (1998). Interference competition and the functional response of oystercatchers searching for cockles by touch. Animal Behaviour 56;639–650. Nunes, P., and J.C.J.M. van den Bergh. (2001). Monetary valuation of biodiversity: Sense or nonsense? Ecological Economics 39(2):203–222. Nunes, P., and J.C.J.M. van den Bergh (2004). Can people value protection against exotic marine species? Evidence from a joint TC-CV survey in the Netherlands. Environmental and Resource Economics 28:517–532. Nunes, P.A.L.D. (2002). Using factor analysis to identify consumer preferences for the protection of a natural area in Portugal. European Journal of Operational Research 140:499–516. Nunes, P.A.L.D., and E. Schokkaert (2003). Identifying the warm glow effect in contingent valuation. Journal of Environmental Economics and Management 45:231–245. Nunes, P.A.L.D., and J.C.J.M. van den Bergh (2002). Measuring the economic value of a marine protection program against invasive exotic species in the Netherlands: Results from a travel cost and contingent valuation survey. Discussion Paper TI 057–3, Tinbergen Institute, Amsterdam, The Netherlands. Olson, L.J., and S. Roy (1996). On conservation of renewable resources with stock-dependent return and nonconcave production. Journal of Economic Theory 70:133–157. Olson, R.L., and R.A. Sequeira (1995). Emergent computation and the modeling and management of ecological systems. Computers and Electronics in Agriculture 12:183–209. Oreskes, N., K. Shrader-Frechette, and K. Belitz (1994). Verification, validation, and confirmation of numerical models in the earth sciences. Science 263:641–646. Ostenfeld, C.H. (1908). On the immigration of Biddulphia sinensis and its occurrence in the North Sea during 1903–1907 series. Plankton I:1–48. Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press, Cambridge, UK. Parson, E.E. (1995). Integrated assessment and environmental policy making. In pursuit of usefulness. Energy Policy 23:463–475. Pascoe, S. (2000). Single species conservation in a multispecies fishery: The case of the Australian Eastern Gemfish. Ecological Economics 32:125–136. Pauly, D., V. Christensen, S. Gu´enette, T.J. Pitcher, R. Sumaila, C.J. Walters, R. Watson, and D. Zeller (2000). Towards sustainability in world fisheries. Nature 408:689–695. Pearce, D. (1998). Auditing the earth. Environment 2:23–28. Pearce, D.W., and R.K. Turner (1990). Economics of Natural Resources and the Environment. Harvester Wheatsheaf, London. Perrings, C. (1997). Reserved rationality and the precautionary principle: Technological change, time and uncertainty in environmental decision-making. In: C. Perrings (ed.), Economics of Ecological Resources. Edward Elgar, Cheltenham, UK. Perrings, C. (1998). Resilience in the dynamics of economy-environment systems. Environmental and Resource Economics 11:503–520. Perrings, C. (2001). Modelling sustainable ecological-economic development. In: H. Folmer and T. Tietenberg (eds.), The International Yearbook of Environmental and Resource Economics 2001/2002. Edward Elgar, Cheltenham, UK, pp. 179–201. Perrings, C. (2002). Biological invasions in aquatic systems: The economic problem. Bulletin of Marine Science 70(2):541–552. Perrings, C., M. Williamson, E.B. Barbier, D. Delfino, S. Dalmazzone, J. Shogren, P. Simmons, and A. Watkinson (2002). Biological invasion risks and the public good: An economic perspective. Conservation Ecology 6(1):1. Perrings C., M. Williamson, and S. Dalmazzone (eds.). (2000). The Economics of Biological Invasions. Edward Elgar, Cheltenham. Peterson, G. (2000). Political ecology and ecological resilience: An integration of human and ecological dynamics. Ecological Economics 35:325–336.

252

REFERENCES

Petraitis, P.S. (1995). The role of growth in maintaining spatial dominance by mussels (Mytilus Edulis). Ecology 76:1337–1346. Pezzey, C.V., C.M. Roberts, and B.T. Urdal (2000). A simple bioeconomic model of a marine reserve. Ecological Economics 33:77–91. Piersma, T., A. Koolhaas, A. Dekinga, J.J. Beukema, R. Dekker, and K. Essink (2001). Long-term indirect effects of mechanical cockle-dredging on intertidal bivalve stocks in the Wadden Sea. Journal of Applied Ecology 38:976–990. Pimm, S.L. (1984). The complexity and stability of ecosystems. Nature 307:321–326. Pindyck, R.S. (1984). Uncertainty in the theory of renewable resource markets. Review of Economic Studies 51:289–303. Pindyck, R.S. (2000). Irreversibilities and the timing of environmental policy. Resource and Energy Economics 22:233–259. Pindyck, R.S. (2002). Optimal timing in environmental economics. Journal of Economic Dynamics and Control 26:1677–1697. Plourde, C.G. (1970). A simple Model of replenishable resource exploitation. The American Economic Review 60:518–522. Popper, K.L. (1959). The Logic of Scientific Discovery. Hutchingson & Co., London. Possingham, H.P., and J. Roughgarden (1990). Spatial population dynamics of a marine organism with a complex life cycle. Ecology 71(3):973–985. Punt, A.E., and D.S. Butterworth (1997). Assessments of the Bering-Chukchi-Beaufort Seas stock of Bowhead whales using maximum likelihood and Bayesian methods. Reports of the International Whaling Commission 47:603–618. Punt, A.E., and R. Hilborn (1997). Fisheries stock assessment and decision analysis: The Bayesian approach. Reviews in Fish Biology and Fisheries 7:35–63. Ragozin, D.L., and G.J. Brown (1985). Harvest policies and nonmarket valuation in a predator–prey system. Journal of Environmental Economics and Management 12:155–168. Rammel, C., and J.C.J.M. van den Bergh (2003). Evolutionary policies for sustainable development: Adaptive flexibility and risk minimising. Ecological Economics 47 (2):121–133. Rappoldt, C., B.J. Ens, E.M. Dijkman, and T. Bult (2003a). Scholeksters en hun voedsel in de Waddenzee. Rapport voor deelproject B1 van EVA II, de tweede fase van het evaluatieonderzoek naar de effecten van schelpdiervisserij op natuurwaarden in de Waddenzee en Oosterschelde 1999–2003. Alrerra-Report 882, Alterra, Wageningen. Rappoldt, C., B.J. Ens, E.M. Dijkman, T. Bult, C.M. Berrevoets, and J. Geurts van Kessel (2003b). Scholeksters en hun voedsel in de Oosterschelde. Rapport voor deelproject D2 thema 1 van EVA II, de tweede fase van het evaluatieonderzoek naar de effecten van schelpdiervisserij op natuurwaarden in Waddenzee en Oosterschelde 1999–2003. Alterra-report 883, Alterra, Wageningen. Rappoldt, C., B.J. Ens, M.A.J.M. Kersten, and E.M. Dijkman (2004). Wader energy balance and tidal cycle simulator WEBTICS. Technical documentation version 1.1. Alterra-Report 869, Alterra, Wageningen. Refsgaard, J.C., and H.J. Hendriksen (2004). Modelling guidelines – Terminology and guiding principles. Advances in Water Resources 27:71–82. Ricker, W. (1954). Stock and recruitment. Journal of the Fisheries Board of Canada 11:559–623. RLG (Wolff, W.J., J. van Noord, and J.W. Koolhaas). (1998). Leven en laten leven: Advies over kustvisserij en natuur in kustgebieden, Raad voor het Landelijk Gebied. http: //www.rlg.nl/adviezen/rlg983.pdf. Roberts, C.M., J.A. Bohnsack, F. Gell, J.P. Hawkins, and R. Goodridge (2001). Effects of marine reserves on adjacent fisheries. Science 294:1920–1923. Rotmans, J. (1998). Methods for IA: The challenges and opportunities ahead. Environmental Modeling and Assessment 3:155–179. Rotmans, J., and B. de Vries (1997). Perspectives on Global Change: The TARGETS Approach. Cambridge University Press, Cambridge, UK. Rotmans, J., and M.B.A. van Asselt (2001). Uncertainty management in integrated assessment modeling: Towards a pluralistic approach. Environmental Monitoring and Assessment 69:101–130. Roughgarden, J., and F. Smith (1996). Why fisheries collapse and what to do about it. Proceedings of the National Academy of science 93:5078–5083.

