Towards Ecosystembased Management: Integrating

Towards ecosystembased management: Integrating stakeholder values in decisionmaking and improving the representation of ecosystems in ecosystem models

Maria Jose Espinosa Romero

B.Sc., Universidad Marista, 2002

A THESIS SUBMITED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF

Master of Science

The Faculty of Graduate Studies

(Resource Management and Environmental Studies)

The University of British Columbia

(Vancouver)

August 2010

Abstract

Ecosystem‐based management (EBM) is increasingly seen as the new paradigm for managing the use of marine resources and ecosystems. Although EBM has been defined in theory, its implementation has faced challenges worldwide. This research aims to examine two approaches to contribute to the operationzalization of EBM by incorporating stakeholder values in the decision‐making process, and by better representing ecosystem dynamics in ecosystem models. First, I illustrate a decision‐making framework for EBM rooted in structured decision‐ making (SDM), a well‐known systematic approach for planning and stakeholder‐consultation processes. SDM helps to identify the values of the constituents and define objectives and indicators that are consistent with those values. I demonstrate how SDM can enable managers to evaluate the performance of management alternatives using indicators specifically chosen to reflect values. This can help managers make more systematic, transparent and informed decisions with respect to the use of marine resources. As a case study, I apply SDM to the marine planning process on the west coast of Vancouver Island (WCVI). Second, as ecosystem models play an important role in EBM, I strive to improve the representation of marine ecosystems using ecosystem models in Ecopath with Ecosim (EwE). I focus on incorporating mediating effects and species reintroductions, both common situations that can strongly influence ecosystem dynamics. These considerations are essential when applying holistic approaches to management but they are not generally included in EwE. I use EwE to model the reintroduction of sea otters (Enhydra lutris) and the mediating effects provided by kelp forests in nearshore ecosystems of the WCVI. Because EwE does not have the functionality to represent reintroductions, I created two scenarios to work around the assumptions of Ecospace on the initial state of the ecosystem. In addition, I demonstrate how mediating effects can be represented using the ‘mediation’ function in Ecosim. These methods and results can contribute to advance EBM on the WCVI and offer insights to other marine planning processes. Both strengths and limitations of this work are presented and analyzed.

Table of contents

Abstract...... ii

Table of contents ...... iii

List of tables...... v

List of figures...... vi

Acknowledgements ...... vii

Co‐authorship statement ...... viii

Chapter 1 Introduction...... 1 1.1 Problem statement...... 1 1.2 Research objectives...... 2 1.3 Thesis outline ...... 3 1.4 Background...... 4 1.4.1 Study area...... 4 1.4.2 Marine planning process on the WCVI...... 6 1.4.3 Decision‐making process for EBM ...... 6 1.4.4 Ecosystem models for EBM: Ecopath with Ecosim (EwE)...... 7 1.5 References...... 11

Chapter 2 Using structured decision‐making for ecosystem‐based management...... 14 2.1 Introduction ...... 14 2.1.1 A need to improve decision‐making processes...... 15 2.1.2 The structured decision‐making process...... 16 2.1.3 Some considerations for the implementation of SDM for EBM...... 19 2.1.4 Case study: Marine spatial planning for the west coast of Vancouver Island...... 20 2.2 Methods...... 21 2.2.1 Applying SDM to design a decision‐making framework for EBM ...... 21 2.3 Results and discussion...... 22 2.3.1 New structure: Fundamental objectives, attributes and indicators...... 23 2.3.2 Insights from the process and applicability for EBM...... 31 2.4 Conclusions...... 33 2.5 References...... 39

iii

Chapter 3 Representing mediating effects and species reintroductions in Ecopath with Ecosim…………...... 43 3.1 Introduction ...... 43 3.1.1 Case study...... 44 3.2 Methods...... 46 3.2.1 Ecopath parameterisation...... 46 3.2.2 Modeling benefits provided by kelp forests: using Ecopath and Ecosim...... 46 3.2.3 Spatial representation of sea otter reintroduction and expansion: Ecospace structure ...... 48 3.3 Results and discussion...... 50 3.3.1 Ecopath results...... 50 3.3.2 Benefits provided by kelp forests over time...... 51 3.3.3 Spatial representation of sea otter reintroduction and expansion...... 53 3.4 Conclusions...... 55 3.5 References...... 67

Chapter 4 Conclusions...... 70 4.1 Discussion ...... 70 4.2 Strengths, limitations and improvements...... 72 4.3 References...... 74

Appendices

Appendix A. Parameters estimation and model balancing …………………………………………………….75

List of tables

Table 2‐1 Original set of objectives and attributes of WCA, based on draft goals for aquatic management on the west coast of Vancouver Island (WCVI)...... 35 Table 2‐2 New structure of fundamental objectives, attributes and indicators for a decision‐ making framework for EBM...... 36 Table 3‐1 Basic inputs for Ecopath...... 58 Table 3‐2 Diet compositions...... 58 Table 3‐3 Results from 100 runs of the Individual‐based model (IBM)...... 59

List of figures

Figure 1‐1 West Coast Aquatic (WCA) management area...... 10 Figure 3‐1 Reintroduction site, current range and sites with optimum habitat for sea otters based on intertidal complexity similarity with the reintroduction site ...... 60 Figure 3‐2 Food web of the nearshore ecosystem on the WCVI...... 60 Figure 3‐3 Culling scenario, with one suitable habitat...... 61 Figure 3‐4 Mixed trophic impacts. Relative impacts between species due to predation and competition...... 61 Figure 3‐5 Results of kelp‐derived detritus biomass for the two scenarios...... 62 Figure 3‐6 Results with and without the ‘mediation’ function. Species biomass over a 100‐year period...... 63 Figure 3‐7 Spatial distribution of sea otters using the Culling scenario and the two types of movement models: diffusion and individual based model (IBM)...... 64 Figure 3‐8 Populations dynamics in a 100 year period...... 65 Figure 3‐9 Sea otter recovery with Ecosim, Diffusion and Individual based Model (IBM)...... 66

Acknowledgements

I want to thank my supervisor Kai Chan for inviting me to be part of his lab and this amazing project and for supporting me in many ways during my studies. I also thank my committee members, Villy Christensen for explaining with patience all about ecosystem modeling and for facilitating the improvements in the software so I could do my model; Tomas Tomascik for providing a practical perspective to my research and valuable contributions about the region and the ecosystem; and Rashid Sumaila for always giving practical advise as well as motivation to complete my studies.

I also thank West Coast Aquatic (WCA), especially to Denise Dalmer for her time and important contributions to my structured decision‐making chapter, and to Tom Okey for his interest on my work, time and invitations to share this work with the staff and stakeholders in the Island.

I feel grateful to have had the opportunity to collaborate with my professors Carl Walters and Tim McDaniels for the main two chapters of my thesis. Their classes and comments on my work were fundamental. I would like to appreciate Steve Martell and Terre Satterfield for having the disposition to help and answer all my questions and from whom I learnt a lot.

I am very thankful to Ed Gregr, Russ Markel, and Becca Martone who supported my whole process with time, interesting discussions, information, and good advise; to Andres Cisneros, Jordan Levine and Megan Bailey for kindly reading and commenting on my documents, to the my lab colleagues for their support and feedback on my work, and to Villy Christensen’s lab, specially Jeroen Steenbeek and Sherman Lai who helped me with the software.

I also thank the IRES department and my sponsors –Consejo Nacional de Ciencia y Technologia (CONACYT), World Wildlife Fund (WWF), The Nature Conservancy and the Faculty of Graduate Studies (FoGS) for making this research possible.

Finally and most importantly, I want to thank and dedicate this research to my family and friends, who were always there, in good and stressful times. Their love and support encourage me to successfully complete this process.

vii

Coauthorship statement

Maria Espinosa Romero designed and performed the research and data analysis of this thesis, prepared the text for the manuscripts, and will be the first author on the publications derived from this work. Several colleagues have contributed to the chapters and the preparation of manuscripts. A version of Chapter 2 will be submitted for publication with Kai Chan (supervisor) and Tim McDaniels, who provided significant contributions regarding Ecosystem‐ based Management (EBM) and decision‐making, respectively. A version of Chapter 3 was presented at the “Ecopath 25 Years” conference and published in the Proceedings with Edward Gregr, Villy Christensen, and Kai Chan. The final version of Chapter 3 will be submitted for publication with Edward Gregr, Villy Christensen, Carl Walters, and Kai Chan as coauthors. Edward Gregr assisted in the development of the first versions of the theoretical model and contributed to the preparation of the manuscript; Villy Christensen (graduate research committee) and Carl Walters participated in model development, results analysis and manuscript revision; and Kai Chan provided important thoughts for the representation of ecological dynamics and the implications/limitations of the model and helped with the manuscript revisions.

viii

Chapter 1 Introduction

1.1 Problem statement

Ecosystem‐based management (EBM) has been suggested as a promising new approach to management, which generally has focused separately on single species and single sectors with insufficient treatment of the challenges facing by marine ecosystems (Lester et al. 2010). EBM considers the linkages between all elements of the ecosystem, including humans (McLeod et al. 2005), as well as the underlying processes that produce the services people want and value (Guerry 2005, Arkema et al. 2006), and the meaningful involvement of stakeholders in management efforts (Guerry 2005, Leslie and McLeod 2007).

Although the principles, key elements and guidelines for EBM have been defined (e.g., McLeod et al. 2005, Leslie and McLeod 2007, Lester et al. 2010) there is still a gap between theory and practice (Arkema et al. 2006, Ruckelshaus et al. 2008, Lester et al. 2010). This is at least partly a symptom of insufficient attention being paid to stakeholder values and decision‐making processes for marine management. Some studies suggest that the main challenges for EBM include: 1) building a collective vision and set of objectives for EBM; 2) defining metrics to evaluate the accomplishment of objectives; 3) designing governance frameworks (Leslie and McLeod 2007); and 4) translating theoretical concepts to operational goals (Arkema et al. 2006). Even when EBM has been explicitly accepted at the management level (Rosenberg and Sandifer 2009), managers still face political, legal, social and scientific difficulties in implementing the complex academic concepts of EBM, which have come to be seen as daunting and expensive (Arkema et al. 2006, Lester et al. 2010, Tallis et al. 2010). Managers require practical conceptual frameworks to develop policies, objectives and goals for effectively implementing EBM (Rosenberg and Sandifer 2009).

As we shift to ecosystem‐based approaches to management, the use of tools capable of representing the possible outcomes of management policies becomes essential (Plagányi 2007). In this sense, ecosystem models play a fundamental role for EBM, as they can integrate and display information pertaining to how ecosystems work (including multiple species and

habitats) and respond to anthropogenic and natural disturbances (Christensen and Walters 2004a, 2004b). Ecopath with Ecosim (EwE) is the most widely used tool to represent marine ecosystems (Christensen and Walters 2004a, 2004b). It has been mainly used for ecosystem‐ based fisheries management (EBFM) to explore fishing impacts and policy scenarios related to fisheries (Plagányi and Butterworth 2004). To expand its use for EBM, EwE models need to represent other ecosystem dynamics such as indirect species interactions called ‘mediating’ effects1 provided by species and the effects caused by species (re)introductions, both of which have shown to have strong influence on ecosystem dynamics (e.g., Grosholz et al. 2000, Dill et al. 2003). Mediating effects can, for example, intensify or reverse trophic interactions, create and expand niches and affect several species (Dill et al. 2003). If such effects are overlooked, as they usually are (Dill et al. 2003, Heithaus et al. 2008)2, ecosystem models will fail to accurately represent the ecosystem (Dill et al. 2003).

Another important aspect that needs to be incorporated into EwE models is the range of human activities that take place in the ocean—not only fishing—which can cause or be impacted by changes in ecosystems.

In light of the recent adoption of EBM as a central management paradigm, as well as the challenges facing its implementation, more research is needed to better formalize how EBM can be successfully used for decision‐making. My research intends to contribute to this gap by demonstrating: (1) how to incorporate EBM into the decision‐making process and (2) how to improve the performance of ecosystem models to advise EBM. As a case study, I use the EBM initiative on the west coast of Vancouver Island (WCVI). A multiple‐stakeholder body, West Coast Aquatic (WCA), is currently leading a marine planning process with an EBM approach in the region. Collaboration with this group has constituted an excellent opportunity to explore the central aspects, and real‐world applications, of my research.

1.2 Research objectives

This project aims to help bridge the gap between theory and practice by investigating and

1“When a change in a property of one species (the initiator) causes a change in the behavior of a second species (the transmitter) and this change in the transmitter has an effect on a third species in the community (the receiver)” (Dill et al. 2003). 2 Heithaus et al. (2008) is specifically on risk effects of mediation impacts on top predators. 2

testing how the EBM approach can be applied to decision‐making for marine management. The two main objectives of this research are as follows:

• Demonstrate how to develop a decision‐making framework for EBM, which can be used to set policy in a systematic and transparent way. This framework integrates the values of the constituents and scientific advice in the evaluation of management alternatives. As a case study, I have used the ongoing marine planning process on the WCVI led by WCA.

• Demonstrate how to incorporate mediating effects and species reintroductions into EwE to improve the representation of ecosystem dynamics–which is essential for EBM. To represent such dynamics I have modeled nearshore ecosystems on the WCVI, where sea otters (Enhydra lutris) have been reintroduced and kelp forests are providing mediating effects to species in the ecosystem. In addition, I have assed the scale of the impacts (on population biomass) of such dynamics on the ecosystem and the ability of EwE to represent them.

This project intends to support the marine planning process on the WCVI and to shed insight for marine planning processes and EBM elsewhere. In addition, this research is part of a bigger project called “British Columbia Coastal Ecosystem Services”, the main objective of which is to develop interdisciplinary research on ecosystem services to assist decision‐making and EBM in coastal British Columbia. The main components of this research include fieldwork, characterization of ecosystem services, modeling, valuation and the integration of all these components into decision‐making processes. This broader project is done in collaboration with stakeholders and in connection with ecosystem‐services research in North America.

1.3 Thesis outline

To address the two main objectives, my research is organized into four chapters. Chapter 1 (this chapter) gives an introduction to the problem, my research objectives, background on the region where this project is applied, and the regional marine planning process. It also explains what structured decision‐making (SDM) is and why it can be used to facilitate EBM. To finalize, as I focused on EwE, I address its strengths and limitations to represent how ecosystems work.

Chapter 2 suggests how SDM can be used to create a decision‐making framework for EBM. This chapter shows the importance of having a systematic process to identify the values of the constituents and bring consistency between these values, objectives and indicators for the evaluation of management alternatives. This process was applied to the marine planning process on the WCVI. As WCA had already stated its objectives, I show how to re‐structure these objectives, without losing or distorting their intended meanings and purposes, to make them operational for managers. In addition, I demonstrate how to define adequate indicators that reflect fundamental objectives and values.

Chapter 3 shows how to incorporate mediating effects and species reintroduction into ecosystem models in EwE. EwE already has the capacity to represent mediating effects through a function called ‘mediation’. Because these effects have been often overlooked in ecosystem models, I explain how to use this function for more thorough results. As EwE does not have a function for representing species reintroductions and range expansion, I create two strategies to represent these phenomena in Ecospace, the spatial module of EwE.

For this latter objective, I develop a model for nearshore ecosystems of WCVI where sea otters have been reintroduced and have altered the ecosystem. This represents an excellent case study, as newly regenerated kelp forests provide mediating effects to some prey‐predator interactions.

Chapter 4 focuses on the findings of this research, the strengths and limitations of the approach, and future avenues for the application of SDM and the use of ecosystem models for EBM.

1.4 Background

1.4.1 Study area

Ecosystems on the WCVI support multiple human activities, such as commercial, recreational and subsistence fisheries, aquaculture, tourism and transportation (Gislason 2007). Multiple and conflicting objectives characterize the use of natural resources in the region. Given that ecological and social systems are coupled, changes in the ecosystem affect the provision of benefits to humans (Ostrom 2009).

An important issue in the area has been the reintroduction of sea otters. In British Columbia, sea otters were extirpated between 1929 and 1930 (Watson 1993). Hunting sea otters for their pelts was a traditional activity for First Nations (Uuathluck 2007, 2009). However, the international fur trade between the 17th and 18th centuries drove sea otters close to extinction (Estes and Palmisano 1974). In 1911, they were protected under the International Fur Deal Treaty, but the population did not recover (Watson 1993).

Sea otters were reintroduced to the WCVI, British Columbia between 1969 and 1972 (Watson 1993). After some years, they have successfully re‐established themselves (Watson 1993, Nichol et al. 2009) and their expansion continues to suitable habitats3 within the region (Gregr et al. 2008).

Sea otters have altered ecosystems and their dynamics in the region. They have shifted an invertebrate‐dominated nearshore system (i.e. urchin ‘barrens’4) into a kelp‐dominated system by releasing macro‐algae (i.e. ‘kelp’ species Macrocystis integrifolia, Nereocystis luetkeana) from grazing pressure. This effect has been observed in other regions of Alaska and California (Estes and Palmisano 1974) and it has motivated studies to determine the trophic cascade effects of sea otter presence in marine ecosystems (Estes and Palmisano 1974, Simenstad et al. 1978, Duggins 1980).

The return of sea otters to the WCVI is controversial. On one hand, they have caused the decline of commercially valuable shellfish populations and reduced the size of target species (e.g., sea urchins, geoduck (Panopea abrupt), clams, crabs), which has negatively impacted those people who depend on these fisheries (e.g., Watson 2000). On the other hand, their return marks a change in ecosystem dynamics that might provide diverse benefits to several species (via kelp forests) as well as benefits to people, through tourism and perhaps finfish fisheries (Markel 2006). Therefore, as sea otters continue their expansion along the WCVI, potential effects in terms of benefits and losses must be anticipated. An EBM approach is essential for evaluating these effects across species and sectors.

