AN INTEGRATED EVALUATION OF STOCK ENHANCEMENT OF RECREATIONAL FISHERIES, APPLIED TO RED DRUM IN FLORIDA By EDWARD VINCENT CAMP A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFIL LMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
Â© 2014 Edward Vincent Camp
To my wife, Genevieve a nd my parents, Peter and Marcia
4 ACKNOWLEDGMENTS I thank t he Florida Fish and Wildlife Commission, the University of Florida Alumni Fellowship, and the National Science Foundation Integrative Graduate Education and Research Traineeship for funding and support that made this work possible. I also thank each of my committee members for the patience, hard work, and guidance they have provided me . I am particularly grateful to Kai Lorenzen for the trust he showed in accepting me as a student . I thank my fellow students for their help, and particularly grateful to Dani el Gwinn for his assistance and friendship. Finally, I thank my wife , Genevieve, for her support throughout my work and life .
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION AND JUSTIFICATION FOR STUDY ................................ ........... 13 2 POTENTIALS AND LIMITATIONS OF STOCK ENHANCEMENT IN MARINE RECREATIONAL FISHERIES SYSTEMS: AN INTEGRATIVE REVIEW OF ENT ................................ ............................ 24 Case Study: Red Drum Enhancement in Florida ................................ ...................... 26 Attributes of the Enhancement Fishery System and Their Role in Determining Key Linkages, Uncertainties, and Recommendations ................................ ............. 44 Discussion ................................ ................................ ................................ .............. 50 3 AN EMPIRICAL ASSESSMENT OF FL EFFORT DYNAMICS: DOES FISH ABUNDANCE DRIVE EFFORT? .................... 60 Methods ................................ ................................ ................................ .................. 63 Results ................................ ................................ ................................ .................... 67 Discussion ................................ ................................ ................................ .............. 70 4 STOCK ENHANCEMENT TO ADDRESS MULTIPLE RECREATIONAL FISHERIES OBJECTIVES: AN INTEGRATED MODEL APPLIED TO RED DRUM Sciaenops Occelatus IN FLORIDA ................................ ............................. 81 Methods ................................ ................................ ................................ .................. 84 Results ................................ ................................ ................................ .................... 93 Discussion ................................ ................................ ................................ .............. 95 5 SOCIOECONOMIC AND CONSERVATION TRADE OFFS IN THE STOCK ENHANCEMENT OF RECREATIONAL FISHERIES ................................ ............ 116 Methods ................................ ................................ ................................ ................ 120 Results ................................ ................................ ................................ .................. 127 Discussion ................................ ................................ ................................ ............ 130
6 6 CONCLUSIONS ................................ ................................ ................................ ... 151 Summar y of Findings ................................ ................................ ............................ 153 Key Contributions ................................ ................................ ................................ . 157 Future Directions ................................ ................................ ................................ .. 158 LIST OF REFERENCES ................................ ................................ ............................. 161 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 176
7 LIST OF TABLES Table page 2 1 Role of attri butes in determining current outcomes. ................................ ........... 53 2 2 Likely influence of attributes on future outcomes. ................................ ............... 54 3 1 Results of ARIMA models ................................ ................................ ................... 75 3 2 Total standard deviations, proportion of variance, and cumulative proportion of variance associated with principle component analyses. ............................... 76 4 1 Parameters and associated values used in integrated quantitative model are described. An asterisk (*) indicates values were estimated. ............................. 103 4 2 Model components and equatio ns are described ................................ ............. 105 4 3 Comparisons between tuned model predictions and region specific observations. ................................ ................................ ................................ .... 109 4 4 Evaluation of cos ts of potential stocking of red drum in Tampa Bay, Florida .... 110 5 1 Description of parameters and parameter values (* indicates estimated). ........ 139 5 2 Description of model components and equations. ................................ ............ 141 5 3 Summary and comparison of model runs. ................................ ........................ 144
8 LIST OF FIGURE S Figure page 1 1 Diagram showing my concept of how natural resource issues are organized in terms of levels affecting outcomes.. ................................ ................................ 23 2 1 Framework for analyzing enhancement fisheries systems used in this study. .... 55 2 2 Red Drum a b undance estimated from the most recent stock assessment, (Murphy and Mu n yandorero, 2009).. ................................ ................................ ... 56 2 3 Estimated Red Drum ta r geted e f fort and catch rates from the most recent stock assessment (Murphy and Mu n yandorero, 2009). ................................ .... 57 2 4 Illustratio n of the ke y relationships essential for producing desired outcomes. .... 58 2 5 Uncertainty associated with attri b utes of Red Drum enhancement and di f ficulty in reducing this uncertaint y . ................................ ................................ .. 59 3 1 Effort predicted from a first order one year lagged mixed effect model. .......... 77 3 2 Estimates of the relationship between effort and a bundance. .......................... 78 3 3 Evaluation of the relationship between changes in estimated fish abundance and estimated effort. ................................ ................................ ........................... 79 3 4 Estima tes of the relationship between effort and principle component regressors. ................................ ................................ ................................ ........ 80 4 1 The structure of the model accounts specifically for phenotypes, life stages and related population processes . ................................ ................................ .... 111 4 2 Alternative assumptions of the response of aggregate fishing effort to the abundance of all fish vulnerable to capture for recreational fishing. ................. 112 4 3 The expected, equilibrium model results.. ................................ ........................ 113 4 4 Evaluation of the effects of effort response on equilibrium model results. ...... 1 14 4 5 Evaluation of the effects of effort response and catch related satisfaction on equilibrium model predicted socioeconomic value. ................................ ......... 115 5 1 Alternative assum ptions of how dynamic angler effort could respond to catchable abundance of fish. ................................ ................................ ............ 145 5 2 The assumed inherent satisfaction from stocking as a function of the numbers of fished and stocked. ................................ ................................ ........ 146
9 5 3 Overall trade offs between number of fish stocked and size at stocking by metric. ................................ ................................ ................................ ............... 147 5 4 Trade offs between conservation and socioeconomic value. ......................... 148 5 5 Tradeoffs between the conservation and socioeconomic value associated with different levels of satisfaction from stocking. ................................ ............. 149 5 6 Comparisons of trade offs under different management strategies. ............... 150
10 LIST OF ABBREVIATIONS AAM Active Adaptive Management ACE Accumulated Cyclone Energy AICc iteria, corrected ANOVA Analysis of Variance AR1 AutoRegressive First Order ARIMA AutoRegressive Integrated Moving Average ASA American Sportfishing Association CSI Consumer Sentiment Index FL Florida FL GDP Florida Gross Domestic Product FMFEI Flo rida Marine Fisheries Enhancement Initiative FWC Florida Fish and Wildlife Conservation Commission FWRI Florida Fish and Wildlife Research Institute IGERT Integrated Graduate Education and Research Traineeship NMFS National Marine Fisheries Service OL S Ordinary Least Squares PCA Principle Component Analysis SDM Structured Decision Making SES Socio Ecological Systems SPR Spawning Potential Ratio WSB Wild Stock Biomass
11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AN INTEGRATED EVALUATION OF STOCK ENHANCEMENT OF RECREATIONAL FISHERIES, APPLIED TO RED DRUM IN FLORIDA By Edward Vincent Camp August 2014 Chai r: Kai Lorenzen Cochair: Sherry Larkin Major: Fisheries and Aquatic Sciences The primary task of fisheries scientists is to make clear the general and specific potential outcomes of alternative management strategies, in terms for social, ecological, gover nance and economic metrics. Using a quantitative approach integrating these components, I have evaluated the potential outcomes of an increasingly common Red Drum fishery a s a case study. I first synthesized available literature describing integrated systems approaches, recreational fisheries, enhancement, and Red Drum to reveal which components and linkages were most important to understand. One of the more potentially im portant but less studied linkages was the dynamics of angler effort, which I then evaluated empirically for key recreational fisheries in Florida. Finding no clearly predictive relationships, I constructed a quantitative integrated model of Red Drum stock enhancement, tuned to empirical data from Tampa Bay Florida. This work showed that even accounting for uncertainty in angler effort dynamics, stock enhancement was almost certain to elicit negative responses for wild Red Drum populations, though socioeco nomic outcomes might be augmented if fish were stocked
12 at larger sizes. I further evaluated this trade off between conservation of wild fish and socioeconomic gain to discover that only changes in stakeholder beliefs regarding stock enhancement could brin g the trade off efficiency of this strategy in line with alternatives, such as habitat restoration or fishing facility improvement.
13 CHAPTER 1 INTRODUCTION AND JUSTIFICATION FOR STUDY The scientific studying of nature holds intrinsic value in its discov ery of patterns and mechanisms. This is one way that natural resource scientists might justify their research. An alternative view is that research serves a utilitarian purpose of forwarding certain outcomes (e.g., sustained fisheries, improved human wel fare, or the existence of coral reefs; Evans et al . 2012). In this view, science is usually justified as providing information essential for management, which is charged wit h actually achieving outcomes of natural resources. The management of natural reso urces is tied to inherent trade offs with their use . The need to both use natural resources now and simultaneously to preserve them for later sets up a trade off, since what is used now often cannot be used in the future. This trade off can also be seen across space (what is used by one region may not be used by another), or resource uses /objectives (e.g., ecological vs. socioeconomic) that conflict (Hilborn 2007). While it is possible to imagine individuals recognizing trade offs and limiting their use to achieve better outcomes for society, this has occurred rarely (though it has in some smaller communities) (Johannes 1981; Pinkerton 2009). Instead the tragedy of the commons has often played out, where groups or individuals follow incentives to use co mmonly available resources before others do, and so impair future opportunities for all. It is in response to this tragedy that natural resource management exists to mediate the trade offs between often conflicting objectives and ensure that resources are available for current and future desired outcomes (Walters and Ahrens
14 management, might not be popular with people whose incentive driven use of resources is constrained. Managemen t can be understood as making decisions about the points at which to realize trade offs among often conflicting objectives, (Walters and Martell 2004; Walters and Ahrens 2009) , but defining good management can be difficult . of trade off decisions depends on the appropriateness of the desired outcomes and the extent to which the desired outcomes occur. However, the appropriateness of outcomes is rarely unanimously agreed upon among the people affected by the management decisi ons (stakeholders). This creates complexity in what the desired trade off points (i.e. desired outcomes) are some stakeholders would desire one, other stakeholders would desire another. A ssessing what combinations of outcomes are achievable is also diff icult because of uncertainty at multiple scales or levels. Seeing a problem in terms of such levels is not a new or uncommon concept, though the distinct levels vary with the viewer (Holling et al. 2002; Holling 2004). I view the world of natural resourc es as having four levels ecological, socioecological, management, and public, all of which contribute uncertainty to outcomes ( Figure 1 1 ). First, there is substantial uncertainty at the level of the natural resource. Often this takes the form of ecologi cal or biological uncertainty in processes and mechanisms that govern population dynamics, species interactions, community structure, etc. At a broader level, there is added uncertainty regarding the interactions between the natural resource and the users of the natural resource (generally humans) sometim es referred to as Socio Ecologi cal Systems (SES) (Brock and Carpenter 2004; 2007). This may take the form of uncertainty in the quantity and
15 type of natural resources being removed, harvested, affected et c., or complex feedback loops and functional relationships between anthropogenic impact and the natural resource (e.g., dynamics between fisher effort or harvest and fish population size or dynamics). I also include relationships between the natural resou rc e and economies in this level. A still broader level is the management of natural resources, and uncertainty here is increasingly recognized. I see this uncertainty in two parts (1) uncertainty in the management paradigm followed (e.g., dogmatic, activ e adaptive management, participatory, etc.) and (2) uncertainty in how the selected paradigm functions. Specifics of some dominant management paradigms are discussed further below, for now it is suffice to acknowledge that different paradigms exist. Unce rtainty in function are often common across paradigms and generally take the form of political and social influences that can affect and bias the process of decisions making (Holling 1998; Ludwig 2001). For example, Walters and Ahrens (2009) describe mult iple pathologies by which managers face moral hazards that influence decisions or lack of decisions. The broadest level of uncertainty is public, which is usually considered spatially to be at region, country, continent or global scales and generally impl ies how and what people impacts on the realization of desired outcomes, and shifts in public opinion to public judgment (Constanza 2001) are largely necessary for whole sy stem change (Osteen). Related, some theories exist focusing on how whole systems, driven by the public, change. Constanza (2001) contended that the public must arrive at some shared judgment of what the future will be, Holling (2004) suggested that episo dic change occurs at certain stages within loops, as understood panarchy framework (Gunderson
16 and Holling 2002) emphasizing metastable regimes, episodic change, resilience and multiple scales and cross scale interactions (Gotts 2007). Others (Wallerstein 1993; Chase approach that emphasis nation states and political power in explaining public change (Gotts 2007). Suggestions of public elements important for achieving desired outcomes o mechanisms may be non linear with respect to any forcing functions (e.g., resource use, education, individual change) (Holling 2004; Evans et al. 2012). So a whole systems view reveal s uncertainty at all levels, though most natural resource researchers have focused on the ecological and socioecological levels. Processes acting at and across these levels produce cumulative uncertainty in outcomes that may be daunting. The management level has received substantial attention, generally focused on (Brock and Carpenter 2007). Each management approach has positives and negatives. Attempts to triumph one manag ement paradigm or framework over the others as a Carpenter 2007). It may be more useful to simply acknowledge that uncertainty exists in which management paradigm to employ, and then to incorporate this uncertainty into assessments of possible outcomes (Holling et al. 1998). It is worth noting that such an embrace of plurality of management paradigms and frameworks can be incorporated into SDM or AAM approaches. Desp ite myriad research and these noted management paradigms/frameworks, many natural resource failures are perceived to occur (Ludwig 2001). Some fisheries
17 are overexploited, coral reefs are declining, freshwater supplies dwindle and globally forests are bei ng reduced quickly (Morgan and Dowlatabadi 1996). While some of failures are considered so because a trade off point exists but is not realized where both outcomes of both conflicting objectives are augmented (Walters and Martell 2004; Ahrens and Walters 2011). Such failures may be due to shortcomings at specific levels such inappropriate ecological or socioecological models, not correctly identifying management syste m objectives (Hilborn 2007; Nicholson and Posingham 2006) or having strong leaders to guide management system processes (Walters and Martell 2004; Hilborn 2007), or using ineffective management paradigms or frameworks altogether (Morgan and Dowlatabadi 199 6). Or, failures may span across the levels, stemming either from incorrectly integrating predictions across levels, or from not considering certain levels (such as public change) at all (Holling 1998). Failures may be unsettling to utilitarian minded na tural resources scientists and have likely eroded public trust in both science and management (which dynamically changes what outcomes may be possible for systems) (Ludwig 2001), but recognizing causes of failure improves understanding of what is possible (Haerling and Parr 1999). Though research has sometimes failed to usher in desired outcomes, researchers can embrace the failure by striving to more honestly describe potential outcomes. Walters and Martell is to describe what is possible, and perhaps by what means that possibility can be reached. This is directly applicable to specific trade offs, which often occur at socioecological levels (e.g. fishing economies now vs conservation of fish stocks for futu re), but it can
18 be applied more broadly to whole system outcomes to imagine what is possible and how that could be achieved. To do this, it will be necessary to provide quantitative predictions (i.e. models) at both multiple levels and across levels to re present whole systems (Evans et al. 2012). Quantitative models are common ecological levels and increasing at socioecological levels, though they must be reshaped for specific natural resource problems. Less research has compared how uncertainty at the m anagement level affects outcomes. It is important to understand what outcomes would be possible under different management specific strategies (e.g., regulation), frameworks (decision making plans) and broader paradigms, as well as how manager moral hazar ds can influence this process. While there is research into how change occurs in public opinion and judgment, it has rarely been adopted into the studies of natural resources. Representing public dynamics or even the range of uncertainty thereof may grea tly augment my ability to say what is possible. Predictive models at each level are valuable, but this value best realized when these components are brought together to predict whole system outcomes with integr ative models that span levels. The challenge and applicability (Starfield 1997; Adkinson 2009). Models must represent important heterogeneity, but also predict whole system outcomes that are meaningful withou t being overwhelmed by uncertainty (Boyd 2012; Evans et al. 2012). This view of natural resource science has several important implications for how research is conducted. The capacity to justify research in terms of fundamental objectives is probably cr itical for producing research that is actually useful, and inability to describe why research matters ultimately for whole system outcomes threatens to
19 further discredit the science I do. Relating research to ultimate outcomes requires predictive and ofte n quantitative models of how whole system outcomes occur (Evans et al. 2012). However, it is not practical for all research and models to encompass the whole system. One solution is to understand the whole system in terms of levels, and to place level sp ecific research in the context of the whole system. In practice, this translates to increased study of certain levels that are sometimes ignored (management, public), acknowledgment of how the dynamics of these levels likely mediate implications of ecolog ical and socioecological research, and finally some focus at integrating studies across levels. Dissertation Case Study My dissertation aims to augment understanding of how whole system outcomes are related to research and management in order to improve those outcomes. To accomplish this I will use case studies, for which I will describe what outcomes are possible, and the paths by which those possibilities are most likely to be achieved. My primary case study is the management of the recreational Red D rum fishery in Florida, particularly as it pertains to potential stock enhancement. This case study is ideal due to the magnitude and complexity (i.e. multiple and conflicting objectives) of recreational fisheries, as well as stock enhancement. The dyn amic complexity of recreational fisheries present challenges in achieving desired outcomes worldwide. Recreational fisheries provide social welfare (e.g. supplying subsistence or satisfaction benefits) and support economies (largely as a result of expendi tures anglers incur while gaining satisfaction from fishing) (Weithman 1999; Cowx et al. 2010). This socioeconomic benefit sometimes comes at an
20 ecological cost, as recreational fisheries can also produce unsustainable fishing levels or even lead to fish population collapse (Post et al. 2002; Figueria and Coleman 2010) through direct (e.g., overharvest) and indirect pathways (e.g., habitat alteration from gear; Cooke and Cowx 2006; Lewin et al. 2006). The complexity of recreational fisheries is largely du e to heterogeneity in how recreational fishers attain utility (e.g., harvesting fish, high catch rates, catching trophy fish, enjoyment of natural surroundings, etc.; Arlinghaus 2006; Johnson et al. 2010). Heterogeneity in utility attainment leads to mult and complicates predicting recreational fishing effort, its impacts, and ultimately stakeholder and public opinion. Such complexity is compounded by spatial (Post et al. 2008) and multiple sp ecies components common to commercial fisheries. Despite this complexity, recreational fisheries can be characterized by two primary objectives from a management perspective (1) maximize socioeconomic benefit from fishing and (2) sustain populations and ecosystems at desired levels or states (Cowx et al. 2010; Koehn 2010). While such objectives are mutually obligate over the long run (Hilborn 2007), they often conflict in shorter time spans (Koehn 2010; Garcia Asorey et al. 2011; van Poorten et al. 2011) . Furthermore, some specific socioeconomic objectives (e.g., high catch rates and high total effort) also conflict in the short term, since effort measurement is usually the denominator of catch rate calculations. These conflicts have led to alternative management strategies designed to satisfy these objectives simultaneously (van Poorten et al. 2011). Often, stock enhancement is seen as a way of achieving this (Halverson 2008; van Poorten et al. 2011).
21 Enhancement of recreational fisheries, defined as releasing hatchery raised fish to augment existing wild populations (Lorenzen 2005), often impacts ecological and socioecological systems, producing complex, feedback driven processes and occasionally unintended outcomes (Lorenzen 2008). The coupled socio economic ecological nature of enhancement systems requires assessment within an integrated framework. Some integrated frameworks have been developed to relate these attributes of enhanced fisheries to overall system outcomes (Blankenship and Leber 1995; M olony et al. 2003; Taylor et al. 2005; Lorenzen 2008 ; Lorenzen et al. 2010). These integrated assessments must often be conducted at broader scales that extend beyond the target stock and fishery, since many recreational systems are composed of multiple a ngler types (e.g., harvest, trophy, catch rate or experience oriented) targeting multiple fish species (Sutton and Ditton 200 5 ; Johnson et al 2010). Thus the socioeconomic and ecological outcomes of enhancing one fishery relate to outcomes of co existing fisheries targeting alternative species and potentially competing leisure economies in the market sector (e.g., hunting, boating, etc.). For example, enhancement related changes in fishing effort may be a re distribution from other fisheries or leisure ac more or join the recreational fishing pool) with the options differing widely in their impact to the broader system outcomes (Loomis and Fix 1998). Both broader and more specific outcom es must be considered in full, integrated assessments of enhancement. Despite the widespread use of enhancement and repeated calls for case specific evaluation, there are few examples of integrated assessments of enhancement programs (Taylor et al. 2005) , especially for recreational fisheries.
22 I propose a number of studies, some of which will be targeted at more specific levels (e.g., ecological investigation of the effect of stocking on density dependent survival of wild Red Drum or socioecological res earch of angler effort dynamics) while I will attempt to construct others at more broad scales, such as assessing the set of decisions, from regulations to management strategies, to management paradigms and philosophies of public change that are likely to maximize trade off points for Red Drum in Florida. Each proposed study will be designed to function alone as a publishable manuscript and to contribute to a broader understanding of how research and management relates to outcomes.
23 Figure 1 1 . D iagram showing my concept of how natural resource issues are organized in terms of levels affect ing outcomes. The size of the box represents the scale of specificity in implications, and thus, in approaches to understanding the uncertainty around the processes inherent to each level. Each level can affect outcomes directly itself (thin arrows) but also dynamically the other levels, the product of which (thick arrows) affects outcomes. The Figure suggests that underrepresenting certain levels, such as management or pubic, could decrease likelihood of accurately describing possible outcomes and the methods by which they may be achieved.