REFERENCES

253

Russ, G.R., and A.C. Alcala (1996). Do marine reserves export adult fish biomass? Evidence from Apo Island, Central Philippines. Marine Ecology Progress Series 132:1–9. Sanchirico, J.N. (2000). Marine protected areas as fishery policy: A discussion of potential costs and benefits. Discussion paper 00–23: Resources for the future, Washington. Sanchirico, J.N., and J.E. Wilen (1999). Bioeconomics of spatial exploitation in a patchy environment. Journal of Environmental Economics and Management 37:129–150. Sanchirico, J.N., and J.E. Wilen (2001). A bioeconomic model of marine reserve creation. Journal of Environmental Economics and Management 42:257–276. Saraydar, E. (1989). The conflation of productivity and efficiency in economics and economic history. Economics and Philosophy 5:55–67. Schaefer, M.B. (1954). Some aspects of the dynamics of populations important to the management of commercial marine fisheries. Bulletin of the Inter-American Tropical Tuna Commission 1:25–56. Schneider, S.H. (1997). Integrated asssessment modeling of global climate change: Transparant rational tool for policy making or opaque screen hiding value laden assumptions? Environmental Modeling and Assessment 2:229–249. Scholin, C., Gulland, F., Doucette, G.J., Benson, S., Busman, M., Chavez, F.P., Cordaro, J., DeLong, R., De Vogelaere, A., Harvey, J., Haulena, M., Lefebvre, A., Lipscomb, T., Loscutoff, S., Lowenstine, L.J., Marin III, R., Miller, P.E., McLellan, W.A., Moeller, P.D.R., Powell, C.R., Rowles, T., Silvagni, P., Silver, M., Spraker, T., Trainer, V., and F.M. Van Dolah (2000). Mortality of sea lions along the central California coast linked to a toxic diatom bloom. Nature 403:80–84. Schweder, T. (1998). Fisherian or Bayesian methods of integrating diverse statistical information? Fisheries Research 37:61–75. Segerson, K., and N. Li (1999). Voluntary approaches to environmental protection. In: H. Folmer and T. Tietenberg (eds.), The International Yearbook of Environmental and Resource economics 1999/2000. Edward Elgar Publishing Ltd., Cheltenham, UK, pp. 273–306. Semmler, W., and M. Sieveking (1994). On the optimal exploitation of interacting resources. Journal of Economics 59:23–49. Sen, S., and J.R. Nielsen (1996). Fisheries co-management: A comparative analysis. Marine Policy 20:405– 418. Settle, C., T.D. Crocker, and J.F. Shogren (2002). On the joint determination of biological and economic systems. Ecological Economics 42:301–311. Shackley, S. (1997). Trust in models? The mediating and transformative role of computer models in environmental discourse. In: M. RedClift and G. Woodgate (eds.), The International Handbook of Environmental Sociology. Edward Elgar, Cheltenham, UK, pp. 237–260. Shaffer, M.L. (1981). Minimum population sizes for species conservation. Bioscience 31:131–134. Sitters, H.P. (2000). The role of night-feeding in shorebirds in an estuarine environment with specific reference to mussel-feeding oystercatchers. Ph.D. thesis, Wolfson Coll. & Edward Grey Inst., Dept. Zool., Univ. Oxford. 301 p. Smaal, A., M. van Stralen, and E. Schuiling (2001). The interaction between shellfish culture and ecosystem processes. Canadian Journal of Fisheries and Aquatic Science 58:991–1002. Smit, C.J., N. Dankers, B.J. Ens, and A. Meijboom (1998). Birds, mussels, cockles and shellfish fishery in the Dutch Wadden Sea: How to deal with low food stocks for eiders and oystercatchers? Senckenbergiana Maritima 29:141–153. Smit, C.J., B.J. Ens, and B. Koks (2000). Afnemende aantallen Scholeksters in de Waddenzee. SOVON-Nieuws 13:16–17. Speakman, J.R. (1987). Apparent absorption efficiencies for redshank (Tringa totanus L.) and oystercatcher (Haematopus ostrlegus L.): Implications for the predictions of optimal foraging models. American Naturalist 130:677–691. Spear, R.C. (1997). Large simulation models: Calibration, uniqueness and goodness of fit. Environmental Modelling and Software 12:219–228. Sprung, M. (1984). Physiological energetics of mussel larvae (Mytilus edulis). 2. Food uptake. Marine Ecology Progress Series 17:295–305. Steelman, A.S., and R.L. Wallace (2001). Property rights and property wrongs: Why context matters in fisheries management. Policy Sciences 34:357–379.