3 Areas with the same sub‐tidal complexity of areas where sea otters were reintroduced and established (Gregr et al. 2008). 4 Urchin dominated communities. 5

1.4.2 Marine planning process on the WCVI

The marine planning process for the WCVI is being led by WCA, a forum for coastal communities and those affected by marine management to participate with governments in the management decision‐making process for the region (WCA 2001). Members of WCA include federal, provincial, local and First Nations governments, representatives of commercial and recreational fisheries, conservation organizations, and aquaculture and tourism industries (WCA 2006).

WCA is currently developing an overarching marine plan for their management area (Figure 1‐ 1) and two spatial plans for the Barkley Sound and Clayoquot Sound regions. WCA is applying an EBM approach to their marine planning processes and have already defined the objectives, as well as the goals and subgoals for achieving those objectives (Denise Dalmer, pers. comm.).

Furthermore, WCA is also in the process of collecting information on indicators of ecosystem health and integrity, as well as developing decision‐making support tools (e.g., EwE models, Marxan) to support their spatial planning process (Tom Okey, pers. comm.).

1.4.3 Decisionmaking process for EBM

At the management level, the operationalization of EBM requires a systematic and participatory decision‐making framework for managers. This framework should integrate a collective vision, operational objectives and indicators for EBM based on the values of stakeholders and the best science available. Such a framework would allow managers to identify the concerns of stakeholders, to build trust in the process as stakeholders can see their values reflected, and to make more informed and systematic, rather than arbitrary or excessively political, decisions (Gregory et al. 2001).

The idea of a systematic decision‐making process has already been identified and suggested by EBM proponents through the Integrated Ecosystem Assessments (IEA)—a decision‐making framework for marine management that has been increasingly seen as central to EBM (Levin et al. 2009, Tallis et al. 2010). This framework is rooted in the decision analysis field. However, IEA implicitly assumes consistency between values, objectives, and indicators—which does not often happen in practice. It suggests but also requires stakeholder involvement. It should be also noted that objectives and indicators do not necessarily integrate stakeholder values and 6

objectives. Whereas, other literature argues that stakeholder engagement and the consistency between values, objectives and indicators are essential components for the implementation of EBM (see chapter 2).

I show how SDM—also called ‘value‐focused thinking’—help advance EBM. SDM is a systematic process that helps individuals and groups to identify values and develop decision‐making frameworks to best satisfy those values (Keeney 1996d, Clemen and Reilly 2001). As this field has been well explored, there are diverse methodologies that guide the identification of values, objectives, attributes and indicators to evaluate management alternatives.

SDM has enough flexibility that it can be applied even when multiple‐stakeholder groups have already defined their objectives, as in the case of WCA. In such circumstances, a facilitator may need to revise the objectives to ensure that they are well articulated, non‐redundant and that they reflect the values of the participants (McDaniels 2000). For this study, I worked in collaboration with the Executive Director of WCA to apply an SDM process to marine planning on the WCVI.

1.4.4 Ecosystem models for EBM: Ecopath with Ecosim (EwE)

Ecosystem models play a fundamental role in EBM, as they can improve our understanding of how complex systems work, enabling us to infer possible outcomes from policies and environmental changes (Christensen and Walters 2004a, 2004b). Traditional stock assessments are insufficient for EBM to explore important management questions such as the ecological impacts of spatial management options (Walters 1999).

Ecopath with Ecosim (EwE) is the most widely used tool to represent aquatic ecosystems (Plagányi 2007). It has been used to predict the impacts of fisheries on target and non‐target species, the effects of marine protected areas and the efficacy of policy interventions (Walters 1999). Advantages of using EwE for decision‐making include its open‐source and user‐friendly interface (Plagányi 2007), its temporal (Ecosim) and spatial (Ecospace) modules, its scenario analysis capabilities (mainly for fisheries policies), the representation of age‐structured populations, the option of fitting models to empirical data (Christensen and Walters 2004a) and the sensitivity analysis (Ecoranger) for main input parameters (biomass, production per unit of biomass, consumption per unit of biomass) (Plagányi and Butterworth 2004). EwE is also an

interactive tool, which allows modelers to draw and explore the effects of changes in topographic features (e.g., climate variations) and policy options (e.g., fishing effort) (Walters et al. 1999). Because there is a growing demand for spatially explicit policies (Walters et al. 1999) such as protected areas, zoning, and fishing areas, Ecospace fills an important niche in marine management.

If we accept McLeod et al. (2005)’s assertion that EBM should account for the whole ecosystem, including humans, it is essential that EwE models incorporate the full range of relevant human activities and improve the representation of ecosystem dynamics. For the purposes of this thesis, with its limited scope, I am focusing only on the latter (ecosystem dynamics), specifically, mediating effects and the (re) introduction of species.

1.4.4.1 Mediating effects The representation of mediating effects in ecosystem models is relevant for EBM because they are always present in ecosystems and they can change ecological dynamics by intensifying or reversing trophic interactions, expanding or contracting niches and causing trophic cascade effects in food webs (Dill et al. 2003). When mediating effects are not considered, ecosystem models can fail the representation of ecosystem dynamics (Heithaus et al. 2008). For this study, I focus on the mediating effects provided by kelp forest to nearshore ecosystems on the WCVI, specifically the effects (i.e. increased feeding areas and food availability for some predators) to some species due to the provision of habitat.

The EwE package already has a function called ‘mediation’ for incorporating mediating effects provided by species (Christensen 2008), but it has been used infrequently. Therefore, to incorporate these effects it will be important to understand how this software function can best be used, and to analyze how it represents mediating effects.

1.4.4.2 Spatial representation of species reintroductions The representation of species (re) introductions in ecosystem models is relevant for EBM due to the potentially significant changes that such phenomena can cause to biological populations and communities (e.g., Watson 1993, Sakai et al. 2001). For this work, I used as a study case the reintroduction of sea otters to the WCVI.

There is currently no built‐in EwE function that accounts for species reintroductions and range expansions. In addition, Ecospace assumes that in year 1, all species in the ecosystem are already distributed among suitable habitats, and for the following years, that species move across the whole study area avoiding unsuitable habitats, which is at odds with situations in which species are new to the ecosystem. Therefore, the representation of species (re) introductions requires the manipulation of the software to work around this assumption of the initial state of the ecosystem.

Although the assumptions of Ecospace can limit the spatial representation of ecosystem dynamics, when used appropriately, EwE can be used to complement traditional management tools such as stock assessments, and can support ecosystem based approaches by providing an integrated understanding of ecosystem structure and dynamics (Plagányi and Butterworth 2004). In addition, EwE is not a static tool as it is under continued improvement.

Figure

Vancouver Clayoquot Sound

region Barkley Sound region

Victoria

Figure 1‐1 West Coast Aquatic (WCA) management area. Three management plans are being developed: one for the whole management area and two regional ones for Clayoquot region and Barkley Sound regions.

1.5 References

Arkema, K. K., S. C. Abramson, and B. M. Dewsbury. 2006. Marine ecosystem‐based management: from characterization to implementation. Frontiers in Ecology and the Environment 4:525‐532. Christensen, V. and C. J. Walters. 2004a. Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling 172:109‐139. Christensen, V. and C. J. Walters. 2004b. Trade‐offs in ecosystem‐scale optimization of fisheries management policies. Bull. Mar. Sc. 74(3):549562. Clemen, R. and T. Reilly. 2001. Structuring Decisions. Pages 43‐110 Making Hard Decisions with Decision Tools. Duxbury Thomson Learning. Dill, L. M., M. R. Heithaus, and C. J. Walters. 2003. Behaviorally mediated indirect interactions in marine communities and their conservation implications. Ecology 84:1151‐1157. Duggins, D. O. 1980. Kelp Beds and Sea Otters: An Experimental Approach. Ecology 61:447‐453. Estes, J. A. and J. F. Palmisano. 1974. Sea Otters: Their Role in Structuring Nearshore Communities. Science 185:1058‐1060. Gislason, G. 2007. Economic Contribution of the Oceans Sector on the West Coast of Vancouver Island. Report for Canada Department of Justice, Vancouver. Gregory, R., T. McDaniels, and D. Fields. 2001. Decision Aiding, Not Dispute Resolution: Creating Insights through Structured Environmental Decisions. Journal of Policy Analysis and Management 20:415‐432. Gregr, E. J., L. M. Nichol, J. C. Watson, J. K. B. Ford, and G. M. Ellis. 2008. Estimating carrying capacity for sea otters in British Columbia. Journal of Wildlife Management 72:382‐388. Grosholz, E. D., G. M. Ruiz, C. A. Dean, K. A. Shirley, J. L. Maron, and P. G. Connors. 2000. The Impacts of a Nonindigenous Marine Predator in a California Bay. Ecology 81:1206‐1224. Guerry, A. D. 2005. Icarus and Daedalus: conceptual and tactical lessons for marine ecosystem‐ based management. Frontiers in Ecology and the Environment 3:202‐211. Heithaus, M. R., A. Frid, A. J. Wirsing, and B. Worm. 2008. Predicting ecological consequences of marine top predator declines. Trends in Ecology & Evolution 23:202‐210. Keeney, R. L. 1996d. Value‐focused thinking: Identifying decision opportunities and creating alternatives. European Journal of Operational Research 92. Leslie, H. M. and K. L. McLeod. 2007. Confronting the challenges of implementing marine ecosystem‐based management. Frontiers in Ecology and the Environment 5:540‐548.

Lester, S. E., K. L. McLeod, H. Tallis, M. Ruckelshaus, B. S. Halpern, P. S. Levin, F. P. Chavez, C. Pomeroy, B. J. McCay, C. Costello, S. D. Gaines, A. J. Mace, J. A. Barth, D. L. Fluharty, and J. K. Parrish. 2010. Science in support of ecosystem‐based management for the US West Coast and beyond. Biological Conservation 143:576‐587. Levin, P. S., M. J. Fogarty, S. A. Murawski, and D. Fluharty. 2009. Integrated Ecosystem Assessments: Developing the Scientific Basis for Ecosystem‐Based Management of the Ocean. PLoS Biol 7:e1000014. Markel, R. W. 2006. Sea otter and kelp forest recovery: Implications for nearshore ecosystem, fishes and fisheries. Parks Canada. McDaniels, T. L. 2000. Creating and using objectives for ecological risk assessment and management. Environmental Science & Policy 3:299‐304. McLeod, K. L., J. Lubchenco, S. R. Palumbi, and A. A. Rosenberg. 2005. Scientific Consensus Statement on Marine Ecosystem‐Based Management. Signed by 221 academic scientists and policy experts with relevant expertise and published by the Communication Partnership for Science and the Sea at http://compassonline.org/?q=EBM.21p. Nichol, L. M., M. D. Boogaards, and R. Abernethy. 2009. Recent trends in the abundance and distribution of sea otters (Enhydra lutris) in British Columbia. Fisheries and Oceans Canada. Ostrom, E. 2009. A General Framework for Analyzing Sustainability of Social‐Ecological Systems. Science 325:419‐422. Plagányi, É. E. 2007. Models for an ecosystem approach to fisheries. FAO Fisheries Technical Paper No. 477 Rome. Plagányi, É. E. and D. S. Butterworth. 2004. A critical look at the potential of Ecopath with ecosim to assist in practical fisheries management. African Journal of Marine Science 26:261‐287. Rosenberg, A. A. and P. A. Sandifer. 2009. What do managers need? Pages 13‐30 in K. L. McLeod and H. M. Leslie, editors. Ecosystem‐based Management for the Oceans. Island Press, Washington, D.C. Ruckelshaus, M., T. Kingler, N. Knowlton, and D. P. DeMaster. 2008. Ecosystem‐based management in practice: Scientific and Governance challenges. BioScience 58:59‐63. Sakai, A. K., F. W. Allendorf, J. S. Holt, D. M. Lodge, J. Molofsky, K. A. With, S. Baughman, R. J. Cabin, J. E. Cohen, N. C. Ellstrand, D. E. McCauley, P. O'Neil, I. M. Parker, J. N. Thompson,

and S. G. Weller. 2001. The population biology of invasive species. Annual Review of Ecology and Systematics 32:305‐332. Simenstad, C. A., J. A. Estes, and K. W. Kenyon. 1978. Aleuts, sea otters, and alternate stable communities. Science 200:403‐411. Tallis, H., P. S. Levin, M. Ruckelshaus, S. E. Lester, K. L. McLeod, D. L. Fluharty, and B. S. Halpern. 2010. The many faces of ecosystem‐based management: Making the process work today in real places. Marine Policy 34:340‐348. Uuathluck. 2007. Nuu‐chah‐nulth's Historical Relationships with Sea Otters. http://www.uuathluk.ca/Sea%20Otter%20onepager2%20final[1].pdf. Uuathluck. 2009. Sea Otters in Nuu‐chah‐nulth Ha‐houlthee. http://www.uuathluk.ca/Uuathluk_Sea_Otters_April_09%20.pdf. Walters, C. J., D. Pauly, and V. Christensen. 1999. Ecospace: Prediction of Mesoscale Spatial Patterns in Trophic Relationships of Exploited Ecosystems, with Emphasis on the Impacts of Marine Protected Areas. Ecosystems. Springer New York 2:15:539‐554. Watson, J. C. 1993. The effects of sea otter (Enhydra lutris) foraging on shallow rocky communities off northwestern Vancouver Island, British Columbia. University of California, Santa Cruz. WCA. 2001. West Coast Vancouver Island. Aquatic Management Board. Terms of Reference. WCA. 2006. http://www.westcoastaquatic.ca.

Chapter 2 Using structured decisionmaking for ecosystembased management5

2.1 Introduction

Ecosystem‐based management (EBM) has been called, by national and international institutions, the new approach to managing marine resources and ecosystems (Leslie and Kinzig 2009). This approach goes beyond traditional management based on single species and single sectors (Leslie and McLeod 2007) and recognizes deep connectivity amongst all elements of the ecosystem—including humans (McLeod et al. 2005)—and the underlying processes of producing the services people need and want (Guerry 2005, Arkema et al. 2006). It is place‐ based and requires a coordinated effort to sustainably manage the human activities that impact ecosystems (Guerry 2005, Leslie and McLeod 2007, Lubchenco and Sutley 2010).

Although there have been various efforts to define the key aspects, principles and guidelines (McLeod et al. 2005, Leslie and McLeod 2007, Lester et al. 2010) of what EBM is and requires, there is still a gap between theory and practice (Arkema et al. 2006, Ruckelshaus et al. 2008, Lester et al. 2010). Managers face political, legal, social and scientific difficulties in implementing the complex concepts of EBM, which has come to be seen as daunting and expensive (Arkema et al. 2006, Lester et al. 2010, Tallis et al. 2010).

More science will not necessarily lead to the implementation of EBM (Lester et al. 2010). This is reflected in research studies that argue that the main challenges for the implementation of EBM include building a collective vision and objectives for EBM, designing metrics to evaluate the accomplishment of the objectives and creating ocean governance frameworks (Leslie and McLeod 2007), as well as bridging the gap between scientific concepts and operational goals (Arkema et al. 2006). Successful initiatives aimed at implementing EBM (e.g, Great Barrier Reef in Australia, Puget Sound in United States, and Raja Ampat in Indonesia) show that meaningful involvement of stakeholders in the definition of objectives and in monitoring processes have

5 A version of this chapter will be submitted for publication. Espinosa‐Romero M., Chan K., and McDaniels T., Structured decision‐making for Ecosystem‐based Management. 14

been key elements for success (Arkema et al. 2006, Tallis et al. 2010). Put differently, environmental management is never an exclusively science‐based undertaking. Human values, articulated and pursued within appropriate governance processes, are at the heart of why EBM is important and they define what it should achieve (Gregory et al. 2006).

As management is the process of making decisions (Walters and Martell 2004), the implementation of EBM requires a systematic and participatory framework to identify the values of the constituents with respect to EBM and to make decisions that best satisfy those values. With this framework, managers can anticipate and address the concerns of stakeholders and make better informed decisions about the use of natural resources (Gregory et al. 2001). In addition, if stakeholders see their values reflected they are more likely to trust the process and/or support its implementation (Gregory et al. 2001).

2.1.1 A need to improve decisionmaking processes EBM proponents have suggested the Integrated Ecosystem Assessments (IEAs) developed by the US National Oceanic and Atmospheric Administration (NOAA) as the most useful decision‐ making framework for marine management that integrates science to assist decision makers (Levin et al. 2009). This framework, increasingly seen as central to EBM, is rooted in the decision analysis field (Levin et al. 2009, Tallis et al. 2010), and it implicitly recognizes the importance of systematic decision‐making.

The six steps of IEA include the following: 1) definition of objectives, threats to ecosystems and ecosystem management drivers; 2) development of indicators for ecosystem state; 3) establishment of thresholds for each indicator; 4) risk analyses to evaluate how indicators respond to human and environmental disturbances and the probability that indicators will reach or remain in an undesirable state; 5) evaluation of management strategies to predict the effects on the indicators; and 6) monitoring management strategy outcomes (Levin et al. 2009, Tallis et al. 2010).