24 CHAPTER 2 POTENTIALS AND LIMITATIONS OF STOCK ENHANCEMENT IN MARINE RECREATIONAL FISHERIES SYSTEMS: AN INTEGRATIVE R RED DRUM ENHANCEMENT Besides control of fishing mortality and habitat protection or restoration, aquaculture based enhancement is a third principal means by which fisheries can be sustained and impr ov ed (Lorenzen et al. 2010). Aquacultur e based fisheries enhancement is a set of management approaches i n v olving the release of cultured o r g anisms to enhance, conser v e, or restore fisheries. This definition c ov ers a great d i v ersity of enhancement fisheries systems, including ranching , ock enhancement , (Bell et al. 2008). Aquaculture based enhancements can, at least in principle, generate a range of benefits, including increasing stock a b undance and fishery yield or catch opportunities, as well as aiding the conser v atio n and restoration of depleted, threatened, and endangered populations (Lorenzen et al. 2012). This may g i v e rise to economic and social benefits, including n e w opportunities for fisheries related l i v elihoods or recreation, and may also pr o vide incent i v es f or act i v e management and better g ov ernance of common pool fisheries resources (Pin k erton 1994; Ar b uckle 2000; Lorenzen 2008). H o w e v e r , ma n y enhancements h a v e f ailed to del i v er significant increases in yield or economic benefits and/or h a v e had deleterious effects on the naturally recruited components of the ta r get stocks (Hilborn 1998; Arnason 2001). Much work on marine enhancements has focused on commercial fisheries, b ut the approach is also and increasingly used in recreational fisheries. Recreational fi sheries are compositionally and dynamically compl e x and present management challenges w orldwide. Constituting an important and sometimes dominant use of fresh and coastal fisheries resources in ma n y countries (C o wx 2002;
25 Arlinghaus and Mehne r 2006), recrea tional fisheries pr o vide social wel f are (e.g., supply subsistence or satis f action benefits) and support economies (la r gely resulting from e xpenditures anglers incur while fishing ) ( W eithman 1999; C o wx et al. 2010). Recreational fisheries can also produce u nsustainable fishing l e v els or e v en lead to population collapse (Post et al. 2002; Figueira and Coleman 2010) through direct (e.g., ov erhar v est) and indirect path w ays (e.g., habitat alteration from fishing gear) (Coo k e and C o wx 2006; L e win et al. 2006). Th e magnitude of these e f fects is dr i v en la r gely by fishing e f fort, which can be especially high in recreational fisheries. Relationships between recreational fishing e f fort and fish populations are often more v ariable than those in commercial fisheries, par tially due to heterogeneity in h o w recreational fishers attain satis f action or utility (e.g., har v esting fish, high catch rates, catching trop h y fish, enj o yment of natural surroundings, etc.; Hunt et al. 2005; Arlinghaus 2006; Johnston et al. 2010). Hetero anglers (Johnston et al. 2010) and complicates predicting recreational fishing effort, its impacts, and fisher satisfaction. Recreational fisheries assessments and management must account for this and other (e.g., multi species ta r geting, spatial, etc.) compl e xities. Despite this compl e xit y , recreational fisheries management can be characterized by t w o primary object i v es (1) maximize socioeconomic benefit from fishing and (2) susta in populations and ecosystems at desired l e v els or states (C o wx et al. 2010; K oehn 2010). While such object i v es are mutually obli g ate ov er the long run (Hilborn 2007), th e y often conflict in shorter time spans ( K oehn 2010; Garcia Asor e y et al. 2011; v an Po orten et al. 2011). Stock enhancement is often seen and promoted in recreational fisheries as a w ay of miti g ating such conflicts in object i v es by sustaining fish
26 populations e v en under v ery high fishing pressure (Hal v erson 2008; v an Poorten et al. 2011). E nhancement of recreational fisheries, defined as releasing hatchery raised fish to augment e xisting wild populations (Lorenzen 2005; Lorenzen et al. 2012), can impact both socioeconomic and ecological systems, producing complex, feedback driven processes a nd occasionally unintended outcomes (Lorenzen 2008). Thus, predicting outcomes of enhanced recreational systems requires integrated assessment. Integrated frameworks have been developed to first understand (generally commercial) enhancements in terms of at tribute groups (e.g., biological, market, stakeholder, etc.) and then by relating these attributes to overall system outcomes (Blankenship and Leber 1995; Molony et al. 2003; Taylor et al. 2005; Lorenzen 2008; Lorenzen et al. 2010). Despite the widespread use of enhancement and repeated calls for case specific ev aluation, there are f e w e xamples of int e grated assessments of enhancement programs ( T aylor et al. 2005), especially for recreational fisheries. Case Study : Red D rum Enhancement in Florida This articl e pr o vides a first int e grat i v e assessment of the role of enhancement in the current and potential future management of a marine recreational fishery: the Florida Red Drum Sciaenops ocellatus fisher y . The potential enhancement of Florida s Red Drum fishery ex emplifies a compl e x recreational enhancement system and is useful as a case stud y . Red Drum is one of the most desired species of Florida s marine recreational fisheries, b ut this recreational fishing imparts substantial mortality on the species (Murp h y and Mu yandorero 2009). Most impo r tantl y , recreational fishing e f fort has been increasing ov er the past decades and is e xpected to continue to rise. This creates a challenge of maintaining sta k eholde r supported management goals for Red
27 Drum populations whil e meeting implicit management goals of sustaining the great socioeconomic utility realized by their exploitation. Stock enhancement is seen by stakeholders as a potential avenue for achieving these goals, which are in conflict in an entirely capture based fishery (Lorenze n 2005). Stock enhancement in Florida enjoys support from many fishing stakeholders, but so far, the efforts have been primarily small scale and research focused. A la r ge r scale marine enhancement initiat i v e is currently being pursued by a public pr iv ate partnership in Florida. A g ainst this background, the object i v e of this w ork w as to synthesize information critical for int e grated assessment of enhanced recreational fisheries through use of a case study: the potential enhancement of Florida s Red Drum fisher y . A nalytical F ramework Undertaking a broad based, int e grat i v e r e vi e w of the role or potential role of enhancement in the fisheries system is an important first step in the recently updated responsible approach to fisheries enhancement (B lankenship and Leber 1995; Lorenzen et al. 2010). To structure this review and analysis, a broad framework is used for analyzing enhancement fishery systems as described in Lorenzen (2008). The framework sets out how situational variables (attributes of th e resource: fishing, aquaculture production, habitat and environment, stakeholders, markets and governance arrangements) influence outcomes of enhancement initiatives through physical biological pathways and through those mediated by stakeholder action ( Fi gure 2 1). Criteria that may be used to evaluate outcomes include biological production, resource conservation, economic benefits and costs, contribution to livelihoods, and institutional sustainabilit y . While not a fully specified model, the fram e w ork pr o vides an aid for thinking through the logic of the fisheries systems and e xploring options for its
28 d e v elopment. This is done in three steps: (1) establishing current outcomes and future scenarios with desired outcomes, (2) r e vi e wing the situational v ariabl es that may impinge current and future outcomes in order to r e v eal the most important dr i v ers of outcomes, and (3) e xploring the dynamics of the most important dr i v ers and the uncertainties associated with them further in order to der i v e management and res earch recommendations. Current outcomes were assessed by r e vi e wing pertinent literature to describe Red Drum population status in Florida, their socioeconomic v alue, and h o w stock enhancement of Red Drum has been used in Florida and nearby areas. This synt hesis w as used to e xplore attri b ute v alues necessary to produce desired outcomes of Red Drum enhancement. From these k e y attri b ute elements, kn o wledge g aps and uncertainties were deduced. Finally, the way in which management of Red Drum enhancement should move forward is suggested first describing potential methods for reducing uncertainty in the key attributes determining system outcomes and then proposing interim recommendations for enhancement in the absence of this uncertainty reduction. Outcomes and S c enarios Current o utcomes : population s tatus Florida s Red Drum stock is considered sustainably fished, i.e., it is not considered ov erfished or subjected to ov erfishing according to the most recent stock assessment (Murp h y and Mu n yandorero 2009). Since 19 8 9 , when commercial sale of nat i v e Red Drum w as outl a wed, estimated numbers of sub adults (ages 1 4 that are typically ta r geted by the fishery) h a v e increased to current l e v els of 2.7 million on the Gulf coast and 1.3 million on the Atlantic coast ( Figure 2 2 ). This a b undance meets the management goal of 40% escapement, though some r e gions are ab ov e or bel o w this
29 number (Murp h y and Mu n yandorero 2009). In this fisher y , escapement (defined as the proportion of estimated number of age 5 fish currently to estim ated number of age 5 fish in unfished conditions) w as used as a proxy for the more traditional biological reference point, sp a wning potential ratio (SPR), due to a dearth of information about sp a wning adults (Murp h y and Mu n yandorero 2009). Socioeconomic o u tcomes Red Drum Crabtree 2001), which is a substantial market with a direct (e.g., license sales) and indirect (e.g., transportation costs) value of $6 billion (American Sportfishing Associ ation [ASA], 2001). The economic contribution specific to Red Drum is difficult to assess (Murphy and Mun yandorero 2009), as Red Drum are generally targeted as part of a multi species inshore fishery. However, targeted Red Drum effort has markedly increase d since 1999 on both Gulf and Atlantic coasts ( Figure 2 3; Murphy and Munyandorero 2009), and this is assumed to result in a proportional increase in economic value. Relative changes in angler satisfaction are unstudied; however, satisfaction is often dire ctly related to catch rates (Ditton and Fedler 1989; Cox et al. 2003; Arlinghaus 2006), which have increased over the last decades ( Figure 2 3; Murphy and Munyandorero 2009). Role of e nhancement Red Drum stocking has occurred on a la r ge production scale el s e where in the southeastern United States (e.g., millions released per year in T e xas), b ut in Florida enhancement has been generally smalle r scale and often research oriented. Research oriented stockings are often smaller scale by design; h o w e v e r , understa nding w h y enhancement in Florida has not progressed to la r ge r scale production is insightful.
30 Enhancement has been discouraged partly by a perception that stoc k ed Red Drum e xperience acute mortality from the stocking e v ent (Serafy et al. 1999; Sher w ood et al. 2004) and v ery l o w surv iv al thereafter ( T rin g ali et al. 2008 B ). This perception la r gely originated from a v ery poor surv iv al of a past enhancement in Biscayne Ba y , Florida, where post lar v al Red Drum were stoc k ed on a la r ge scale (millions) into what w as later determined to be quite poor post lar v al habitat ( T rin g ali et al. 2008 B ). Concerns that stocking might not be e f fect i v e may not h a v e been allayed by recent research oriented enhancement in T ampa Ba y , Florida, which suggested surv iv al of stoc k ed R ed Drum in certain areas to be similar to that of wild fish ( T rin g ali et al. 2008 B ). H o w e v e r , la r ge scale Red Drum stocking has persisted els e where (e.g., T e xas) despite li k ely l o w surv iv al of stoc k ed fish (Scharf 2000). Alternat i v el y , lack of production l e v el enhancement in Florida may be superficially due to a current lack of production l e v el hatching and rearing f acilities. This, h o w e v e r , w ould suggest ultimately a lack of sta k eholder mot iv ation to ad v ance production l e v el enhancement. It is possible th at sta k eholders are satisfied with the obser v ed increase in wild Red Drum a b undance ov er the last 20 years, as well as the a b undance of alternat i v e species ta r geted by recreational anglers (Sutton and Ditton 2005). Future Scenario, Desired Outcomes, and Ma nagement Options Scenarios The most recent stock assessment for Red Drum suggests future scenarios for ecological outcomes li k ely include declining Red Drum populations. Increases in recreational fishing e f fort for Red Drum are projected to lead to increas es in total mortality of 20% ov er a fi v e year period (i.e., 4%/year for 2007 2012) on the Atlantic
31 coast and approximately 50% (i.e., 10%/year for 2007 2012) on the Gulf coast, according to the last stock assessment (Murp h y and Mu n yandorero 2009). Ecologic al o utcomes These increases in total kill were e xpected to lead to declines in escapement on the Atlantic coast to roughly 40% b ut to approximately 22% on the Gulf coast (i.e., bel o w the 40% escapement management goal; Murp h y and Mu n yandorero 2009). While e f fort may not h a v e increased as sharply as predicted due to a national economic d o wnturn, the prediction that future Red Drum escapement will li k ely f all bel o w the management threshold is v alid. The desired future outcome is that wild Florida Red Drum pop ulations remain ab ov e the 40% escapement threshold, which is anticipated to meet fundamental conser v ation object i v es as well as promote long term sustainable socioeconomic v alue. The distinction of wild fish is not e w ort h y because wild fish are e xpected to maintain maximal fitness and genetic d i v e r sity (Lorenzen et al. 2012). Socioeconomic o utcomes Red Drum fishery appear well met currently, they may not be in the future. Increasing effort may ou tstrip angler satisfaction) if effort is stable and independent of fish abundance/catch rates, or it could decrease effort if effort is related to abundance or catch rates. If effort is weakly related to catch rates, both might decrease following declining Red Drum populations. Regardless, decreased effort or catch rates would result in decreased socioeconomic value. Conversely, the desired socioeconomic outcomes are to maintain or increase both economic effects and social satisfaction.
32 Options and the role of e nhancement Enhancement of Red Drum in Florida is seen as a potential w ay to maintain or increase socioeconomic v alue of the Red Drum fishery without harming wil d populations. This may be accomplished in three ways: (1) stocking increases catch rates when effort is stable, (2) stocking maintains stable catch rates when effort is increasing due to external drivers (e.g., human population growth), or (3) stocking mo tivates initially greater catch rates and ultimately increased effort. In all cases, socioeconomic value of the fishery would increase as a result of stocking increasing total Red Drum abundance. In order for this increase in economic value to be sustainab le, the increase in v alue must e xceed the costs associated with the enhancement program increasing fish abundance. Alternatively, enhancements may support positive changes in socioeconomic outcomes even when a direct impact of stocking on Red Drum abundanc e or fishing effort is not discernible if stakeholders either believe fishing quality is improved or value stocking as a form of active resource stewardship. Enhancement is not the only option by which the projected pressures could be handled. Others inclu de (1) switching to a catch and release fishery and (2) restricting access to the fisher y . Reducing or eliminating har v est, i.e., making the fishery predominantly or e xclus i v ely catch and release, all o ws maintaining fishing mortality and escapement in the f ace of increasing e f fort. This approach is limited in scope by the f act that e v en released fish su f fer increased mortality (discard mortality i.e., mortality of released fish due to capture related injuries), b ut nonetheless, a change t o w ard catch and rel ease could absorb some l e v el of e f fort increase. Such a change could be mandated in r e gulations b ut could and often is made v oluntarily by anglers, for
33 e xample, in the Florida B ass (fresh w ater) and S nook (inshore costal) fisheries. Restricting access to th e fisher y , in theor y , is the most e f fect i v e w ay of limiting fishing mortality and it could be done at di f ferent l e v els to maintain a desired escapement l e v el while all o wing or not all o wing some l e v el of har v est. Attributes of the Enhancement F ishery S ystem and Their R ole in D etermining Current and Future Outcomes Biological Attributes Several biological attributes of Red Drum are important in influencing fisheries and enhancement outcomes: basic life history and ontogenetic shifts in habitat use (which infl uence vulnerability of life stages to fishing and to habitat degradation and inform release strategies for hatchery reared juveniles) and the strength and ontogenetic pattern of compensatory density dependence (which influences the extent to which stocking can raise abundance and its impacts on the wild population component). Red Drum are a large (maximum weight 30 kg), long lived (maximum age 40 60 years) marine fish of the southwest Atlantic and Gulf of Mexico that, as adults, occupy near and offshore areas. As juveniles and sub adults (ages 0 5), Red Drum loosely associate year round with structural habitat (e .g., sea grass beds, oyster bars) in estuaries and inshore areas where they grow rapidly to large size ( 5 8 kg), feeding on small fish, shrimp, and invertebrates. These inshore areas are where recruitment dynamics take place, which are critical to all enh anced system outcomes. Red Drum probably exhibit strong compensation and populations are relatively abundant (Murphy and Mu n yandorero 2009). F or enhancement to potentially augment wild populations, stocking must occur either after highly compensatory surv i v al stages
34 or when a b undances are l o w enough for potential g ains in total recruitment (Lorenzen 2005). H o w e v e r , Florida Red Drum compensation with age/length has not been well characterized. In T e xas, Scharf (2000) found compensation w as substantial throug h the end of the first yea r . Alternat i v el y , Bacheler et al. (2008) suggested that, in North Carolina, yea r class strength w as set shortly after lar v al settlement. St e w art and Scharf (2008) similarly suggested that Red Drum recruitment in South Carolina w as set shortly after settlement in the first yea r , although some un e xplained v ariation suggested later compensatory processes might also occu r . The spatial scale of density dependent processes also influences enhancement outcomes locally defined, l o w recruit ment areas hold potential to av oid competition between wild and stoc k ed fish. Studies from T e xas and North and South Carolina suggest that recruitment probably v aries on som e what local (10s 100s km) scales (Bacheler et al. 2008; St e w art and Scharf 2008), t hough it may be e v en estuary specific (Scharf 2000). Locally defined recruitment li k ely translates into local populations of (catchable) sub adults. Studies suggest sub adult Red Drum exhibit high site fidelity (Reyier et al. 2011) and move little outside of their nursery estuary (Adams and T remain 2000; Rooker et al. 2010), though they are capable of large scale (100s of kilometers) movements. However, adult Red Drum are quite mobile. In Florida, populations are assessed on coast wide scales, and females a re likely to spawn within 500 600 km of their natal estuaries (Gold et al. 1999; Murphy and Crabtree 2001; Gold 2008). Genetic structure is likely commensurate to these spawning areas, with evidence existing of genetic differentiation between Atlantic and Gulf stocks ( M. T rin g ali, personal communication). In concert, these studies suggest generally well mi x ed adult populations sp a wning at la r ge spatial scales along each of the
35 Gulf and Atlantic Florida coasts b ut probably more discreet sub adult populations e xisting at local estuarine scales. Red Drum biological attri b utes may relate to outcomes through indirect feedback loops. F or e xample, Red Drum recruitment processes (i.e., the need to raise fish la r ge enough to bypass compensatory surv iv al) a f fects the cost of raising fish, and thus the opportunity cost (in terms of alternative management) of stocking. Red Drum biological attributes also affect system outcomes by route of governance and fisheries attributes. Specifically, the ontogenetic shift to offshor e waters where they become semi pelagic ensures the spawning population is largely invulnerable to recreational anglers (it is also protected from commercial fishing by a regulation that bans harvest in federal [offshore] waters). Perhaps the most obvious feedback is how biological attributes affect the population effect of stocking, with population changes influencing fishing effort and, therefore, Red Drum mortality. Many other feedback loops are possible, because nearly all outcomes of enhanced recreatio nal systems are routed in some fashion through the biological attri b utes of the ta r get species. Fishery Attributes Technical fisheries attributes directly influence both socioeconomic and ecological outcomes of enhanced recreational fisheries by determinin g catch and fishing mortality (Taylor et al. 2005; Lorenzen 2008), which are functions of effort, catchability, and discard mortality. Effort can change related to stock abundance (Loomis and Fix 1998; Walters and Martell 2004; van Poorten et al. 2011), fi shing regulations (Beard et al. 2003), alterations in stakeholder attitudes or typologies (Johnston et al. 2010), or knowledge of sto cking efforts (Baer and Brinker 2007). Although Red Drum fishing effort in Florida has increased over the last several d eca des (Murphy and Munyandorero
36 2009) with growing fish and human populations and knowledge of the species, causality has not been determined. Due to the aforementioned ontogenetic shifts in habitat use, and restriction of recreational fishing for Red Drum to nearshore waters, Red Drum catchability in Florida is limited to juveniles below the ages of 4 6 and, therefore, is generally described as dome shaped with respect to age. Recent estimates of Red Drum catchability and effort indicate moderate fishing mort ality rates tively (Murphy and Munyandorero 2009). Harvest and discard mortality combine to produce an estimated 0.8 million Red Drum killed annually (Murphy and Munyandorero 2009; Reyier et al. 2011). While recent Florida Red Drum assessment models assume discard mortality rates of 5% per catch and release event (Murphy and Munyandorero 2009), other estimates from the Gulf and southeast Atlantic have ranged from 0 to 44% (Muoneke and Ch ildress 1994). Because even low discard mortality can profoundly affect total mortality when effort or catchability are high (Coggins et al. 2007), discard mortality may mediate enhancement outcomes as the mechanism for increased wild fish mortality, throu gh stocking induced effort increases. Through har v est and discard mortalit y , e f fort and catchability directly dr i v e ecological system outcomes and indirectly a f fect socioeconomic outcomes dependent on fish a b undance. Additionall y , e f fort relates directly t o mar k et outcomes by dictating the economic e f fects accrued by the fisher y . Satis f action is also strongly influenced by catch rate oriented metrics, and so depends in part on e f fort (Ditton and Fedle r 1989; Arlinghaus 2006). In turn, e f fort dynamics are re lated to biological attri b utes (as
37 pr e viously described), and li k ely also to attri b utes of sta k eholders (often anglers) who find Red Drum substitutable for and by other species (Sutton and Ditton 2005). Fishing is unselective with respect to wild and stock ed fish, but harvesting could be made selective to hatchery fish if these could be identified through tagging. Selective harvesting of stocked fish could be allowed to satisfy harvest oriented anglers while discouraging or outlawing harvest of wild Red Dru m . Selective harvesting of hatchery fish could also reduce their ecological and genetic interactions with the wild stock. Selective harvest policies in an enhanced Red Drum fishery could reduce some, but certainly not all, conflict between socio economic a nd ecological objectives. Technical Attributes of Aquaculture and Release Technical attributes of aquaculture and release (including feasibility and efficiency of mass culture, domestication effects, size at release, microhabitat of release, season/tide of release) are particularly related to biological outcomes by determining the stocked fish survival, health, and contribution to total population dynamics (Leber et al. 1998, 2005; Lorenzen 2008). Technical expertise required to spawn, hatch, and rear Red D rum exists, though production level aquaculture still faces challenges. In Florida, all juveniles for stocking have been spawned from wild brood stock and reared intensively in tanks as well as extensively in ponds (Tringali et al. 2008b). Red Drum post st ocking survival has been shown to increase with size at release (Willis et al. 1995; Tringali et al. 2008b ); however, cost of hatchery production also increases. Understanding the trade off between cost and survival is important to produce the greatest pot ential for population increase per unit cost (Leber et al. 2005). Stocked Red Drum may also experience high mortality immediately post release regardless of stocking size (Sherwood et al. 2004). The causes of such mortality are
38 unclear, but in other specie s, immediate post release mortality has been related to microhabitat, tide, and season of release, as well as pre release acclimation (Leber et al. 1997, 1998; Brennan et al. 2006). Low survival of stocked fish may also be related to domestication effects (Leber 2002; Lorenzen 2008). Hatchery Red Drum react more slowly to food and predators than do wild fish, and anti effect on these behaviors (Stunz and Minello 2001; Beck and Rooker 2007, 2012). Additionally, domesticatio n effects may result in less fit hatchery fish that may contribute deleterious gene s to wild populations (Lorenzen 2008; Lorenzen et al. 2010). While past Red Drum enhancement projects in Florida posed little risk of genetic swamping (Tringali et al. 2008b ), this is in part due to low absolute survival of stocked fish, and potential genetic risks of large scale stocking programs still exist. It must be understood that given Red Drum life history and the small size of stocked Red Drum , low absolute survival is expected (i.e., wild fish also experience low survival at this size/stage). While some stocking events have exhibited much lower survival than expected, the survival rates attributable to other stockings suggest the current enhancement technology and cu lture methods can result in some recruitment of hatchery fish the challenge will be producing the desired quantity at desired sizes. Habitat A ttributes Understanding stocked fish habitat needs is broadly acknowledged (Molony et al. 2003; Lorenzen 2008), as structural habitat can mediate density dependent surviva l processes (Walters and Juanes 1993) and impact the population level effects of enhancement activities. Unfortunately, controlled stocking experiments testing habitat effects on Red Drum survival ar e lacking. However, broader studies suggest habitat is important to Red Drum and thus may mediate stocking success. Juvenile and sub adult
39 Red Drum are strongly associated with structural habitat (Rooker et al. 1999; Murphy and Munyandorero 2009), and thes e habitats are predicted to be influential to total Red Drum pop ulation growth (Levin and Stunz 2005), though recruitment, growth, and mortality may be related to non structural habitats, such as river d ischarge (Purtlebaugh and Allen 2010). Stakeholder At tributes Stakeholder attributes can directly alter outcomes by determining fishing effort and social utility dynamics. Because these dynamics are largely functions of primary stakeholder (i.e., angler) attitudes and values, assessing overall outcomes requi res characterizing stakeholders into different typologies to better predict responses to enhancement (Johnston et al. 2010). Typologies may be characterized in terms of coarse motivations (e.g., outdoor recreation general versus fish ing specific; Fedler an d Ditton 1994; Arlinghaus 2006), as well as more specific motivation differences (e.g., species targeted or trophy oriented anglers versus harvest oriented anglers ; Sutton and Ditton 2005; Johnston et al. 2010). Because diverse motivations require modeling multiple typologies to assess enhancement outcomes, understanding the breadth of motivations is critical. Florida Red Drum f other inshore species (e.g., Spotted S eatrout , F lounder and Common S nook) as potential substitutes for Red Drum (Sutton and Ditton 2005) suggests that a generalist typology may be common. While studies of Red Drum anglers in Texas (Oh and Ditton 2006) suggested different typologies (in terms of expected/desired angling experiences) existed, little is known of Florida Red Drum typologies.