254

REFERENCES

Steins, N.A. (1999). All hands on deck: An interactive perspective on complex common-pool resource management based on case studies in the coastal waters of the Isle of Wight (UK), Connemara (Ireland) and the Dutch Wadden Sea. Ph.D. Thesis, Wageningen University, Wageningen. Stephenson, R.L. (1999). Stock complexity in fisheries management: A perspective of emerging issues related to population sub-units. Fisheries Research 43:247–249. Sterman, J.D., E.J. Rykiel Jr., N. Oreskes, K. Belitz, and K. Shrader-Frechette (1994). The meaning of models. Science 264:329–331. Stichting Odus. (2001). Uit de Schulp: Visie op Duurzame Ontwikkeling van de Nederlandse Schelpdiervisserij. Stichting Odus, Yerseke. Stillmann, R.A., J.D. Goss-Custard, A.D. West, S.E.A.L.V.D. Durell, S. McGrogerty, R.W.G. Caldow, K.J. Norris, I.G. Johnstone, B.J. Ens, J. van der Meer, and P. Triplet (1996). Models of shellfish populations and shorebirds: Final report. November 1996. Report to Commission of the European Communities, Directorate-General for Fisheries, Institute of Terrestrial Ecology (Natural Environment Research Council), Wareham, Dorset, UK. Stillmann, R.A., J.D. Goss-Custard, A.D. West, S.E.A.L.V.D. Durell, S. McGrogerty, R.W.G. Caldow, K.J. Norris, I.G. Johnstone, B.J. Ens, J. van der Meer, and P. Triplet (2001). Predicting shorebird mortality and population size under different regimes of shellfishery management. Journal of Applied Ecology 38:857–868. Strazzera, E., M. Genius, R. Scarpa, and G. Hutchinson (2003a). The effect of protest votes on the estimates for use values of recreational sites. Environmental and Resource Economics 25:461–476. Strazzera, E., R. Scarpa, P. Calia, G.D. Garrod, and K.G. Willis (2003b). Modelling zero values and protest responses in contingent valuation surveys. Applied Economics 35:133–138. Strobele, W.J., and H. Wacker (1995). He economics of harvesting predator–prey systems. Journal of Economics 61(1):65–81. Strogatz, S.H. (1994). Nonlinear Dynamics and Chaos. Addison-Wesley, Reading, Mass. Sumaila, U.R. (1998). Protected marine reserves as fishery management tools: A bioeconomic analysis. Fisheries Research 37:287–296. Sumaila, U.R., S. Guenette, J. Alder, and R. Chuenpagdee (2000). Addressing ecosystem effects of fishing using marine protected areas. ICES Journal of Marine Science 57:752–760. Sutherland, W.J., B.J. Ens, J.D. Goss-Custard, and J.B. Hulscher (1996). Specialization. In: J.D. Goss-Custard (ed.), The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford, pp. 56–76. Sutinen, J, A. Rieser, and J. Gauvin (1990). Measuring and explaining non-compliance in federally managed fisheries. Ocean Development and International Law 21:335–372. Sutinen, J.G., and K. Kuperan (1999). A socio-economic theory of regulatory compliance. International Journal of Social Economics 26:174–193. Swennen, C. (1976). Populatie-structuur en voedsel van de Eidereend Somateria m. mollissim in de Nederlandse Waddenzee. Ardea 64:311–371. Swennen, C. (1991). Ecology and Population Dynamics of the Common Eider in the Dutch Wadden Sea. Ph.D. Rijksuniversiteit Groningen, Groningen. Swennen, C. (2002). Development and population dynamics of Common Eider Somateria mollissima colonies in The Netherlands. Danish Review of Game Biology 16:63–74. Symes, D. (1997). Fisheries management: In search of good governance. Fisheries Research 32:107– 114. Tjelmeland, S. and B. Bogstad (1998). MULTSPEC – a review of a multispecies modelling project of the Barents sea. Fisheries Research 37:127–142. Townsend, R.E. (1986). A critique of models of the American lobster fisheries. Journal of Environmental Economics and Management 13:277–291. Triplet, P., R.A. Stillmann, and J.D. Goss-Custard (1999). Prey abundance and the strength of interference in a foraging shorebird. Journal of Animal Ecology 68:254–265. Tu, P.N.V., and E.A. Wilman (1992). A generalized predator–prey model: Uncertainty and management. Journal of Environmental Economics and Management 23:123–138. Turnbull, B.W. (1976). The empirical distribution function with arbitrarily grouped, censored and truncated data. Journal of the Royal Statistical Society 38:290–295. Turner, R.K., J.C.J.M. van den Bergh, and R. Brouwer (eds.). (2003) Managing Wetlands: An Ecological Economics Approach. Edward Elgar Publication, Cheltenham.