Although IEA suggests stakeholder involvement (Levin et al. 2009, Tallis et al. 2010), it does not require it; and although it suggests a systematic decision‐making process it does not pay detailed attention to some aspects of the process (e.g. definition of objectives, and of indicators specifically related to stakeholder values or objectives). Lack of attention to these aspects may

contribute to the observed inconsistency between values, objectives, indicators and management decisions. When stakeholders are not involved, important values and objectives can be ignored, and stakeholders might not trust the process and might reject the implementation of a decision/management plan (Gregory et al. 2001). When stakeholders are involved but there is no systematic process, important objectives can be dismissed (Gregory et al. 2001) or stated in a complex way that is not operationalizable and cannot guide decision‐ making (Keeney 1996b). The lists of indicators suggested by scientists to measure ecosystem state may not be useful for managers to make decisions, or may not even reflect the fundamental management objectives and stakeholders values (Failing and Gregory 2003, Gregory et al. 2006). A decision‐making framework must identify objectives based on values and define indicators for those objectives (Gregory et al. 2001).

Structured decision‐making (SDM) refers to applied decision analysis conducted with stakeholders and technical specialists to gain insight of and guide specific management decisions; it is a systematic process that can help stakeholders and managers to construct a framework to create and evaluate alternatives and make more informed decisions to satisfy a collective set of values (McDaniels 2000). It pays special attention to the consistency between values, objectives, indicators and alternatives. Since it is a well‐explored field for multiple stakeholder planning processes, it provides methodologies and approaches for each stage of the process. Therefore, SDM can be used to integrate stakeholder values and science in a decision‐ making framework for EBM, oriented to satisfy collective values and objectives.

We summarize the SDM process below and present the case study of West Coast Aquatic (WCA)—a multiple‐stakeholder regional management board that is using the SDM approach to implement EBM on the west coast of Vancouver Island (WCVI), British Columbia, Canada. Specifically, we give insights on how to define operational objectives that reflect the values of the constituents and derive performance measures based on those objectives so as to facilitate the process of decision‐making within an EBM context.

2.1.2 The structured decisionmaking process As a first step, the decision to be made must be defined (McDaniels 2000). For this context, the decision represents the design of a decision‐making framework for EBM. The second step is to define what matters to stakeholders in this EBM context and why, as well as the objectives that

need to be achieved (McDaniels 2000). These can be done by asking stakeholders directly what matters, why, and what objectives are wanted; or indirectly, by asking them what alternatives and consequences are desirable or undesirable and why, or by reviewing the stated goals, objectives, constraints and guidelines (mission statements, policy guidelines, strategic plans) of the group (Keeney 1996d). Generally, a list of wants, desires, and concerns that reflect stakeholder values and their collective vision on a particular issue or situation is derived at this stage. This list needs to be categorized to identify collective objectives that reflect stakeholder values (McDaniels 2000).

Subsequently, objectives are separated into fundamental and means objectives. Fundamental or end objectives are those that are important because they directly reflect the values of the participants, while means objectives are those that are important because they contribute to the achievement of fundamental objectives (Clemen and Reilly 2001). For example, for a company, working fewer hours may be seen as an objective, but its importance relies on allowing employees to spend more time with their families, to relax or do hobbies. Therefore, minimizing work hours would be a means objective for maximizing employee’s spare time (Clemen and Reilly 2001). End objectives should be identified to guide decision‐making as they reflect what is important to the constituents (Keeney 1996c). When means objectives are mistaken for end objectives, the risks are that management may achieve means objectives in a manner that fails to achieve the true end objective, or constraint the creation of better alternatives. In the previous example, efforts to minimize work hours might fail to achieve desired end. In addition, other alternatives that could be better for employees, such as flexibility on the work schedule, might not be identified if minimizing the work hours is mistaken as the end objective.

Fundamental objectives should be non‐redundant to be clear and concrete; measurable to facilitate the evaluation of alternatives and the achievement of objectives; and meaningful to those who are going to use them to ensure their applicability for decision‐making and the engagement of stakeholders (McDaniels 2000).

Once fundamental objectives are defined, the next step is to define the attributes or specific objectives for each of them. The attributes represent the meanings or the context in which the fundamental objectives are perceived (Failing and Gregory 2003). For example, if the objective is to maximize economic benefits, the attributes can be the net profits, the distribution of

benefits and loss among stakeholders, or the local retention of benefits. In a multiple stakeholder process this stage is crucial to have a collective understanding of the fundamental objectives (McDaniels et al. 2006). The clear articulation of attributes is essential to define appropriate indicators for the fundamental objectives (Failing and Gregory 2003).

The next step is to define performance measures or indicators for the attributes of the fundamental objectives. The attributes and performance measures make the shift from fundamental objectives into an operational decision‐making framework (McDaniels 2000). The importance of measuring the achievement of EBM objectives through the use of indicators has been widely recognized (Guerry 2005, Tallis et al. 2010) and great efforts have been made to define multiple indicators for ecosystem conservation, not all of which are useful for decision makers, and which collectively may not reflect stakeholder values and objectives (Failing and Gregory 2003). In a SDM approach, the only indicators that are accepted are those based on fundamental objectives that can guide the decision‐making process. Technical assistance and stakeholder involvement is required for this stage of the process (see Keeney and Gregory 2005).

Once indicators are defined, the following step in SDM is to create alternatives based on fundamental objectives and evaluate the performance of those alternatives based on the selected indicators. Because choosing between alternatives involves tradeoffs, a key strength of SDM is that such tradeoffs are made explicit (a relevant aspect of EBM); stakeholders are able to understand what tradeoffs each alternative entails. This can contribute to a more transparent decision‐making process. The identification of fundamental objectives, attributes and indicators often spurs participants to create more innovative alternatives that better satisfy the full range of objectives (Gregory et al. 2001).

Stakeholders often end up choosing amongst only a small set of alternatives, as those that poorly satisfy the objectives are quickly eliminated. When there is disagreement on a group’s preferred alternative, stakeholders could, for instance, be asked to choose their preferred alternative, and then present the reasons for choosing it, as well as the expected winners, losers, pros and cons (Gregory et al. 2001). After this, stakeholders will likely be more amenable to agree on a preferred alternative. If they do not, the whole SDM process can nonetheless provide managers with good insights about the advantages and disadvantages of the alternatives when

making a decision (Gregory et al. 2001). In addition, SDM can help managers identify decision opportunities to better satisfy the objectives, rather than just responding to the problems as they come (Keeney 1996c).

Decision makers face multiple and complex decisions; a decision‐making framework can therefore help managers to make consistent decisions oriented to satisfy the same fundamental objectives (Keeney 1996a).

2.1.3 Some considerations for the implementation of SDM for EBM When organizations or multiple stakeholder groups have already defined their initial set of objectives to achieve EBM, these objectives should be revised to ensure that they reflect the values of the participants and that they are clearly articulated to guide decision‐making. If objectives are not appropriate, well defined, and well thought out, decisions may not satisfy the fundamental objectives regardless of the rest of the process.

In the case of pre‐defined objectives, it will be essential to verify with stakeholders the attributes of each fundamental objective to define appropriate indicators for the objectives. In SDM, the definition of indicators is usually a technical process, involving subsequent consultation with stakeholders to ensure expert‐chosen indicators reflect the objectives and are meaningful to them. In the case of EBM, communication between stakeholders and scientists is essential to define a set of indicators that reflect what matters and but also how the ecosystems work.

Adaptive management—treat policies as experiments to gather information on complex systems in the face of high uncertainty (Walters 1986)—has been identified as a key element of EBM (Arkema et al. 2006, Tallis et al. 2010). With regards to SDM, it has been suggested to include adaptive management as an explicit objective to motivate and recognize the importance of learning and flexibility from the beginning of the process (McDaniels and Gregory 2004). This objective can help stakeholders understand the limitations of current knowledge, recognize opportunities to reduce uncertainties, and build trust in the decision‐making process (McDaniels and Gregory 2004).

2.1.4 Case study: Marine spatial planning for the west coast of Vancouver Island The west coast of Vancouver Island (WCVI) is a large and diverse area that supports multiple human activities such as commercial, recreational and subsistence fisheries, aquaculture, tourism and transportation (Gislason 2007). Therefore, diverse and conflicting objectives characterize the use of the natural resources in the area.

West Coast Aquatic (WCA) is a forum for coastal communities and those affected by marine management decisions to participate with governments in the decision‐making process for managing the area (WCA 2001). Multiple stakeholders are currently involved and work together to develop long‐term objectives and operational plans for the use of aquatic ecosystems of the WCVI. Members of the WCA include Federal, Provincial, Local and First Nations governments, representatives of commercial and recreational fisheries, the aquaculture and tourism industries, and conservation organizations, among others (WCA 2006).

The ethical principles under which WCA operates include management based on an ecosystem approach, conservation, precaution, adaptivity, sustainability, shared responsibility, inclusivity, benefits, and flexibility (WCA 2009).

WCA has initiated a planning process for an overarching marine plan and two spatial regional plans, for which they want to apply the EBM approach. They have already defined eight objectives that reflect both “what matters” and “what EBM entails” to the stakeholders involved as well as the goals and sub‐goals required to achieve the eight main objectives (Denise Dalmer, pers. comm.).

The eight objectives are (1) integration and collaboration; (2) sustainable economic benefits; (3) healthy ecosystems; (4) healthy, prosperous and safe communities and waterways; (5) First Nations, reconciliation and relationships‐strengthening; (6) collection of knowledge, information and technology; (7) capacity building; and (8) good governance (WCA 2009).

2.2 Methods

2.2.1 Applying SDM to design a decisionmaking framework for EBM The original set of objectives defined by WCA (Table 2‐1) involved important aspects for EBM; however, they were not stated in an operational way to support decision‐making. We therefore review them and suggest how to make them operational without losing or distorting their meanings.

We separated fundamental from means objectives, using the “why each objective is important” test defined by Clemen and Reily (2001). When an objective was important because it contributed to the achievement of another objective, we defined that objective as a means objective. When an objective was important because it represented WCA values we defined it as fundamental objectives.

We looked at the goals and sub‐goals documents (see WCA 2009) to understand the intended meanings or attributes of each of the eight objectives. We then combined any attributes with the same meaning, and moved any attributes that better fit with other fundamental objectives to those respective objectives. For those objectives that were broadly described, we suggested specific attributes based on a literature review and other case studies that expressed similar objectives.

We then suggested performance measures or indicators for the attributes of fundamental objectives. For example, for ‘distribution of benefits across stakeholders’, we suggest ‘the percentage of benefits retained by each sector or stakeholder’. For attributes that were not informative for measures, such as ‘vibrancy’ of communities, we reviewed the performance measures used in other case studies to measure those attributes. We do not intend to provide an exhaustive list of indicators for each attribute in this work, but instead focus on indicators that reflect fundamental objectives and can be useful for WCA decision‐making processes.

The Executive Director of WCA (the chief proponent of applying SDM to the organization’s management strategy) revised, complemented and confirmed the process. This helped to ensure that the new structure reflected the original values and was consistent with their goals, and that performance measures were appropriate and meaningful to them.

2.3 Results and discussion

From the list of eight main objectives defined by WCA for their marine planning process, four were considered fundamental objectives: (a) foster economic benefits; (b) foster healthy ecosystems; (c) foster healthy communities; and (d) foster good governance. Adaptive management was included as a fifth fundamental objective. It represents an ethical principle for WCA and was part of one fundamental objective (“healthy ecosystems”), but from conversations with WCA it was agreed to include it as a separate objective due to its importance for the whole process rather than only for that particular fundamental objective.

Three of the eight original objectives—‘integration and collaboration’, ‘knowledge, information and technology’, and ‘capacity building, engagement and communications’—were identified as means objectives, as their importance relies on their contribution to the five fundamental objectives.

Three attributes of the ‘integration and collaboration’ objective (i.e. shared responsibilities, collaboration with other plans, and participatory management) were re‐identified as falling under the fundamental objective of ‘good governance’.

The attributes of the objective ‘First Nations reconciliation and relationships strengthening’ proved to fit within other fundamental objectives. We thus ensured these attributes were made explicit within the other fundamental objectives. For example, ‘respect aboriginal and treaty rights’ was re‐grouped within ‘healthy communities’; and ‘participation in decision‐making’ was grouped under ‘participatory management’ as part of the ‘good governance’ objective.

Compacting and re‐grouping attributes of one fundamental objective into another does not mean that some attributes are ignored, changed or left out, but rather provides a way to organize them to facilitate operationalizing and monitoring of objectives. Attributes may appear in more than one objective or be stated as fundamental objectives, sometimes because stakeholders strive to elevate particular interests as much as possible in the objectives. The result is that operationalizing the objectives is very difficult. When groups of decision‐makers want to emphasize their preference for certain objectives, or attributes thereof, they do not need to repeat them; instead they can assign higher weights to them.

Below, we describe the new structure (Table 2‐2) of fundamental objectives and attributes based on the revision of WCA objectives, goals and sub‐goals, the feedback from WCA. We also suggest indicators based on the attributes of fundamental objectives that can help measure the performance of different alternatives.

2.3.1 New structure: Fundamental objectives, attributes and indicators

2.3.1.1 Foster economic benefits This objective includes generating benefits derived from the ecosystem and their fair distribution across present and future generations. Benefits can include profits and employment, but also the goods or services themselves, because trading is a traditional practice in the region. The retention of benefits by local communities, specifically First Nations, was stated twice in the original document of objectives, goals and sub‐goals, and also emphasized in conversations with WCA representatives.

The WCA representative mentioned as part of this objective the balance of ecological, social and economic components as well as sustainable management; one would think that both are related and represent what the decision‐making process should entail rather than attributes of this particular objective. They also mentioned sustainable fisheries and aquaculture, monitoring, enforcement and regulations, which represent means objectives and may apply not only to this objective. Conservation as the main priority for fisheries management was included here; however this represents a value or preference towards certain alternatives. First Nations access to natural resources was included as part of the ‘healthy communities’ objective. a) Attributes Attributes for this objective include economic benefits (profits, employment and the goods themselves), and the distribution of benefits and loss across stakeholders and over time. b) Suggested indicators A commonly used indicator for profits is the net present value—aggregated benefits minus aggregated costs, discounted over time. Discount rates are applied to estimate the present value of future revenues or costs; these rates can vary among individuals or social levels (Sumaila and Walters 2005). Net present value is a well‐known indicator but insufficient alone, because it

does not capture the distribution of benefits and costs among stakeholders (Timko and Satterfield 2008a).

Income per capita and person‐years of employment (the equivalent to one person employed for a full working year) have been also used to represent benefits (Hobbs and Horn 1997, McDaniels et al. 2006). For the case of employment, we suggest accounting for skilled, unskilled, temporary and permanent jobs derived from each alternative (Timko and Satterfield 2008a, Timko and Satterfield 2008b).

The proportion of these benefits (net revenues, income per capita, skilled, unskilled, temporary, and permanent jobs) and losses retained by each stakeholder group (adapted from Philcox 2007) and over time (Timko and Satterfield 2008a, Timko and Satterfield 2008b) can help to measure the distribution of benefits. WCA could also measure the proportion of benefits and losses retained among local communities or First Nations to evaluate if the most vulnerable groups are retaining benefits (e.g., McDaniels et al. 2006).

2.3.1.2 Foster healthy ecosystems To foster healthy ecosystems, WCA will focus on the integrity of ecosystems. Based on the approach used by the Department of Fisheries and Oceans Canada (DFO) in the Eastern Scotian Shelf (DFO 2007), WCA defined the followings three main aspects of integrity: diversity, productivity and environmental quality. Because most of the alternatives to be evaluated represent changes in human activities, this objective is to minimize the adverse effects of human activities on these aspects of ecosystem integrity.

Resilience was mentioned as part of this objective. Ecological resilience is a complex concept and the ways to measure it are still under research. It is generally understood as the system’s ability to absorb shocks or disturbances while maintaining its function, structure, identity and feedbacks (Walker et al. 2004). Resilience is often assumed to be good, as people often assume one desirable state and the ability of the system to maintain or go back to this state after disturbances. However, it has been demonstrated that many systems, including marine ecosystems, have multiple states or attractors (e.g. from kelp‐dominated ecosystems to ‘urchin barrens’). Some states may be undesirable for societies; maintaining the resilience of these states may therefore not be desired. In addition, the resilience of an ecosystem is influenced by

the feedbacks between social and ecological systems (Leslie and Kinzig 2009), a fact that applies not only to this objective but also to others such as ‘economic benefits’ and ‘healthy communities’. We consider the appropriate management of resilience to be a means objective for the fundamental objectives, and it will be important for societies or group to identify the desirable states for which resilience is sought.

WCA also included as part of this objective other means objectives such as the protection of species at risk, the establishment of marine protected areas, development of plans to respond to natural disturbances, and adaptation to climate change. They also mentioned to consider conservation as a first priority, which represent implicit preferences towards objectives and alternatives. They included EBM and the integration of values in the planning and decision‐ making processes, which represent the overall objective; and adaptive management, which was included as a separate fundamental objective. They also included the application of the precautionary principle, which we moved to the ‘good governance’ objective. a) Attributes WCA will focus on the three main elements of integrity: diversity, productivity and marine environmental quality. Diversity includes species, populations and communities; productivity refers to primary and secondary productivity, as well as trophic structure and population productivity; and marine environmental quality involves physical, chemical and habitat quality (DFO 2007, Okey 2009 ). b) Suggested indicators • Diversity: species, populations and communities Species richness—number of species—within defined boundaries such as communities or habitats (Gray 1997) and evenness—distribution of species biomass—have been suggested to measure diversity However, the composition of (biological) communities and habitats is important: while species richness may show a high number of species in the region, this number may also include introduced species, which are not members of the native community and can cause negative impacts to native species and ecosystems.

Because it is very difficult to focus on all species in the community or habitat, it is necessary to select species or groups whose characteristics represent attributes of other species, the

ecosystem and environmental conditions (Lindenmayer et al. 2000); those groups that play an important role ecologically or culturally such as endemic species, species at risk, etc.