40 A key stakeholder attribute that broadly affects outcomes is stakeholder social 2008; Lorenzen et al. 2010). This investment is critical, since this is generally essential for broader system changes (Ostrom 1990; Oakerson 1992), such as enhancements (Lorenzen 2008; Lorenzen et al. 2010). The importance of Florida Red Drum stakeholder for enhancement is illustrated by the changes in the Biscayne Bay Red Drum stocking where stocking changed from experimentally stocking a wide range of fish sizes to mass production of small post larval fish, based on stakehold er demand for increasing stocking density (Tringali et al. 2008b). While predicting stakeholder response to potential future Red Drum enhancement is difficult, support or buy in generally requires participation of stakeholders in the management process (Lo renzen et al. 2010; Miller et al. 2010). Non angler stakeholders are also affected by and may influence enhancement indirectly (Arlinghaus 2006) and so should also be considered (Lorenzen, 2008), but little is known of how such groups view potential enhanc ement in Florida. Given past Red Drum stocking in Florida, it is reasonable to assume that stakeholder opinions and support will play a large role in future enhancement outcomes and that such support is largely dependent on the extent and method by which s takeholders participate in enhancement decisions. Market Attributes Market attributes influence enhanced recreational fisheries by directly affecting economic and social outcomes and by indirectly alterin g ecological outcomes (Lorenzen 2008; Lorenzen et al . 2010). While the Florida Red Drum fishery is valuable as part of a multi billion dollar fishery (ASA, 2001), the absolute economic or social value is difficult to asc ertain (Murphy and Munyandorero 2009). Absent absolute values, both economic
41 and social values are generally considered proportional to fishing effort (Cox et al. 2003; Walters and Martell 2004). While it may be fair to assume economic impact is somewhat positively correlated to effort, the ratio between effort and economic impact (e.g., 1:1, 10:1) is unknown. Fishing effort response is likely related to the social satisfaction anglers expect to attain from fishing (Arlinghaus 2006), which is also important to gauge as a metric of social value. Expected satisfaction is a function of the motiva tions of anglers, and can vary by stakeholder typologies, but commonly responds to catch rates, crowding, facilities, etc. Though it is assumed that satisfaction of Red Drum anglers is positively related with catch rate, the shape of this relationship and how it may be mediated by other elements (e.g., harvest versus catch and release, boat ramps and facilities, congestion, etc.) is unknown. The effect of scale should be very important for understanding the impact of changing market values but has not been well studied. For example, the extent to which potential increases in Red Drum effort are redistributions from others substitutable species is unknown. While the existence of substitutes (Sutton and Ditton 2005) suggests that demand for Red Drum might be m ore elastic, Red Drum may be preferred over some of these substitutes (because of consistent availability, etc.). The lack of clarity of how recreational fishing and specifically Red Drum fishing is valued compared to alternatives (i.e., relative demand) m akes it difficult to assess the gains possible from increasing Red Drum abundance via enhancement. Furthermore, the cost and funding Red Drum fisheries must be understood to evaluate the effect of enhancement on market values. Currently, it is not known how hatcheries and enhancement activities will be funded. Depending on funding sources,
42 enhancement may be viewed in terms of opportunity costs, such as habitat restoration, facility augmentation, etc. Public (i.e. , general taxpayer) funding should be evaluated in a broader view of opportunity costs. This may requires assessing stakeholder response to these alternative actions, in addition to enhancement. Governance Governance attributes impact system outcomes prima rily by directly controlling the type of enhancement allowable (Lorenzen 2008; Lorenzen et al. 2010). Red Drum recreational fisheries are regulated open access throughout the United States, with each state having autonomy to create its own regulations for harvest and for enhancement. Marine stock enhancement in Florida is organized jointly by the state management agency Florida Fish and Wildlife Conservation Commission (FWC) through the Fish and Wildlife Research Institute (FWRI) and the stakeholder led Flo rida Marine Fisheries Enhancement Initiative (FMFEI). The structure allows FMFEI to exert some influence on governance, which it does through raising funds for enhancement . As such, governance and ultimately the outcomes of Red Drum enhancement in Florida is dependent on stakeholder opinions and actions. However, past and present enhancement initiatives in Florida have remained essentially unconnected with fisheries managem ent, i.e., actual or potential stocking has not been considered in fisheries management plans, nor have enhancement initiatives made any specific claims as to desired changes in fishing regulations related to stocking. It is unclear whether the enhancement initiative is ideologically driven, i.e., stocking is fundamentally favored, or whether other motivations, e.g., greater angler satisfaction per trip, are the true fundamental objectives. The former limits governance actions, under the current co -
43 manageme nt system, to shaping enhancement activities, while the latter allows for assessment of enhancement in the context of alternative management actions (e.g., habitat restoration, traditional fisheries regulations). Governance attributes may directly control enhancement or may affect enhancement indirectly by influencing alternative management strategies and ultimately impacting fish abundance and angling effort (Aas 2008; Lorenzen 2008; Lorenzen et al. 2010). Florida management goals of 40% escapement is mana ged solely by size and bag limits, which, until recently, allowed for one Red Drum between 18 27 inches per day per person to be harvested with no closed season. The lack of a closed season probably elevates the value of this fish to stakeholders (especial ly fishing guides). On 01 February 2012, the bag limit increased to two Red Drum per day per person in northeast and northwest Florida waters. This increase both allows for the potential of increased harvest oriented satisfaction that might occur if enhanc ement effectively increases catch rate and calls into question the necessity for stock enhancement in impacts on the Red Drum fishery and its potential enhancement. For exa mple, recent closures of the offshore reef fisheries could potentially shift additional effort toward inshore fisheries like Red Drum , although Fisher and Ditton (1994) found these inshore species to be considered poor substitutes by some offshore anglers. Similarly, an increased fishing effort or even harvest directed toward Red Drum , since these fish were thought of as substitutes (Sutton and Ditton 2005).
44 Key Linkages , U ncertainties , and Recommendations Outcome Linkages and Requirements Enhancement outcomes are a function of dynamically linked attributes (Lorenzen 2008), but in many recreational systems (e.g., this case study), key linkages exist where outcomes particular ly hinge on certain attributes ( Figure 2 4). In this case, socioeconomic outcomes (i.e., stakeholder satisfaction and local economies) depend directly on market, stakeholder, governmental, and fisheries attributes and indirectly on changes in Red Drum abun dance ( Tables 2 1 and 2 2 , Figure 2 1 ). Ecological outcomes (i.e., Red Drum population status) are direct functions of biological and habitat/environmental attributes, but also are a function of fisheries attributes (namely the effort response; Tables 2 1 and 2 2 , Figure 2 1 ). Thus, socioeconomic outcomes depend partially on Red Drum abunda nce, which are closely related to ecological outcomes, and ecological outcomes are in part functions of fishing effort/mortality, which is closely related to socioeconomic outcomes. In this way, socioeconomic and ecological outcomes depend on each other. H owever, realizing each objective also requires additional specific attribute values. Desired socioeconomic objectives of Red Drum enhancement may be realized through multiple pathways, but each require certain attribute values. The primary pathway is for e nhancement to augment Red Drum populations, leading to increased fishing effort and/or catch rates and finally increased economic impact and/or stakeholder satisfaction ( Figure 2 4 ). Each of these steps has key requirements. For wild populations to be increased (the first transition in Figure 2 4 ), Red Drum must be stocked when survival compensation (and competition with wild fish) is low and in areas where adequate habitat is present. For augmented popu lations to translate into higher
45 catch rates (the second transition in Figure 2 4 ), catchability must remain unchanged and effort must increase proportionally less than Red Drum abundance (i.e., effort must not be too responsive to higher catch rates). Alternatively, for total Red Drum effort to increase (second transition in Figure 2 4 ), effort must be responsive to increased fish abundance or persist following temporarily higher catch rates. Finally, f or total socioeconomic value to increase on a broad scale (third transition in Figure 2 4 ), increases in effort and satisfaction must be organically created, i.e., not redistributions from other fisheries or even other economic s ectors, and increased effort must not cause substantial decreases in satisfaction (such as might occur from crowding). Since in Florida, substitutions and redistributions among at least inshore sport fish species are likely, economics and satisfaction of R ed Drum enhancement should really be judged at a larger scale (e.g., total inshore effort/satisfaction). If the values of all of these attributes are favorable, stocking may achieve the socioeconomic outcomes desired. However, since these outcomes are the product of such a linked system, simply one misaligned value may jeopardize the desired outcomes. Alternatively, some desired socioeconomic outcomes (higher satisfaction) may be achieved if Florida stakeholders gain satisfaction in response to the act of e nhancement alone, regardless of how Red Drum populations change. Similarly, if enhancement facilitates greater inclusion or investment of stakeholders in management processes, some socioeconomic objectives may be met without enhancement altering the availa bility of fish. Meeting the desired ecological outcome of sustained wild Red Drum populations further constrains attribute values. For wild Red Drum exploitation/mortality to not increase, catchability must not increase and effort must not respond in great er
46 magnitude than any increase in abundance. However, responsive effort is critical for achieving desired socioeconomic outcomes through the primary pathways. Additionally, aquaculture attribute values must result in avoidance of deleterious genetic effect s; that is, less fit stocked fish must not spawn in appreciable numbers with wild fish. Ensuring that stocked fish are similarly fit to wild fish is not realistic, since this would require stocked fish go through full selective (i.e., compensatory) process es, which essentially negates the potential for population augmentation. Accordingly, two of the most critical links between attributes and outcomes are (1) the Red Drum recruitment dynamics, which are likely to mediate how population size could change wit h stocking, and (2) angler catch rate and effort dynamics, specifically as they relate to changing Red Drum population size ( Figure 2 4 ). Simply, the values required for socioeconomic outcomes conflict with those required for eco logical outcomes. Socioeconomic outcomes require motivating increased fishing effort and having high stocked fish survival and are likely to negatively impact desired ecological outcomes of sustained wild populations through higher fishing mortality, incre ased competition with stocked fish, deleterious genetic impacts, or a combination of all of these. It is not clear how this trade off between socioeconomic and ecological outcomes can be avoided. Uncertainties Several of the values of key requirements for achieving ecological or socioeconomic outcomes are uncertain, which ultimately results in uncertain outcomes of Red Drum enhancement. The main uncertainty in biological attributes surrounds Red Drum recruitment dynamics specifically the timing, extent, and mediation by habitat of compensatory periods and spatial distribution of sub adult (catchable) Red Drum . Another key uncertainty is market value attributes, specifically the responsiveness of
47 effort to increased abundance/catch rates, and the source of th is effort (redistribution versus original recruitment). Finally, there is substantial uncertainty in stakeholder attribute values, which affects stakeholder satisfaction functions, as well as how enhancement might affect stakeholder investment in managemen t. So while many attribute values are uncertain, the most critical uncertainty surrounds stakeholder, biological, fisheries, and market attributes ( Figure 2 5 ). Viewing these in concert with key linkages between attributes and ou tcomes, recruitment dynamics, angler effort dynamics, and stakeholder investment emerge as key uncertainties to reduce. Recommendations For responsible progression of Red Drum stock enhancement in Florida, reduction and accounting of uncertainty (not inher ent variability) surrounding recruitment, effort, and stakeholder dynamics are needed. This may best be achieved using an active adaptive management approach that combines quantitative modeling, mensurative and manipulative experiments, and monitoring in a n a priori designed structure t o learn from uncertainty (Leber 2002; Walters and Martell 2004). In this framework, modeling and small scale experiments reduce some uncertainty (e.g., recruitment timing) to inform initial enhancement, which, when carried ou t stocking). Multiple modeling efforts should be designed general quantitative models addressing both the feasibility of achieving desired outcomes and spatially expli cit models useful for predicting acceptable stocking locations and strategies. Pursuing enhancement (Lorenzen et al. 2010) and detailed information of conditions in which enh ancement is most likely to succeed. Specifically, age structured population models
48 including multiple young of year Red Drum stages or types (e.g., hatchery, wild) can be used to simulate outcomes of alternative timing, intensities, and habitat mediation o f compensation. Small scale experimental stocking should be designed to provide starting estimates of uncertain variables in the quantitative models, such as the magnitude and variability of density dependent survival of stocked Red Drum and variation in r elease microhabitat mediated survival. Further compensation inferences may be gleaned from analysis of existing monitoring data, similar to Scharf (2000). Similarly, analyses of existing effort and catch rate data (e.g., Marine Recreational Information Pro may be used to estimate effort response dynamics. Effort responses might also be evaluated through actual experimental stocking manipulations, as has been done in some freshwater systems (Baer et al. 2007), or inferences of redistribution of effort among fisheries gained from construction of random utility models, similar to Schuhmann (1998). Broader, multi market economic analysis may be necessary to assess the total net effect of enhance ment on the economics of the state. To reduce uncertainty in the stakeholder response to enhancement through active adaptive management, it is critical to first develop and nurture existing participatory stakeholder involvement in future enhancement decisi ons. This should improve ability to detect changes in stakeholder satisfaction with experimental stocking, though this process will likely require initiation by managers, as has been previously done in Florida (Tringali et al. 2008a). An emphasis should be placed on finding strong leaders of this process who are trusted by stakeholders (Ostrom et al. 1999) and can help develop participatory approaches that function to guide and in fact design
49 enhancement activities. Accomplishing this may be time and effort intensive but likely will result in a group of invested stakeholders who understand and participate in analysis of enhancement, as described in Miller et al. (2010). Such an approach seems to offer the best chance for enhancement to meet stakeholder objec tives and should also result in stakeholder engagement sufficient for evaluating enhancement in the adaptive management framework. Synthesizing information from various stakeholder, angler effort, and recruitment dynamic studies should provide a basis for designing an active adaptive management process by which enhancement is experimentally conducted and the results monitored to reduce uncertainty and inform future enhancements. It is worth acknowledging that few examples exist of successful active adaptive management and that informing the explicit a priori design of this process is costly. It may not be possible for all useful studies (particularly those including field work) to be completed with scarce time and funds prior to any enhancement, especially i f stakeholder demands for stocking increase. Even in this case, key steps should be followed to maximize benefits per risk. First, existing knowledge of attribute values and uncertainty can be incorporated into translucent, quantitative models that predict ranges of outcomes and so inform any potential enhancement. Second, any potential enhancement can be well monitored, particularly with respect to key uncertainties. For example, if a certain region is stocked, perhaps creel surveys can be increased in tha t region and adjacent, unstocked regions to allow greater power to detect angler effort responses. Finally, any enhancements should be particularly adverse to risk to wild Red Drum populations, because any harm to them may be quite long term, given their l ife
50 span, and incur high socioeconomic costs, given the value of the fishery. To minimize the probability of deleterious impacts on wild Red Drum populations and to maximize potential socioeconomic gains, the largest possible Red Drum i.e., the closest to catchable size should be stocked in high fishing pressure areas. Such fish are likely to have better post release survival based on size dependent mortality, are less likely subject to high density dependent survival, and may not be as sensitive to habitat as smaller fish. Stocking larger sized fish in areas where abundance of wild Red Drum is low due to high fishing effort increases probability of augmented catch as a result of stocking, while possibly decreasing likelihood of negative genetic effects thro ugh interbreeding with wild fish (though Ryman Laikre effects might be more prevalent). To allow the greatest potential of learning, such stockings would be completed experimentally (e.g., replicated within a blocked design) such that system uncertainties may be studied (e.g., Leber et al . 1997, 1998). Furthermore, all, or at least a large, known proportion of the fish stocked should be individually marked so that the effects of the stocking can be m onitored (Blankenship and Leber 1995; Walters and Martell 2004; Lorenzen et al. 2010). While stocking in such a way may have a high financial cost of production per stocked fish, stocking large fish where wild fish are and anglers are not is likely to maximize catch of stocked fish per negative impact on wild pop ulations. Discussion Synthesis of past studies and current understanding shows that socioeconomic and ecological outcomes of potential enhancement are reflexively related and likely depend on combinations of multiple linked attributes. However, the simulta neous accomplishment of socioeconomic and ecological outcomes is doubtful, since trade offs are clear stocking that would achieve greater angler effort and satisfaction likely
51 increases the probability of negative effects on wild populations. This trade of f can be perhaps better understood and more translucently presented by reducing uncertainty regarding recruitment/compensatory dynamics, angler effort dynamics, and stakeholder investment patterns. To measure the full impact of enhancement, ecological and especially socioeconomic outcomes must also be viewed in a much broader context that extends beyond the specific enhanced species. These assessments may be best made by combining simulated populations models, small scale stocking experiments, and analyses of existing data. Regardless of these analyses, how and if Red Drum stock enhancement occurs may be determined by stakeholder interests, which play a large role in enhancement outcomes. Accordingly, fostering stakeholder understanding and investment may be the most crucial next steps for Red Drum stock enhancement in Florida. Red Drum fishery are largely universal to recreational fisheries, and so outcomes of this assessment may be widely applicable. Despite this overlap, few stu dies, regardless of species, have assessed the outcomes of enhancement, particularly in an in tegrated framework representing the complex coupled socio ecological system of recreational fisheries. In most enhancem ent systems, key linkages will be how enhan cement affects total population size and wild fish (i.e., recruitment dynamics), how fishing effort responds to population size (i.e., angler effort dynamics), and how stakeholders view and shape the entire process (i.e., stakeholder investment). Similarl y, the trade offs exhibited in this case study likely exist in nearly all recreational enhanced systems, since almost all recreational enhancement is an attempt to augment fishing effort and satisfaction while sustaining populations. It
52 follows then, that the key recommendations of this case study coupled quantitative and experimental evaluations and a greater emphasis on stakeholder inclusion in management may be useful in other systems.
53 Table 2 1. Role of attributes in determining current outcomes . Attribute Ecological outcome Socioeconomic outcome Role of enhancement Current outcome Population slightly more abundant than management threshold. Increasing effort and catch rate lead to high current socioeconomic value. Sporadic trial stockin gs have occurred but production or adaptively managed stockings not imminent. Biology Red Drum are relatively robust to recruitment overfishing, largely due to an ontogenetic shift to quasi pelagic, offshore habitats. Year round availabi lity of sub adults makes Red Drum a seasonally consistent sportfish in Florida, which is valuable. High compensation may have led to low survival of stocked fish. Fishery Moderate fishing mortality results in current sustainable exploitatio n, but increasing effort and non negligible discard mortality leads to mounting concern. High effort/catch rates lead to high socioeconomic value. Dome shaped vulnerability leads to a small temporal window for recapturing stocked fish, medi ating the observed catch related benefits of stocking. Aquaculture Because little Red Drum stocking has occurred, ecological impacts are minor. Because little Red Drum has occurred, socioeconomic impacts are minor. Occasional high survival o f stocked fish suggests that issues in the culture techniques are unlikely. Habitat May mediate recruitment and thus affect population status outcomes. Habitat mediates recruitment alterations and in turn socioeconomic and enhancement outcome s. May interact with compensation to result in locally poor survival in trial stockings. Stakeholder Stakeholder engagement has motivated management thresholds (e.g., escapement) to be relatively high. Stakeholder perceived preference relative to other inshore species drives value high for Red Drum. Historically, stakeholders have encouraged stocking by the state and influenced stocking practices but not provided or helped to secure funding for wider support or enhancement. Market Increasing total demand for Red Drum has indirectly led (realized through effort) to sustainability concerns. Perceived high value of the Red Drum fishery is due to the high and increasing effort and catch rate. Uncertainty in enhancement fund ing is likely related to lack of current production and perhaps stakeholder investment. Governance Red Drum regulations limit harvest to roughly one year, and this is likely responsible for current abundance of Red Drum. No closed season pr omotes year round consistency in Red Drum value, which is vital for businesses reliant on the fish sector. Governance designed stakeholder involvement in enhancement has directly driven current outcomes.