REFERENCES

255

Turner, R.K., J.C.J.M. van den Bergh, T. S¨oderqvist, A. Barendregt, J. van der Straaten, E. Maltby, and E.C. van Ierland (2000). Ecological-economic analysis of wetlands: Scientific integration for management and policy. Ecological Economics 35(1):7–23. United Nations. (1998). International fisheries instruments with Index. Division for ocean affairs and the Law of the Sea, New York. van Asselt, M.B.A., and J. Rotmans (2002). Uncertainties in integrated assessment modelling. Climatic Change 54:75–105. van Dalen, C.E., L. Dresen, and M.A. Janssen (2002). The roles of computer models in the environmental policy life cycle. Environmental Science and Policy 238:1–11. van den Bergh, J.C.J.M. (1991). Dynamic Models for Sustainable Development. Ph.D. Thesis, Tinbergen Institute, Thesis Publishers, Amsterdam. van den Bergh, J., and P. Nijkamp (1994). An integrated model for economic development and natural environment: An application to the Greek Sporades Islands. The Annals of Operations Research 54:143– 174. van den Bergh, J.C.J.M., and P. Nijkamp (2000). Marine resources and coastal zones: A challenge to the research community. International Journal of Development Planning Literature 15(1): 21–32. van den Bergh, J.C.J.M. (2004). Evolutionary modelling. In: J. Proops and P. Safonov (eds.), Modelling in Ecological Economics. Edward Elgar, Cheltenham, pp. 9–35. van den Bergh, J.C.J.M., A. Barendregt, and A. Gilbert (2004). Spatial Ecological-Economic Analysis for Wetland Management: Modelling and Scenario Evaluation of Land-Use. Cambridge University Press, Cambridge, UK. van den Bergh, J.C.J.M., A. Barendregt, A. Gilbert, M. van Herwijnen, P. van Horssen, P. Kandelaars, and C. Lorenz (2001). Spatial economic-hydroecological modelling and evaluation of land use impacts in the Vecht Wetlands area. Environmental Modelling and Assessment 6(2):87–100. van den Bergh, J.C.J.M., P.A.L.D Nunes, H.M. Dottinga, W.H.C.F. Kooistra, E.G. Vrieling, and L. Peperzak (2002). Exotic harmful algae in marine ecosystems: An integrated biological-economic-legal analysis of impacts and policies. Marine Policy 26(1):59–74. Van der Have, T. (2002). Birds and shellfisheries in the Dutch Wadden Sea: A case of adaptive management. In: W.J. Wolff (ed.), Proceedings of the 10th International Wadden Sea Symposium. November 2000, Groningen. van der Heide, C.M., J.C.J.M. van den Bergh, and E.C. van Ierland (2002). Towards an ecological-economic theory of nature policy. In: S. Dovers, D. Stern, and M. Young (eds.), New Dimensions in Ecological Economics: Integrative Approaches to People and Nature. Edward Elgar, Cheltenham. van der Meer, J. (1997). A Handful of Feathers. Faculty of Natural Sciences. Ph.D. Thesis, Rijksuniversiteit Groningen, Groningen. van der Meer, J., and B.J. Ens (1997). Models of interference and their consequences for the spatial distribution of ideal and free predators. Journal of Animal Ecology 66:846–858. van Stralen, M. (1998). De ontwikkeling van het mosselbestand in de Waddenzee en Oosterschelde na 1992. Rapport c.006.98, 5–31. RIVO, Yerseke. van Stralen, M. (2002). Inventarisatie van het wilde mosselbestand in de Waddenzee in het najaar van 2002. MarinX notitie 2002.28, 1–5. Scharendijke, marinX. Vasconcellos, M. (2002). An analysis of harvest strategies and information needs in a purse seine fishery for the Brazilian sardine. Fisheries Research 1390:1–16. Virtala, M., S. Kuikka, and E. Arjas (1998). Stochastic virtual population analysis. ICES Journal of Marine Science 55:892–904. Visser, H., R.J.M. Folkert, J. Hoekstra, and J.J. de Wolff (2000). Identifying key sources of uncertainty in climate change projections. Climatic Change 45:421–457. Wacker, H. (1999). Optimal harvesting of mutualistic ecological systems. Resource and Energy Economics 21:89–102 Wade, R. (1987). The management of common property resources: Collective action as an alternative to privatization or state regulation. Cambridge Journal of Economics 11:95–106. Walters, C. (1986). Adaptive Management of Renewable Resources. MacMillan Publ. Co., New York. Walters, C., and J.F. Kitchell (2001). Cultivation/depensation effects on juvenile survival and recruitment: Implications for the theory of fishing. Canadian Journal of Fisheries and Aquatic Science 58:39–50.

256

REFERENCES

Weeks, H., and S. Berkeley (2001). Uncertainty and precautionary management of marine fisheries: Can the old methods fit the new mandates. Fisheries Management 25(12):6–14. Weitzman, M.L. (2002). Landing fees vs. harvest quotas with uncertain fish stocks. Journal of Environmental Economics and Management 43:325–338. Weninger, Q., and R.E. Just (2002). Firm dynamics with tradable output permits. American Journal of Agricultural Economics 84(3):572–584. Wilson J., B. Low, R. Costanza, and E. Ostrom (1999). Scale misperceptions and the spatial dynamics of a social-ecological system. Ecological Economics 31:243–257. Wilson, J.A., J.M. Acheson, M. Metcalfe, and P. Kleban. (1994). Chaos, complexity and community management of fisheries. Marine Policy 18:291–305. Winsemius, P. (1986). Gast in Eigen Huis: Beschouwingen over Milieumanagement. Tjeenk Willink, Alphen a/d Rijn. Wirl, F. (1995). The cyclical exploitation of renewable resource stocks may be optimal. Journal of Environmental Economics and Management 29:252–261. Wolff, W.J. (ed.). (1983). Ecology of the Wadden Sea.A.A. Balkema, Rotterdam, The Netherlands. Wolff, W.J, N. Dankers, B.J Ens, A.C Smaal, and M.O van Wijk ( 2001). Op zoek naar Win-Win situaties voor natuurbescherming en schelpdiervisserij in de Waddenzee, uitgebreide rapportage. Groningen University in corporation with Alterra, LEI and RIVO. Yodzis, P. (1989). Introduction to Theoretical Ecology. Harper & Row, New York. Yodzis, P. (1994). Predator–prey theory and management of multispecies fisheries. Ecological Applications 4:51–58. Young, O.R. (1998). Institutional uncertainties in international fisheries management. Fisheries Research 37:211–224. Zwarts, L. (1991). Seasonal variation in body weight of the bivalves Macoma balthica, Scrobicularia plana, Mya arenaria and Cerastoderma edule in the Dutch Wadden Sea. Netherlands Journal of Sea Research 28:231–245. Zwarts, L., and A.M. Blomert (1996). Daily metabolized energy consumption of oystercatchers Haema ostralegus feeding on larvae of the crane fly Tipula paludosa. Ardea 84A:221–228. Zwarts, L., J.T. Cayford, J.B. Hulscher, M. Kersten, P.M. Meire, and J. Triplet (1996a). Prey size selection and intake rate. In: J.D. Goss-Custard (ed.), The Oystercatcher: From Individuals to Populations. Oxford University Press, Oxford, pp. 30–55. Zwarts, L., and R.H. Drent (1981). Prey depletion and the regulation of predator density: Oystercatchers (Haematopus ostralegus) feeding on mussels (Mytilus edulis). In: N.V. Jones and W.J. Wolff (eds.), Feeding and Survival Strategies of Estuarine Organisms. Plenum Press, New York, pp. 193–216. Zwarts, L., B.J. Ens, J.D. Goss-Custard, J.B. Hulscher, and S.E.A.L.V.D. Durell (1996b). Causes of variation in prey profitability and its consequences for the intake rate of the oystercatcher Haema ostralegus. Ardea 84A:229–268. Zwarts, L., B.J. Ens, J.D. Goss-Custard, J.B. Hulscher, and M. Kersten (1996c). Why oystercatchers Haematopus ostralegus cannot meet their daily energy requirements in a single low water period. Ardea 84A:269– 290.