Endemic species for example are important because they only occur in specific places, regions, ecosystems or communities (Reid et al. 1993); and their populations are usually small and vulnerable to extinction (Lamoreux et al. 2006). Some studies have demonstrated that the conservation of sites with high levels of endemism can capture large proportions of all identified species of a region (Lamoreux et al. 2006). Species at risk, can also be useful in the sense that it highlights biodiversity components that might be lost, but this measure can also be limited because of the political process of listing, and because simple species counts are far removed from ecological integrity. Endangered species lists might be derived from those listed under the Species at Risk Act (SARA) or those proposed by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC). We suggest the latter, because not all species at risk suggested by COSEWIC are listed under SARA due to political, social and economic factors.

It is advisable to look at species richness and abundance across the selected groups, as well as the historical trends to evaluate if their populations are increasing, stable or declining (Tallis et al. 2010).

The mean trophic level (TL) (Pauly et al. 1998, Pauly and Watson 2005) is a well‐known indicator that can also be used as a proxy for the community composition. It is calculated by assigning species to trophic levels and using information on species’ catch and diet composition. The TL has been often used to indicate the impacts of fisheries (Pauly et al. 1998, Pauly and Watson 2005); however, it can be also used to analyze the trophic structure of an ecosystem by including non‐target species abundance and diet composition.

In terms of biodiversity, it has also been suggested to pay attention to community and habitat diversity (DFO 2007) because their conservation can ensure the conservation of species (Gray 1997).

• Productivity: primary, secondary, trophic level and population productivity. Abundance per trophic level can be used as an indicator of the productivity at trophic level (Jamieson et al. 2001, O'Neill et al. 2008) as well as for primary and secondary production.

Historical catch and biomass of target species can also be used to measure productivity of species and populations (DFO 2007) and provide insights on the status (declining, stable, or decreasing) of those species.

• Marine environmental quality: physical, chemical and habitat quality This can be accounted by evaluating the concentration of toxics in the water, habitats (DFO 2007, O'Neill et al. 2008) and species (e.g., harbour seals, pelagic and benthic fish, clams, mussels and juvenile salmon) (PSP 2009) as well as the generation of noise and atmospheric pollution (DFO 2007) derived from each of the alternatives.

For habitat quality, WCA could select those habitats or communities they consider important to conserve, and evaluate the potential impacts of alternatives on those habitats. This can be done by assessing the total area of ‘selected’ habitat impacted (e.g., Gregory et al. 2001) or by identifying the main threats for those habitats (e.g. trawlers) and the magnitude of the particular threat (e.g., number of trawlers).

2.3.1.3 Foster healthy communities This objective refers to avoiding adverse effects on the health, safety and vibrancy of local communities. Vibrancy is understood as those things that make locals stay in their communities as well as the survival of groups and traditions over time.

We include as attributes of vibrancy the ‘respect of First Nations rights and title’, which was part of the original objective ‘First Nations reconciliation and relationships’, and ‘First Nations access to natural resources’, which was an attribute of the ‘economic benefits’. a) Attributes Attributes for minimizing the adverse effects on health and safety can be related to people, private and public property (McDaniels et al. 1999).

Attributes for vibrancy can include cultural practices, access to natural resources (Timko and Satterfield 2008a, Timko and Satterfield 2008b), the maintenance and respect of First Nations treaties (Dalmer, pers. comm.) and aesthetics (adapted fromMcDaniels et al. 2006).

b) Suggested indicators Effects on human health and safety, specifically on people, private and public property can be measured by identifying risks, their magnitudes and probabilities, who (or what, in the case of private and public property) is exposed, and to what extent. This can be based on science and stakeholders’ perceptions (Hobbs and Horn 1997). Illness and deaths that could be associated with marine resources—in this case those associated to the management alternatives—have been suggested as a health indicator (PSP 2009).

Indicators for the attributes of vibrancy can be evaluated based on constructed scales. For the cultural practices, WCA can identify the traditional and ceremonial practices that are important for locals (e.g., potlatch, festivities, transmission of traditional knowledge, language) and evaluate the impacts of different alternatives on these practices (adapted from McDaniels et al. 2006, Philcox 2007, PSP 2009). When a cultural form can be lost or negatively impacted it will be important to evaluate if those who are affected consider the compensation fair (Timko and Satterfield 2008b).

To monitor access to natural resources including land, WCA can evaluate if the conditions of access are perceived locally to be fair, if the conditions of access to locals will be impacted (negatively and positively), and when negatively impacted, if the compensation is fair (adapted from Timko and Satterfield 2008b). To avoid the adverse effects on First Nations ‘rights and title’, WCA can evaluate if the treaties and title are respected and to what level of satisfaction according to local perception (adapted from Timko and Satterfield 2008b). For aesthetics, WCA could evaluate the visual, odor and water quality impacts in the region (Satterfield, pers. comm.).

2.3.1.4 Foster good governance Good governance includes participatory management, shared responsibilities, compatibility with other plans, social agreements, and public accountability. This objective has been identified as a key element and a challenge for the implementation of EBM (McLeod et al. 2005, Leslie and McLeod 2007). We included the precautionary principle as an attribute of this objective.

Other aspects originally included in this objective were the identification of priorities, the responsible use of financial resources, and the evaluation of WCA performance on a regular basis, each of which represent means objectives. a) Attributes According to WCA, participatory management refers to the meaningful involvement of stakeholders. Attributes of participatory management include the following: representativeness, fairness and competence (Renn 2004). Representativeness means that all stakeholders are represented when making decisions. Fairness refers to the equitable access to participatory processes, equal opportunities to make and reject claims during deliberation, and the consideration of different sources of information. Competence means that all participants have a sufficient level of understanding of the consequences of each alternative, including knowledge, uncertainties and ambiguities (Renn 2004). Precautionary principle states that one should take action to mitigate possible harmful effects on ecosystem and human health despite scientific uncertainty of those effects.

Shared responsibilities, alignment with other plans through the collaboration with agencies and other organizations, social agreements between First Nations and other governments, precautionary principle and public accountability are also attributes of this objective. b) Suggested indicators The attributes of this objective can be measured qualitatively using constructed scales (e.g., 1 to 5; low, medium, high) to answer questions determined by WCA. Here, we suggest some questions helpful for scoring each attribute, considering the components mentioned in the previous section.

For the participatory management objective: Were all stakeholders represented in the process? Were there opportunities (e.g., forums, presentations, meetings) for the public and stakeholders to participate in the process? Was there a capacity building program and training opportunities for locals and other stakeholders to improve their competence (Timko and Satterfield 2008a, Timko and Satterfield 2008b)?

For the responsibilities distribution: Do all the members agree with the distribution of tasks? Do the members have the capacity to do the work they were tasked with?

For alignment with other plans: Were synergies and new partnerships with other organizations made during the process?

For resolutions between First Nations and other governments: To what extent is there a respectful relationship and agreement between First Nations and other governments (Timko and Satterfield 2008a, Timko and Satterfield 2008b)?

For the precautionary principle: Was there a comprehensive consideration of possible negative effects of human actions and activities, including those with considerable uncertainty? Do proponents of potentially actions/activities have responsibility to demonstrate small likelihood of major negative effects? Were alternatives considered to reduce harm on ecosystems and human health? Was a monitoring plan adopted to evaluate the harm on ecosystems and human health?

For public accountability: Was information on financial resources available for all members of the board? Was information on the impacts of management alternatives in relation to the objectives and attributes available for all members? Are there sufficient measures in place to ensure transparency?

2.3.1.5 Foster learning through adaptive management Adaptive management represents the ability to learn from the process as new information arrives (McDaniels et al. 1999) and treating policies as experiments to explore possible outcomes (Walters 1986). a) Attributes Adaptive management can include the following: learning from stakeholders, learning by treating policies as experiments to explore potential outcomes, and learning from the decision‐ making process (Walters 1986, McDaniels and Gregory 2004).

b) Suggested indicators Most attributes of these objectives can be also qualitatively measured with constructed scales to answer questions defined by WCA. Here, we present some examples.

For learning from other participants: Did the board learn from conflict resolution (adapted from Timko and Satterfield 2008b)? Did stakeholders learn from each other during the process?

For learning through treating policies as experiments: Were key sources of uncertainty indentified? Were opportunities to reduce uncertainties identified? Did members consider means for applying such opportunities in policy/management alternatives? Were those means implemented?

For learning from the process: Are members satisfied with the process? Was new information identified and integrated to the process? Were there opportunities to review and adjust agreements and policies (Timko and Satterfield 2008a,b)? Were there learning opportunities over time (Gregory et al. 2001)?

2.3.2 Insights from the process and applicability for EBM The implementation of EBM is of interest of scientists, managers and stakeholders; therefore a collective vision of what is wanted from EBM to achieve is very important. Understanding and articulating in a proper way our values can help direct our decisions to what we want. We intend with this work to help WCA articulate the initial set of objectives in an operational way to guide decision‐making processes.

The process of re‐structuring objectives requires the involvement of stakeholders for two main reasons. First, to ensure that stakeholder values are well understood and reflected in the new structure. And second, for stakeholder to understand the process and realize that their objectives were not changed, left out or ignored, but only re‐structured to make them operational in order to guide the decision‐making for their own benefits. If stakeholders are not involved, they may resist the changes to the new structure. For this particular case, we consulted and confirmed the new structure with the Executive Director of WCA. However, for its implementation, the results will have to be consulted with all WCA members.

Little attention has been paid to the definition of indicators for EBM that reflect stakeholder values and guide decision‐making. Through this work, we emphasize the importance of linking indicators to fundamental objectives as well as involving stakeholders and scientists in the definition of indicators to integrate value judgments and technical information.

The resulting proposed set of indicators for the case study that may seem extensive; however, it reflects WCA fundamental objectives. It is essential that WCA analyze and explore them using particular decisions and select the most helpful and convenient ones according to their goals, priorities and constraints.

EBM involves complex decisions, conflicting objectives and massive uncertainties. This framework (objectives and indicators) has significant potential as it can be repeatedly used for any decision even in the most complex situations (Keeney 1996b) and can encourage integrated and clear thinking of the consequences of each decision based on the values of the participants. However, this framework will have to be flexible enough as particular decisions might be different from each other, and the weighting of objectives might require re‐negotiations for each decision, in part because the people strongly affected by particular decisions will differ.

The new structure of objectives, attributes and indicators presented here can be considered as a first step in building this framework. Once it is reviewed and confirmed by all members, WCA could use it for any complex issue of concern—from the marine zoning initiative, to the unregulated float homes in Barkley Sound, to the water pollution—in which multiple objectives, emotions and history are involved (Denise Dalmer, pers. comm.). The framework will help to work (in a systematic way) through these issues with stakeholders (Denise Dalmer, pers. comm.), create better alternatives, evaluate between alternatives, identify information needs and opportunities to reduce uncertainties (Keeney and McDaniels 2001).

SDM gives the opportunity to bring together stakeholders, scientists and managers for the implementation of EBM. Stakeholders need to be able to explicitly state what matters to them in detail and their values must be incorporated in the decision‐making process; decision‐makers need to be able to express their constraints on management/monitoring; and scientists need to be able to express the differences between metrics and suggest adequate indicators to measure

the objectives given the values and management constraints. This can be best done through a communication involving representatives of the three groups.

SDM can also bring consistency to the process as all the decisions are meant to achieve the same fundamental objectives; it can bring transparency, as tradeoffs are made explicit; and it can build trust in the process as stakeholders and their values are involved since the beginning. Stakeholders might support decisions that they would not have supported otherwise, by understanding the trade‐offs and the benefits of each alternative in terms of the fundamental objectives.

2.4 Conclusions

This work provides an example of how to apply SDM for the implementation of EBM. We specifically showed how to restructure objectives to make them operational and how to integrate values, objectives and indicators in a consistent framework. Managers can use these frameworks to create and evaluate alternatives, to make more informed and consistent decisions based on stakeholder values, as well as to identify data gaps and opportunities to reduce uncertainties. By no means do we intend to provide the EBM framework to be applied for the WCVI, but rather insights and a first step towards building it. For this case study, WCA will need to explore and decide what is most appropriate for their values, needs and goals.

With this work we demonstrate that stakeholders‐derived objectives and values can help shape a collective vision and objectives for EBM; however, a systematic process and the communication between stakeholders, managers and scientists will be required to identify fundamental objectives and indicators that are consistent with stakeholder values.

Without a systematic approach, objectives will be often stated in a redundant and complex way making their operationalization very difficult, and indicators may also be defined independently of stakeholder objectives and the decision‐making process. SDM should therefore result in decision‐making processes for EBM that are more consistent and transparent.

SDM can also help build trust in the process. Stakeholders are likely to feel more engaged with the planning process of EBM and its implementation if they are involved and see their values reflected since the beginning of the process.

Tables

Table 2‐1 Original set of objectives and attributes of WCA, based on draft goals for aquatic management on the west coast of Vancouver Island (WCVI). FN means First Nations.

Ecosystem approach to management Healthy, Knowledge, Sustainable First Nations Healthy prosperous information Capacity Good Integration and economic reconciliation ecosystems and safe and building governance collaboration benefits and communities technology relationships and waterways • Responsible and • Opportunities • Healthy, • Safety • Respect • Integrate • Education for • Establishment participatory for locals diverse and (modernized aboriginal information stakeholders of priorities decisions • Sustainable resilient transportation, treaties and and to participate • Responsibility • Shared social, ecosystems infrastructure, rights knowledge for in aquatic and responsibilities cultural and • Ecosystem and response • Participation ecosystem conservation, accountability • Efficient economic productivity services) of FN in health use and • Public communication benefits • Genetically • Health decision‐ • Expertise and management reporting • Sustainable and • Balance diverse salmon (ecosystem making knowledge • New industries holistic ecological, and other health and • Ensuring from diverse • Strong First management social and species community benefits for sources Nations consistent with economic • Integrity of fish health) First Nations • Training for culture traditional values aspects and habitat • Vibrancy • Dispute users, • Improve First • Collaboration • Future • Water waste (diversified resolutions managers, Nations with other states generations management economies, between First stewards, economic self and global links • Conservation (water cultural Nations and community sufficiency and • Integrated and first in pollution and practices) other capacity community participatory fisheries disposal • Partnerships governments • Passing on stability management management management) • Integrate • FN sharing the traditional • Safety and • Sustainable • Precautionary traditional wealth with knowledge efficiency of management principle knowledge into marine • Stewardship marine • FN access to approach decision‐ resources efforts transportation natural • Conservation making) • Clear • Equipment and shipping resources as a first understanding and • Modernize the • Sustainable priority of the needs of technology Canadian Coast fisheries and • Species at risk FN • Education Guard fleet aquaculture protection • FN are the • Gather • Monitoring, • Network of second information enforcement marine priority after and protected conservation regulations areas • FN access to • Plans for natural natural resources disasters • Adaptation to climate change • EBM and values in planning and decision‐ making • Adaptive management

Table 2‐2 New structure of fundamental objectives, attributes and indicators for a decision‐making framework for EBM. CS means constructed scales. Fundamental Attributes Attributes Performance measures References objective components Foster Benefits Net benefits Net present value (Sumaila and Walters economic 2005) benefits Income Net income per year (Hobbs and Horn 1997, McDaniels et al. 2006) Employment Skilled, unskilled, (Timko and Satterfield temporal and permanent 2008a, Timko and jobs Satterfield 2008b) Goods Weight WCA criteria Distribution of benefits and Proportions of benefits (Philcox 2007) loss across stakeholders and loss retained by stakeholder groups Proportion of benefits (McDaniels et al. 2006) retained across First Nations Distribution of benefits and Benefits and loss (Timko and Satterfield loss over time distributed in time 2008a, Timko and Satterfield 2008b) Foster healthy Minimize Diversity Species richness and (Jamieson et al. 2001, ecosystems adverse evenness within defined O'Neill et al. 2008) effects on boundaries ecosystem Community composition: (DFO 2007) integrity Species richness and abundance of selected groups Community composition: (Pauly et al. 1998, Mean trophic level Jamieson et al. 2001, Pauly and Watson 2005, O'Neill et al. 2008)(Jamieson et al. 2001, O'Neill et al. 2008) Community/habitat (DFO 2007) diversity Productivity Biomass per trophic level (Jamieson et al. 2001, O'Neill et al. 2008) Target species catch and (DFO 2007) biomass Marine Concentrations of toxics (Jamieson et al. 2001, environment in sediments, water and DFO 2007, PSP 2009) al quality biota Habitat quality (area (DFO 2007) impacted, or magnitude of the activity that impact the habitat) Noise (DFO 2007) Atmospheric pollution (DFO 2007)

Fundamental Attributes Attributes Performance measures References objective components Foster healthy Avoid People Number of people (Hobbs and Horn 1997) communities adverse exposed, magnitude and effects on probability of the risk health and safety Illness and death (PSP 2009) associated to marine resources or the alternative Private and Number of private or (Hobbs and Horn 1997) public public property exposed, property magnitude and probability of the risk Avoid Cultural CS: Impacts on identified (Timko and Satterfield adverse practices cultural practices 2008a) effects on CS: Is the compensation (Timko and Satterfield vibrancy fair? 2008a) Access to CS: Are the conditions of (Timko and Satterfield natural access fair? 2008a) resources CS: Are the conditions of (Timko and Satterfield access impacted? 2008a) CS: Is the compensation (Timko and Satterfield fair? 2008a) First Nations CS: Are the treaties and (Timko and Satterfield rights and rights respected? 2008a) title CS: To what level of (Timko and Satterfield satisfaction? 2008a) Aesthetics CS: Perceived level of the Satterfield pers. comm. visual, odor and water quality impacts Foster good Foster Representati CS: Were all stakeholders (Timko and Satterfield governance participatory veness represented in the 2008a, Timko and management process? Satterfield 2008b) Fairness CS: Were there (Timko and Satterfield opportunities for the 2008a, Timko and public and stakeholders Satterfield 2008b) to participate in the decision‐making process? Competence CS: Was there a capacity (Timko and Satterfield building program and 2008a, Timko and training opportunities for Satterfield 2008b) locals? Foster shared responsibilities CS: Do all members agree WCA criteria with the distribution of tasks? CS: Do all members have (Timko and Satterfield the capacity to do the 2008a, Timko and work they were tasked Satterfield 2008b) with? Foster alignment with other CS: Were synergies and WCA criteria plans partnerships made during the process?