54 Table 2 2. Likely influence of attribute s on future outcomes . Attribute Ecological outcome Socioeconomic outcome Role of enhancement Future or desired outcome s Sustained healthy wild populations above management threshold. Maintained or increased effort and catch rates leading to increa sed or sustained socioeconomic value. Accomplish both desired socioeconomic and ecological outcomes at low cost relative to alternative management. Biology High site fidelity and vulnerability of sub adults mean catchable Red Drum may become s easonally locally depleted, particularly if effort increases. Reliability and popularity suggest an increase Red Drum targeted effort is possible . Region wide population augmentation requires stocking after highly compensatory life stages, thou gh local depletions may reduce compensation and allow for successful small scale stockings. Fishery High, increasing, and/or responsive effort suggests positive socioeconomic outcomes and negative population status outcomes are likely. High, i ncreasing, and/or responsive effort suggests positive socioeconomic outcomes and negative population status outcomes are likely. Reduced discard mortality can mitigate negative impacts of stocking induced effort on wild populations. Aquaculture Larger, post compensatory Red Drum may be technically possible and pose a greater threat to wild genetic integrity. Desired outcomes more likely if culture methods are developed to produce larger, post compensatory Red Drum. Producing high number of advanced fingerlings requires space, that is difficult and costly to acquire in coastal Florida. Habitat Likely to mediate recruitment and thus population status outcomes if habitat is altered. Changing habitat could alter desired outcomes indirectly through population status changes. Likely to mediate survival of stocked fish and thus enhancement outcomes (socioeconomic and ecological). Stakeholder Stakeholder behavior and opinions indirectly drive ecological outcomes by determi ning effort, percent harvest, Satisfaction may be strongly influenced by perceived involvement in, and buy in, to the management process. Future enhancements will likely require strong stakeholder engagement and are unlikely to b e initiated and sustained without at least moderate support. Market Demand and elasticity drive effort and thus population and ecological outcomes, but these (and thus effort dynamics) are largely unknown. Demand and elasticity drive total effo rt directly and likely satisfaction indirectly and thus socioeconomic outcomes, but these (and thus effort dynamics) are largely unknown. Costs of stocking and funding sources relative to alternative management (i.e., habitat restoration) likel y to impact enhancement. Governance Recent alterations in regulations may alter population status but, because of long adult spawning life, may not be noticeable soon. Regulation change may increase satisfaction in short term, potential to de crease in long term if populations falters. Regulations (or their absence) that result in lower populations and satisfaction may motivate stakeholders to invest/embrace enhancement.
55 Figure 2 1 . Framework for analyzing enhancement fisheries s ystems used in this study. Operational interactions between elements are shown as solid lines and determine outcomes in the short term when the situational variables are fixed. In dynamic interactions, shown as dashed lines, situational variables are modif ied in response to the outcomes of operational interactions (from Lorenzen 2008).
56 A B Figure 2 2. Red Drum a b undance estimated from the most recent stock assessment , (Murphy and Mu n yandorero, 2009). A) Atlantic coast red drum, all ages (solid line) and catchable fish, aged 1 4 (dashed line). B) Gulf coast red drum all ages (solid line) and catchable fish, aged 1 4 (dashed line).
57 A B Figure 2 3 . Estimated Red Drum ta r geted e f fort and catch rates from the most recent stock assessment (Murph y and Mu n yandorero, 2009). A) Atlantic coast estimated effort (solid lines) and catch rates (dashed lines). B) Gulf coast estimated effort (solid lines) and catch rates (dashed lines). from the most recent stock assessment (Murphy and Mu n yandorero, 2009) .
58 Figure 2 4 . Illustration of the ke y relationships essential for producing desired outcomes. K e y dynamics that are particularly uncertain and associated requ i rements of these aspects (t e xt lists) are pr o vided.
59 Figure 2 5. Uncertainty associ ated with attributes of Red Drum enhancement and difficulty in reducing this uncertainty. Radial distance is proportional to uncertainty or difficulty in reducing uncertainty.
60 CHAPTER 3 DYNAMICS: DOES FISH ABUNDANCE DRIVE EFFORT? Recreational fisheries are increasingly recognized to involve dynamic relationships between anglers and fish populations (Johnston et al. 2010; van Poorten et al. 2011; Allen et al. 2013). Understanding these relationshi ps facilitates achieving the broad management goals inherent to recreational fisheries sustaining satisfactorily abundant fish populations and achieving immediate desired socioeconomic benefits. (Cox et al 2002; Cowx et al. 2010). Recreational fisheries provide substantial social utility -i.e. satisfaction obtained by fishers from simply fishing, catching fish, experiencing natu re, etc. (Arlinghaus and Mehner 2005), and support economies through expenditures fishers make attaining this satisfaction (Props t and Gavrilis 1987; Cox and Walters 2002; Ihde et al. 2011). Accordingly, these socioeconomic effects are often considered directly related to aggregate fishing effort (Cox et al. 2002). Of course, aggregate recreational fishing effort also affects fish populations, primarily through harvest or discards mortality (Lewin et al. 2006). Intensive fishing effort can lead to fish population decline or even collapse (Post et al. 2002; Post et al. 2008), and in turn, a reduction in social utility (Cox et al. 2 003; van Poorten et al. 2011). Recreational fishing effort is largely unrestricted in the mostly open access systems of North America (Cox et al. 2003), where ag gregate effort is non constant and dynamic (Post et al. 2008; Allen et al. 201 3 ). Understandin g the dynamics of fishing effort is critical to producing quantitative predictions of fisheries system outcomes under alternative management scenarios, and critical to well managed recreational fisheries (Cox and Walters 2002; Walters and Martell 2004).
61 Factors influencing (e.g., catch rate, access) effort allocation of individual anglers are relatively well studied and theoretically should lead to well defined effort abundance relationships (Cox et al. 2003; Arlinghaus 2006), but phenomenological studies evaluating the observed strength of this relationship are less common (Fenichel et al. 2013). Angling effort is often presumed proportional to fish abundance, analogous to predator prey dynamics (Johnson and Carpenter 1994). This can be expected to prod uce a stabilizing effect on fish populations, as falling fish abundance will lead to decreased effort and presumably eventual recovery of fish (Allen et al. 2013). However, recreational fisheries differ foundationally from this example in that fishing is usually considered a leisure activity that need return neither physical sustenance nor economic benefit, but instead must provide satisfaction (Cox et al. 2003). The derivation of satisfaction from fishing is likely heterogeneous due to different motivati ons for fishing (Fedler and Ditton 1994), such as harvesting fish, catching many fish or large sized (trophy) fish, crowding of anglers, and aesthetics of the fishing site, as well as perceived opportunity costs of fishing (Arlinghaus 2006; Johnson et al. 2010; Garcia Asorey et al. 2011). Anglers may also target multiple species of fish, may switch target species, and may themselves increase in population or change either their motivation (Arlinghaus 2006). These complexities, combined with little effort restriction in North America, allow fishing effort to be potentially de coupled from fish abundance. Angling effort that responded instead to some exogenous factor (e.g. time, human population) would not be expected to have any stabilizing effect on fish populations and could lead more quickly to fish population decline absent protective management. Despite the potential for decoupling, some studies have shown apparently strong relationships between effort
62 and fish abundance (Johnson and Carpenter 1994; C arpenter et al. 1994; Cox et al. 2002; Post et al. 2008), even suggesting that recreational fisheries may self regula te (Pereira and Hansen 2003). Coupled angler effort responses to fish abundance are now largely assumed, and included in many modeling sim ulations studying recreational fisheries (Johnson et al. 2010; van Poorten et al. 2011; Allen et al. 2013). Interestingly, there have been remarkably few studies that have actually evaluated if aggregate angler effort is well predicted by fish abundance (or the change in fish abundance or the change in fish abundance over time). Nearly all studies empirically evaluating effort abundance relationships have focused on discrete inland and often northern lakes, most of which are dominated by a single target species. Further, few (Johnson and Carpenter 1994; Carpenter et al. 1994) assessed a time series of a given fishery over time the others consider multiple discrete fisheries (e.g., different lakes) over a short temporal window and so emphasize angler choi ce characteristics. Evaluations of time dependent effort abundance relationships are particularly rare in complex recreational f isheries, especially in larger , m ulti species marine fisheries. The lack of empirical evaluations hinders assessments of poten tial recreational fisheries socioeconomic and ecological outcomes, since it is difficult to quantitatively predict changes in fishing effort (upon which most outcomes partially depend) without knowing how tightly effort is coupled to fish abundance (or oth er metrics). To better understand the extent of coupling between angling effort and fish population abundance, and strength of relationships between angler effort and fish abundance, I assessed observed angler effort dynamics for in a multispecies marine fishery. The
63 specific objectives of my work were to (1) assess empirical support for the hypothesis that fishing effort is a function of fish abundance and (2) evaluate additional variables that could predict fishing effort. Taken together, these objecti ves provide insight into not only the shape (i.e., slope) of the relationship between effort and abundance, but also how important this relationship is for predicting effort, relative to other potential predictors. I considered two null and research hypot heses corresponding to the study objectives. I first tested the null hypothesis that angling effort was not related to fish abundance, with my research hypothesis being that abundance predicted effort well. To consider my second objective, I tested the n ull hypothesis that additional variables did not improve predictions of angling effort, with my research hypothesis that these variables were related and co uld be used to predict effort. M ethods Study S ystem Hypotheses were evaluated using data collected o which is among the most popular and socioeconomically important systems worldwide. Within Florida , the fishery includes popular freshwater, marine offshore, and marine insho re fisheries which combined create over $6 billi on in market activity (ASA 2001). The most popular component of this system is the marine inshore fishery which has attracted on average more than 14 million trips annually since 2000 (NMFS 2013). These trips are almost certainly made by anglers with var ying motivations, forming different typologies . Typologies may differ by target specific species, or even techniques and motivations for the same species (Camp et al. 2013). This complex setting is ideal for this study in that it provides a challenging t est of whether effort -
64 abundance relationships, which are theoretically reasonable and enjoy some empirical support in simple systems, can be observed in data from complex fisheries. Data Data used for this study were estimated annual catchable fish abund ance (i.e. the number of fish vulnerable to recreational anglers ) (individuals) and estimated targeted angling effort (trips) as reported in stock assessments completed by the state agency, the Florida Fish and Wildlife Conservation Commission. These esti mates are particularly appropriate to use, because they represent the best available information used to inform regulatory decisions. This study focuses on four popular inshore species for which recent sto ck assessments were available. The species includ ed were Red Drum , Sciaenops occellatus (Murphy and Muyandorero 2008), S potted S eatrout, Cynoscion nebulosus (Murphy et al. 2011), C ommon S nook, Centropmus undecimalis (Muller and Taylor 2012), and P ompano, Trachinotus carolinus (Murphy et al. 2008). Estim ates of both effort and abundance for these species were available as a time series of at least 25 continuous years between 1981 and 2010, though not every year was available for each species. To evaluated additional variables useful for predicting aggre gate effort, I investigated time series of local economic measures: Florida Gross Domestic Product (FL GDP) (US EIA 2010), Florida Unemployment (BLS 2013), Florida real gas prices, in 2005 dollars (Harter 2011), measures of human density Florida populatio n (EDR 2013) and annual visitors (FL DOT 2013) to the state, a measure of national economy Consumer Sentiment Index (CSI) (Thompson Rueters 2013), and weather metrics annual Florida rainfall (NOAA 2013) and annual Accumulated Cyclone Energy (ACE) in
65 the At lantic basin (NOAA 2013). These data revealed various time series trends, for which multi c olinearity might be expected. Approach I constructed linear mixed effect models to assess both research objectives. Due to the time series nature of the data, and initial tests for homoscedasticity of variance (Durbin Watson tests), I explicitly assessed the autocorrelation using AutoRegressive Integrated Moving Average (ARIMA) models. I estimated best fit ARIMA models (estimating lag[ p ], differencing[ d ] and moving average[ q ]) for each species by coast combination, using the auto.arima function (Package forecast, Program R). These results were used to determine if and what autoregressive structures should be incorporated in the model. I then incorporated appropria te autoregressive struct ure into mixed effect models. Linear Mixed M odels with AR1 L ag The linear models considered relationships between effort and abundance for each species as completely independent, but an alternative assumption is that species effects are actually related. This can be accomplished by considering the species:coasts groups as random effects, i.e. each species:coast effect is theoretically drawn from a normal distribution, and fit using linear mixed effect models (Pin h e i ro and Bates 2000 ). Such an approach has an advantage over simple linear models in that it relaxes assumptions of independence of responses of different species and allows the inferences to be extended beyond the specific groups (in this case species:coasts) that were te sted. Effort i j(i) j(i) j(i) i j(i) j(i)
66 Âµ = (Âµ a , Âµ b ) (E q 3 1) i 2 ) , w here z(abun), z(lag1_effort) is the z transformed (i.e. resc aled to a standard normal) estimated fish abundance and lagged effort, and i are normally distributed residuals, and variance covariance matrix associated w ith variance of rando m effects . The model ( Equation 1 ) was run with the lmer function from th e lme4 package in Program R, and was fit with a restricted maximum likelihood (Kery 2010). I compared the parsimony of this mixed effects approach to the lagged linear model via AICc . While R 2 is not directly assessable for non Ordinary Least Square (OLS) models, goodness of fit was assessed by calculating the R 2 value of a linear model regressing observed data (effort) a gainst the model predictions. Adherence to the assumption of n ormality was assessed via a Shapiro Wilk test. To evaluate if evidence of first order autocorrelation remained, model residuals were regressed against first order lagged model residuals and the latter evaluated for sig nificance as a predictor. As a furth er evaluation of the strength of this relationship, I contrasted the above model with one representing an autonomous effort function (i.e. effort was unrelated to abundance), by simply replacing z transformed abundance with z transformed year and excluding any information of fish abundance. Linear Difference M odels To further assess the strength and potential causality of the relationship between angling effort and fish abundance, I constructed a linear difference model regressing the
67 differences in annua l effort year to year against the corresponding differences in z transformed fish abundance. I designed candidate difference models using the various species:coast groupings, and selected the best fit model based on AIC scores. I evaluated the selected m odel in terms of overall significance, significant of parameters, and goodness of fit as measured by R 2 conformity to assumptions. Principle Component Analysis For objective two I assessed how effort might be bes t modeled by additional variables representing economic, human, and weather components which theoretically could alter the effort responses. While each of these variables might reasonably be expected to relate to annual fishing effort, many of them are co rrelated, and so including them all in any model would likely produce problems of multi colinearity. To address this, I conducted a principle component analysis (PCA) on all of the variables, and use d the reco vered components as regressors in the model. I then compared alternative linear mixed effect models and selected a model based on AIC scores, goodness of fit, and reasonable parameter estimates. Model a ssumptions regarding normality we assessed as described above. R esults ARIMA M odels Fishing effort was autocorrelated over time. The ARIMA fitting procedure showed that for most species:coast groups, a first order autoregressive error structure, with no differencing or moving average component (p=1, d=0, q=0) was best ( Table 3 1 ). While there was evidence that alternative structure might be better for certain groups, I made a simplifying assumption to include first order autoregressive error structure (AR1). The
68 ARIMA models also suggested directionally similar coefficient s for fish abundance as those from the simple linear models. Overall, these results suggest at the inclusion of a lagged effort term (i.e. effort from last year) would be appropriate. Simple AR1 Linear Mixed Effect M odels The linear mixed effect model p roduced a well fit model (R 2 of 0.901) to the z transformed abundance data ( Figure 3 1 ). While there was little evidence of first order autoregressive error structure in the residuals (lagged residual p val 0.063), the residuals were non normal (Shapiro Wilk W of W = 0.9709, p value < 0.001). Visual inspection of the residuals suggested that this model did not account for some of the outlying residuals, leading to the departure from normality (i.e. data were over dispersed with r espect to the assumed normal error structure). Overall, fishing effort was positively related to z random effects differed and had broad confidence intervals ( Figure 3 2 ). Interestingly, the results of the linear mixed effect model using year instead of fish abundance ( Figure 3 1 , right panel) were quite similar to the model using fish abundance in terms of the R 2 values (0.89267 vs. 0.901 44, respectively), though AIC was lower (suggesting a more parsimonious model) for the model using fish abundance (5470.704 vs 5446.195, respectively). Linear Difference M odels Despite high R 2 values for models linking fishing effort to fish abundance, the re is evidence such linkages are simply correlative and not causal. The results of the linear difference models suggested that there were weak relationships between differences in angling effort and z transformed abundance. Using AICc scores, the simples t difference model was selected, which included z transformed differences as the only
69 regressors . However, this model was only marginally significant (F statistic 5.296 on DF 1,198 , p value 0.022). Additionally the adjusted R 2 value was quite low (0.021), revealing a generally poor model fit ( shown in Figure 3 3 ) and the overall model marginally significant. The residuals departed from normality (Shaprio Wilk W = 0.959, p value = 1.554e 05), though there was no evidence of autoc orrelation (Durbin Watson DW = 2.5629, p value = 0.9999). Essentially, these results show that changes in fish abundance may only weakly predict angling effort on a year to year time scale, even if the trends in absolute values for abundance and effort ar e closely correlated . Principle Component Analysis The PCA on all eight potential covariates explained 83.6% of variance with three components ( Table 3 2 ), and these three components were retained for use in the linear mixed ef fect model. The first component had highest loadings for FL GDP (0.529), FL population (0.520), and total visits (0.526), and relatively lower loadings for all other variables. Accordingly, this component can be understood as characterizing both the stat e level economy and human density. The second axis was primarily loaded from national consumer sentiment (0.638), and the third axis was largely loaded via rainfall (0.864) and to a lesser extent unemployment (0.309 ). PCA AR1 Linear Mixed Effect M odels Au toregressive linear mixed effect models for which principle components were considered yielded results similar to simpler models. The three principle components were included in a linear mixed effect model, and the results were compared to an alternative linear mixed effect model containing only the first and third components. While the model containing all three components had a lower AIC score (5377 versus 5396, delta AIC =19), the second component had a very low t value ( 0.600), and as
70 such was consid ered unimportant. Accordingly, I selected the reduced model (Equation 3 2), Effort i j(i) j(i) j(i) j(i) j(i) i ( eq 3 2 ) This model assumed identical specification of random effects as Equation 3 1 , and retaining only principle components 1 and 3. This model produced a good (R 2 of 0.91) fit similar to simple AR1 mixed effect models. The fixed effects of z transformed abundance, z transformed, lagged effort, and principle component 1 were meaningful ( Figure 3 4 ), and estimates of random effects similar to the s imple AR1 mixed effect model ( Figure 3 2 ). Residuals of this model were still non normal (Shaprio Wilk W = 0.980, p value = 0.007), though regressions of residuals shows no AR1 autocorrelation. This suggests that while there is some statistical evidence for including additional PCA components in a model, such inclusion does not substantially alter the conclusions drawn from other, more simple analyses that there are correlative but not necessarily causal linkages between fish abu ndance and angling effort. D iscussion My analysis of several of inshore fisheries indicates a significant and positive relationship between measured an gling effort and fish abundance . However, the comparably good fits produced with alternative variables (e.g., year) coupled with the poor fit of difference models supports correlative more than causal relationships. The positive correlation found in this study is similar to previous work (Johnson et al. 1994; Cox et al. 2002; Post et al . 2008), but the inference of weakness causality is largely novel and similar to only Loomis and Fix (19 98). The apparent lack of causal relationships between effort and abundance has serious implications . If angler effort is in fact decoupled from fis h abundance, preserving
71 desired levels of fish abundance or angler catch rates will likely require active effort control. Such notions are not new (Cox et al. 2003), but take on added importance in fisheries such as those studied, where effort (and human populations) continue to increase and the market activity this increasing effort brings is so highly valued . It is also possible that key differences exist between this study and previous work indicating more causally linked effort abundance relationships. Effort abundance relationship may be more significant across space than across time, as found by Loomis and Fix (1998), perhaps because attributes of effort (e.g., angler motivations, satisfaction functions and typologies) should be constant at a given t ime over time . Attributes such as acceptable catch rates, economic variables or the number and demographics of potential anglers will likely change over time, and this could obscure strong effort abundance linkages that may have existed t emporarily . Additionally, while previous studies have generally focused on single species fisheries, the multi species nature of the fisheries studied here may allow anglers to substitute one species for another, further obscuring effort abundance linkage s. The existence of multiple angler typologies targeting the same fishery, as is expected for Red Drum (Camp et al. 2013 ), or prominent non catch related motivations (Beardmore et al. 2011) could certainly de couple the effort abundance relationship. Add itionally, the closed system nature of previous studies (Carpenter et al 1994; Johnson and Carpenter 1994, etc.) may have allowed differences in populations to be more clearly recognized by fishers, particularly for systems where annually stocked fish comp rised a substantial component of populations (Cox et al. 2002; Post et al. 2008). Distinguishing among these alternative explanations is difficult because it would require information about the
72 mechanisms driving effort, something largely absent in Florid a ( Camp et al. 2013 ) and scarce elsewhere (Schumann 1998; Johnson et al. 2010; Arlinghaus et al. 2013). A simple explanation of weak effort abundance relationships is that other, additional variables influence effort. My study supports this, as indicated by the significant of economic/human population and weather components in the analysis. Though the few studies looking at effort abundance time series did not consider these variables, site choice models indicate costs associated with fishing are importan t to anglers (Hunt 2005; Arlinghaus 2006; Johnston et al. 2010) It is also possible that variables that drive effort were not included in my best fit model predicting effort. For example, facilities and access points are strongly and positively related t o fishing effort on discrete water bodies (Hunt 2005; Post et al. 2008), but this information was not available corresponding to the years of this study. The apparently good fit of the PCA mixed models could be misleading. Instead, it is possible that t he inclusion of multiple continuous variables as regressors in this the regressions implicitly assume each data point is true and measured without observation error, but the re is substantial observation error associated with both the estimates of effort and that of the regressors, particularly the estimates of fish abundance. A further littl e contrast, and show a noisy, consistent trend over time. Such data are common in fisheries where regulations often aim to maintain a certain level of biomass, and often prove difficult to interpret (Walters and Martell 2004). The lack of contrast, in
73 co mbination with unaccounted for process error in the data are likely responsible for instability encountered parameterizing the models. The shortcomings of these data and the resultant failure to recover evidence of a causal effort abundance relationships is an important insight. Integrating ecological/fisheries and socioeconomic models hinges upon angler effort dynamics, and often the effort abundance relationship (Arlinghaus et al. 2013). These findings suggest such models could reasonably assume posit ive relationships between effort and abundance, but that the high uncertainty in this relationship must be considered. It may then be best to avoid making strong absolute predictions of effort in response to book outcomes assuming more and less responsive effort abundance relationships, as well as exploring the assumption that effort will increase autonomously with time. Alternatively, assumptions of strong effort abundance relat ionships should be nuanced with inclusion of additional regressors, some of which are more predictable (human population) than others (economic trends). Models that explicitly consider the uncertainty apparent in this relationship will likely produce more realistic predictions of outcomes, and prove more useful for making management decisions in the context uninformative data. If the available data are not sufficiently informative, alternative data may be needed to draw more specific inferences of angler effort dynamics. While large scale manipulations of fish abundance are usually prohibitively expensive, they can be conducted via stock enhancement (Blankenship and Leber 1995). Careful planning of adaptive management and use of natural experiments may provide similarly useful
74 information as a decreased social and financial cost (Lorenzen et al. 2010). In Florida and elsewhere, marine recreational stock enhancement is increasingly considered. Whether or not such enhancement is the optimal management st rategy, enhancement could be designed in the context of an adaptive management experiment to maximize learning, especially regarding angler effort dynamics (Camp et al. 2013). Planned stocking and monitoring of fish populations and angling effort in areas stocked and unstocked should allow more direct testing of angler effort relationships, which would be very useful for predicting outcomes to future stocking scenarios (Aske y et al. 2013). A different approach involves using stochastic environmental event s (e.g., red tide, cold kills, etc.) as natural experiments to evaluate (1) if angling effort responds to decreased abundance and (2) how effort potentially redistributes among alternative fisheries. Finally, integrating adaptively designed enhancement an d natural experiments with human dimensions studies designed to assess angler motivations may provide key information as to why certain fisheries may have stronger or weaker relationships between effort and abundance (Fenichel et al. 2013; Camp et al. 2013 ).