INDEX

abiotic factors, 54 adaptive management, 4, 23, 30, 70 adult stage, 41, 42, 117 adult–juvenile competition for space, 134 Advisory Committee of Fisheries Management (ACFM), 84, 85, 87 Advisory Group Wadden Sea Policy, 207 ageing, 152, 155 aging of shellfish, 155 algal monitoring program, 231 aquaculture, 56 area policy measures, 211 ash-free dry mass (AFDM), 153 attitudinal questions, 210, 215, 222 attribute levels, 213, 217 attribute-based methods, 211 attributes, 10, 211–215, 223, 233, 234, 239 availability of space to recruit, 122 Baie de Somme, 149, 167 ballast water, 2, 11, 229–231, 239 ballast water disposal treatment, 11, 231 ballast water program, 231 ballast water standards, 229 bang-bang approach, 40 bang-bang control, 8, 33, 108–114 bang-bang cycle, 102, 108, 109, 111 banning, 214, 219 banning policy, 219 Barents Sea, 64 Bayesian estimation methods, 59, 70 Belgium, 192, 193 Bellman equation, 67, 143 Bendixson–Dulac criterion, 99 Beverton and Holt model, 39 bifurcation analysis, 96 biodiversity loss, 51 bioeconomic equilibrium, 31, 32, 36, 41 bioeconomic modeling, 3, 8, 61, 71, 91, 133 bioeconomic models, 2, 6, 15, 21, 29, 30, 54, 62, 70, 71, 117, 119, 133, 137, 138, 201

biologists, 77, 94, 95, 210 birds, 98–100, 102–110, 113–115, 149–154, 156–163, 165–169, 171–173, 176, 177, 179–181, 186, 189, 210–214, 218–221, 223–225, 227 Bloemendaal, 234, 235, 238, 239 blue mussels, 1 boundaries, 82, 83 Bowhead whales, 60 Brazilian sardine fishery, 61 Britain, 87 bury inlet, 149 calibration, 6, 17, 24, 25, 149 capacity use, 3 capelin, 64 capital invested, 47, 48, 73 capital theory, 36 carrying capacity, 20, 31, 41, 94, 119, 122, 125, 130–134 catastrophic events, 54, 55, 139, 140 catch per unit effort (CPUE), 57 catchability, 30, 32, 35, 36, 42, 49, 58, 61, 119 cellular automata, 150 centralized government control, 75 characteristics of integrated models, 15, 20 civil servants, 210 classification of economic damages, 231 climatic change, 15, 19 climatological changes, 54 coastal nations, 74 cockle fishery, 10, 22, 70, 118, 185, 186, 188, 190, 194, 195, 198, 199, 202, 203, 205, 206, 209–212, 214, 217, 219, 222–224 cockle landings, 189 cockle-fishermen, 219, 221, 223 cockles, 1, 2, 9, 10, 149, 151, 154, 155, 159, 164–169, 171, 177–179, 185, 186, 189, 194, 197, 202, 205, 206, 209, 210–213, 216–219, 223, 224, 226 cod, 58, 62, 64, 68, 85, 140 cohort, 39, 40

257

258

INDEX

co-management, 7, 10, 75, 78–81, 83, 84, 89, 90, 186, 207 common pool resources, 4 compensating variation, 215 competition, 43, 44, 73, 74, 119, 121, 134, 149, 168, 193, 194, 195, 199, 201 complex models, 23–25 complicated models, 23, 26 concentration profile, 35–37 conceptual motivations, 222 conditional logit model, 217 conflict, 1, 6, 10, 17, 18, 19, 21, 85, 86, 87, 88, 93, 185, 186, 191, 199 conservation, 1–10, 63, 64, 74, 75, 77, 81, 89, 93, 94, 113, 114, 133, 138, 140–143, 146–149, 185, 187, 188, 196, 197, 206 conservation of biodiversity, 197 conservationists, 10, 19, 56, 186–188, 191, 197, 198, 200, 205–207 consultants, 187, 200 consultation framework, 210 contingent valuation (CV) method, 212, 215, 229–231, 236, 238, 240 continuous recruitment function, 126 Cormas, 150 cost function, 32, 109, 110 cost of fishing, 8, 32, 41, 42, 43, 55, 56, 63, 65, 66, 102, 103, 110, 114, 133 Council of State, 208 critical, 36, 37, 65, 88, 93, 115, 225 Danish, 75, 76, 149, 161, 196 decommissioning schemes, 201 demand curve, 46, 47, 235 demersal and flatfish species, 203 demographic, 54, 55, 59, 63, 68, 71, 233, 234 Denmark, 1, 75, 185 density of fish population, 35 depensation, 6, 36, 37, 39 descriptive approach, 22 design matrix, 175, 176 deterministic economic models, 29 DIALOGUE, 24 dichotomous-choice elicitation question format, 236 differences between individual birds, 180 digestive bottleneck, 163, 180 discount rate, 8, 114 discrete density-dependent recruitment function, 121 discrete optimization problem, 118, 134 discrete time models, 37 dispersal, 6, 40, 41, 55, 123, 159 dredging, 54, 197, 202 drift, 66

Dutch Bird Conservation Society, 187 Dutch citizens, 209, 217–221, 224–226 Dutch Natural Gas Company (NAM), 207 Dutch province of Zeeland, 185, 186 Dutch shellfishery, 141, 199 dynamic optimization, 3, 108 dynamics, 2, 6–9, 17, 18, 21, 22, 30, 33, 35, 40–42, 53–60, 62, 63, 65–71, 73, 77, 81, 82, 89, 93, 94, 117–120, 124, 127, 128, 132–134, 137, 138, 140–143, 147–150, 153, 156, 161, 168, 171, 174, 179–181, 191, 201, 204–206 EC Bird Directive, 196 ecological sustainability, 3, 187, 197, 198, 211 ecological uncertainty, 54, 55–58, 62, 65, 67, 90 ecological values, 185, 186, 188, 196, 200, 205, 207 ecologically related species, 42, 93 econometric robustness, 224 economic incentives, 51, 76, 79, 80 economic uncertainty, 55, 62 economic valuation method, 210 ecosystem behavior, 96 effect of effort on recruitment, 121, 125, 126, 130, 131 effect of effort on shellfish habitat, 131 efficiency, 2, 3, 9, 66, 67, 77, 110, 114, 162, 181, 189, 236 effort, 3, 22, 23, 26, 30–37, 39, 43, 44, 45, 49, 50, 52, 53, 57, 61, 62, 64, 66–68, 71, 75, 84–86, 93, 117–121, 124–127, 130–134, 137, 139, 141, 203, 204, 221 Eider ducks, 1, 163 emigration of birds, 157, 171 empirical models, 20, 21 employment, 4, 56, 77, 83, 195, 198, 199, 201 end state, 8, 94, 101–103, 106, 107, 108–111, 113–115 energy intake, 9, 94 enforcement and monitoring, 76, 188 English fishermen, 87 entrance fee, 221, 231 entry and exit behavior of fishermen, 66 environmental economics, 212 environmental organization, 81, 185, 208, 209 environmental variance, 55 environmentalist’s pressure, 56 escapement, 38, 49, 67 EU, 7, 8, 75, 76, 80, 83, 84, 85, 86, 87, 88, 89, 90, 196, 203, 207 EU law, 88 European Community (EC), 78 European Council of fishery ministers, 85 EVA II, 196, 209, 210 evolutionary models, 3, 4, 52