Fundamental Attributes Attributes Performance measures References objective components Foster good Foster Resolutions CS: To what extent there (Timko and Satterfield governance social between First is respectful relationship 2008a, Timko and agreements Nations and between First Nations Satterfield 2008b) other and other governments? governments Promote Learning from other CS: How did the board (Timko and Satterfield learning participants solve the conflicts? 2008b) through CS: How did the board (Timko and Satterfield adaptive learn from conflict 2008b) management resolution? CS: Did stakeholders (McDaniels and Gregory learn during the process? 2004) Learning from the process CS: Are the members (McDaniels and Gregory satisfied with the process 2004) experience? CS: Was new information (McDaniels and Gregory identified and integrated 2004) to the process? CS: Were there (Timko and Satterfield opportunities to review 2008a, Timko and and adjust agreements Satterfield 2008b) and policies? CS: Were there learning (Gregory et al. 2001) opportunities over time? Treating policies as CS: Were key sources of (Walters 1986, experiments uncertainty identified? McDaniels and Gregory 2004) CS: Were opportunities (Walters 1986, to reduce uncertainty McDaniels and Gregory identified? 2004) CS: Did members define (Walters 1986, means of applying McDaniels and Gregory policies to realize such 2004) opportunities? Were those policies (Walters 1986, implemented? McDaniels and Gregory 2004)

2.5 References

Arkema, K. K., S. C. Abramson, and B. M. Dewsbury. 2006. Marine ecosystem‐based management: from characterization to implementation. Frontiers in Ecology and the Environment 4:525‐532. Clemen, R. and T. Reilly. 2001. Structuring Decisions. Pages 43‐110 Making Hard Decisions with Decision Tools. Duxbury Thomson Learning. DFO. 2007. Easter Scotian Shelf Integrated Ocean Management Plan: Strategic Plan. Oceans and Habitat Branch. Fisheries and Oceans Canada. Dartmouth, Nova Scotia. 68 pp. Failing, L. and R. Gregory. 2003. Ten common mistakes in designing biodiversity indicators for forest policy. Journal of Environmental Management 68:121‐132. Gislason, G. 2007. Economic Contribution of the Oceans Sector on the West Coast of Vancouver Island. Report for Canada Department of Justice, Vancouver. Gray, J. S. 1997. Marine biodiversity: patterns, threats and conservation needs. Biodiversity and Conservation 6:153‐175. Gregory, R., L. Failing, D. Ohlson, and T. L. McDaniels. 2006. Some Pitfalls of an Overemphasis on Science in Environmental Risk Management Decisions. Journal of Risk Research 9:717 ‐ 735. Gregory, R., T. McDaniels, and D. Fields. 2001. Decision aiding, not dispute resolution: Creating insights through structured environmental decisions. Journal of Policy Analysis and Management 20:415‐432. Guerry, A. D. 2005. Icarus and Daedalus: conceptual and tactical lessons for marine ecosystem‐ based management. Frontiers in Ecology and the Environment 3:202‐211. Hobbs, B. F. and G. T. F. Horn. 1997. Building public confidence in energy planning: a multimethod MCDM approach to demand‐side planning at BC gas. Energy Policy 25:357‐375. Jamieson, G., R. O'Boyle, J. Arbour, D. Cobb, S. Courtenay, R. Gregory, C. Levings, J. Munro, I. Perry, and H. Vandermeule. 2001. Proceedings of the National Workshop on Objectives and Indicators for Ecosystem‐based Management. Page 142. FIsheries and Oceans Science, Sidney, Bristish Columbia. Keeney, R. L. 1996a. Identifying and Structuring Objectives. Pages 55‐98 in R. L. Keeney, editor. Value‐focused thinking: A Path to Creative Decisionmaking. Harvard University Press, Cambridge, Massachussets.

Keeney, R. L. 1996b. The Framework of Value‐Focused Thinking. Pages 29‐52 in R. L. Keeney, editor. Value‐focused thinking: A Path to Creative Decisionmaking. Harvard University Press, Cambridge, Massachussets. Keeney, R. L. 1996c. Thinking about values. Pages 3‐28 in R. L. Keeney, editor. Value‐focused thinking: A Path to Creative Decisionmaking. Harvard University Press, Cambridge, Massachussets. Keeney, R. L. 1996d. Value‐focused thinking: Identifying decision opportunities and creating alternatives. European Journal of Operational Research 92. Keeney, R. L. and T. McDaniels. 2001. A Framework to Guide Thinking and Analysis Regarding Climate Change Policies. Risk Analysis 21:989‐1000. Lamoreux, J. F., J. C. Morrison, T. H. Ricketts, D. M. Olson, E. Dinerstein, M. W. McKnight, and H. H. Shugart. 2006. Global tests of biodiversity concordance and the importance of endemism. Nature 440:212‐214. Leslie, H. and A. Kinzig. 2009. Resilience Science Pages 55‐73 in K. McLeod and L. Heather, editors. Ecosystem‐Based Management for the Oceans. Island Press, Washington. Leslie, H. M. and K. L. McLeod. 2007. Confronting the challenges of implementing marine ecosystem‐based management. Frontiers in Ecology and the Environment 5:540‐548. Lester, S. E., K. L. McLeod, H. Tallis, M. Ruckelshaus, B. S. Halpern, P. S. Levin, F. P. Chavez, C. Pomeroy, B. J. McCay, C. Costello, S. D. Gaines, A. J. Mace, J. A. Barth, D. L. Fluharty, and J. K. Parrish. 2010. Science in support of ecosystem‐based management for the US West Coast and beyond. Biological Conservation 143:576‐587. Levin, P. S., M. J. Fogarty, S. A. Murawski, and D. Fluharty. 2009. Integrated Ecosystem Assessments: Developing the Scientific Basis for Ecosystem‐Based Management of the Ocean. PLoS Biol 7:e1000014. Lindenmayer, D. B., C. R. Margules, and D. B. Botkin. 2000. Indicators of Biodiversity for Ecologically Sustainable Forest Management. Conservation Biology 14:941‐950. Lubchenco, J. and N. Sutley. 2010. Proposed U.S. Policy for Ocean, Coast, and Great Lakes Stewardship. Science 328:1485‐1486. McDaniels, T., H. Longstaff, and H. Dowlatabadi. 2006. A value‐based framework for risk management decisions involving multiple scales: a salmon aquaculture example. Environmental Science & Policy 9:423‐438. McDaniels, T. L. 2000. Creating and using objectives for ecological risk assessment and management. Environmental Science & Policy 3:299‐304.

McDaniels, T. L. and R. Gregory. 2004. Learning as an objective within a structured risk management decision process. Environmental Science & Technology 38:1921‐1926. McDaniels, T. L., R. S. Gregory, and D. Fields. 1999. Democratizing risk management: Successful public involvement in local water management decisions. Risk Analysis 19:497‐510. McLeod, K. L., J. Lubchenco, S. R. Palumbi, and A. A. Rosenberg. 2005. Scientific Consensus Statement on Marine Ecosystem‐Based Management. Signed by 221 academic scientists and policy experts with relevant expertise and published by the Communication Partnership for Science and the Sea at http://compassonline.org/?q=EBM.21p. O'Neill, S., C. Bravo, and T. Collier. 2008. Environmental Indicators for the Puget Sound Partnership: A Regional Effort to Select Provisional Indicators (Phase 1). Summary Report‐December 2008. Okey, T. 2009 Abstract: Ecosystem Elements of Barkley Sound and the West Coast of Vancouver Island. Barkleypedia. http://www.westcoastaquatic.ca/barkleysymposium/content/ecosystem‐elements‐ barkley‐sound‐and‐west‐coast‐vancouver‐island. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres, Jr. 1998. Fishing Down Marine Food Webs. Science 279:860‐863. Pauly, D. and R. Watson. 2005. Background and interpretation of the ‘Marine Trophic Index’ as a measure of biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences 360:415‐423. Philcox, N. 2007. Literature review and framework analysis of non‐market goods and services provided by Bristish Columbia's ocean and marine coastal resources. PSP. 2009. Identification of Ecosystem Components and Their Indicators and Targets. Puget Sound Partnership. Reid, W. V., J. A. McNeely, D. B. Tunstall, D. A. Bryant, and M. Winograd. 1993. Biodiversity indicators for policy‐makers. WRI/IUCN/UNEP Global Biodiversity Strategy. Renn, O. 2004. The Challenge of Integrating Deliberation and Expertise: Participation and Discourse in Risk Management. Pages 289‐366 in T. McDaniels and M. Small, editors. Risk Analysis and Society. An Interdisciplinary Characterization of the Field. Cambridge University Press, New York. Ruckelshaus, M., T. Kingler, N. Knowlton, and D. P. DeMaster. 2008. Ecosystem‐based management in practice: Scientific and Governance challenges. BioScience 58:59‐63.

Sumaila, U. R. and C. J. Walters. 2005. Intergenerational discounting: a new intuitive approach. Ecological Economics 52:135‐142. Tallis, H., P. S. Levin, M. Ruckelshaus, S. E. Lester, K. L. McLeod, D. L. Fluharty, and B. S. Halpern. 2010. The many faces of ecosystem‐based management: Making the process work today in real places. Marine Policy 34:340‐348. Timko, J. and T. Satterfield. 2008a. Criteria and indicators for evaluating social equity and ecological integrity in national parks and protected areas. Natural Areas Journal 28:307‐ 319. Timko, J. and T. Satterfield. 2008b. Seeking Social Equity in National Parks: Experiments with Evaluation in Canada and South Africa. Walker, B., C. S. Hollin, S. R. Carpenter, and A. Kinzig. 2004. Resilience, adaptability and transformability in social‐ecological systems. Ecology and Society 9:9. Walters, C. 1986. Adaptive Management of Renewable Resources. New York: John Wiley & Sons. Walters, C. J. and S. J. D. Martell. 2004. Fisheries Ecology and Management. Princeton University Press, Princeton. WCA. 2001. West Coast Vancouver Island. Aquatic Management Board. Terms of Reference. WCA. 2006. http://www.westcoastaquatic.ca. WCA. 2009. AMB Draft Goals for Aquatic Management in the WCVI area http://www.westcoastaquatic.ca/barkleysymposium/content/key‐documents‐1.

Chapter 3 Representing mediating effects and species reintroductions in Ecopath with Ecosim6

3.1 Introduction

Ecosystem models play a fundamental role in Ecosystem‐based management (EBM) because they improve our understanding of complex systems and allow us to integrate and visualize the trophic and non‐trophic interactions between multiple species within an ecosystem. They do this while accounting for the impacts of environmental forcing factors and human activities. They also support scenario analysis to predict policies outcomes and the effects of environmental changes (Christensen and Walters 2004a, 2004b).

Ecopath with Ecosim (EwE) is widely used to represent aquatic ecosystems (Christensen and Walters 2004a, 2004b) because of its user‐friendly interface and continued improvements to the software (Plagányi 2007). It remains the most popular tool for exploring the impacts of fisheries on ecosystems (Plagányi 2007), and has been largely recommended for supporting EBM (Kaufman et al. 2009).

To make ecosystem models in EwE widely applicable for EBM, they must account for those indirect mediating effects that have been shown to have notable impacts on community dynamics (Dill et al. 2003, Heithaus et al. 2008), sometimes overshadowing direct predation mortality (Heithaus et al. 2008). Mediating effects for instance can amplify or counteract predator‐prey relationships, facilitating or inhibiting risk to predation and competition for food and space (Dill et al. 2003, Heithaus et al. 2008). For example, kelp forests provide refuge for species like juvenile fish, and this reduces their risk to predation (negative effect for predators of juvenile fish); by providing increased feeding areas, kelp forests may also increase food availability for some predators through prey retention. Such mediating effects have been largely overlooked in ecosystem models (Dill et al. 2003, Heithaus et al. 2008).

6 A version of this chapter has been submitted for publication. Espinosa‐Romero M., Gregr E., Christensen V., Walters C., and Chan K. Representing mediating effects and species reintroductions in Ecopath with Ecosim. 43

Another important process to represent in ecosystem models is the introduction of species (e.g., reintroduction of extirpated species, or the arrival of exotic species) to an ecosystem. When a new species enters an ecosystem, dramatic changes can occur, such as with the re‐introduction of sea otters (Enhydra lutris) to the west coast of Vancouver Island (WCVI), British Columbia, Canada (Watson 1993), or the arrival of green crab (Carcinus maenas) to Bodega Bay Harbour, California (Grosholz et al. 2000).

Here, we examine how existing EwE functionality can be used to incorporate the mediating in ecosystem models. We also explore how to represent species reintroductions and range expansion in the Ecospace module of EwE, which does not directly allow such representation – Ecospace assumes that all species populations are distributed in suitable habitats in the initial state of the ecosystem (Christensen et al. 2008). We present as a case study, the reintroduction of sea otters to nearshore ecosystems on the WCVI.

3.1.1 Case study

Historically, sea otters were part of the nearshore ecosystem throughout the Pacific Northwest (Watson 2000). First Nations people hunted them and dressed in otter pelts to signify authority (Uuathluck 2007, 2009). Later, the pelts became highly prized in Asian markets and the international fur trade between the 17th and 18th centuries drove the North Pacific population close to extinction (Estes et al. 1978). In British Columbia, sea otters were extirpated between 1929 and 1930 (Watson 1993).

In the 1970s, sea otters were re‐introduced to British Columbia (Estes 1990, Watson 2000) specifically to Checleset Bay, WCVI where they have successfully re‐established themselves (Watson 1993, Nichol et al. 2009). They have reached their carrying capacity in the central portion of their range (Figure 3‐1) and their expansion continues, particularly to suitable habitats (Figure 3‐2) along the WCVI (Gregr et al. 2008).

Sea otters are widely regarded as a keystone species, able to structure nearshore marine environments by releasing macro‐algae (i.e., ‘kelp’ species Macrocystis integrifolia, Nereocystis luetkeana) from grazing pressure (Estes and Palmisano 1974). In otter‐free areas, invertebrate grazers—primarily sea urchins—are abundant and brown algae (which form kelp forests) become rare because of extensive grazing, often leading to so‐called ‘sea urchin barrens’ (Estes

and Palmisano 1974, Watson 1993). In many areas where sea otters are present, they regulate grazer invertebrates’ abundance (Estes and Palmisano 1974, Estes et al. 1978, Duggins 1980), leading to a kelp‐dominated ecosystem, often a more productive (Duggins et al. 1989) and diverse system than urchin barrens (Estes and Palmisano 1974, Simenstad et al. 1978).

Kelp forests provide a number of important ecological benefits to nearshore ecosystems (e.g., Duggins et al. 1990, Markel 2006). They contribute significantly to primary and secondary production (Duggins et al. 1990). Only a small portion of their net production enters the food web through grazing; rather, most of the net production is captured by other organisms in the form of detritus or dissolved organic matter, a high quality source of carbon for many species (Duggins et al. 1989, Duggins and Eckman 1997). The three‐dimensional structure of kelp forests provides a complex habitat (Duggins et al. 1989), which can increase feeding areas, refuges, and food availability for diverse species (Anderson et al. 1997). By disrupting the water flow, kelps can attenuate high velocity currents, which has been suggested to facilitate larval retention, survival, and recruitment (Duggins et al. 1990, Eckman and Duggins 1991). These effects of kelp are important for the consideration of ecosystem changes due to the return of sea otters.

The recovery of sea otters along the WCVI is dominated by controversy (Watson 2000). To some stakeholders, such as shellfish fishers, sea otters represent a threat as voracious predators of species with high commercial (e.g., sea urchins, geoduck, clams) and cultural (e.g., abalone) value (Watson 2000). To others, sea otters provide ecological and social benefits in terms of biodiversity and ecosystem productivity (because of kelp forest expansion) as well as opportunities for finfish fisheries (e.g., rockfish) and tourism industries (Markel 2006, COSEWIC 2007).

Thus, as the sea otter populations expand their range along the WCVI, a comprehensive science‐ based understanding of the potential effects might be useful for decision‐making. We therefore built an ecosystem model using Ecopath with Ecosim (EwE) to represent potential effects from sea otter reintroduction and expansion in the area. We had three specific objectives for this model: to demonstrate how to represent a) the contribution of kelp‐particulate detritus to primary and secondary production, b) the mediating effects provided by kelp by increasing feeding areas and food availability for some species; and c) the sea otter reintroduction and

expansion across the study area to display these effects across space and over time.

3.2 Methods

We used three modules (Ecopath, Ecosim and Ecospace) of EwE version 6 (EwE6) to address our three objectives.

3.2.1 Ecopath parameterisation

The conceptual food web contains 14 functional groups of commercially and ecologically important species (Figure 3‐3) living in the nearshore ecosystem, which are likely to be most affected by sea otter reintroduction. This food web was adapted from Simenstad et al. (1978), Halpern et al. (2006), Markel (2006), and Markel (pers. comm.).