75 Table 3 1 . Results of ARIMA models for each time series of each species for each coast, with coefficients and the standard errors associated with fish abundance reported. ARIMA structure refers to the estimated autoregressive structure (p), differenc e (d) and moving average (q). Spec:coast Red:Gulf Red:Atl Cosn:Gulf Cost:Atl Pomp:Gulf Pomp:Atl SST:Gulf SST:Atl Coef (s.e.) 0.521 (0.073) 0.440 (0.074) 0.217 (0.230) 0.471 (0.362) 0.255 (0.072) 0.322 (0.057) 0.007 (0.015) 0.018 (0.0310 ARIMA (p,d,q ) (0,0,1) (1,0,0) (0,1,0) (1,1,0) (0,1,0) (1,0,0) (2,1,2) (1,0,0)
76 Table 3 2. Total standard deviations, proportion of variance, and cumulative proportion of variance associated with principle component analyses (PCA), where each column refers to the v ariance explained by varying numbers of axis (1 8). PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 St. Dev 1.8491 1.4673 1.0575 0.8067 0.67513 0.42648 0.14248 0.02885 Prop of Var 0.4274 0.2691 0.1398 0.08135 0.05697 0.02274 0.00254 0.0001 Cum. Prop. 0.4274 0.6965 0. 8363 0.91765 0.97462 0.99736 0.9999 1
77 A B Figure 3 1. Effort predicted from a first order one year lagged mixed effect model . A) Observed effort (open points, color = species) and predicted (solid points, color = species) using standard normalize d (z transformed) fish abundance. B) Observed effort (open points, color = species) and predictions (solid points, color = species) using standard normalized calendar year .
78 Figure 3 2. Estimates of the relationship between effort and abundance. Cond itional random effect estimates (points) and confidence intervals (lines) on the coefficient of the relationship between effort and z transformed fish abundance, for each species:coast combination, where sst = spotted seatrout, red = Red Drum , pomp = pompa no, cosn = common snook, gul = Gulf coast and atl = Atlantic coast. Grey colored points and lines show estimates from a simple mixed effect model without additional covariates, and black colored points and lines show estimates from a model including princ iple components that accounted from human population/visitor/economic and weather effects.
79 Figure 3 3 . Evaluation of the relationship between changes in estimated fish abundance and estimated effort. Open circles represent observed effort (color = species targeted), and solid points represented effort predicted from standard normalized (z transformed) differences in abundance (color = species targeted).
80 Figure 3 4. Estimates of the relationship between effort and principle component regress ors. Fixed effect estimates (points) and confidence intervals for a linear mixed effect model including principle components that accounted from human population/visitor/GDP (pc1) and weather effects (pc3), as well as standard normalized (z transformed) e ffort from the preceding year (z(lageff1)), and standard normalized fish abundance (z(abun)), allowing for different intercepts for each species.
81 CHAPTER 4 STOCK ENHANCEMENT TO ADDRESS MULTIPLE RECREATIONAL FISHERIES OBJECTIVES: AN INTEGRATED MODEL APPLI ED TO RED DRUM Sciaenops O ccelatus IN FLORIDA A primary goal of fisheries scientists is to evaluate what outcomes are possible under various management decisions and to identify key uncertainties that affect the consequences of policy decisions (Walters an d Martell 2004). Assessing outcomes of well as those common to most natural resources (Cowx et al. 2010). Recreational fisheries are increasingly recognized for both their ecological consequences and socioe conomic importance (Post et al. 2002; Lewin et al. 2006; Fenichel et al. 2013). The economic activity (e.g. revenue and jobs) related to recreational fisheries is substantial in many areas (Arlinghaus and Cooke 2009 ) and is widely considered directly related to aggregate numb er of fishing trips (Cox et al. 2003), since it results from expenditures that fishers incur attempting to attain some social utility from fishing (Propst and Gavrilis 1987; Edwards 1991; Weithma nn 1999). Anglers may attain this utility (consumer surplus) in different ways owing to heterogeneous motivations for fishing, including enjoyment of nature, solitude and most notably, high catch rates (Hunt 2005; Arlinghaus 2006; Johnston et al. 2010). T ogether, objectives of increased economic activity and stakeholder utility potentially conflict in the short term with ecological conservati on oriented objectives (Hilborn 2007; Arlinghaus and Cooke 2009), since high fishing effort and/or catch rates poten tially dep lete fish abundance (Cox et al. 2003), sometimes leading to collapsed fish populations (Post et al. 2008). To deal with such potentially conflicting objectives, alternatives to traditional management tools (e.g.,
82 effort or harvest control) are so ught, includ ing stock enhancement (Lorenzen 2005; Naish et al. 2007; Camp et al. 2013). Stock enhancement, defined as regular release of hatchery raised fish to augment existing, naturally recr uiting populations (Bell et al. 2008; Lorenzen et al. 2012), is sometimes considered a means to improve or maintain socioeconomic objectives (aggregate effort or catch rates) without depletin g fish populations (Camp et al. 2013). Management via enhancement is ingrained in some fisheries (Engstrom Heg 1971; Halverson 2 008), perhaps due to its apparent popularity with stakeholders (McEacheron and Daniels 1995; Walters and Martell 2004), and its related entrenchment in management (Washington and K ozoil 1993; Naish et al. 2007). However, enhancements incur substantial fina ncial investment and operating costs for collecting or maintaining brood stock, raising and stocking out juveniles (Hilborn 1998; Lorenzen et al. 2012). In the U.S., where enhancements are often carried out by state or federal fisheries agencies, such prog rams can comprise a substantial share of agency budgets (Johnson and Martinez 2000). In addition to financial costs, enhancement may have unintended and potentially adverse consequences to wild fish populations (Washington and Kozoil 1993; Cowx 1994; Naish e et al. 2007). Stocked fish may negatively interact with wild fish though predation or competition, which at the population level may lead to partial replacement of wild by hatchery raised fish (Kennedy and Strange 1986; Petrosky and Bjornn 1988; Washingt on and Kozoil 1993; Lorenzen 2005). Widespread concern also exists over potential deleterious effects of stocking on the genetic structure, diversity and fitness of fish populati ons (Araki et al. 2007; Tringali et al. 2008; Lorenzen et al. 2012). It is als o possible that enhancement
83 can lead to increases in fishing effort which in turn may exert additional pressure on the wild po pulation component (Baer et al. 2007). Many studies have considered enhanced recreational fisheries, but very few of those studies have actually evaluated how enhancement could affect these fisheries economic, social and ecological outcomes (Larkin 1974; Cowx 1994; Camp et al. 2013). Among studies that have investigated broader outcomes, most studies actually consider entirely cultur e based fisheries (where no stock of the same species exists) rather than enhancements, or are generally focused on ecological consequences (Flecker and Townsend 1994; Ham and Pearsons 2001); though some consider economic effects (Loomis and Fix 1998; John son and Martinez 2000). The paucity of quantitative assessments of enhancement in an integrated socioeconomic and ecological framework provides little information to decision makers (Rogers et al. 2010; van Poorten et al , 2011). This work aims to provid e a framework for evaluating potential recreational stock enhancement in terms of multiple and sometimes conflicting management goals. This framework was applied to describe a range of outcomes possible from a specific enhancement in order to provide pract ically useful information to decision makers. Here the potential enhancement of red drum Sciaenops ocellatus recreational fishery was considered. Likely outcomes of this enhancement were assessed in terms of four response metri cs conservation of wild fish, aggregate effort, social satisfaction (i.e. utility) per trip, and total socioeconomic value of the fishery and outcomes were evaluated with respect to uncertain attributes, such as angler effort dynamics and relationships bet ween catch rates and satisfaction . The costs (in terms of
84 resources and finances) of achieving proposed enhancement objectives were also assessed. Finally, the applicability of these findings to other fisheries is discussed, specifically with respect to m onitoring and experimental management. M ethods Study S ystem The recreational S. ocellatus fishery in Florida is extremely popular, with generally more angler trips targeting this species than any other marine species in the state in recent years (NMFS 2013 ). This high aggregate effort translates into substantial rine recreational fishery (NMFS S. ocellatus fishery neither overfished nor to be undergoing overfishing, fishing effort is increasing and there is concern that the fishing mortality (from both harvest and mortality from discard) may and Muyandorero 2008). In the management of this fishery, escapement is used as a proxy for spawning potential ratio, and is defined as the ratio of S. ocellatus survivorship at 5 years of age under current conditions to that expected under unfished conditions. While ove rfishing is a conservation concern (both ecologically and for long term socioeconomic value), the current popularity of this fishery causes any effort restriction to have a (perceived) high economic and social cost, in terms of potentially lost fishing tri ps and stakeholder satisfaction. In this context, enhancement is currently being considered as a means to avoid socioeconomically costly effort restriction without further depleting the wild fish population(s). Such enhancements would initially likely be r elegated to relatively small discrete areas (e.g. Tampa Bay, a large estuary popular with recreational anglers). In a recent pilot study, enhancement goals were stated as
85 increasing S. ocellatus abundance in Tampa Bay by 25% (Tringali et al. 2008) though i t is likely smaller increases would be satisfactory. Integrated Quantitative M odel The potential enhancement of this system was modelled by integrating fish population dynamics (represented with a biological sub model), socioeconomic dynamics (represented with a sub model including satisfaction, aggregate effort and value calculations) and management actions (represented by simulating different stocking scenarios). All parameters and equations used for the integrated model are fully specified in Tables I a nd II, respectively, to allow replication of the modelling approach. The sub components of the model and tuning procedure are described below and reference the full model description in Tables 4 1 and 4 2 . The general approach followed was to first represent the recreational S. ocellatus fishery in Florida, requiring an integrated model. This model was then scaled to a specific region, Tampa Bay. This allowed the assessment of outcomes of various stock enhancement scenarios, in terms of conservation and socioeconomic oriented metrics. Fish P opulation M odel The biological sub model was constructed as a discrete annual time step, age structured, numbe rs dynamic population model (Equation 4 1 throug h 4 6, 30 36, Table 4 1 ), similar to those commonly employed in fisheries assessments (Walters and Martell, 2004; Haddon, 2010). The model was then extended to allow analysis of fisheries enhancements following the approach of Lorenzen (2005). The enhanced stock is differentiated into components according to genotype and origin ( Figure 4 1 ). The three components of the total stock (subscript t ) considered are wild (wild genotype, naturally recruited, s ubscript w ), hatchery (hatchery genotype, naturally recruited,
86 subscript h ) and stocked (hatchery genotype, stocked, subscript s ). This differentiation allows a range of questions to be addressed, including the contributions of stocking and natural recruit ment to yield. It is expected that sub stocks may differ in life history traits such as survival (Stunz and Minello 2001). In this study, stocked fish are assumed to have slightly lower survival relative to wild fish during the stage 2, density dependent mortality part of recruitment as described by ( Table 4 1 and Equation 4 12 and 4 13, Table 4 2 ). It is assumed that post recruitment, stocked, wild and hatchery fish experience the same age dependent survival ( S a , Eqequation 4 5, Table 4 2 ). Interactions between wild, stocked, and hatchery fish are limited to these fitness modified survival rates and density modified survival in the pre recruit phase of the life cycle, i.e. no explicit predation or competition between sub population components is assumed, such that the components are impacted symmetrically by density de pendent processes (Lorenzen 2005). Once released, stocked hatchery fish and their offspring are subject to natural selection which can be expect ed to result in the fitness of hatchery type fish increasing over generations in the wild and eventually approaching the fitness of wild fish ( Figure 4 1 ). The model mimics this effect of natural selection by allowing some hatche ry type fish to transition into the wild type component at a rate equivalent to the heritability of fitness traits ( , Table 4 1 ) (Lorenzen 2005). R ecruitment In line with the normal convention used in fisheries models, recruitment is defined as the number of late juveniles entering the fishable stock following a period of highly density dependent (compensatory) mortality. Explicit consideration of the processes that affect mortality rates during the juvenile (pre recruit) stages are critically important to outcomes of enhanced fisheries, since most fish are stocked at a life stage and si ze
87 when survival is density depende nt and size dependent (Lorenzen 1996; Lorenzen 2000; Lorenzen 2005; Hazlerigg et al . 2012). Both density and size dependence in life, pre recruitment mortali ty into multiple stages ( Lorenzen 2005). This allowed the representation of stocked fish experiencing some density dependence in survival during the pre recruit period following release (such that the amount of density dependent survival depended on the si ze of stocking), and importantly allows consideration of stocked, hatchery and wild fish in the same recruitment process, but with modified survival for each (since it may be that stocked fish experience greater mortality that wild fish). Methods for accou nting for specific components of juvenile mortality prior to recruitment to sub adult stages are described in detail in Lorenzen (2005), and so are summarized in text and included tables (Equation 4 7 through 4 21, Table 4 2 ). It is assumed that the entire early life history period, from eggs to recruits (sub adults), is composed of two stages ( Figure 4 1 ). Stage 1 represents roughly represents the larval life history stage, from hatching until settlem ent, where mortality is assumed density independent, following Lorenzen (2005). This assumption is particularly reasonable given the offshore, pelagic spawning behaviour of S. ocellatus (Murphy and Muyandorero 2009). Stage 2 represents the juvenile life history stage, from settlement until recruitment to sub adults, and mortality in this stage is assumed density dependent. For the purpose of modelling the stocking of juvenile fish, this juvenile stage is further sub divided into two consecutive phases of density dependent mortality, below and above the size at stocking. The first phase accounts for density dependent survival of wild fish ( w ) and hatchery offspring ( h ) from size at settlement ( L 0 , Table 4 1 ) until size at whic h stocking
88 occurs ( L s ), and second phase accounts for density dependent survival of wild ( w ), hatchery ( h ) and stocked ( s ) fish from size at which stocking occurs ( L s , Table 4 1 ) until size at recruitment to sub adult stages ( L r , Table 4 1 ), i.e. the cessation of density dependent mortality. Overall base survival rates for both phases are size and growth de pendent (Equation 4 9 and 4 10, Table 4 2 ), but are then modified by relative fitness o f stocked and hatchery fish (Equation 4 12 and 4 13, Table 4 2 ) and by the dens ities of fish in each phase (Equation 4 14, Table 4 2 ). Survival is dependent on the combined dens ity of wild, stocked and hatchery fish and density dependence is assumed to act symmetrically on all stock components. Thus total recruits from each wild, hatchery, and stocked sub population to sub adult life stages are represented by two Beverton Holt st ock recruit functions (Walters and Korman 1999) the first accounts for stage 1 (density independent mortality) and phase 1 of stage 2 (density dependent mor tality) ( Figure 4 1 and Equation 4 16 through 4 18, Table 4 2 ), while the second accounts for phase 2 of stage 2, during which mortality is density dependent ( Figure 4 1 and Equation 4 19 through 4 21, Table 4 2 ). In this case, it is assumed that S. ocellatus recruit to sub adults at 0.75 years, so for convenience of accounting for ages at whole year increments, the numbers of fish entering the first year of life are further reduced by a back scaled, size dependent, density independent mortality r ate ( S 0.5 , Table 4 1 and Equation 4 22 through 4 24, Table 4 2 ). Angler E ffort D ynamics Aggregate fishing effort each year ( F t ) is modeled to respond to harvestable fish abundance through a log istic function (Equation 4 29, Table 4 2 ) (Walters and Martell 2004; Allen et al. 201 3 ). Varying the Table 4 1 ) of this function allows consideration of different strengths of effort abundance relationships, including a
89 strong (sharp) response, a proportional (moderate) response, and no response (flat) as shown in ( Figure 4 2 ). While a moderate effort abundance relationship is commonly thought to exist in recreational fisheries (Allen et al. 201 3 ; Johnston et al. 2010), too few empirical estimates exist, particularly for this f ishery, for it to be assumed with certainty (Camp et al. 2013). Uncertainty of this key relationship is accounted for by evaluating enhancement outcomes under alternative scenarios of strength of effort response. Socioeconomic E lements Aggregate fishing e ffort (total number of fishing trips) was considered a proxy for the economic (i.e. market) activity, as has been commonly assumed for recreational fisheries (Cox et al. 2003). This simplifying assumption is reasonable because the regional revenue generate d from a recreational fishery depends on the cost paid by fishers to fish, including variable costs (cost of gas, ice, bait, terminal tackle, boat maintenance, etc.) and probably fixed costs (boat, equipment, etc.) (Propst and Gavrilis 1987). Over the lon g run, it is reasonable to consider these total costs to be a function of number of trips (Cox et al. 2003). It is important to acknowledge that not all trips produce the same economic activity, e.g., trips of longer duration or travel distance will likely incur greater cost and thus produce greater market activity, ceteris paribus . However, for this work the simple assumption is made that the average, expected value of cost per trip did not change substantially with the advent of stocking, thus justifying the use of aggregate effort as a proxy. Total social utility (subscript t) was represented by calculating total satisfaction per trip ( , Table 4 1 ), which was calculated as the sum of two components weighted catch r ate related satisfaction ( , Table 4 1 ) and non catch rate related satisfaction
90 ( , Table 4 1 ). Non catch rate related satisfaction, ( , T able 4 1 ) represents the utility experienced by an angler simply going fishing, and so is considered a constant per trip , with a magnitude similar catch rate related satisfaction attributed to moderate catch rates. Catch rate related satisfaction ( , Table 4 1 ; Equation 4 39, Table 4 2 ) is, of course, largely a function of catch rate, per Cox et al ., (2002). Representing angler utility as these two components recognizes that (1) utilit y depends on more than simply catch rate and (2) catch rate related satisfaction is a large component of satisfaction that is most likely to vary between trips (Finn and Loomis 2001; Arlinghaus, 2006). Summing the catch rate related and non catch rate rel ated components yiel ded an overall satisfaction (Equation 4 39, Table 4 2 ). Total socioeconomic value was calculated as the product of social utility (per trip) and the total number of trips, aggregate effort (Equation 4 40, Table 4 2 ), following Cox et al ., 2002. This produces a socioeconomic metric represented in units of utility. The assumptions that effort is driven by abundance of fish, that market activity is proportional to effort, and that sa tisfaction is driven by catch rate are simplistic but useful, since they are optimistic relative to the management strategy evaluated (stock enhancement). Accordingly results from this work should be interpreted as what is possible with stock enhancement. Uncertainty in outcomes owing to these assumptions was explored with various model runs. Model T uning The model described above was tuned to represent the S. ocellatus recreational fishery of Tampa Bay Florida in terms of both parameters and output. Key li fe history parameter values (e.g., natural mortality ( M , Table 4 1 ), maximum age ( A m , Table 4 1 )
91 for this model were borrowed from the most recent S. ocellatus stock assessment (Murphy and Muyandore ro 2009) and from other studies describing Florida S. ocellatus populations (Peters and McMichael 1987; Murphy and Taylor 1990). Other biological parameters not known at this small spatial scale (specifically R 0 recruitment at unfished conditions, q catcha bility and (1 k ) voluntary release rate, Table 4 1 ), were estimated by minimizing a negative log likelihood summing log deviances based between observed predicted effort, catch, and harvest for the counties surrounding Tampa Ba y (NMFS MRIP 2013) and the model predictions. The likelihood weighted evenly the deviances from of catch, harvest, and effort. Deviances for each were calculated using a single value (the average of the last three years from MRIP data). This tuning falls s hort of fitting the model to the full time series, which was not necessary in this case since (1) the research objectives were not to perform a stock assessment and (2) much information from the stock assessment was already incorporated into the model and (3) the data available are not considered by managers to be sufficient for conducting a stock assessment at this spatial scale. However, this tuning did ensure that important but unknown parameter values (e.g. strength of dynamic effort response) used in m y model would produce at reasonable representation of a Tampa Bay fishery. More importantly, by fitting my simulation model to observed current (and thus pre stocking) conditions, I can establish a ceteris parabus baseline to which simulated effects of sto cking can be compared. Model A nalyses Three general categories of results of the simulation runs are described: an evaluation of overall outcomes in terms of multiple objectives, outcome sensitivity to uncertain parameters, and an assessment of resource c ommitments necessary to
92 achieve the stated stock enhancement objective. Each of these analyses considered at least one of five response metrics: wild spawning biomass ratio, relative to that at unfished conditions, aggregate angler effort ( F t ), angler sat isfaction ( A t ), total (stocked, hatchery and wild) vulnerable fish and overall socioeconomic value ( V t ) . For the first analysis, to evaluate expected outcomes, the response metrics considered were satisfaction per trip, fishing effort, wild spawning biomas s and total vulnerable fish over a range of number of fish stocked ( T t ) and sizes of stocked fish ( L s ) and assuming a moderate angler effort response and full satisfaction from catch rate ( Figure 4 3 ). The second analysis evaluat ed how the response metrics satisfaction, effort, and wild spawning biomass were sensitive to assumptions of effort response ( Figure 4 4 ), and then assessed overall value assuming different combinations of effort response and sat isfaction from catching fish ( Figure 4 5 ). For the latter only value is reported, because these differing assumptions of catch related satisfaction have no effect on wild spawning biomass. Response metric values reported for th e first two analyses are final values from 100 year model runs, which are considered equilibrium values. These values are given to illustrate the expected behavior of recreational enhanced systems. The final analysis considered what resources would be ne cessary to realize a given increase in total vulnerable fish in or around the Tampa Bay estuary ( Table 4 4 ) in ten years of stocking. This response depends in part on the effort response assumed, and so multiple responses are c onsidered. Here these resources are represented in terms of both numbers of fish that would need to be stocked and the approximate cost in dollars of stocking them (Young, personal communication).