INDEX Exe estuary, 149 exotic species, 2, 239 exotic, harmful algae, 2 experimentation, 2 externalities, 2, 4, 18 extinction, 16, 18, 35–37, 39, 44, 51, 55, 71, 82, 93, 107 extreme policy scenario, 214 factor analysis, 222, 227 factor loadings, 222 feedback, 6, 16, 70 feeding time, 162, 163, 180 fields of sea grass, 190 financial costs, 230, 236 fisheries habitat, 68 fisheries law, 196 fisheries management scenarios, 81 fishery effort, 3, 31 fishery regulation, 32, 48, 49, 52 fishing, 1, 6–10, 26, 29, 30, 32, 33, 35–43, 45, 46, 48, 49, 50–57, 61–71, 73–90, 97, 99, 100, 102, 103, 107, 109, 110, 111, 114, 117–122, 124–131, 132–134, 137–142, 146–148, 150–152, 155, 168, 176, 185–191, 193–213, 217, 219, 221, 222, 224–226 fishing by hand, 209 fishing capacity, 73, 84, 85–87, 186, 188, 198, 199, 201–203, 205, 206 fishing gear restrictions, 52 fishing industries, 29 fishing pressure, 9, 57, 61, 64, 82, 83, 85, 86, 137, 140, 141, 147, 148, 185, 186, 189, 200, 201 fishing rights, 65, 66, 67, 77, 206 fishing sector, 7, 8, 56, 69, 70, 71, 74, 84, 87–90, 185, 187, 188, 190, 191, 193–195, 198, 199, 201, 202, 204, 219 flagellates, 229 flatfish fishery, 85, 204 food intake by birds, 151, 153 foraging behavior, 149, 188 France, 74, 149, 192, 221 Friesland, 215 full factorial design, 213 functional response, 94–96, 153, 165–167 gastrointestinal disorders, 230, 231 Gause model, 44 generalized TC, 233, 235 genetic inbreeding, 71 genetic uncertainties, 55 geometric Brownian motion, 66 German, 149, 196 golden rule, 6, 33, 34, 37, 38, 43, 45–47, 63, 93, 104, 107, 121, 124, 134, 146

259

Gordon–Schaefer model, 30, 32, 35, 36, 39, 41, 44, 45, 47, 50, 51 government, 7, 8, 10, 32, 50, 65, 67, 70, 73–84, 86, 89, 90, 141, 186–188, 191, 195–197, 200, 201, 203–209, 211, 232, 236 Green’s theorem, 99 grid cell, 150–157, 159, 167, 169, 171, 177, 180, 181 Groningen, 215 Habitat Directive, 196 habitat effects, 117 Hamiltonian, 33, 49, 100, 101, 110, 114, 125, 134 handling time, 8, 99, 118, 158, 159, 169, 170, 172, 182, 183 harmful algal blooms (HABs), 10, 11, 229–231, 233, 239, 240 harp seals, 64 harvesting, 6, 8, 16, 33, 34, 37, 39, 40, 43, 48, 50, 57, 62, 63, 67–70, 93, 94, 96, 100–105, 107, 109–111, 113–115, 146 harvesting cost, 33, 63, 100, 110 herbivore–plant systems, 94 herring, 39, 64, 140 Holling’s disc equation, 95 Holling’s four-box model, 4 homoclinic orbit, 96 Hopf bifurcation theorem, 108 Hotelling rule, 34 human health, 11, 229–231 hyperdepletion, 58 hyperstability, 58 immediate recruitment, 123 immigration, 151, 152, 159 implementation error, 57 income, 5, 94, 104, 110, 114, 216, 232–235, 237–239 indirect effect, 104, 117–119, 121, 125, 130, 140 individual transferable quotas (ITQ), 51, 65, 67, 77 information, 4, 7, 10, 15, 30, 51, 53, 54, 57, 59, 60, 64, 69, 70, 74, 77, 84–86, 89, 114, 134, 137, 140, 146, 150, 161, 164, 165, 201, 211, 212, 216, 219, 227, 230, 233–235, 237, 238 instruments, 4, 78, 204, 217 integrated assessment, 1, 3, 6, 15, 19 integrated models, 6, 15, 17, 18, 19, 20, 24–26, 29 interest groups, 2 International Whaling Commission, 56, 60 investment, 3, 6, 47, 48, 66, 67, 73, 198 irreversible, 2, 4, 18, 23, 48, 69, 197, 209 islands, 1, 156, 185, 210, 215 Ito’s lemma, 67, 143

260

INDEX

juvenile growth, 41 juvenile stage, 117 laisser faire, 3 larvae, 41, 42, 64, 95, 119, 121, 123, 128, 141, 150, 204 latent underlying motivations, 215 law on nature protection, 196 level, 3, 4, 7, 8, 16, 17, 19, 20, 21, 24, 27, 31–34, 36–40, 44–46, 48, 49, 51, 57, 62, 63, 67, 69, 75, 76, 78, 82–90, 102, 107, 118, 119, 121, 124, 130, 137, 141, 143, 158, 168, 169, 177, 196, 199, 208, 212–214, 218–221, 224–226, 236 life stages, 54, 61, 117, 139, 140 Likert-scale, 222 limit access, 139 limit cycle, 96, 99, 102, 103, 108, 109, 113 lobby groups, 210 lobsters, 50, 51 local and family traditions, 77 local fishing communities, 87 local politicians, 210 local populations, 40, 68 local residents, 10, 209, 211, 215–220, 223, 224, 226, 227 Lotka-Volterra model, 43 low interest loans, 73 Lyapunov stability, 97 management options,, 69, 181, 211 mandatory, 201 marginal WTP, 219, 220, 221 marine habitat, 53, 54 marine protected areas, 201, 204, 205 Marine Reserve Creation, 8, 9, 68, 137, 138, 141, 147 marine reserves, 9, 16, 18, 68, 137–140, 147, 188, 204, 206 marine tourism, 230 market intervention, 7, 77 Markov Chain Monte Carlo method (MCMC), 60, 61 mass mortalities, 181 mass/handling time ratio, 154, 167 matrix of factor scores, 222 maximum harvesting, 101, 102, 109, 110 maximum likelihood techniques, 60 maximum sustainable frontier, 44 maximum sustainable yield, 31, 32, 65 mechanized cockle fishery, 185, 190, 199, 205 mechanized fishing techniques, 185 Meijer committee, 207 MERGE, 24 meta-analysis, 59, 60 metapopulations, 148