We parameterized our model for the nearshore waters along the Pacific coast of Canada, using data from previous EwE models (Ainsworth et al. 2002) and existing databases (Markel 2006). Because the models and database sources had different spatial extents and functional groups than our analysis, we adjusted the reported parameters (Table 3‐1 and 3‐2). Parameter‐ estimation and balancing procedures are described in Appendix A.

3.2.2 Modeling benefits provided by kelp forests: using Ecopath and Ecosim

We accounted for two effects provided by kelp forests in the model: a) increased primary and secondary production via kelp‐derived detritus and b) increased feeding areas and food availability for some species through the retention of prey.

We modeled the contribution of kelp to primary and secondary production by adding a detritus functional group for kelp‐derived detritus. Detritus groups in EwE6 represent the biomass of all the species that are not eaten, but that is retained in the system and decomposed into organic matter. When a group of detritus is added to the model and linked to particular species, the organic matter of those species will go directly to the new detritus group.

The kelp‐derived detritus group was linked to the canopy and understory kelp biomass and included in the diets of filter feeders (Duggins and Eckman 1997), benthic invertebrates, sea urchins (Duggins et al. 1989, Markel 2006), forage fish (Hand and Berner 1881) and

zooplankton (Markel 2006, Markel pers. comm.) as they can feed on drift kelp or kelp‐derived organic matter. In this way, in the model, organic matter derived from kelp goes directly to the new detritus functional group, and it is directly consumed by these species.

The mediating effects of kelp in augmenting feeding areas and food availability for some species due to prey retention was modeled using the Ecosim ‘mediation function’. The mediation function allows the explicit representation of the effects of one species/group in facilitating or inhibiting prey‐predator relationships of two other species. By default, it affects the trophic flow rate (TFR), the biomass per time step consumed by a predator (Christensen et al. 2008), which is defined as follows:

* ai, j TFR(biomass/time) = ( )"Vi, j " Pj (1) Ai, j

where a*i,j is the effective area searched by each unit of predator (Pj) for prey per unit time , Ai,j ! is the foraging area by predator j on prey i, Vi,j is the vulnerable prey biomass per unit area, and

Pj the effective predator abundance per unit area. The vulnerable prey biomass is defined by an

exchange rate v’i,j which defines the proportion of the total prey biomass vulnerable to predator j (Christensen et al. 2008).

The mediation function uses multipliers from 0 to 1 scale to affect the parameters a*ij, Aij and v’ij of the trophic flow rate. Modelers can graphically define the function in the software or choose the relation between the multiplier and the biomass of the mediating organism from four different relationships (linear, sigmoid, hyperbolic and exponential). Up to five mediation functions per prey‐predator relationship are possible (Christensen et al. 2008).

In our model, we applied only one type of mediation where the multiplier was used to change

the parameters Ai,j and v’i,j together. This represents the concept that in the presence of kelp,

some species increase their foraging arenas or feeding areas (Ai,j) in which the vulnerable

biomass of some prey (V’i,j) becomes more available. We assumed a sigmoid relationship between kelp biomass and the multiplier; meaning that increased kelp biomass will increasingly affect the multiplier until a point where kelp biomass will not make significant changes in the multiplier. To define the trophic flow rates or prey‐predator relationships affected by the mediation function (Figure 3‐3) we consulted with experts in the regional nearshore systems

(Markel and Martone, pers. comm.). To see the impacts of the mediation in the ecosystem, we analyzed the results in intervals of 10 years.

Vulnerabilities for each functional group in Ecosim define how the biomass of different groups is controlled in the system (top‐down or bottom‐up). Low vulnerabilities (close to 1) indicate that prey only becomes vulnerable to predation when it leaves its refuge. High vulnerabilities occurs in situations where prey has no refuge and is easily encountered by its predator, meaning that predator biomass defines how much is eaten of a prey (top‐down) (Christensen et al. 2008). We kept the default vulnerability value of 2.0 for most groups, except for sea urchins eaten by sea otter—sea urchins are supposed to be easier to capture by sea otters compared with other prey—kelp forests and kelp detritus which were set to 10.0.

3.2.3 Spatial representation of sea otter reintroduction and expansion: Ecospace structure

Sea otters were reintroduced in one location on the WCVI and after 40 years they have re‐ established themselves in the area of reintroduction and continue to expand along the coast (Gregr et al. 2008). To correctly represent this ecological history, we wanted our model to demonstrate how the reintroduction of sea otters to one specific area (in the model, one cell) can lead to their expansion across the whole study area.

To explore how this dynamic can be represented in Ecospace, we began with a linear map defined by a single row of 30 10x10 km2 cells. This map hypothetically represents almost 300 km2 of the nearshore ecosystem on the WCVI, a region that includes both the area where sea otters were reintroduced, and the area into which they are expanding.

To configure the Ecospace model, habitats need to be defined for the model cells and suitable habitats assigned to each functional group. Habitat suitability is defined by assigning values to relative dispersal, vulnerability and feeding rates. When set to 1, these parameters are disabled, making all habitats equally suitable.

On initialization, Ecospace automatically calculates the initial biomass per functional group per cell by dividing the Ecopath biomass of each group among the cells within suitable habitats for that group, assuming that all species are already distributed in the ecosystem. After this initial

distribution, organisms move across cells according to assigned parameters for dispersal rates, food availability, and vulnerability to predation. Species avoid cells identified as unsuitable habitats and/or move out faster when they end up in those cells (Christensen et al. 2008).

We created two scenarios to work around the assumption of initial species distribution in the system, to represent the range expansion of sea otters following the reintroduction to one site.

In the Habitatbased scenario, we initialized the model with two types of habitat. We assigned the first habitat to a single edge cell and defined it as suitable for sea otters. We did this to concentrate all the biomass of this species in cell 1 for the first year, simulating the reintroduction to a specific area. We assigned the second habitat to the 29 remaining cells and made it equally suitable for year 2 to the end of the time period by assigning a value of 1 to parameters for relative dispersal, vulnerability and feeding rates for sea otters. We did this to allow sea otter expansion to the cells of the second habitat after the first year.

In the Culling scenario, we defined the whole area as a single suitable habitat for sea otters. We then defined cell 1 as a marine protected area (MPA) and culled sea otters with high fishing effort for the first 10 years of a 100‐year simulation (Figure 3‐4) in the remaining 29 cells. In this way we retained the sea otter biomass in cell 1, and drove the population to 0 in cells 2 to 30. We assumed that after 10 years the population in cell 1 would be the only surviving one, and that it would become the source of the expansion to the other cells.

Ecospace represents species movement in two ways: Diffusion and Individual based model (IBM). Diffusion assumes functional‐group movement from one cell to the four adjacent cells per time step according to set dispersal rates, vulnerability to predation and feeding rates. For the cells at the border of the map, Ecospace assumes diffusion from the outside, a logical assumption since the study area is not assumed to be isolated from a larger geographic area. This leakage across the study area boundaries can be avoided by adding land limits to the map. We thus explored both Habitatbased and Culling scenarios with and without land limits to investigate how the diffusion from areas outside affected the spatial representation of sea otter expansion. To analyze the impacts of diffusion, we sampled the results every 10 years.

The IBM requires more than one life history stage (called stanzas in EwE6) to be defined for a

species. It assumes that each life history stage is comprised of groups of organisms with different body masses, each of which moves randomly across cells. Because IBM involves a random component, we ran the model 100 times to analyze how this randomization influenced the results. To analyze the variation of the results, we divided the whole area into eight regions and calculated the average biomass, standard deviation and coefficient of variation per region.

We compared both types of movement models, diffusion and IBM, to assess which provided a better spatial representation of sea otter expansion through the study area. To improve comparisons of the results obtained from both types of models, we ran the models using two stanzas (juveniles and adults) for sea otters and the same dispersal parameters. The models account for stanzas biomass in different ways, described in Walters et al. (2010). We also compared the results obtained from Ecosim with those from Ecospace, to examine how the spatial representation influenced the predicted sea otter recovery.

We defined hypothetical base dispersal rates to run the two scenarios: 2 km per year for sea otters and fish populations and 1 km per year for invertebrates and primary producers. We assumed mortalities of sea otter juveniles and adults to be 0.15 and 0.20 respectively. For the Culling scenario we set the initial fishery landings to 0.015 ton/km2/year with a high fishing pressure for the first 10 years.

3.3 Results and discussion

3.3.1 Ecopath results

The network analysis of our food web demonstrates the interactions between species within the ecosystem (Figure 3‐5). Sea otters had negative effects on sea urchins and crabs due to predation and positive effects on canopy and understory kelp by releasing kelp from the sea urchin grazing pressure. This sea otter‐urchin‐kelp relationship has been demonstrated empirically in the Aleutian Islands, Alaska (Estes and Palmisano 1974, Duggins et al. 1989) and the WCVI (Watson 1993). The negative effects of sea otters on crab abundance have been less explored, but there is some evidence of crab population declines in Alaska as a result of increased sea‐otter abundance (Garshelis and Garshelis 1984). Results also suggest positive impacts of sea otter abundance on lingcod and forage fish, previously described by Markel (2006).

Sea otters caused indirect positive effects on benthic invertebrates and filter feeders by releasing the grazing pressure on kelp, an important food source for these groups. This is consistent with empirical information provided by Duggins et al. (1989).

Sea otters negatively affected themselves due to intra‐specific competition, which is likely to occur in regions where they have reached their carrying capacity (Gregr et al. 2008). Indirect negative effects were shown on surfperches due to increased predation by lingcod, consistent with effects described by Markel (2006).

3.3.2 Benefits provided by kelp forests over time

To assess the change in kelp abundance as otters recovered, we compared the final biomass with the biomass at year 20 (0.39 ton/km2), the year of lowest sea otter abundance (.01 ton/km2). Sea urchins in this time period decreased 44% and 69% with the Habitat‐based and Culling scenario respectively. Canopy kelp increased over 125% and understory kelp by almost 50% after sea otters recovered under the Habitatbased scenario. Under the Culling scenario the increases were higher (142 and 57% respectively).

Including kelp detritus as a functional group allowed us to analyze its contribution to primary and secondary production in the ecosystem. In the presence of sea otters, kelp forests, largely through kelp‐derived detritus, contribute to the biomass of benthic invertebrates, filter feeders and sea urchins (Figure 3‐4). The biomass for kelp‐derived detritus increased 66% (from 0.09 to 0.16 ton/km2) in the Habitatbased scenario and 77% (from 0.09 to 0.17 ton/km2) in the Culling scenario between year 20 and year 100 (Figure 3‐5). We expected the contribution from kelp forests to primary production to be even higher. However, we realize that the ecotrophic efficiency (the ratio of biomass flow out of a group to the biomass that flows in) of kelp‐derived detritus was 0.99 in year 1. This means that virtually almost all the kelp‐derived detritus produced is consumed and could explain the limited increase of kelp‐derived detritus in subsequent years. The inclusion of empirical data would improve the representation of kelp and kelp‐derived detritus in the model, and allow a more accurate representation of kelp contribution to the food web.

An important dynamic not represented in the model was that sea urchins feed on kelp forest holdfasts often leaving big proportions of the blades available for grazers or decomposition. On

one hand, this dynamic intensifies the negative impacts of sea urchins on kelp forest abundance and on the other, contributes to the production of kelp detritus. Because this dynamic was not represented in the model, the impact of sea urchins on kelp and kelp‐derived detritus abundance may be limited (underestimated).

The application of the mediation function to represent kelp‐derived benefits (i.e. increased feeding areas for predator and risk vulnerability for prey) introduced changes in ecosystem dynamics, specifically the biomass of some functional groups (Figure 3‐6). The results with the mediating effects show that lingcod and rockfish biomasses increased as sea otters and kelp forests increased. However, surfperch biomass declined due to lingcod predation. While the effects of lingcod on surfperches have been identified (e.g., Markel 2006), it will be important to compare predicted results with observed data to better characterize the magnitude of these effects. Benthic invertebrates’ biomass increased 17% with the mediation function due to the increased availability of kelp‐derived detritus and decreased predation from rockfish and lingcod—the first 50 years of time period the biomass of these fish populations were low due to low biomass of kelp forest and sea otters. Some functional groups showed slight changes in abundance. For example, crab biomass increased up to 2.8% due to the increased availability of benthic invertebrates (prey) and decreased predation of rockfish. The decrease in zooplankton biomass (up to 1.5%) likely contributes to the observed increases in phytoplankton (up to 2.4%), kelp‐derived detritus (up to 22%) and detritus (up to 1.9 %) biomasses.

Mediation effects under the Habitatbased scenario were not available in Ecosim. Ecosim drives all the species to equilibrium in year 1. Then, when a model does not have any drivers of dynamic change (e.g., fisheries and fishing effort) or an unstable initial condition (e.g., biomass accumulation assumed from outside the map), Ecosim shows no change in the species biomasses for the subsequent years. In addition, the mediation function is applied after year 1. Thus, in the Habitatbased scenario, species were driven to equilibrium and mediation effects could not be observed because there was no change in kelp forest biomass. The spatial representation (Ecospace) was necessary to observe mediating effects for this scenario. For the Culling scenario, which includes a fishery, mediating effects could be observed in both Ecosim and Ecospace.

3.3.3 Spatial representation of sea otter reintroduction and expansion

With the Habitatbased scenario, neither sea otter reintroduction nor expansion across the area correspond to observations/expectations. Sea otters were supposed to be concentrated in cell 1 (at the edge of the linear region) from which the expansion was intended to proceed. Instead, the population started to expand from the centre to the sides, an unlikely situation given that the dispersal rate would not allow them to move that fast to the centre of the map.

While the use of land limits eliminated the diffusion from the outside, the results demonstrated little differences in the species biomass results over time. Except for surfperches whose biomass variation was on average 1.6% with the diffusion, the rest of the functional groups variation was lower than 1%. Despite the land limits, the same spatial dynamic was observed: sea otters expanded from the centre to the sides.

Further analysis revealed that Ecospace, to initialize the model, distributes the initial biomass assigned to functional groups in Ecopath not only across the cells with suitable habitat but also across cells with unsuitable habitat. It concentrates high biomass in the former and low biomass in the latter (not 0 as intended). The underlying assumption of the software is that the low biomass in unsuitable habitats will move rapidly to suitable habitats using a gradient in Ecospace, intended to make species leave the unsuitable habitats quickly (Christensen et al. 2008). In the Habitatbased scenario, as we made both habitats equally suitable, the low initial biomass assigned by Ecospace to the second (unoccupied) habitat did not leave this habitat, instead started to increase due to the high food availability. While one might think that species at low biomasses may not survive; in Ecospace, species are only driven to 0 biomass with predation or fishing pressure.

With the Culling scenario we avoided the initial biomass assignment to the cells where we wanted sea otter biomass to be 0 by applying high fishing effort for sea otters for the first 10 years. The population retained in the protected area established in cell 1, providing the population for the subsequent expansion to the other cells.

When using the diffusion model, the expansion of otters from year 0 to year 70 was uniform (from the first cell onwards) and well represented. However, otters seemed to expand faster than the dispersal rates, and dominated the whole area by year 100 (Figure 3‐7). The reason

might be that Ecospace assumes a normal distribution of dispersal rates around the given value, allowing species biomass to move faster and slower than the given rate. Furthermore, the diffusion model does not distinguish the position that the functional groups have within a cell. Therefore, once a species occupies a cell, it can move to another one in the next time step, even if the cell is considerably larger than the stated dispersal rate.

When adding the two stanzas (sea otters juvenile and adults) to compare diffusion and IBM models under this scenario, we observed that biomass estimates with the IBM were more conservative than those with the diffusion model (Figure 3‐8). The two dispersal models represented the spatial expansion differently (Figure 3‐7). The diffusion model showed a uniform distribution of sea otters while the IBM model showed a random distribution. Contrary to the diffusion model, IBM does recognize the position that species occupy within the cells which makes the sea otter expansion slower.

The diffusion model showed similar dynamics with and without the stanzas. Sea otters started to move across the whole study area after year 70. However, the diffusion model with stanzas showed that at the end of the time period (year 100), sea otters did not dominate the ecosystem. Their biomass was high in the first half of the map and medium to low in the second half of the map (Figure 3‐7).

The IBM showed an initial expansion starting from cell 1, but after some years the otter expansion appeared to be random in individual runs. At the end of the simulations, some cells showed a high biomass of sea otters, while others showed a low biomass or sometimes no biomass at all (Figure 3‐7). The analysis of the influence of the randomization in the IBM showed high standard deviations and coefficient of variations in most regions (Table 3‐3). The model shows exponential decay in distribution from the source point to the rest of the cells at first when sea otters are expanding. Then, the population in areas further from the starting point of expansion should increase until there is a uniform distribution of the biomass, as our model assumes homogeneous habitat and primary production. Because our analysis only includes a period of 100 years, the predicted otter biomasses are somewhat even in regions 3 to 8, but show a major concentration of biomass in regions 1 and 2. We expect that over a longer period of time, the biomass will be more evenly distributed.

Because the distribution of sea otters is driven by food availability and consequently on suitable habitat (Gregr et al. 2008), the integration of habitat information and habitat preferences of each species will improve the representation of sea otter expansion with both models. If the ‘patchiness’ of species habitat is included in the model, we expect that the spatial distribution of sea otters predicted by both models (diffusion and IBM) will be patchier and perhaps more similar to observations.

Our comparison of the results obtained from the Ecosim and the two Ecospace models showed that the recovery of sea otters with no space (Ecosim) starts in year 30 with rapid population growth from year 40 to year 70. With the spatial representation, the recovery starts in year 10; however, population growth is slower with diffusion and slowest with IBM (Figure 3‐9). This suggests that models that better represent exploitation and depletion of the species across a study area (Ecospace models) provide more conservative estimates of population growth and biomass than those models that assume widespread populations with full access to vulnerable food resources (Ecosim models).