93 R esults Model T uning Tuning this model to recent point es timates of the Tampa Bay, Florida S. ocellatus fishery output produced equilibrium conditions that represent the current state of the fishery, with deviances for all metrics < 6% ( Table 4 3 ). The estimate of voluntary release d iscard (73%) was surprisingly high, though not unreasonable given the 1 fish bag limit in Tampa Bay, and so for the purposes of this study (and in the absence of direct empirical estimates) is considered adequate. The catchability estimate was an order of magnitude than the current stock assessment estimate (5.1x10 6 vs 9.0x10 7 ). coast or, more likely, the fact that nuisance parameters such as catchability routinely absorb error in fisheries model, such as incorrect assumptions effort dynamics. The latter case is not particularly troublesome, because I explicitly consider how uncertainty in this assumption could affect outcomes. Model A nalyses My results suggest tha t certain scenarios of S. ocellatus stock enhancement may increase the socioeconomic objectives associated with this recreational fishery, though at a cost. This is illustrated by the increases in aggregate fishing effort (related to economic impact) and s atisfaction (driven largely by catch rates) under certain stocking scenarios, related to increases in total vulnerable fish ( Figure 4 3 A,B,D ). However, increases in the vulnerable fish and the related effort and satisfaction are only possible when fish are stocked at larger sizes. In fact, even very high number of fish stocked at small sizes failed to produce increases in vulnerable fish or associated socioeconomic metrics ( Figure 4 3 A,B,D ). One of the most important results is that any amount of
94 stocking will cause decreases in wild abundance of fish, and these decreases are exacerbated by increasing numbers of fish stocked, regardless of the size of fish stocked ( Figure 4 3 C ). This suggests that stocking small fish will cause declines in wild fish abundance without returning socioeconomic benefits. In concert, evaluations of these metrics suggest an apparent trade off between the conservation objectives of not hurting wild f ish populations and the socioeconomic objectives of continued consumer benefit and regional economic impact. Evaluations of alternative angler effort responses revealed a secondary trade off between satisfaction and effort. A flat effort response can cause the greatest increases in satisfaction (via increased catch rate), since total vulnerable fish increase but effort does not change ( Figure 4 4 A ). Conversely, sharp effort responses can actually lead to satisfaction, if effort has increased at proportionally more than vulnerable fish ( Figure 4 4 C ). Aggregate effort (and related economic impact) has the greatest potential to increase with sharply responsive effort ( Fig ure 4 4 F ), and unresponsive effort will obviously lead to no increases regardless of how many fish are stocked of any size ( Figure 4 4 D ). The broader conservation socioeconomic trade off was not generally sen sitive to alternative assumptions of effort responses. Wild fish are still likely to decline under any effort response ( Figure 4 4 G I ) indicating that the biological effects of stock enhancement on wild fish (e.g. competition in recruitment) are substantial enough to cause declines even in the absence of attracted fishing effort. However, the substantial declines in wild fish populations occur at lower stocked numbers under sharp angler effort response ( Figure 4 4 G ) relative to flat ( Figure 4 4 I ) or moderate ( Figure 4 4 H ), suggesting that the attracted effort still affects wild fish.
95 These results suggest that increases in overall socioeconomic val ues will be greatest when effort is responsive and substantial satisfaction is derived from increased catch rates ( Figure 4 5 ). However, some increases are possible as long as effort is not unresponsive while simultaneously satis faction is unrelated to catch rate ( Figure 4 5 B I ). It should be noted that the alternative assumptions of satisfaction alter the magnitude of the value, but do not affect the wild spawning biomass. Therefore the losses of wild fish would be the same as seen in Figure 4 4 , changing with the assumptions of effort response. As a whole, these results suggest that alternative assumptions regarding particularly uncertain elements of recreational fisheries do little to alleviate the trade off between increased socioeconomic objectives and sustained wild fish populations caused by stock enhancement. The resource commitments necessary to produce the stated objective (25% increase in harvestable S. ocellatus 10 years after commencement of stocking) are substantial, but driven by size at which fish are stocked ( Table 4 4 ). This work suggests that stocking quite small S. ocellatus (25mm) is incomprehensively costly (quadrillions of fish stocked and dollars spent per year), whereas stocking larger fish is at least potentially possible (700,000 1,000,000 fish stocked at a cost of < $400,000/year). Resources needed were generally greater if effort responded sharply to fish abundance, presum ably because the catch rate gains from stocking are quickly lost to growing effort requiring more stocking to attain the stated goal. D iscussion The greatest utility of this work arises from its integration of individual components populations dynamics of stocking enhancement, angler effort dynamics, and socioeconomic value to produce a quantitative model broad enough to represent
96 important conservation and socioeconomic objectives, but straightforward enough to be readily described and tuned to commonly a vailable data. This tuning is an important step forward, as it lends greater empirical credibility to predictions, which can be produced at scales appropriate for addressing specific management questions. One formidable challenge to constructing integrat ed models is describing appropriate linkages between components, such as linkages between population dynamics an d socioeconomic value (Lorenzen 2008). Two primary options exist for creating these linkages either individual site choice models (common in eco nomic literature) or broader scale models where the aggregated population effects of choices are represented (Walters and Martel 2004; Fenichel et al. 2013). The latter was chosen due to its transparency, but account for uncertainty inherent to this appro ach by (1) investigating a range of strengths of linkages between both fish populations and fishing effort as well as between catch rates and satisfaction and (2) describing likely implications of additional assumptions. This approach has yielded several specific findings that improve the socioecological understanding of stock enhancement of recreational fisheries. The most important finding of this work is the trade off realized under stock enhancement between conservation and socioeconomic objectives, wh ere greater socioeconomic value may be achieved with stocking, but at a cost to wild fish populations. Surprisingly, this loss of wild fish was largely insensitive to the strength in angler effort response. This is likely due to the short time S. ocellat us are highly vulnerable to harvest or capture (one and three years, respectively) relative to their lifespan (30 40 years). While not commonly recognized in the literature (but see van
97 Poort en et al. 2011), socioeconomic conservation trade off should be e xpected with stock enhancement (Camp et al. 2013). Stocking smaller or larger fish can lead to replacement of wild fish directly or via stocked fish offspring, and stocking larger fish that more feasibly increases overall abundance (and can lead to greate r socioeconomic value) may lead to increased overall effort and fishing mortality on wil d and stocked fish (Baer et al. 2007). This work also made explicit a secondary trade off realized under stock enhancement, between social utility and aggregate effort, related to market activity. The trade off here should also be anticipated, since higher effort will drive catch rates (affe cting utility) down (Cox et al. 2002). A key implication from the recognition of both socioeconomic conservation and effort catch r ate trade offs is the importance of explicitly describing to stakeholders the likely outcomes of stock enhancement in terms of expected costs (wild fish, aggregate fishing trips/market activity or catch rate related satisfaction). Failure to do so may forw ard the unrealistic expectation of stock enhancement as a panacea that can simultaneously buoy wild fish population, catch rates, and effort (van Poorten et al. 2011), whereas explicit recognition of the specific type of socioeconomic gains and potential w ild fish losses supports the view of stock enhancement as a potentially useful management tool, provided the costs are acceptable. More reasonable and explicit objectives regarding enhancement should promote greater success at their achievement (Washington and Kozoil 1993; Naish et al. 2007) and help avoid the primary pitfall of poorly defined management objectives (Walters 1986; Possingham et al. 2001; Martin et al. 2009).
98 This work specifies the influence on enhancement outcomes of the size of fish at st ocking, relative to the size at recruitment. Fish stocked at a size prior to most density dependent survival in the wild (e.g., S. ocellatus stocked at 25mm) were unlikely to increase the total population or lead to socioeconomic benefits. At best (worst) , the result will simply be a replacement of wild fish with stocked fish. This finding is based in tautological theory of recruitment dynamics so long as recruitment overfishing is minimal it is unlikely stocking reasonable numbers of fish prior to cessati on of strong density dependent mortality could augment tota l numbers of fish (Rogers et al. 2010; Camp et al. 2013). While this seems evident, it is not always recognized, and many hatchery programs still stock ver y small fish (McEacheron et al. 1998; Sera fy et al. 1999; Scharf 2 000; Tringali et al. 2008) that may replace wild fish if they survive through the density dependence stage following release. This understanding of the recruitment dynamics depends upon the foundational theory of sustainable fishing and has a clear implication for how enhancements are monitored and evaluated. Traditionally, percent contribution (proportion of stocked fish divided by proportion total fish recaptured) has been used as a proxy for survival of stocked fish and thus relat ive success of enhancement at augmenting the fishery. However, it is clear that such a ratio depends as much upon the numbers of wild fish available for capture as it does the survival of stocked fish. If such a metric is used, any replacement of wild by s tocked fish would likely be misunderstood as greater success of enhancement, rather than loss of a wild population and little change in the overall abundance. Preferably, monitoring would measure actual changes in fish populations following enhancement (re presented by
99 total vulnerable fish in this work), or directly measure the changes in socioeconomic objectives (e.g., effort, satisfaction). Achieving the desired 25% increase in the overall abundance of harvestable S. ocellatus is possible at an estuarine scale and would be expected to lead to greater market activity from increased trips or satisfaction related to greater catch rates, provided larger fish are stocked and the costs (in terms of wild fish and dollars) are amenable. However, it is critical t o recognize that this 25% increase in catchable fish would likely cause only slight increases in these socioeconomic metrics relative to the many fold increases described in the theoretical results (which corresponded to many fold increases in catchable ab undance), and such an increase is only possible after ten consecutive years of successful stocking. Given inherent variability in natural recruitment, angler catch rates or numbers of trips, and hatchery rearing, this should temper expectations of enhancem ent dramatically altering the fishery. The outcomes of both the theoretical and practical analyses are also subject to several additional uncertainties. The positive socioeconomic outcomes associated with stocking larger S. ocellatus may be optimistic bec ause fish were assumed to experience minimal immediate post stocking mortality (i.e. due to transit shock, etc.) and survive after recruitment as well as wild fish, and they may not (Lorenzen 2000; Melnychuk et al. 2013). Further, because this fishery is m anaged via a bag limit, if abundance and catch rate increase disproportionately greater than effort, then on average a greater proportion of captured fish will be released and subject to discard mortality, potentially dampening gains in socioeconomic valu e. Additionally, this work did not account for the potential of hatchery programs to increase risk of disease in wild fish (Washington and
100 Kozoil 1993; Naish et al. 2007; Lorenzen et al. 2012). Further, while density dependent growth effects of enhancement are possible (Baer and Brinker 2008), none were assumed in this model and such effects are not intuitive because slower growth could prolong the time period for which S. ocellatus are harvestable as well as delay their reaching harvestable size. Additiona lly, while crowding can negatively a ffect angler satisfaction (Hunt 2005; Schuhmann and Schwabe 2004), it was not assessed in this model. Of course, positive socioeconomic effects of stocking could also exist, such as educational opportuni ties at hatcherie s (Camp et al. 2013), but these are not well documented and so are not represented in this model. Finally, this work is useful for understanding both ecological and socioeconomic outcomes of enhancement in a single area in absentia of other regions, but t his represents an unrealistic abstraction. In reality, the presence of alternative fisheries or fishing sites suggests that some increase in effort and overall socioeconomic value from stocking are simply redistributions of effort from other areas (Sutton and Ditton 2005; Askey et al. 2013). Given what has not been represented in this work, the results are best interpreted as an optimistic assessment of what may be possible with enhancement and an explicit accounting of some minimum costs of achieving thi s. Despite limitations, this work has clear implications for the biological and socioeconomic conditions under which enhancement may be most successfully S. ocellatus fishery. Recreational fisheries systems where effort is responsiv e and satisfaction is driven by catch rates may experience the greatest gains in overall socioeconomic value, since stocking (at high enough rates) can potentially cause absolute increases in both effort and catch rate related satisfaction
101 (though the trad e off will limit the amount of satisfaction increase possible). Positive socioeconomic may be most noticeable if the areas stocked already have low abundances of wild fish but sufficient juvenile habitat, such as areas where wild S. ocellatus have been ess entially recruitment overfished. Alternatively, if high fishing effort causes rapid depletion of vulnerable fish, through harvest or release mortality (Cox and Walters 2002), enhancement of larger juvenile S. ocellatus could temporarily increase catch rate experienced by anglers. Due to dynamic effort, this increased catch rate would be expected to dissipate quickly as anglers are attracted from alternative areas or fisheries, resulting ultimately in little increase in catch rate but greater aggregate effor t (unless enough fish are stocked to satiate effort) (Baer et al. 2007). If effort satiation is not possible (which is likely), high catch rate enhanced fisheries can be achieved only throug h effort limitation (Cox et al. 2003). Such limitation is expected to be quite undesirable in marine fisheries, but may be entertained experimentally in smaller, semi discrete regions. A mostly unrecognized situation where enhancement may be particularly useful is in an experimental adaptive management framework designe d to address key uncertainties in socioecological systems of recreational fisheries (Walters 1986; Walters and Holling 1990). In fact, perhaps the greatest benefit from stocking may stem from cooperation between managers, researchers, and stakeholders to u nderstand better angler behaviours and the relationships linking ecological and social systems. Stocking represents a rare opportunity at large scale manipulations. If stocking are planned in a framework to answer questions (i.e. an active adaptive managem ent framework), and if stocked fish are discernible from wild and adequate monitoring is in place (Lorenzen et
102 al. 2010) stocking may be useful for reducing the very uncertainties (e.g., effort, satisfaction) considered in this and other studies. Caution h ere is warranted not all experiments may be worth doing when ecological or social costs are high (Walters and Ahrens 2009). I am grateful to P. Medley for developing the recr uitment unpacking algorithm (Equation 4 8 and 4 14) implemented in this model. Thi s work was supported by th e Integrated Graduate Education and Research Traineeship (IGERT) program in Quantitative Spatial Ecology, Evolution and Environment at the University of Florida and id in Sport Fish Restoration Project F 136 R.
103 Table 4 1. Parameters and associated values used in integrated quantitative model are described. An asterisk (*) indicates values were estimated. Symbol Description Units Value *R 0 Recruitment at unf ished conditions fish 450,371 Length at infinity mm 934 K Metabolic in growth equation yr 1 0.46 t 0 Age at length=0 yr 1 0.26 w a Weight length constant g 0.000000617 w b Weight length constant g 3.09 W m Weight at maturity kg 10.084 M Instanta neous mortality yr 1 0.113 L m Reference length mm 730 C l Allometric exponent of length mortality relationship constant 0.9 A m Maximum age years 40 Recruitment compensation parameter ratio 11 L 0 Length at entering recruitment period mm 20 L s Length at stocking mm 25 175 L r Length at leaving recruitment period mm 180 d 1 Duration of density dependent mortality recruitment phase, from size L0 to Lr years 0.75 d 2 Duration of the second stage of the density dependent mortality recruitment stage Prop ortion Calculated M 1 Natural mortality year 1 of 10mm fish year 15 S r Cumulative base survival for the recruitment period rate Calculated h Fitness (or survival) of hatchery relative to wild, stage 1 rate 1.0 s Fitness (survival) of stocked relative to wild, stage 2 rate 0.8 Share of hatchery eggs inheriting wild characteristics % 0.2 T t Number of fish stocked each year fish 0 4.5m S 0.5 Back scaled mortality to fish size midway between 0.75 yrs and 1 yrs yr 1 0.86 , Fish length for vulnerability to capture: low, high mm 400, 850 , , , F ish length for vulnerability to harvest: low, high mm 457, 686 , , *k Proportion of harvestable fish killed % 0.27 Standard deviation of logistic constant 0.05 2000
104 Table 4 1. Continued. Symbol Des cription Units Value F m , F o Minimum effort and effort an unfished stock size trips 200,000; 600,000 *q Catchability coefficient rate 0.0000051 D Discard mortality rate 0.08 A c Satisfaction from catch rate Calculated Magnitude similar to average satisfaction from catch constant 6.5 CPUE where sat cat = 0 rate 0.05 Ratio of catch to non catch related satisfaction ratio 0 1 Slope of the relationship between catch related satisfaction and CPUE constant 3
105 T able 4 2. Model compone nts and equations are described. For all equations, italicized subscript ( t ) represents time dynamics (years), italicized subscript ( a ) represents age dynamics (years), subscript ( w ) represents wild fish, subscript ( s ) represents stocked fish, and subscrip t ( h ) represents hatchery fish and subscript ( t ) represents total combined fish. Eq Component Equation Life History Characteristics of Stock: E q . 4 1 Growth: Length L at age a E q . 4 2 Size: Weight (kg) to length ratio Eq. 4 3 Fecundity ( f at age a ) E q . 4 4 Survival (year 1 ) ( S at age a ) E q . 4 5 Survivorship ( l at age a ) E q . 4 6 Eggs per recruit Unpacked Recruitment Dynamics: E q . 4 7 Beverton Holt a , b, and re parameterized b , , E q . 4 8 Duration of phase 1 of recruitment stage 2, DD E q . 4 9 Linear growth rate (year 1 ) for recruitment stage 2, V E q . 4 10 Base su rvival of recruitment phase 1 and 2, respectively of recruitment stage 2, S1, S2 , E q . 4 11 Survival rate of larvae for entire recruitment,
106 Table 4 2. Continued. Eq Component Equation E q . 4 12 Survival for phase 1, stage 2, modified by relative fitness of sub populations, a 1 , a 1 h , E q . 4 13 Survival for phase 2, stage 2, modified by relative fit ness of sub populations, , , , , E q . 4 14 Density dependent component of survival for stage 2, phase 1 and 2, respectively, b1 and b2 , E q . 4 15 Total eggs in beginning year t, which is sum of wild ( w ) and hatchery eggs ( hat ) where , Number of fish surviving stage 1 and phase 1 of stage 2: E q . 4 16 wild E q . 4 17 hatchery E q . 4 18 Total fish entering the phase 2 of rec. stage 2 Number of fish surviving stage 2 of phase 2 and thus leaving recruitment period: E q . 4 19 Wild E q . 4 20 Hatchery
107 Table 4 2. Continued. Eq Component Equation E q . 4 21 Stocked Number of fish entering age 1, modifi ed by back calculated survival from 0.75 to 1 yr of age: E q . 4 22 Wild numbers at age a E q . 4 23 Hatchery numbers at age a E q . 4 24 Stocked numbers at age a Fishery Characteristics: E q . 4 25 Vulnerability to capture at age a where E q . 4 26 upper lower , E q . 4 27 Vulnerability to harvest, which is where E q . 4 28 upper lower , E q . 4 29 Effort ( F t ) Time Dynamics: E q . 4 30 Wild Spawning Biomas s (WSB) E q . 4 31 Exploitation rate E q . 4 32 Numbers at age E q . 4 32 Numbers at age E q . 4 33 Survival of all types from: E q . 4 34 Harvest E q . 4 35 Discard of non legal catch
108 Table 4 2. Continued. Eq Component Equation E q . 4 36 Volu ntary discard of legal catch Socio economic Dynamics: E q . 4 37 Total catch E q . 4 38 Catch per un it effort (CPUE) E q . 4 39 Satisfaction per trip where E q . 4 40 Overall socioeconom ic value
109 Table 4 3. Comparisons between tuned model predictions and region specific observations from the Marine Recreational Information Program (MRIP). Asterisks (*) indicate leading estimated parameters from tuning procedure . Model was fit assuming a flat (non responsive) effort respo nse. Output MRIP for Tampa Bay region Model Percent Deviance (MRIP Model)/MRIP Effort (trips) 535,618 534,965 <0.01 Catch (numbers of fish) 827,059 828,997 < 0.01 Harvest (numbers of fish) 93,011 91,316 0.06 CPUE (catch/effort) 1.54 1.55 < 0.01 KPUE (harvest/effort) 0.17 0.17 0.01 Unfished recruitment (Ro) * NA 464,685 NA Catchability (q) * NA 5.10x10 6 NA Kept proportion * NA 0.27 NA Escapement NA 0.35 NA
110 Table 4 4. Evaluation of costs of potential stocking of red drum in Tampa Bay, Flori da. Costs are shown in terms of how many S. ocellatus would need to be stocked, and what it would cost to realize a 25 % increase in S. ocellatus abundance 10 years after commencing stocking, under different recruitment and effort responses. These estimat es include some instant mortality from stocking and slightly lower survival of stocked fish relative to wild. Cost basis of $0.15/25mm, $0.46/100mm and $1.58/post recruit, per Chris Young, unpublished data. Size at stocking Effort response sharp moder ate flat 25mm 4.6e 17 ($6.9e 16 ) 4.5e 17 ($6.9e 16 ) 4.6e 17 ($6.9e 16 ) 100mm 712,624 ($327,807) 727,565 ($334,679) 721,870 ($332,060) 175mm 189,784 ($299,860) 190,242 ($300,582) 188,537 ($297,888)
111 Figure 4 1. The structure of the model accounts sp ecifically for phenotypes, life stages and related population processes.
112 Figure 4 2. Alternative assumptions of the response of aggregate fishing effort to the abundance of all fish vulnerable to capture for recreational fishing. The solid line ( ) ) represents moderately responsive effort and the dotted line ( ) represents sharply responsive effort.
113 Figure 4 3. The expected, equilibrium model results. A) Satisfaction per trip. B) Ag gregate number of trips per year. C) Wild spawning biomass and D) Total vulnerable fish per year. All results depend upon the size of fish at stocking (y axis) and the number of fish stocked per year (x axis). Darker colors show greater values of each re sponse metric corresponding to the size and number of fish stocked, and cont our lines show actual values.
114 Figure 4 4. Evaluation of the effects of effort response on equilibrium model results. A) Satisfaction per trip assuming a flat effort respo nse. B) Satisfaction per trip assuming a moderate effort response. C) Satisfaction per trip assuming a sharp effort response. D) Total annual trips assuming a flat effort response. E) Total number of trips assuming a moderate effort response. F) Total number of trips assuming a sharp effort response. G) Wild spawning biomass assuming a flat effort response. H) Wild spawning biomass assuming a moderate effort response and I) Wild spawning biomass assuming a sharp effort response. For all panels, d ark er colors show greater values of each response metric corresponding to the size and number of fish stocked, and cont our lines show actual values.