migratory species, 140 minimum fish size rules, 50 minimum population abundance requirement, 138, 142 Ministry of Agriculture, 216 Ministry of Environment, 216 Minke whales, 64 mixed lognormal model, 238 models of renewable resources, 120 monetary valuation, 1, 2, 5, 6, 9, 11, 183, 186, 188, 190, 191, 194, 196, 198, 200, 202, 204, 206, 208, 210, 212, 214, 216, 217, 220, 222, 224, 226, 230, 232, 234, 236, 238, 240 monitoring and enforcement costs, 196 monitoring system, 4 mono-disciplinary models, 17 Multi Annual Guidance Programs (MAGP), 84–86, 203 multilevel governance, 56 multiple use, 4, 43, 149 multiple use conflicts, 4 multispecies fisheries, 2, 63 multivariate regression, 240 mussel beds, 185, 190, 197, 205, 206 mussel fishery, 186, 198 mussel seed, 186, 187, 188, 189, 191, 193, 195, 201, 202, 205, 206 mutational meltdown, 55 mutualistic systems, 43, 44 myopic behavior, 34 national property regime, 74 National Union for the Conservation of the Wadden Sea, 187 natural equilibrium, 99 natural scientists, 10, 209–211, 215–220, 223, 226, 227 natural system, 16, 25 nature monuments, 187 nature policy, 4, 8 Netherlands, 1, 10, 11, 70, 76, 149, 185, 192, 193, 198, 199, 208–210, 216, 223, 229, 232, 240 Netherlands Organization of Scientific Research, 216 neurotoxins, 230 newborn, 152 no choice, 214 NOAA guidelines, 231 no-fishing season, 117, 121, 122, 126, 127 non-critical depensation, 36 non-targeted species, 53, 56, 138, 197 non-use value, 5, 68, 93, 139, 230, 231 Noord-Holland, 215 North Sea, 1, 2, 9, 76, 85, 139, 180, 189, 194, 230 North Sea fisheries, 76

INDEX North-Sea herring, 39 North-East Atlantic Cod, 68 objectives of the stakeholders, 188 observation errors (OE), 57, 237 on-site expenditures, 239, 248 Oosterschelde, 149 open-access, 6, 29–37, 41, 49–51, 65, 73, 74 open-access fishery, 30, 49, 50, 65 opportunity cost, 50, 105, 121, 133 opportunity cost of waiting, 105 optimal harvesting, 37, 69, 94, 100, 104, 113 optimization models, 3, 22 ownership rights, 206 oyster population, 185 oystercatchers, 1, 9, 26, 149, 151, 154, 156, 158, 161–173, 177–181, 210 payment mechanism, 231, 236 perturbation, 69 Plackett–Burman design, 9, 175, 176 Plaice Box in the North Sea, 139 Poisson process, 63 polder-model, 210 policy evaluation, 22 policy failure, 10, 185, 192 policy makers, 10, 110, 134, 204, 209–211, 215–220, 223, 226, 227 policy measures, 212 policy optimization, 6, 15, 22 policy-orientation, 2 policy process, 15, 17, 23 political pressure, 70 population abundance requirement, 138, 142, 143, 147 population dynamics, 2, 6–9, 18, 40–42, 53, 54, 56–59, 60, 62, 63, 65, 67, 69, 70, 71, 73, 77, 82, 117, 118, 120, 124, 127, 128, 133, 134, 137, 138, 140, 141, 143, 147, 148, 150, 156, 161, 168, 180, 204, 205 port authorities, 229, 230 positive spillover, 68 precaution, 2, 10, 207 precautionary approach, 3, 70, 71, 204, 207 precautionary principle, 69, 139, 196, 203, 206, 207 predator–prey system, 8, 43 predators, 8, 43, 93, 94, 96, 104, 114, 152, 153, 159 prescriptive method, 21 present value, 32, 34, 37–39, 42, 44, 48, 62, 63, 65–67, 119, 143, 221 pretests, 214 prey, 1, 8, 9, 43, 44, 54, 64, 93–97, 107, 113, 114, 149, 150–155, 157, 159, 162, 164, 166–168, 173, 177, 178, 180, 181

261

price change, 45 price uncertainty, 66 pristine area, 140 pristine state, 114 Producers Organizations (PO), 80, 187, 197, 202, 226 production of chicks, 161 profit, 6, 20, 30–32, 34, 36, 37, 40, 44, 45, 48, 54, 62, 65, 66–68, 70, 71, 75, 77, 81, 102, 104, 108, 118–121, 124, 133, 134, 138, 141–143, 149, 191, 192, 194, 198, 199, 205 profitable business, 198 project specific warm glow, 225 property rights, 7, 29, 32, 74, 77, 78, 89, 230 protest responses, 236 provinces, 210, 215 quality of the ecosystem, 212, 213, 223 questionnaire, 11, 210, 212, 215, 216, 222, 227, 229–232, 238, 239 quota, 4, 6, 16, 19, 26, 48–52, 65–67, 71, 76, 84–88, 110, 137, 202, 204, 211, 213, 218, 219, 224, 226 quota policy measures, 211 Raad van State, 208 Ramsar convention on Wetlands, 196 random utility (RU) model, 214, 217 ranking of policies, 227 real options theory, 66 recolonization, 55 recreation demand function, 233 recruitment, 37, 61, 117, 121, 125, 126, 133, 151, 155, 168, 176, 179 recruitment density, 120, 122, 129, 134 reducing fishing effort, 75, 139 regional and local concerns, 76 regulation of cockles fishery, 213 researchers, 61, 187, 188, 200 reserve, 8, 9, 41, 68, 117, 137–143, 146–148, 186, 204, 238 reserved rationality, 69 resilience, 3, 4, 6, 20, 22, 69, 137 resource damage assessment, 230 Ricker stock recruitment curves, 128 risk premium, 63, 66, 69 risk-averse heterogeneous fishermen, 66 rotation, 10, 211–213, 218–220, 222–224, 226, 227 rotation policy measures, 211 Rotterdam harbor, 11, 231 SC attributes, 213 SC question, 212, 213, 219, 223 scale, 8, 25, 40, 42, 51, 53, 54, 56, 73, 76, 77, 82, 83, 85, 88, 90, 140, 156, 180, 204, 215, 222, 237, 239