3.4 Conclusions

With this work, we show how to represent mediating effects in EwE6 and species reintroductions in Ecospace, as well as their effects on the dynamics of a particular ecosystem.

Our model of nearshore ecosystems on the WCVI represented some trophic cascades effects due to sea otter reintroduction and expansion. In the presence of otters, kelp forests, lingcod and rockfish biomasses increased, while sea urchin and crab biomasses decreased due to sea otter predation. Surfperches were affected negatively due to lingcod predation.

With the model we demonstrated how to account for kelp forest benefits by using an additional detritus functional group and the mediation function available in EwE6. This explicit accounting for kelp‐derived detritus showed the contribution of kelp forests to primary production and primary consumers of kelp‐derived detritus.

The mediation function represented the mediating effects provided by kelp forests to species such as finfish (specifically to lingcod, rockfish and forage fish groups), crabs, and benthic

invertebrates. We found that such mediation effects can only be seen in Ecosim if the model includes a fishery or an unstable initial condition (e.g., biomass accumulation). Otherwise, Ecosim leads the functional groups to equilibrium in year 1 showing no differences between species biomass over time. Under such conditions, mediation effects can only be observed in Ecospace, where a spatial representation is required.

We only used one mediation function to represent the increased feeding areas and food availability for some predator. A better representation of the benefits provided by kelp will require the integration of the other mediating effects caused by this species such as the provision of refuge for some species. We applied the mediation function to hypothetical prey‐ predator relationships based on consultation with experts. However, this mediation function may apply to other prey‐predator relations in the food web. In addition, more mediating effects caused by other species may take place in the ecosystem.

Our model demonstrates how to incorporate these effects in ecosystem models and the possible implications for the results. The identification of mediating effects and the extent to which these effects alter prey‐predator relations are often hard to identify even with relevant empirical data. In addition, modeling these effects requires numerous assumptions that will be reflected in changes of ecosystem dynamics but that do not stem directly from empirical data— accordingly, modelers should use caution when applying and interpreting these effects in ecosystem models.”

Our Culling scenario presented a way to model the reintroduction of a species to a specific region and its expansion across the whole study area. We found that the definition of habitats and habitat suitability for species (i.e., Habitatbased scenario) does not allow such representation (reintroduction of a species) due to the initial biomass assignment by Ecospace to unsuitable habitats. The Culling scenario helped to avoid the assignment of biomass to the cells where 0 was intended and showed a coherent representation of the sea otter expansion.

The two types of movement models (diffusion and IBM) demonstrated similar trends in the total biomass of functional groups for most cases but represented the sea otter expansion in two different manners. The diffusion model led to a uniform and more rapid dispersion while the IBM result in a more heterogeneous distribution with slower dispersion. Although the

results of IBM may seem closer to reality, as some variability among areas for high and low biomass of sea otters can be expected, such heterogeneity would in any case accompany a more realistic representation with the incorporation of heterogeneous habitat (habitat was assumed homogenously suitable).

The comparison of the results obtained with Ecosim and Ecospace models showed that models that represent species exploitation and depletion spatially provide more conservative estimates of population growth and biomass (being IBM the most conservative) than non‐spatial models, which assume widespread species with full access to vulnerable food sources (e.g., Ecosim models).

There are several model enhancements necessary to provide a more realistic representation of how the nearshore ecosystem on the WCVI responded to sea otter reintroduction. These include incorporating fisheries information, reviewing functional groups, and fitting the model with empirical information. In addition, the spatial representation of the ecosystem would be improved by including habitat types and species habitat preferences, and by accounting for any significant sex‐specific differences, such as dispersal rates between male and female sea otters.

Although more information will help to better characterize the ecosystem, our hypothetical model is nevertheless able to represent key ecological dynamics—empirically demonstrated— such as direct and indirect kelp‐derived benefits, spatial expansion of sea otters, and the trophic cascade effects resulting from the re‐introduction of a keystone species.

Tables

Table 3‐1 Basic inputs for Ecopath. B means biomass, P/B production per unit biomass, Q/B consumption per unit biomass and EE ecotrophic efficiency. Parameters were adapted from Ainsworth et al. (2002), Martell (2002), (Preikshot 2005), Heymans (2005a, 2005b) and Froese and Pauly Editors (2010). See Appendix A for details on how parameters were estimated. Functional Trophic B(t⋅km2) P/B(year1) Q/B(year1) EE P/Q groups level Sea otters 3.49 0.127 0.15 75.00 Lingcod 4.00 .0650 1.00 3.30 Rockfish 3.62 1.50 0.26 3.36 Forage fish 2.66 0.60 8.40 0.95 Surfperches 3.42 1.00 0.70 10.82 Crabs 3.36 8.00 3.00 0.275 Benthic 2.55 43.00 1.90 12.70 invertebrates Filter feeders 2.00 20.00 1.20 0.20 Sea urchins 2.00 30.00 0.40 0.25 Zooplankton 2.11 68.00 16.50 62.50 Phytoplankton 1.00 22.00 179.0 Canopy kelp 1.00 9.00 5.30 Understory kelp 1.00 18.00 5.30 Kelp detritus 1.00 0.11 Detritus 1.00 10.00

Table 3‐2 Diet compositions. Predators are listed in column heading, preys are listed in rows. Parameters were adapted from Ainsworth et al. (2002), Martell (2002), (Preikshot 2005), Heymans (2005a, 2005b), and Markel (unpublished data). See Appendix 1 for details on how parameters were estimated.

hes Functional groups Sea otters Lingcod Rockfish Forage fish Surfperc Crabs Benthic invert. Filter feeders Sea urchins Zooplankton Sea otters Lingcod Rockfish 0.25 0.001 Forage fish 0.01 0.50 0.02 0.10

Functional groups Sea otters Lingcod Rockfish Forage fish Surfperches Crabs Benthic invert. Filter feeders Sea urchins Zooplankton Crabs 0.35 0.25 Benthic 0.10 0.429 0.70 0.50 invertebrates Filter feeders 0.30 0.20 Sea urchins 0.33 Zooplankton 0.30 0.60 0.30 0.20 0.50 0.10 Phytoplankton 0.30 0.40 Canopy kelp 0.10 0.40 Understory 0.10 0.40 kelp Kelp detritus 0.10 0.01 0.10 0.15 0.01 Detritus 0.49 0.15 0.05 0.49

Table 3‐3 Results from 100 runs of the individual‐based model (IBM), indicating sea otters average biomass, standard deviation and coefficient of variation per region. Each region includes four cells except for region 1 which includes only one cell. Cell 1 represents the area where sea otters where reintroduced. Region Cells included Average Standard Coefficient of predicted deviation variation biomass (t⋅km2) (t⋅km2) 1 1 0.1530 0.0136 .0888 2 2,3,4,5 0.1510 0.0093 .0616 3 6,7,8,9 0.0893 0.0441 .4938 4 10,11,12,13 0.0860 0.0392 .4558 5 14,15,16,17 0.0917 0.0421 .4591 6 18,19,20,21 0.0935 0.0408 .4364 7 22,23,24,25 0.0857 0.0361 .4212 8 26,27,28,29 0.0786 0.0459 .5840 Total 1‐29 0.0972 0.0003 .0309

Figures

Figure 3‐1 Reintroduction site, current range and sites with optimum habitat for sea otters based on intertidal complexity similarity with the reintroduction site (Gregr et al. 2008). Published with the permission of the authors.

Lingcod Sea otter Rockfish

Surfperch Crabs

Forage fish Benthic invertebrates Filter Sea urchins (micro and small grazers) feeders (red & green)

Zooplankton

Phytoplankton Understory Canopy kelp kelp

Detritus Kelp-derived detritus

Figure 3‐2 Food web of the nearshore ecosystem on the WCVI. Trophic interactions adapted from Simenstad et al. (1978), Halpern et al. (2006), Markel (2006), and Markel (pers. comm.). Grey arrows represent the hypothetical prey‐ predator relationships affected by mediating benefits provided by kelp forests (i.e. increased feeding areas and food availability through prey retention) based on Markel and Martone (pers. comm.).

PA Cull Figure 3‐3 Culling scenario, with one suitable habitat. To simulate the sea otter reintroduction, we worked around the Ecospace’s assumption on the initial distribution of biomass. Sea otter biomass (in yellow) was assigned to all the cells in the map, one protected area (PA) was set in the first cell and a cull in the form of high fishing effort for the first 10 years in the 29 remaining cells. The intention was to concentrate sea otter biomass in cell 1 and drive the population down to 0 in the rest of the cells.

Impacted group Impacting group Sea otters Lingcod Rockfish Forage fish Surfperches Crabs Benthic invert. Filter feeders Sea urchins Zooplankton Phytoplankton Canopy kelp Understory kelp Kelp detritus Detritus Sea otters Lingcod Rockfish Forage fish Surfperches Crabs Benthic invert. Filter feeders Sea urchins Zooplankton Phytoplankton Canopy kelp Understory kelp Kelp detritus Detritus

Figure 3‐4 Mixed trophic impacts. Relative impacts between species due to predation and competition. Black bars above a line indicate positive effects, and grey bars below a line indicate negative effects.

Figure 3‐5 Results of kelp‐derived detritus biomass for the two scenarios. The solid line represents the results for the Habitat based scenario and dotted line the results for the Culling scenario.

Sea Otter Lingcod Rockfish Forage fish Surfperch 0.4 17.5 1.04 0.070 1.50 17.0 1.02 0.3 0.065 1.45 16.5 1.00 0.060 0.2 16.0 0.98 1.40 0.055 15.5 0.96 0.1 0.050 1.35 15.0 0.94 Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) 14.5 0.92 0.0 0.045 1.30 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80

Years Years Years Years Years Crab Benthic invert. Filter feeders Sea urchin Zooplankton

48 26 8.5 68.2 25 30 8.4 46 24 25 68.0 8.3 23 67.8 20 44 22 8.2 67.6 21 15 42 67.4 Biomass(ton/km2) 8.1 Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) 20 10 19 67.2 Biomass(ton/km2) 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80

Years Years Years Years Years Phytoplankton Canopy kelp Understory kelp Kelp-derived det. Detritus 22.4 0.18 24 10.10 22.3 16 0.16 22.2 14 22 10.05 22.1 12 20 0.14 10.00 22.0 10 18 0.12

Biomass(ton/km2) 21.9 Biomass(ton/km2) 8 Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) 9.95 16 0.10

0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 Year Years Years Years Years Years

Figure 3‐6 Results with and without the ‘mediation’ function. Species biomass over a 100‐year period. Solid lines show the results with the mediation function which incorporate the indirect impacts of kelp on other on species, by providing increased feeding areas and more vulnerable prey to some predators.

Figure 3‐7 Spatial distribution of sea otters using the Culling scenario and the two types of movement models: diffusion and individual based model (IBM). Yr represents the years of the results. Blue represents low biomass and red high biomass. The results of IBM were taken from a one single run.

Sea Otter Lingcod Rockfish Forage fish Surfperch 1.55 1.06 0.20 0.070 17.0 1.50 1.04 0.065 16.5 0.15 1.02 0.060 1.45 16.0 1.00 0.10 0.055 1.40 15.5 0.98 0.050 15.0 0.05 1.35 0.96 Biomass(ton/km2) Biomass(ton/km2) 0.045 Biomass(ton/km2) Biomass(ton/km2) 14.5 Biomass(ton/km2) 0.94 0.00 1.30 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 Years Years Years Years Years Crab Benthic invert. Filter feeders Sea urchin Zooplankton 8.35 48 24 68.0 8.30 30 8.25 47 23 67.8 8.20 46 22 25 67.6 8.15 21 45 67.4 8.10 20 44 20 Biomass(ton/km2) 8.05 Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) 67.2 19 8.00 43 15 Biomass(ton/km2) 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 Years Years Years Years Years Phytoplankton Canopy kelp Understory kelp Kelp-derived det. Detritus

22.4 10.12 14 22 0.15 10.10 22.3 10.08 12 20 0.14 10.06 22.2 10 0.13 10.04 18 10.02 22.1 8 0.12 10.00 Biomass(ton/km2) Biomass(ton/km2) Biomass(ton/km2) 16 Biomass(ton/km2) Biomass(ton/km2) 0.11 9.98 22.0 0 40 80 0 40 80 0 40 80 0 40 80 0 40 80 Year Years Years Years Years Years

Figure 3‐8 Populations dynamics in a 100 year period. Solid lines represent the results with the Diffusion model and dotted lines the results with the Individual based Model (IBM).

0.20

0.15

0.10 otters

Biomass(ton/km2) 0.05

0.00

0 20 40 60 80 100 Year Figure 3‐9 Sea otter recovery with Ecosim, Diffusion and Individual year based Model (IBM). The solid line represents the results with Ecosim which assumes a single homogeneous area. The dashed line represents the results with the Diffusion model, and the dotted line the results with Individual based Model (IBM).

3.5 References

Ainsworth, C., J. J. Heymans, T. J. Pitcher, and M. Vasconcellos. 2002. Ecosystem models of the Northern British Columbia for the time periods 2000, 1950, 1900 and 1750. Fisheries Center. University of British Columbia, Canada. Anderson, R. J., P. Carrick, G. J. Levitt, and A. Share. 1997. Holdfasts of adult kelp Ecklonia maxima provide refuges from grazing for recruitment of juvenile kelps. Marine Ecology Progress Series 159:265‐273. Beaudreau, A. 2009. The Predatory Role of Lingcod (Ophiodo elongatus) in the San Juan Archipelago, Washington. University of Washington, Seattle. Christensen, V. and C. J. Walters. 2004a. Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling 172:109‐139. Christensen, V. and C. J. Walters. 2004b. Trade‐offs in ecosystem‐scale optimization of fisheries management policies. Bull. Mar. Sc. 74(3):549562. Christensen, V., C. J. Walters, D. Pauly, and R. Forrest. 2008. Ecopath with Ecosim version 6. Lenfest Ocean Futures Project. COSEWIC. 2007. COSEWIC assessment and update status report on the sea otter Enhydra lutris in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa. vii + 36 pp. (www.sararegistry.gc.ca/status/status_e.cfm). Dill, L. M., M. R. Heithaus, and C. J. Walters. 2003. Behaviorally mediated indirect interactions in marine communities and their conservation implications. Ecology 84:1151‐1157. Duggins, D. O. 1980. Kelp Beds and Sea Otters: An Experimental Approach. Ecology 61:447‐453. Duggins, D. O. and J. E. Eckman. 1997. Is kelp detritus a good food for suspension feeders? Effects of kelp species, age and secondary metabolites. Marine Biology 128:489‐495. Duggins, D. O., J. E. Eckman, and A. T. Sewell. 1990. Ecology of understory kelp environments. II. Effects of kelps on recruitment of benthic invertebrates. Journal of Experimental Marine Biology and Ecology 143:27‐45. Duggins, D. O., C. A. Simenstad, and J. A. Estes. 1989. Magnification of Secondary Production by Kelp Detritus in Coastal Marine Ecosystems. Science 245:170‐173. Eckman, J. and D. Duggins. 1991. Life and death beneath macrophyte canopies: effects of understory kelps on growth rates and survival of marine, benthic suspension feeders. Oecologia 87:473‐487. Estes, J. A. 1990. Growth and Equilibrium in Sea Otter Populations. Journal of Animal Ecology 59:385‐401.

Estes, J. A. and J. F. Palmisano. 1974. Sea Otters: Their Role in Structuring Nearshore Communities. Science 185:1058‐1060. Estes, J. E., N. S. Smith, and J. F. Palmisano. 1978. Sea Otter Predation and Community Organization in the Western Aleutian Islands, Alaska. Ecology 59:822‐833. Froese, R. and D. Pauly. Editors. 2010. FishBase. World Wide Web electronic publication. www.fishbase.org, version (03/2010). Garshelis, D. L. and J. A. Garshelis. 1984. Movements and Management of Sea Otters in Alaska. The Journal of Wildlife Management 48:665‐678. Gregr, E. J., L. M. Nichol, J. C. Watson, J. K. B. Ford, and G. M. Ellis. 2008. Estimating carrying capacity for sea otters in British Columbia. Journal of Wildlife Management 72:382‐388. Grosholz, E. D., G. M. Ruiz, C. A. Dean, K. A. Shirley, J. L. Maron, and P. G. Connors. 2000. The Impacts of a Nonindigenous Marine Predator in a California Bay. Ecology 81:1206‐1224. Halpern, B. S., K. Cottenie, and B. R. Broitman. 2006. Top‐down vs. bottom‐up effects in kelp forests ‐ Response. Science 313:1738‐1739. Hand, C. and L. Berner. 1881. Food of the Pacific Sardine (Sardinops caerulea). Fishery Bulletin of the Fish and Wildlife Service 164 60:14. Heithaus, M. R., A. Frid, A. J. Wirsing, and B. Worm. 2008. Predicting ecological consequences of marine top predator declines. Trends in Ecology & Evolution 23:202‐210. Heymans, S. J. J. 2005a. Ecosystem model of the eastern Aleutians and central Gulf of Alaska in 1963. Pages 83‐105 in S. Guenette and V. Christensen, editors. Food web models and data for studying fisheries and environmental impacts on Eastern Pacific ecosystems. UBC Fisheries Center, Vancouver. Heymans, S. J. J. 2005b. Ecosystem models of the Western and Central Aleutian Islands in 1963, 1979 and 1991. Pages 8‐82 in S. Guenette and V. Christensen, editors. Food web models and data for studying fisheries and environmental impacts on Eastern Pacific ecosystems. UBC Fisheries Center, Vancouver. Kaufman, L., L. B. Karrer, and C. H. Peterson. 2009. Monitoring and Evaluation. Pages 115‐128 in K. L. McLeod and H. M. Leslie, editors. Ecosystem‐based Management for the Oceans. Island Press, Washington, D.C. Markel, R. W. 2006. Sea otter and kelp forest recovery: Implications for nearshore ecosystem, fishes and fisheries. Parks Canada.