115 Figure 4 5. Evaluation of the effects of effort response and catch related satisfaction on equilibrium model predicted socioeconomic value. A) Socioeconomic value assuming full satisfaction from catch rate and a flat effort response. B) Socioeconomic value assuming full satisfaction from catch rate and a moderate effort response. C) Socioeconomic value a ssuming full satisfaction from catch rate and a sharp effort response. D). Socioeconomic value assuming half satisfaction from catch rate and a flat effort response. E). Socioeconomic value assuming half satisfaction from catch rate and a moderate effo rt response. F) Socioeconomic value assuming half satisfaction from catch rate and a sharp effort response. G ). Socioeconomic value assuming no satisfaction from catch rate and a flat effort response. H) Socioeconomic value assuming no satisfaction fro m catch rate and a moderate effort response. I) Socioeconomic value assuming no satisfaction from catch rate and a sharp effort response. For all panels, d arker colors show greater values of each response metric corresponding to the size and number of fis h stocked, and cont our lines show actual values .
116 CHAPTER 5 SOCIOECONOMIC AND CONSERVATION TRADE OFFS IN THE STOCK ENHANCEMENT OF RECREATIONAL FISHERIES The management of most natural resources can be characterized by objectives of two primary themes long er term conservation and shorter term socioeconomic benefits (Shea 1998). Conservation objectives include valuing a resource for some intrinsic purpose (e.g., endangered species, rare habitats) or future benefit (e.g., sustained harvest, unrealized benefit s) (Cowx et al. 2010). Socioeconomic objectives can be defined as valuing the immediate use of a natural resource for the utility such use generates. Over the long run these objectives are of course complimentary (Hilborn 2007), but in the short term they often conflict (Koehn 2010). This conflict results in a present time trade off that is common to nearly all natural resource management problems (Gregory and Keeney 2002), where achieving more of one objective may come at the cost of an alternative objecti ve (Walters and Martel 2004). Selecting the optimal combination of satisfying both objectives is the primary occupation of most natural resource managers (Gregory and Keeney 2002; Walters and Ahrens 2009). To this end, management systems often employ vario us frameworks (e.g. command control, stakeholder driven), strategies (harvest limitation vs. resource enhancement), and scenarios (intensity of a strategy) to reach a Pareto efficient solution where both objectives are best satisfied (Lackey 1998; Lester e t al. 2013). To provide information are possible under given management frameworks, strategies, and scenarios (Lackey 1998; Walters and Martel 2004; McNie 2007). Perha ps the most effective way to conservation socioecological tradeoffs (Cheung and Sumalia 2008).
117 Across disciplines an increasing emphasis is placed on evaluating decisions ( especially trade offs) in the context of their alternatives (Halpern et al. 2011; Lester et al. 2013). In economics this is well developed by portfolio and production theories, which essentially compare return or production per risk or cost, revealing impl ications for various and often diversified strategies (Markowitz 1959), and nearly all bioeconomic models implicitly or explicitly consider socioeconomic conservation trade offs. This approach has been adapted by natural resources scientists describing how choices can be made concerning complex systems given knowledge of alternatives (Walters and Martel 2004; Sanchirico et al. 2008; Gaydon et al. 2012). Some have discussed on a broader, holistic scale the need to entertain multiple management paradigms and even worldviews (e.g., Ostrom 2009), and ecologists have more mechanistically adapted portfolio theory to evaluate strategies for maintaining biodiversity (Figge 2004; Schindler et al. 2010). Recently, the field of decision making has triumphed explicit pr ocesses to compare alternatives (e.g. structured decision making) (Gregory and Keeney 2002; Kiker et al. 2005). Underlying this emphasis are two related ideas. The first is the notion that not all alternative management strategies are equal, but specifical ly that some are affect conservation and socioeconomic objectives (Gregory and Keeney 2002; Walters and Martel 2004; Lester et al. 2013). The second is that science supp orting natural resource management has sometimes failed to sufficiently improve outcomes (Starfield 1997; Ludwig 2001), in large part due to inadequate communication of information useful for making decisions between alternative actions (Shea 1998; Possing ham and Shea 1999; NcNie 2007). While explicitly assessing trade off shape and nature to
118 clearly and directly illustrate potential outcomes of alternative management actions has long been recommended (e.g., Sylvia and Cai 1995; Walter and Martel 2004), suc h assessments are not common (Lester et al. 2013), particularly in recreational fisheries. This work quantitatively assesses the trade offs in recreational fisheries between conservation objectives of sustaining healthy, wild populations of fish for future use and the enjoyment and socioeconomic benefits of catching or harvesting those fish in the short term (Cowx et al. 2010). Conservation objectives may be more complex, but can be represented simply by the ratio of wild fish biomass achieved with the mana gement strategy of interest relative to wild fish biomass at unfished (and non stocked) conditions, referred to as the relative wild spawning biomass (WSB). Socioeconomic objectives are probably best represented by the total economic value (i.e. social uti lity attained per trip scaled up by the total number of trips taken) (Propst and Gavrilis 1987). While conservation objectives could certainly be achieved by limiting fishing effort, this would impinge on the total socioeconomic value and are unpopular wit h anglers and management agencies (Cox et al. 2002; Cox et al. 2003). In an attempt to protect wild populations of fish without incurring these high socioeconomic costs, alternative management strategies are increasingly being considered and used. One of t hose alternative strategies is stock enhancement: the stocking of hatchery fish to augment fish population abundance. It is often implicitly assumed that stocking of hatchery fish can maintain high fish stock abundance even under very high fishing effort a nd corresponding exploitation rates, thereby mitigating or circumventing the fundamental, short term socioeconomic/conservation tradeoff. However, the efficacy of enhancement
119 as a means to circumvent trade offs between conservation and socioeconomic goals in recreational fisheries has not been well demonstrated for recreational fisheries. Recreational stock enhancement programs can cause a suite of biological and human responses that may undermine the intended alleviation of conservation and socioeconomic tradeoffs. Hatchery rearing influences the biology of fish through developmental and genetic mechanisms and often results in fish that are less fit in natural environments than their wild conspecifics (Lorenzen et al. 2012) such that released hatchery fis h and their offspring are not, in general, fully equivalent to wild fish (Araki et al. 2008; Fraser 2008). Moreover, once released, hatchery fish may interact biologically with wild fish through competition, predation and reproduction. Through those mechan isms released hatchery fish may partially or fully displace existing wild fish (Lorenzen et al. 2005). It is, therefore, important to distinguish wild and hatchery origin population segments in the analysis of fisheries enhancements. If enhancement does au gment fish populations, socioeconomic value depends on how anglers respond to this increase since any change in fishing effort is likely to affect wild populations simultaneously. Accordingly, the ability of enhancement to address trade offs in fisheries s hould not be assumed. Additionally, in order for managers to make good decisions regarding management strategies to optimize this trade off, outcomes of enhancement should be compared to alternative strategies, the efficacy of can be assessed by evaluating the nature and shape of these trade offs. This work evaluates the shape of stock enhancement trade offs, and those of alternative strategies using the potential stocking of red drum in Florida, USA, as a case study.
120 Methods Approach The overall objectiv e of my work was to improve the understanding of the shape and nature of trade offs realized under stock enhancement, relative to alternative management strategies. I believe this information provides a more explicit understanding of the efficacy of manag ement strategies and provides managers a simple means of broadly comparing alternative strategies. While such an approach may be useful for comparing or scenarios relative to each other (e.g., where do alternative stocking scenarios fall on a trade off fr ontier, or how do frontiers of alternative strategies compare), this approach is much less useful for calculating absolute costs and benefits of a proposed management action. Accordingly, this work should not be confused with a benefit cost analysis rathe r I have constructed an integrated model that I argue represents well (if not precisely) the primary socioeconomic objective of aggregated angler satisfaction (akin to consumer surplus and directly related to socioeconomic value) as well as the primary con servation objectives of preserving wild fish populations. Construction of such broad scale, integrated models requires substantial abstraction, and I describe the implications of this in the discussion. Case Study Fishery Recreational fishing is an imp ortant social and economic activity for the state of Florida, with the marine recreational fishery alone valued at $5.7 billion annually (NMFS 2010) and fishing one of the most popular recreational activities overall (FL DEP 2011). In particular, the Red D rum is the most popular recreational fishery in terms of total targeted trips (NMFS 2013). While the Florida Red Drum population is currently
121 characterized as not overfished or undergoing overfishing, there is concern that high effort could lead to overfis hing (Murphy and Mu n yandorero 2009). While overfishing is a conservation concern (both ecologically and for long term social value), the current popularity of this fishery causes any effort restriction to have a (perceived) high cost, in terms of the poten tial for lost revenue from fewer fishing trips as well as decreased stakeholder satisfaction. In an attempt to support fishing effort while sustaining wild fish populations (i.e. circumvent the conservation socioeconomic trade off), stock enhancement is be ing considered. Integrated Quantitative Model The quantitative model integrates the dynamics of a biological sub model (fish population) with a socioeconomic sub model (driven by angler effort and satisfaction from fishing) and compares alternative managem ent outcomes, including different stocking scenarios. The biological sub model is a discrete annual time step, age structured, numbers dynamic population model as commonly employed in fisheries assessments (Walters and Martell 2004; Haddon 2010) that has b een extended to allow analysis of fisheries enhancements following the approach of Lorenzen (2005). This model explicitly accounts for three sub components of a single stock. Recruitment of fish into each sub component is described by a Beverton Holt stock recruit relationship, dependent components of survival. This allows consideration of multiple stock enhancement scenarios (i.e., number and size of fish stocked). The mo del has been tuned (calibrated to current fisheries estimates) to represent the Red Drum fishery of Tampa Bay, Florida. The biological sub model is linked to the socioeconomic sub model through dynamic angler effort, which is assumed proportional to harves table fish per
122 Allen et al. (2012). The socioeconomic sub model consists of calculations of proxies for economic effects, social welfare (utility) and socioeconomic value that are modified from Cox et al. (2003) to include catch rate related satisfaction ( utility), non catch rate related satisfaction, and satisfaction inherent from stocking. All parameters and equations used for the i ntegrated model are defined in T ables 5 1 and 5 2 , respectively. Co mponents of the model that either differ from these cited studies or require additional descriptions are detailed below. Wild, Stocked and Hatchery Fish Population Components and Interactions The model explicitly accounts for three components of a single stock: wild fish, stocked fish (those that are stocked in each year) and hatchery fish (those fish that are offspring of hatchery fish). Sub stocks may differ in life history traits such as survival. In my study, stocked fish are assumed to have a slightl y lower survival during recruitment (Stunz and Minello 2001) and may experience some immediate, post release or transport mortality (Sherwood et al. 2004) as described in ( fit st , Table 5 1 ). I assume that post recruitment, stoc ked, wild and hatchery fish experience the same survival ( S a , Table 5 2 ). All population components interact through density dependent processes in the pre recruit phase of the life cycle. I assume that these processes are depe ndent on the combined density of all three population components and that the components are impacted symmetrically by density dependent processes (Lorenzen 2005). Once released, stocked hatchery fish and their offspring are subject to natural selection wh ich can be expected to result in the fitness of hatchery type fish increasing over generations in the wild and eventually approaching the fitness of wild fish. The model mimics this effect of natural selection by allowing hatchery type fish to transition i nto the wild type
123 component at a rate equivalent to the heritability of fitness traits ( , Table 5 1 ) (Lorenzen 2005). Recruitment Recruitment processes are critically important to outcomes of enhanced fisheries since most fish are stocked at a life stage and size when survival is both density and size d ependent (Lorenzen 2005). I accounted for both density and size dependent This allows stocked fish to experience density dependence in survival during the pre recruit period following release and allows consideration of stocked, hatchery and wild fish in the same recruitment process but with modified survival for each. Specifically, the life history of sub adult fish is assumed to begin with a density independent stage lasting from hatching until settlement and then a density dependent stage from settlement until recruitment. Survival is size dependent during both stages. In the unpacking, the Beverton Holt stock recruitment function is split into two consecutive densit y dependent phases (Beverton and Holt 1957; Walters and Korman 1999) with the first stage from the size of fish entering the recruitment period until the size of fish at the time of stocking, and the second from the size at stocking until recruitment. Over all density dependence in the recruitment phase is apportioned to the before and after stocking component. In this study I rearrange the standard Beverton Holt equation to consist of two components: rate and density. The rate component is assumed to be de termined by the both the proportion of time (size) elapsed in each consecutive phase of density dependence as well as differential mortality for stocked and wild fish; the density component is assumed the same for each fish type (wild, stocked) specific to density entering each stage. As shown in Unpacked Recruitment Dynamics of T able 5 -
124 2 , these methods allow calculation of hatchery, wild, and stocked recruits which can then be graduated into separate age structured models. An gler Effort Dynamics Aggregate fishing effort ( E t ) is assumed to potentially respond to harvestable fish abundance through a logistic function following Walters and Martell (2004). By varying I are able to consider different strengths of effort abundance relationships, including strong (sharp response), a proportional (moderate) response, and no response (flat) as described in Alle n et al. (2012) ( Figure 5 1). While a moderate effort abundance relationship is commonly thought to exist in recreational fisheries ( Johnston et al. 2010 ; Allen et al. 201 3 ), too few empirical estimates exist, particularly for this fishery, for it to be as sumed with certainty. I account for uncertainty of this key relationship by evaluating trade offs under alternative scenarios of dynamic response strength. Angler Satisfaction and Socioeconomic Value Social utility was represented by calculating satisfact ion per trip ( ), which was calculated as the sum of three components: weighted catch rate related satisfaction ( ), non catch rate related satisfaction ( ), and inherent satisfaction from stoc king ( ). Non catch rate related satisfaction represents the utility experienced by an angler simply going fishing, and so is considered a constant per trip, with a magnitude similar catch rate related satisfaction attributed to mode rate catch rates. Catch rate related satisfaction is, of course, largely a function of catch rate, per Cox et al. (2002). Inherent satisfaction from stocking was considered as a hypothetical function that would be mathematically expected to alter the trade off curve,
125 but also captures some realism in the fact that stocking has proven to be quite popular within recreational fisheries, sometimes even when no evidence exists of its impact to the overall fish population (Scharf 2000). In this case I represented the inherent satisfaction with a saturating function such that satisfaction initially increased dramatically with any stocking then asymptotes with greater numbers of fish stocked per year ( Figure 5 2 ). This represents the hypot hetical situation in which stakeholders gain substantial satisfaction simply from knowing the state is stocking fish. Summing the catch rate related, non catch rate related and inherent from stocking satisfaction components yielded an overall satisfaction. Aggregate fishing effort was considered a proxy for economic impacts of recreational angling because the cost paid by fishers (including gas, bait, tackle, transportation, etc.) is directly related to the number of trips taken (Cox et al. 2003). Total s ocioeconomic value ( was calculated as the product of social utility (per trip) and total aggregate effort (Cox et al. 2002), which produces a socioeconomic metric in units of satisfaction that is scaled up by the number of trips in the fishery. In total, the assum ptions that effort is driven by fish abundance, and that fishing satisfaction is driven in part by catch rate, and that overall socioeconomic value can be represented by this scaled up satisfaction are simplistic but useful for considering a stock enhancem ent program. Uncertainty in outcomes owing to these assumptions was accounted for with various model runs. Model Calibration The simulation model was calibrated for the Red Drum recreational fishery of Tampa Bay, Florida, using biological parameters (e.g., natural mortality, M , and maximum age, Amax ) from the most recent stock assessment (Murphy and
126 Muyandorero 2009) and from other studies describing Florida Red Drum populations (Peters and McMichael 1987; Murphy and Taylor 1990). Remaining biological param eters that were not known at this spatial scale (recruitment at unfished conditions, R 0 ) and fishery parameters (catchability, q , and voluntary release rate, k ), were estimated by minimizing a negative log likelihood summing log deviances between observed effort, catch, and harvest for the counties surrounding Tampa Bay (NMFS 2013) and the model predictions. The likelihood weighted evenly the deviances from of catch, harvest, and effort. Deviances for each were calculated using a single value (the average o f the last three years from MRIP data). This tuning ensures that important but realistic initial conditions. More importantly, by fitting the simulation model to the obse rved current (pre stocking) conditions, I can establish a ceteris paribus baseline to which simulated effects of stocking can be compared. Analyses and Scenarios The analyses of results involved three tasks: a description of the likely shape of the conse rvation socioeconomic trade off realized with stocking, an assessment of the nature of the trade offs between conservation and socioeconomic value, and a comparison of likely trade off shapes resulting from stocking under alternative management strategies. For each task, I used terminal values resulting from simulating the fishery for 70 years, which was sufficient to produce stable outcomes. To describe the shape of the trade off, I evaluated likely outcomes in terms of multiple conservation and socioecono mic objectives under numerous stocking scenarios. Here scenarios were defined by the components most controllable by mangers, namely stocking size and intensity. These results were described in absolute terms and in terms of
127 opportunity costs, which reveal ed the shape of the trade off and its frontier. To consider the nature of this trade off shape, I first assessed how the trade off shape changed or shifted under various assumptions of two key uncertainties: how responsive angler effort was and the relativ e importance of catch rate related satisfaction. I then assessed what assumptions would be necessary to realize a dramatically different shape of the trade off frontier. Finally, to compare the trade off shape realized from stocking to that of alternative management strategies, I simulated system outcomes of various scenarios of catch and release fishing, habitat restoration, and direct modification of resource facilities (e.g., improved ramps, bathrooms, etc.). Various scenarios of catch and release fishin g were considered by simply changing the rate of voluntary release. Scenarios of habitat restoration were simulated by running the model forward under different assumptions of recruitment at unfished conditions that essentially represents different carryin g capacities of juveniles, which habitat restoration would be expected to affect. Infrastructure improvements were simulated by considering outcomes under various levels of non catch rate related satisfaction. The specific values of varied and constant par ameters, along with response variables considered to complet e these tasks, are detailed in Table 5 3 . Results Nature of Trade offs An inverse trade off between conservation and socioeconomic metrics occurred when stocking was used as a management strategy. Specifically, the stocking scenarios (in terms of fish size and number stocked) that provided the greatest potential improvements in socioeconomic outcomes were the same as those likely to cause the greatest loss of wild fis h populations, which is the primary conservation objective
128 ( Figure 5 3 ). Wild spawning biomass is best preserved by not stocking fish at all. Overall value is maximized by stocking a high number of larger fish with essentially no increase in overall value if smaller fish are stocked. The outcomes of these scenarios translate into generally convex trade off shapes ( Figure 5 4 ). These shapes are un advantageous in that they show gains of socioeconomic obje ctives are associated with quite high conservation costs. Some stocking scenarios (e.g., stocking at small sizes) reveal a nearly straight line (slope ~0), where stocking more fish provides nearly no increase in socioeconomic value but causes a substantial in conservation ( Figure 5 4 A ). In all cases, substantial gains in socioeconomic objectives are expected only after substantial decreases in conservation. Seen another way, if or when conservation becomes substantially compromis ed (<0.4), gains in socioeconomic value may be realized with relatively little additional conservation loss. The overall results is that the shape of the trade off with any stocking scenario is not one that should be expected to circumvent inherent conserv ation socioeconomic trade offs, but rather shapes that imply strong, particularly costly tradeoffs. Sensitivity of Trade offs The patterns of trade offs were generally resistant to changing assumptions of rate driven satisfact ion ( sat cat ). The overall shape itself, a convex shape at larger stocking sizes and nearly flat at small stocking sizes, remained particularly constant under different assumptions of responsiveness of angler effort ( Figure 5 4 B ) . As a whole, trade offs shifted left and up with flat effort (signifying greater socioeconomic outcomes at the cost of conservation objectives), but right and off meaning that
129 both a greater range of outcomes under various stocking strategies (from high conservation outcomes to high socioeconomic outcomes) were possible. The shapes of trade offs changed more markedly under various assumptions of the importance of catch rate oriented satisfaction ( Figure 5 4 C ). When catch rate oriented satisfaction was low, no scenarios of stocking were substantially able to increase overall socioeconomic outcomes though conservation concerns were similarly compromised unde r all satisfaction assumptions. Interestingly, no tested values of these two primary uncertainties of integrated recreational fishery systems (effort response and satisfaction from catch rate) were able to substantially alter the un advantageous shape of t he observed trade offs. Altering the actual shape of the trade off curve required assuming substantial inherent satisfaction from stocking. In order to achieve a concave shape of the trade off curve, the inherent satisfaction due to stocking had to give an initial increase in satisfaction roughly equal to the combined satisfaction from catch and non catch related satisfaction, and then quickly asymptote ( Figure 5 5 ). The resultant concave shape reveals an optimal stocking scenario of stocking minimal numbers of small fish. Stocking in this way would do little to increase the overall population of fish (since fish were stocked at the beginning of the density dependent period) and would cause minimal replacement of wild fish (since f ew were stocked). Essentially, this would satisfaction from operating a stocking program, but not to cause any real biological effects. Accomplishing this requires stocking sma ll fish as the prominence of the concave shape decreased with increasing size of fish stocked, such that at larger sizes, only a convex shape remained ( Figure 5 5 ).