262

INDEX

Schiermonnikoog, 215 schooling fish, 36, 39 scientific interests, 140 scientists, 10, 15, 17, 57, 74, 75, 77, 79, 88, 187, 200, 201, 209, 210, 211, 215, 216, 217, 218, 219, 220, 223, 226, 227 Sea and Coastal Fisheries Policy (SCFP), 10, 10, 185, 186, 187, 188, 189, 190, 191, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 204, 205, 206, 207 seagrass, 185–191, 206 seasonal changes in climatic conditions, 181 second-best policies, 4 sedentary species, 2, 8, 117, 137, 140, 204 seed spatfall, 191, 205 selection processes, 4 sensitivity analysis, 9, 26, 150, 174–178, 181 shadow price, 50, 100–102, 105, 106, 108 shellfish, 1, 2, 8–10, 19, 41, 42, 70, 94–111, 117–201, 205, 207, 210, 212, 231 shellfish bed, 117–121, 123, 125, 132, 153, 159 shellfish habitat, 118, 132 shellfish stock sustainability, 186, 198 shellfish-processing sector, 195 shorebirds, 2 simulation grid, 151, 176 singular state, 102, 109, 110 sink, 8, 9, 141, 142, 146–148 skin irritations, 230 social costs, 10, 65 social welfare function, 62, 65, 94, 100, 110, 114 socio-economic system, 16, 19 sole-owner model, 30, 33, 44, 47, 48 source, 8, 9, 57, 59, 60, 77, 128, 141, 142, 146–148, 157, 163–165, 176, 189, 190, 192–199, 202, 203, 209, 230, 233 sovereign rights, 74 space, 2, 6, 8, 17, 18, 21, 42, 99, 117, 118, 121–129, 133, 134, 150, 156, 171, 174, 180, 235 Spain, 74, 87, 194, 199 Spanish, 87, 192–195, 210 spatial–temporal model, 149 spawning, 38, 53, 61, 67, 82, 85, 117, 121, 123, 131, 138, 139 spawning stock biomass (SSB), 85, 86 special protection area, 196 species interaction, 2, 43, 54, 149 Spisula, 1, 194, 198 Spisula fishery, 194, 198 stability, 4, 18, 22, 44, 69, 84–88, 93, 96–99, 108, 138, 161 stability analysis, 96, 97, 108 stages in political process, 17 stakeholders, 3, 5, 10, 15, 19, 76, 78, 186–188, 206–211, 213, 215, 217, 219, 220, 223–227

stakeholder analysis, 3, 5, 6, 9, 183, 186, 188, 190, 194, 196, 198, 200, 202, 204, 206, 208, 210, 212, 214, 216, 220, 222, 224, 226, 230, 232, 234, 236, 238, 240 stated choice method, 10, 211 stated preference techniques, 5 statistical parameter inference, 60 stochastic differential equation, 62 stochastic recruitment, 150, 178, 179, 181 stochastic systems, 23 stock, 3, 6, 9, 16, 19, 29, 32, 34–39, 41, 43, 44, 54–71, 75, 85, 93, 95, 100, 104, 105, 107, 109, 114, 117–121, 128, 129, 133, 138, 139, 146–148, 150–155, 168, 169, 171, 176, 179, 181, 185, 186, 198, 200–202 stock assessment models, 57, 70, 71 stock effects, 104 stock–recruitment relationship, 61, 62, 71 stock sustainability, 3, 186, 198 sublittoral stock, 202 subsidies, 73, 79, 80, 83 suction dredges, 185, 197, 199, 202 sudden collapse, 37 summer, 150, 151, 152, 155, 156, 168, 177, 178, 215, 238 supply curves, 46 supra-national management, 88 survey, 6, 29, 85, 209, 210, 212–216, 221, 227, 229–232, 240 sustainability, 2, 3, 10, 23, 69, 71, 74, 77, 186, 187, 195, 197, 198, 203, 204, 206, 207, 211 sustainability: ecological vs. stock, 3, 198 sustainable fisheries, 3, 7, 78, 80, 89, 133, 201 tax benefits, 73 taxes, 4, 6, 50, 235 technological advances, 56, 73 technological progress, 4 Terschelling, 215 testing for warm glow effect, 222 theoretical models, 3, 20, 21, 24 threshold, 23, 36, 48, 157, 158 time scales, 2 timeframe, 82 timing of adopting policy, 64 total allowable catch (TAC), 49, 64–67, 84–86, 186, 195, 199, 201, 202 total cost (TC), 30–32, 36, 102, 120, 226, 229, 230, 233, 234, 235, 238, 240 total direct and indirect marginal effect, 125, 130, 131 total revenue, 10, 209, 211, 215–227, 232 total value, 230 tourists, 11, 218, 220, 224–236, 241 trace condition, 108

INDEX tradeoffs, 212 transboundary nature, 88 travel cost models, 5 travel time, 230, 232–235, 239 Trilateral Wadden Sea Plan, 196 tropical fisheries, 138 Turnbull estimator, 238 type II functional response, 94–96, 167 typology of uncertainty, 54 UK, 86, 149 uncertainty, 1, 3–9, 17, 19–21, 29, 30, 53–74, 79, 82, 85, 89, 90, 137, 139, 142, 147, 148, 185, 188, 201, 204–206 uncertainty in modeling, 56 United Nations Convention on the Law of the Sea (UNCLOS), 74, 81, 89 Uruguayan tuna fishery, 64 user-groups, 7, 8, 78–90 validation, 6, 17, 24–26, 181 varimax rotation estimation procedure, 222 vertical integration, 199

263

vessels, 3, 47–49, 56, 73, 86, 87, 194, 199, 201, 202 violating regulations, 76 voluntary agreements, 79 voluntary schemes, 201 Wadden Sea (Waddenzee), 1, 2, 8–10, 21, 94, 95, 118, 149, 155, 156, 161, 163, 173, 180, 181, 185–191, 194–196, 198–227 Waddenvereniging, 209, 211 warm glow, 10, 210, 215, 222–227 Wash, 149 welfare measure, 230 Westerschelde, 149 Wiener process, 62, 63, 66, 68, 143, 145 Wilde Kokkels, 211 willingness to pay (WTP), 212, 214, 215, 219–221, 223, 226, 227, 230, 231, 236–240 winter, 150–169, 173, 177, 178, 186, 188, 238, 239 winter bird population, 188 win–win situation, 205, 206 WTP function, 239 yield–effort curve, 31