Martell, S. J. D. 2002. Variation in pink shrimp populations off the west coast of Vancouver Island: Oceanographic and trophic interactions. University of British Columbia, Vancouver. Nichol, L. M., M. D. Boogaards, and R. Abernethy. 2009. Recent trends in the abundance and distribution of sea otters (Enhydra lutris) in British Columbia. Fisheries and Oceans Canada. Plagányi, É. E. 2007. Models for an ecosystem approach to fisheries. FAO Fisheries Technical Paper No. 477 Rome. Preikshot, D. 2005. Data sources and derivation of parameters for generalised Northeast Pacific Ocean Ecopath with Ecosim models Pages 179‐216 in S. Guenette and V. Christensen, editors. Food web models and data for studying fisheries and environmental impacts on Eastern Pacific ecosystems. UBC Fisheries Center, Vancouver. Simenstad, C. A., J. A. Estes, and K. W. Kenyon. 1978. Aleuts, sea otters, and alternate stable communities. Science 200:403‐411. Uuathluck. 2007. Nuu‐chah‐nulth's Historical Relationships with Sea Otters. http://www.uuathluk.ca/Sea%20Otter%20onepager2%20final[1].pdf. Uuathluck. 2009. Sea Otters in Nuu‐chah‐nulth Ha‐houlthee. http://www.uuathluk.ca/Uuathluk_Sea_Otters_April_09%20.pdf. Walters, C., V. Christensen, W. Walters, and K. Rose. 2010. Representation of multistanza life histories in Ecospace models for spatial organization of ecosystem trophic interaction patterns. Bulletin of Marine Science 86:439‐459. Watson, J. C. 1993. The effects of sea otter (Enhydra lutris) foraging on shallow rocky communities off northwestern Vancouver Island, British Columbia. University of California, Santa Cruz. Watson, J. C. 2000. The effects of sea otters (Enhydra lutris) on abalone (Haliotis spp.) populations. Pages 123‐132 NRC Canada. Canadian Special Publications of Fisheries and Aquatic Science.

Chapter 4 Conclusions

4.1 Discussion

Given the broad acceptance of EBM as a paradigm for the management of marine ecosystems and the need for its implementation, this study contributes to the operationalization of EBM by addressing two key issues: (1) the decision‐making process and (2) the improvement of the representation of marine ecosystems using ecosystem models.

Successful experiences aimed at implementing EBM have shown that the meaningful involvement of stakeholders is key in the management process (e.g., Arkema et al. 2006, Tallis et al. 2010). This study shows that the involvement of stakeholders requires a systematic process to identify the collective values, and define fundamental objectives and ecological and socio‐economic indicators that are consistent with those values (Chapter 2). Without this systematic process, important values and information can be dismissed (Gregory et al. 2001), objectives can be stated in a complex way that makes their operationalization very difficult, and selected indicators may not reflect objectives and values nor be useful for the decision‐making process (Failing and Gregory 2003).

In this thesis, I presented a successful example of how to apply SDM to a multiple stakeholder planning process for EBM, in order to make values, objectives and indicators consistent. Once this is done, managers can use this information as a framework to create and select the alternative that best satisfies the collective objectives and values (McDaniels 2000). I argue that, in order to best represent both (1) what ‘matters’ and (2) how the ecosystem works, quality communication between scientists and stakeholders is essential. This interaction between scientists and the public can inform the process of defining indicators and evaluating alternatives as a method for the needed integration of value judgments and technical information (Chapter 2).

Improving how ecosystem models represent ecosystem dynamics is important for the operationaliztion of EBM, especially because these models will more often be used for representing marine ecosystems and as decision‐support tools. These models can play an 70

important role on predicting the performance of each alternative in terms of the selected indicators during the SDM process.

In this study, I show how to incorporate mediating effects and species reintroductions to ecosystem models in Ecopath with Ecosim (EwE). Both these factors (mediating effects and species reintroductions) have been shown to significantly alter ecosystem dynamics (Grosholz et al. 2000, Dill et al. 2003). For instance, mediating effects can amplify and counteract trophic (prey‐predator) relationships; therefore, when they are not included, ecosystem models can fail in the representation of ecosystem dynamics (Dill et al. 2003). As an example, tuna and seabirds share school fish prey; therefore, one would expect that the sea bird population would benefit from tuna decreases. However, tuna predation causes schooling pelagic fish to move closer to the surface, where these fish become more available to sea birds. Therefore, decreases in tuna negatively affect sea bird populations (Dill 2003).

This study demonstrates how to use the ‘mediation’ function in EwE in order to incorporate mediating effects. For my case study—the nearshore ecosystems of the WCVI—mediating effects provided by kelp forests include the provision of habitat and feeding areas, which are relevant services the ecosystem provide and which should therefore be considered for EBM. The incorporation of mediating effects provided by species (like kelp forests) that provide habitat in ecosystem models can give a better understanding to managers of the role and the potential impacts of alterations of these species.

The assumptions of Ecospace in EwE do not allow the explicit representation of species reintroduction, as the software assumes that species are distributed in suitable habitats in the first year of the given time period. Therefore, I created one scenario to work around this assumption on the initial state of the ecosystem.

I also demonstrated how both types of movement models—Diffusion and Individual‐based models (IBM)—in Ecospace represent the expansion of the introduced species, which I believe has not been explored before. The diffusion model rendered a uniform distribution of the introduced species, while IBM produced a random and patchy distribution. Adding habitat heterogeneity to the model (to better reflect reality) would show a patchier distribution of

species, and perhaps increase similarities in the results obtained from each type of model (Chapter 3).

4.2 Strengths, limitations and improvements

This thesis contributes to the operationalization of EBM through the better integration of science and stakeholder values in the decision‐making process and through the representation of ecosystems in EwE models. My case study of the marine planning process and models of the nearshore ecosystems on the WCVI is sufficiently connected to the EBM initiatives in the region to be of tangible benefit. The numerous general findings and innovations also offer insight and guidance for marine planning processes elsewhere.

The suggested SDM approach to define the decision‐making frameworks is well‐known, and has been successfully applied to planning and multiple stakeholder processes (e.g., McDaniels et al. 1999, Gregory et al. 2001, McDaniels et al. 2006). Its application to EBM, however, is novel and can be very useful.

In this thesis, I have shown how to apply the SDM for a particular marine planning process, demonstrating how stakeholder values and science can be integrated to evaluate and decide amongst management alternatives. This is an essential part of what EBM means, yet the actual translation from organizations stated values into scientifically appropriate measures is currently missing from much of the literature and practice. Scientists generally develop EBM objectives without making the link with stakeholder values, and performance measures often include long lists of indicators not clearly connected to management objectives. I demonstrated how to deal with complex and non‐informative objectives (when attributes are not clear), and how to choose a manageable number of appropriate indicators among many to enable decision‐ making and on‐going management that reflect stakeholder values.

This process was revised and complemented by the Executive Director of WCA and represents a first step for WCA’s systematic decision‐making framework. Limitations of this work are that, (1) the process was not conducted in consultation with the full WCA board (which, ideally, would be required before implementation); and (2) that any framework needs to be applied to see what works, what does not work, and what is most appropriate for the constituents

depending on their values, objectives and priorities. By no means have I intent to provide a full elucidation of the exact management framework to be used on the WCVI, but rather I have provided insights on how to build it.

The ecosystem model I have developed shows ways to address and account for two particularly important drivers of dynamic ecosystem change—mediating effects and species introduction— that I believe are applicable beyond this particular case study, as these factors often present themselves in diverse ecosystems. I have shown how to include these factors in EwE, which is a widely used tool, with an expanding applicability for EBM.

The ecosystem model provided demonstrates how to represent mediating effects, including the representation of habitat provision, and refuge and feeding areas, which are crucially relevant considerations for ecosystem‐based approaches to management (e.g., Francis et al. 2007). I also explored the representation of species recovery after introduction and compared the results that one can obtain with Ecosim (the temporal module of the software) and the two movement models in Ecospace (the spatial component of the software).

Given the numerous assumptions and significant uncertainties involved in ecosystem models, it is important to use them carefully, they are designed for broad policy screening to recognize those alternatives that require further attention (Walters et al. 2010). Ecosystem models are not meant to substitute current single species models, but to complement them.

The model I have built is theoretical. Therefore, the results should not be used directly for the current planning process on the WCVI. They have to be revised and fitted to empirical data. Currently, the model does not include all the mediating effects that must occur in the ecosystem, habitats and fisheries information, which are essential to better representing how the ecosystems work. In addition, there are numerous important concerns of WCA (e.g. salmon’s life cicle) as well as diverse human activities that are not incorporated. Nevertheless, my contribution constitutes the first model for the region and a crucial foundation for future efforts. It has provided a structure for the integration of empirical information, helped to identify data requirements and evaluated alternative approaches for dealing with two important ecological processes not generally addressed by exiting models.

4.3 References

Arkema, K. K., S. C. Abramson, and B. M. Dewsbury. 2006. Marine ecosystem‐based management: from characterization to implementation. Frontiers in Ecology and the Environment 4:525‐532. Dill, L. M., M. R. Heithaus, and C. J. Walters. 2003. Behaviorally mediated indirect interactions in marine communities and their conservation implications. Ecology 84:1151‐1157. Failing, L. and R. Gregory. 2003. Ten common mistakes in designing biodiversity indicators for forest policy. Journal of Environmental Management 68:121‐132. Francis, R. C., M. A. Hixon, M. E. Clarke, S. A. Murawski, and S. Ralston. 2007. Ten Commandments for Ecosystem‐Based Fisheries Scientists. Fisheries 32:217‐233. Gregory, R., T. McDaniels, and D. Fields. 2001. Decision Aiding, Not Dispute Resolution: Creating Insights through Structured Environmental Decisions. Journal of Policy Analysis and Management 20:415‐432. Grosholz, E. D., G. M. Ruiz, C. A. Dean, K. A. Shirley, J. L. Maron, and P. G. Connors. 2000. The Impacts of a Nonindigenous Marine Predator in a California Bay. Ecology 81:1206‐1224. McDaniels, T., H. Longstaff, and H. Dowlatabadi. 2006. A value‐based framework for risk management decisions involving multiple scales: a salmon aquaculture example. Environmental Science & Policy 9:423‐438. McDaniels, T. L. 2000. Creating and using objectives for ecological risk assessment and management. Environmental Science & Policy 3:299‐304. McDaniels, T. L., R. S. Gregory, and D. Fields. 1999. Democratizing risk management: Successful public involvement in local water management decisions. Risk Analysis 19:497‐510. Tallis, H., P. S. Levin, M. Ruckelshaus, S. E. Lester, K. L. McLeod, D. L. Fluharty, and B. S. Halpern. 2010. The many faces of ecosystem‐based management: Making the process work today in real places. Marine Policy 34:340‐348. Walters, C., V. Christensen, W. Walters, and K. Rose. 2010. Representation of multistanza life histories in Ecospace models for spatial organization of ecosystem trophic interaction patterns. Bulletin of Marine Science 86:439‐459.

Appendix A. Parameters estimation and model balancing

1. Parameter estimation a) Sea otters Sea otters are assumed to be at their carrying capacity in some areas on the west coast of Vancouver Island (WCVI). The biomass was arbitrarily set to 0.127 ton/km2, based on the carrying capacity divided by the 300km2 in the area. Sea otter mortality (0.13 year‐1), Q/B (101.5 year‐1), and diet was adapted from Ainsworth et al. (2002).

The diet was described as 50% epifaunal inverts, 20% small crabs, 1% large crabs, and 29% shallow benthic fish, juvenile pollock, and squid. While otters have not yet been observed eating finfish on the WCVI, a small level of predation on forage fish (.01) and surfperch (.01) was introduced to allow diet switching and reduce pressure on other preys. b) Lingcod Lingcod’s parameters were adapted from Preikshot (2005) and Martell (2002). Biomass of juveniles (0.031 ton/km2) and adults (0.034 ton/km2) were summed. P/B (1 year‐1) was taken from the range 0.8 year‐1 (P/B for adults) to 1.5 year‐1 (P/B for juveniles). Diet composition was adapted from Preikshot (2005), Ainsworth et al. (2002) models and from Beaudreau (2009) Markel (pers. comm.). c) Rockfish Rockfish biomass (1.5 ton/km2) was estimated as the sum of Pacific Ocean Perch and Other Rockfish groups in (2005). The ratios P/B (0.18 year‐1) and Q/B (3.36 year‐1) were based on the average of Ainsworth et al. (2002) inshore, piscivorous, and planktivorous rockfish groups. The diet composition was taken from Markel (unpublished data) on diet composition for black and copper fish in the area. d) Forage fish This group includes herring, hake, sardines. We adapted the values from inputs parameters were based on herring values used by Ainsworth et al. (2002), using a P/B of 0.6 year‐1 (herring natural mortality) and a Q/B of 8.4 year‐1 calculated by the average of adults and juveniles. Following Ainsworth et al. (2002), an EE of 0.95 was used to estimate the biomass for this

group. The diet composition based on zooplankton and phytoplankton was adapted from Ainsworth et al. (2002), and Preikshot (2005). We included a small proportion (.10) of kelp‐ derived detritus in their diet because forage fish are filter feeders with opportunistic foraging behavior (Hand and Berner 1881), therefore, they feed on what is available in the water column, such as the decaying kelp. e) Surfperches Mortality and consumption parameters were obtained for the four species that occur in BC (Froese and Pauly Editors. 2010) FishBase and averaged them for the functional group (Z = 0.70 year‐1; Q/B = 10.825 year‐1). Biomass (1 ton/km2) was arbitrarily estimated as 2/3 of rockfish biomass. It was assumed they eat both plankton and benthic invertebrates. f) Crabs Biomass (3.8 ton/km2) was taken from Preikshot (2005) To estimate the initial P/B (2.5 year‐1), large and small crabs P/Bs taken from Ainsworth et al. (2002) were averaged. Q/B ratio was estimated using a P/Q ratio (0.275). A broad diet was assign to this species based on the same sources. g) Urchins Initial urchin parameters (B=20 ton/km2; P/B=0.3 year‐1; P/Q=0.25 year‐1) including diet composition were estimated based on Preikshot (2005) echinoderm’s and benthos’ group. h) Filter feeders bivalves Abundance (7.7 ton/km2) and ratios (P/B=0.9 year‐1, P/Q=0.2 year‐1) were based on Preikshot (2005) bivalves group. The diet was adapted from Preikshot (2005) and Ainsworth et al. (2002). i) Benthic invertebrates Initial estimates of this group were based on Preikshot (2005) benthos group. After assigning 20 ton/km2 units to urchins, the remaining (from a total benthos of 57.8 ton/km2) 37.8 ton/km2 were used for the biomass of this group. The ratios P/B (1.9 year‐1) and Q/B (12.7 year‐1) were based on the average of the benthos groups defined by Heymans (2005a, 2005b). Diet composition was adapted from the same sources.

j) Zooplankton The biomass was estimated (68 ton/km2) as the sum of Preikshot (2005) herbivorous, carnivorous, and other zooplankton groups. P/B (16.5 year‐1) and Q/B (62.5 year‐1) ratios were based on the average of Preikshot (2005) values for euphausiids and copepods. The diet was assumed to be composed by zooplankton, phytoplankton and detritus. k) Phytoplankton Initial phytoplankton biomass (22 ton/km2), P/B (179 year‐1) and diet composition were based on Preikshot (2005). l) Canopy kelp Canopy kelp biomass (9 ton/km2) and P/B (5.3 year‐1) were estimated from Preikshot (2005) macrophytes group that included bull kelp, giant kelp, seaweeds and sea grasses. m) Understory kelp Understory kelp biomass was assumed to be 18 ton/km2, double the value of canopy kelp. P/B was assumed the same for canopy kelp. n) Kelpderived detritus Detritus kelp biomass was estimated using the ratio of kelp to all biomass in the ecosystem. This proportion (0.11) was moved to this functional group from the detritus group. o) Detritus We initially used Preikshot (2005) value, but reduced it marginally after adjusting for the kelp detritus pool.

2. Balancing the model The model did not balance when the initial parameters described above were entered. Basic parameters were adjusted starting with those groups that were the most out of balance.

Balance problems were related with the three otter diet groups (crabs, benthic invertebrates and filter feeders). Sea otter mortality was increased from 0.13 year‐1 to 0.15 year‐1 and

consumption (Q/B) decreased from 101.5 year‐1 to 75 year‐1. This improved reduced the predation pressure on those three groups, but they still did not balanced.

Biomass and P/B values were changed first for crab, we increased the biomass from 3.8 ton/km2 to 8.0 ton/km2, and P/B from 2.5 year‐1 to 3.0 year‐1. This finally created sufficient crab biomass for sea otter diet.

To balance filter feeders, their proportion in the crab diet was reduced to 0.2, the biomass was increased from 7.7 ton/km2 to 20 ton/km2, and P/B was increased from 0.9 year‐1 to 1.2 year‐1. Benthic invertebrates biomass was changed from 37.8 ton/km2 to 43 ton/km2. Similar changes were necessary to balance urchins (biomass changed from 20 ton/km2 to 30 ton/km2; and P/B from 0.3 year‐1 to 0.4 year‐1).

Finally, to explore the impacts of kelp‐derived detritus on the system, available kelp‐derived detritus was also added to benthic invertebrates, filter feeders, sea urchins and zooplankton diets.