130 Alternative Management Strategies My comparisons revealed that s tocking in general was un advantageous compared to alternative management strategies. Stocking could produce socioeconomic outcomes at or even above the maximum value generated under alternative strategies (slightly less than satisfaction improvement, more than habitat restoration or catch and release fishing). However, whereas these alternatives at their worse did not decrease the conservation objectives (in this case, wild fish spawning stock biomass), stocking realized the substantial gains in socioecono mic value at great conservation losses. In contrast, satisfaction improvement revealed a straight positive increase in socioeconomic value with no decrease in conservation ( Figure 5 6 ), whereas both habitat restoration and catch and release fishing showed positive linear increases (maximizations) where more intense scenarios led to greater socioeconomic and conservation outcomes. Discussion This study supports the admonishments from Walters and Martell (2004) and more recent fin dings from Lester et al. (2013) that evaluating trade offs is a useful method for addressing conflicting objectives in natural resource management. My results extend these findings, showing how considering the nature of trade offs can reveal not only impor tant information of which parameters outcomes are sensitive to but also can be used to accumulate evidence for potential mechanisms driving the efficiency of a given strategy. Perhaps most importantly, assessing trade off shape and nature can summarize imp ortant information in a cost effective and easily interpreted manner to facilitate discussion regarding natural resource decisions. Augmenting such discussions is perhaps the most meaningful and responsible contribution natural
131 resource science can make (P ossingham and Shea 1999; McNie 2007). Any such small contribution stemming from my results may be particularly useful, since my work has been statistically calibrated to assess outcomes of decisions currently being made ional fishery. Specifically, my results suggest that stock enhancement, a common and popular management strategy, is (under most conditions) an inefficient tool for addressing coupled wild stock conservation and socioeconomic objectives in recreational f isheries, due largely to the un advantageous tradeoff between measures. The results are novel to the extent that enhancement trade off shape and nature have not been previously assed, but are also logical. Improved socioeconomic outcomes of recreational fi sheries from enhancement usually require increased angler effort, increased catch rates, or both. Such increases will almost tautologically negatively affect wild fish populations in the form of attracted angler effort (Baer and Brinker 2010) or competitio n with stocked and hatchery fish. The convex nature of this inverse relationship ( Figure 5 2 ) likely occurs for two primary reasons: (1) increases in socioeconomic outcomes requires stocking larger fish, and little socioeconomic value is generated by stocking very few or very small fish, and (2) any substantial increase in overall (wild plus stocked) abundance of a fish population (that is not severely overfished) will entail partial replacement of wild with stocked fish, due to t he action of compensatory density dependence in the pre recruit stage (Lorenzen 2005). While my results suggest stocking fish large enough to bypass much of the density dependent survival recruitment period is critical to realizing higher overall fish abun dance, many stock enhancement programs stock small fish that are likely susceptible to extensive density
132 dependent mortality (McEacheron 1995; McEacheron et al. 1998; Tringali et al. 2008). Such stocking should result in some displacement of wild fish (con servation losses) but no real increase in socioeconomic value. If stocked fish are unmarked and not distinguishable from wild fish, such replacement may be cryptic (Lorenzen et al. 2010). If larger fish are stocked, they may not be stocked in sufficient nu mbers to lead to a noticeable increase because natural recruitment varies substantially. Better monitoring of enhancement will be crucial in future work designed to test the predicted trade off shape and nature. The shape of the tradeoffs generated from my models renders stock enhancement an inefficient management strategy. Compared to alternative strategies (e.g., habitat restoration, facilities improvement, etc.), enhancement carries greater uncertainty of achieving desired socioeconomic outcomes and gr eater risk of compromising conservation outcomes. Increasing socioeconomic outcomes with enhancement is addressed indirectly by first requiring an increase in fish abundance. This increase is far from certain (Scharff 2000; Tringali et al. 2008; Rogers et al. 2010) due, in part to domestication issues (Lorenzen et al. 2012), density dependent survival (Lorenzen 2005), and the propensity of management to stock small fish. Uncertainty can also exist in the second step necessary for enhancement to achieve soci oeconomic objectives (van Poorten et al. 2011), which is that angler effort dynamics should be responsive to overall fish abundance ( Figure 5 3 ) and/ or that satisfaction must have a strong catch rate component ( Figure 5 4 A ). Thus, success at both steps is uncertain but necessary for enhancement to augment socioeconomic outcomes. And regardless of success at both steps, risk to wild fish is inherent through competition with stocked
133 fish, attracted angler effort, or both. It is more efficient to employ management strategies that directly address satisfaction (e.g., facilities improvements), which engender less uncertainty (by not requiring an increase if fish populations), and are not predicted on increase d risk to wild fish (though effort could respond positively to these actions as well). Alternatively, steps taken to directly augment wild fish populations (e.g., habitat restoration, catch and release fishing) are subject to the same uncertainty (increasi ng fish to achieve a satisfaction/effort responses of anglers), but come without the conservation risk. From this perspective, the popularity and resiliency of recreational fisheries stock enhancement programs appears perplexing. My results also suggest th at even when trade offs are generally un advantageous (convex), some specific scenarios exist in which that strategy could be most useful. An interesting notion from the general convex trade off is that when one objective has been substantially compromised ; the other objective can be increased with lower marginal opportunity costs. In my case study this is represented by a system where wild fish populations have decreased and where additional intense stocking of larger fish can augment socioeconomic objecti ves at near point (0.4, 0.4) of Figure 5 2 . If such outcomes were experienced over a broad spatial scale, there could be particularly negative consequences if greater value placed on rare species (e.g., Endangered Species Act). H owever, if smaller spatial regions are viewed as part of a particularly if it may be mediated by better achieved conservation objectives in another area. An alternative us eful stocking scenario was found to be stocking small numbers of very small fish when inherent satisfaction from stocking was high, allowing greater
134 socioeconomic value (driven by increased satisfaction) with little impact to wild fish ( Figure 5 4 C ). This result is intriguing because it is the scenario (few small fish) that appears least useful in Figure 5 1 ; under different assumptions of stakeholder opinion, however, this scenario produces a more adva ntageous, concave trade off shape. The strong tradeoffs between socioeconomic value and wild stock conservation identified here are predicated on the strong assumption that only the wild population component contributes to the conservation objective. If the hatchery (naturally recruited fish of hatchery parentage) or even stocked fish population components contribute to conservation objectives, the tradeoffs are less strong than implied here. This may be the case for example where hatchery production resu lts in fish that are indistinguishable from wild fish in their genetic makeup and fitness (a feat that can be achieved in some conservation hatchery programs, but is typically very expensive), or where the wild population is deemed of limited conservation interest (e.g. in introduced species). In some situations, the stocked and hatchery components may be viewed as contributing to conservation objectives but at a lower weight than the wild stock component. Stakeholder opinions contribute to the sha pe of the socioeconomic conservation trade offs. In my study this is illustrated by the following: the futility of stocking if catch rate oriented satisfaction is negligible ( Figure 5 4 A ); the efficient (concave) trade off shape s realized if inherent satisfaction from stocking is high ( Figure 5 4 C ); and the substantial gains to both objectives if voluntary catch and release of fish becomes an acceptable management strategy. The effect of alternative as sumptions of satisfaction found in this study underscores the importance of accurately assessing the motivations or antecedents of satisfaction, a message in congruent with recent work both
135 specifically within fisheries (Hunt 2005; Arlinghaus and Cooke 200 9) and more broadly in work integrating social science and natural resources (Constanza 2000). A less commonly discussed, albeit intuitive implication of this is the idea that attempting to directly alter opinions may be an effective management strategy. I t is possible that shifts in management paradigms to those that more explicitly and meaningfully involve stakeholder processes may result in greater satisfaction and overall socioeconomic outcomes, potentially even with little change in the actual natural resource. Regardless of the exact mechanism, increasing stakeholder satisfaction would likely be an efficient Figure 5 5 . These concepts should translate across most natural resource trade offs where stakeholder satisfaction and opinion is important. Predicting changes in stakeholder satisfaction in response to potential shifts in governance paradigms requires not only representing the stakeholders but als o the management or governance component, which may have its own dynamics. In fact, the dynamics of governance systems may be partly responsible for why some management strategies that may be inefficient (e.g., enhancement) continue. Governance systems gen erally act to support their own legitimacy in part to maintain power. In the case of politicians this usually involves some catering to constituents. In the case of natural resources, where mangers generally answer to politicians, this may entail favoring strategies that expand budgets or personnel. It is possible that the management strategy of stock enhancement remains so widespread for the same reasons. While my results do not provide direct inference of such processes they do suggest that enhancement is a management strategy that is almost tautologically inefficient
136 regarding inherent natural resource objectives. Perhaps the most reasonable explanation for the increase in enhancement programs given my results is that there are other elements of the syste m that substantially affect outcomes that are not well considered in my models. Aligning my model with the persistence of enhancement may not be difficult since the inherent satisfaction that stakeholders were hypothesized to experience from stocking ( Figure 5 4 A ) could be reimagined as benefits enjoyed by governance/management. Adequately understanding the outcome landscapes of natural resource systems will likely require more explicit consideration of how the governance process itself affects management decisions. While this study may be useful as a model for considering trade offs in natural resources, it is not complete. I present equilibrium results expected under average conditions, and do not account for time streams for realized benefits. My work also does not explicitly consider dynamics of the governance process either feedback loops whereby management scenarios (e.g., size of stocked fish or intensity) might be modified as monitoring information is received, or the po tential benefits to governance of one strategy over another. Further, I have not considered financial capital and operating costs of alternative strategies, which could alter the location of the trade off frontiers relative to alternatives. For example, if habitat restoration entailed an order of magnitude greater costs per unit augmented conservation objective, this maximization curve might shift downward such that it would no longer be clearly superior to stock enhancement. Future case specific work would need to consider these costs as well as potential implications that alternative strategies have for satisfaction/utility experienced by stakeholders. A substantial but implicit assumption of my study is that trade offs are
137 meaningful in a single spatial a rea that is managed without regard to other areas. This may be common in certain situations (e.g., angling clubs in Europe, community based resource reef fisheries or forestry initiatives, subsistence hunting, etc.) where conservation and socioeconomic obj ectives must be obtained from essentially the same spatial region. Again, perhaps the most pervasive assumption of my work is that conservation objectives are best represented by spawning biomass of wild fish. While such an approach has broad support (Holl ing and Meffe 1996), alternative understanding of the conservation value of hatchery originated fish would obviously affect my results, as previously described. As increasing realism and complexity is added to representations and evaluations of natural re source trade offs perhaps the greatest challenge will be determining the balance of generality and specificity in the models used to represent these trade offs (Boyd 2012; Evans et al. 2012). In summary, my study employed a basic model designed to capture biological dynamics of fishery with wild and hatchery fish and evaluate conservation and socioeconomic tradeoffs given different assumptions on anger behavior and preferences. The analysis yielded several general and case specific conclusions. Generally, e valuating trade offs with quantitative models may be useful for abetting management discussions by comparing alternative strategies. These models are likely to be most useful when various stakeholder opinions and management preferences are explicitly consi dered. One potential but overlooked strategy may be modifying the management process itself. Specific to my case study of recreational stock enhancement, four primary messages emerge: (1) nearly any recreational stock enhancement is almost certain to have negative consequences for wild fish, (2) most
138 implemented enhancement programs (those that stock small fish) are unlikely to achieve substantial socioeconomic outcomes regardless of costs to wild fish, (3) more efficient strategies for attaining socioecono mic and conservation objectives likely exist, but (4) stock enhancement could still be a useful strategy if wild fish populations are greatly degraded in spatially explicit areas or if stakeholders (or managers) simply prefer stocking programs. Since these specific findings will be useless if they are not understood clearly by decision makers (Possingham and Shea 1999), the ultimate utility of my work will depend on the veracity of my general statements that descriptions of trade off shape and nature consti tute a clear form of communication of complex and conflicting objectives of natural resources.
139 Table 5 1. Description of parameters and parameter values (* indicates estimated). Symbol Description Units Value *R 0 Recruitment at unfished conditions (num ber of fish) 450,371 Length at infinity mm 934 K Metabolic in growth equation yr 1 0.46 t 0 Age at length=0 yr 1 0.26 a Weight length constant g 0.000000617 b Weight length constant g 3.09 W mat Weight at maturity kg 10.084 M Instantaneous mortality yr 1 0.113 LR R eference length mm 730 lorenzC Allometric exponent of length mortality relationship 0.9 Amax Maximum age years 40 Recruitment compensation parameter ratio 11 L0 Length at entering recruitment period mm 20 Ls Length at stocking mm 25 175 Lr Leng th at leaving recruitment period mm 180 tt Duration of recruitment period years 0.75 M1 Natural mortality year 1 of 10mm fish year 15 S1S2 Cumulative base survival for the recruitment period rate Calculated fit hat Relative fitness (or survival) of hatc hery to wild, stage 1 rate 1.0 fit st Relative fitness (survival) of stocked to wild, stage 2 rate 0.8 Share of hatchery eggs inheriting wild characteristics % 0.2 ST t Number of fish stocked each year mil. 0 4.5 S 0.5 Back scaled mortality to fish size midway between 0.75 yrs and 1 yrs yr 1 0.86 , Fish length for vulnerab ility to capture: low, high mm 400, 850 , Standard deviations of vulnerability to capture , , Fish length for vulnerability to harvest: low, high mm 457, 686 , S tandard deviations of vulnerability to harvest , *k Proportion of harvestable fish killed % 0.35 Standard deviation of logistic 0.05 2000 Emin, eo Minimum effort and effort an unfished stock size trips 200,000; 600,00 0 *q Catchability coefficient rate 0.0000026 d Discard mortality rate 0.08 CPUE where sat cat = 0 rate 0.05
140 Table 5 1. Continue d. Symbol Description Units Value Ratio of catch to non catch related satisfaction ratio 0 1 Slope of the r elationship between sat cat and CPUE 3 Magnitude similar to average sat cat 6.5 B Inherent satisfaction from stocking 0, 1 Inherent satisfaction from stocking at low stocking 10 Shape parameter of inherent satisfaction i ncrease 0.08
141 Table 5 2. Description of model components and equations. Component Equation Life History Characteristics of Stock: Growth: Length L at age a Size: Weight (kg) to length ratio Fecundity ( f at age a ) Survival (year 1 ) ( S at age a ) Survivorship Eggs per recruit Unpacked Recruitment Dynamics: Beverton Holt a , b, and re parameterized b , , Duration of second stage of recruitment Linear growth rate (year 1 ) through recruitment period V Base survival of recruitment stage 1 and 2 , Survival rate of larvae from eggs until beginning of recruitment period f Stage 1 survival of fish a1 , Stage 2 survival of fish a2 , , Density dependent component of survival for stage 1 and stage 2 ( b1 and b2) , Total eggs in beginning year t, which is sum of wild ( w ) and hatchery eggs ( hat ) where ,
142 Table 5 2. Continued. Component Equation Number of wild fish surviving the first recruitment stage Number of hatchery fish surviving the first recruitment stage Total fish entering the second rec. stage Number of fish leaving recruitment period: Wild Hatchery Stocked Number of fish entering age 1: Wild Hatchery Stocked Fishery Characteristics: Vulnerability to capture, which is where up low , Vulnerability to harvest, which is where up low , Effort ( E t ) Time Dynamics: Wild spawning biomass Exploitation rate
143 Table 5 2. Continued. Component Equation Numbers at age Survival of all types from: Harvest Discard of non legal catch Voluntary discard of legal catch Socio economic Dynamics: Total catch Catch per unit effort (CPUE) Satisfaction per trip where , Overall socioeconomic value
144 Table 5 3. Summary and comparison of model runs. Parameter values Figure Analysis Response metric Size at stocking ( Ls ) Effort response ( ) Satisfaction: Catch ( ) and or Stocking ( B ) 1 Overall outcomes WSB t , E t , sat t , val t Variable (25 175mm) Moderate ( 1) High catch, no stocking ( =1, B =0) 2 Overall trade offs WSB t , val t Variable (25 175mm) Moderate ( 1) Hig h catch, no stocking ( =1, B =0) 3 Effort response WSB t , val t Variable (25 175mm) Variable ( 0.05, 1, 2000) High catch, no stocking ( =1, B =0) 4 Catch/non catch satisfaction WSB t , val t Variable (25 175mm) Variable ( 0.05, 1, 2000) Variable c atch, no stocking ( =0,0.5,1; B =0) 5 Nature of satisfaction from stocking sat t NA NA NA 6 Trade off with satisfaction from stocking WSB t , val t Variable (25 175mm) Moderate ( 1) High both ( =1, B =1) 7 Alternative mgmt. strategies WSB t , val t Moderate (100mm) Moderate ( 1) High catch, no stocking ( =1, B =0) Note: All scenarios were based on the following stocking assumption: ST t = 0: 10* R 0.
1 45 Figure 5 1. Alternative assumptions of how dynamic angler effort could respond to catc hable abundance of fish.
146 Figure 5 2. The assumed inherent satisfaction from stocking as a function of the numbers of fished and stocked.
147 Figure 5 3. Overall trade offs between number of fish stocked and size at stocking by metric. A) Wild s pawning biomass. B) Satisfaction per trip. C) Total fishing effort and D) Overall socioeconomic value. For each panel, the y axis represents the size of fish stocked and the x axis represents the numbers stocked.
148 Stocking Response Effort R esponse A B Catch Rates C Figure 5 4. Trade offs between conservation and socioeconomic value . A) The overall trade off as a result of stocking. B). The trade offs resulting from stocking with different effort responses. C) The trade offs resultin g from stocking with different satisfaction from catch rates, assuming a moderate effort response .
149 Figure 5 5. Tradeoffs between the conservation and socioeconomic value associated with different levels of satisfaction from stocking.
150 Figure 5 6. Comparisons of trade offs under different management strategies. Tradeoffs between conservation objective ( x axis, represented by wild spawning biomass) and socioeconomic value objective ( y axis, represented by socioeconomic value ) for alternative man agement strategies.
151 CHAPTER 6 CONCLUSIONS The management of most natural resources is ultimately interested in producing favorable outcomes of the inherent immediate trade off between conserving a resource and using that resource for some socioeconomic gain (Holling 1978; Cheung and Sumalia 2008; Walters and Ahrens 2009). A primary job of scientists is then to understand what outcomes specific and generalized are likely to result from alternative management strategies (Walters and Martell 2004). Assess ing these outcomes is often challenging because natural resource are generally composed of various ecological, social and economic components that must be taken into account (Ostrom 2009). Here an integrated approach that considers the most important elem ents and interactions of biological, stakeholders, market/economy, environment, and governance systems is often necessary to understand the potential effects of management strategies (Oakerson 1992; Lorenzen 2008). An increasingly common management strate gy is stock enhancement (Lorenzen 2005), especially in recreational fisheries (Askey et al. 2013). Understanding the effects of enhancement on the socioecological recreational fishery system requires accounting for the interactions between at least fish p opulations, the actions, opinions and beliefs of anglers and non anglers alike (stakeholders) , the economic impact and consumer surplus generated from fishing, the technological capacity for rearing fish in hatcheries, and how all of these are affected by the overall management paradigms (Lorenzen 2008; Camp et al. 2013). A number of studies have evaluated enhancement, but an integrated and quantitative assessment of enhancement of recreational fisheries has been largely Red Drum fishery as a case study, I have designed my
152 dissertation to conduct such an assessment to improve the understanding of the efficacy of enhancement as a management strategy for recreational fisheries and simultaneously provide information useful to managers of this fishery. I used multiple methodologies to complete four distinct studies that together improve the integrated understanding of stock enhancement for recreational fisheries. One of the greatest challenges in an integrative framework is representin g all the full spectrum of important aspects of a system without representing so many aspects that the system is not comprehensible (Lorenzen 2008). I addressed this in my first chapter by synthesizing available literature describing integrated systems ap proaches, recreational fisheries, enhancement, and Red Drum to identify which components and linkages were most important to understand. I found one of the more potentially important but less studied linkages was the dynamics of angler effort, which was t he topic of my second chapter. In this work I used an empirical approach to evaluate relationships between recreational fishing effort targeting key marine species in Florida and potential covariates, inc luding estimated fish abundance, human abundance , a nd calendar year (which could represent an unexplained temporal trend) . Results from this work helped inform the study topic investigated in my third chapter, a quantitative simulation predicting biological and socioeconomic outcomes of Red Drum stock enh ancement in Florida. One of the pri mary results of this work was the finding of a conservation socioeconomic trade off produced by stock enhancement. I extended my simulation approach to investigate the nature and shape of this trade off further with my fourth chapter. This chapter c onsidered how stakeholder values and opinions could
153 alter the shape of this trade off, and compared the efficiency trade offs realized with stock enhancement to those of alternative management strategies. Summary of F indings Chapter 1 The synthesis of Red Drum stock enhancement in Florida represented one of the first a priori integrative evaluations of recreational stock enhancement, and revealed linkages likely important is most enhanced recreational fisheries. This work hig hlighted the influence that recruitment dynamics are likely to play in enhancement outcomes, and the need to understand the size or age of fish at which dens ity dependent mortality subsides , as well as the spatial scale at which recruitment is regulated (e .g. coast wide or estuary specific). These recruitment dynamics, including potential competitive interactions between wild and stocked fish, tautologically define the actual effect of enhancement on the overall stock abundance, and thus affect nearly all system outcomes. Additionally, my synthesis showed that angler effort dynamics represented a substantial gap in knowledge. Specifically, it was not known how strongly aggregate fishing effort would respond to potential increases in abundance of catchable Red Drum (combined wild and stocked fish). This was considered important because both economic impacts (e.g., jobs and revenue, both quite important to political elements of governance systems), as well as fishing mortality depend directly on fishing eff ort levels. A final and important linkage was how stock enhancement might affect the stakeholder opinions and sense of ownership for the resource. While this synthesis was directed towards Red Drum in Florida, these linkages would be expected to be impor tant in nearly all enhanced recreational fisheries.
154 Chapter 2 Aggregate angler effort (how many fishing trips are made) has been for years implicitly assumed to be directly related to abundances of fish, but there has been fe w empirical evaluations of th is (Post et al. 2008). My work assessing the empirical predictors of recreational fishing effort in Florida was most useful for what it did not show a clear predictor of fishing effort. Essentially it was possible to predict targeted species specific fis hing effort nearly equally well with either estimated fish population abundance or estimated human population in Florida. The inability to distinguish between these potential drivers of aggregate fishing effort is concerning in the context for the resilie nce of the fish population (though perhaps not the recreational fishery system). Aggregate effort that is linked directly to fish populations will behave more like a coupled predator prey system, such that declining fish populations will eventually lead t o decreased fishing effort and probably fishing mortality, allowing the recovery of the population (Johnson and Carpenter 1994; Carpenter et al. 1994; Allen et al. 2013). Conversely, if aggregate effort responds exogenously to human population or time, fi sh populations and probably catch rates will decline absent effort limitation. The lack of a clear signal from this analysis suggests (1) evaluations of stock enhancement must consider a range of potential angler effort dynamics, from non responsive to qu ite responsive and (2) it may be necessary and worthwhile to explore natural or manipulative experiments to better assess the responsiveness of angler effort. Chapter 3 This work is particularly novel in its integration of the pertinent biological elemen ts of population dynamics of enhanced fisheries (Lorenzen 2008) with metrics most important for understanding socioeconomic outcomes (Propst and Gavrilis 1987;
155 Cox et al. 2003), and doing so in a tractable quantitative framework that can be tuned to an act ual fishery. The quantitative evaluation of the likely and unlikely responses of Red Drum stock enhancement in Florida was designed to explicitly address management questions important to the state, but also revealed relationships likely to be general to other fisheries. The primary finding was that nearly any enhancement strategy (size or number of fish stocked) would be likely (under equilibrium conditions) to elicit a negative response to wild fish populations, via replacement of wild with stocked fis h or their offspring in the recruitment phases or additional fishing mortality from increased angling effort. While this finding is straightforward, it has been rarely communicated clearly (van Poorten et al. 2011) and this has probably led to widespread misunderstanding of the effects of enhancement within fisheries science and management. My results specifically showed stocking fish at larger sizes allowed for increased overall fish populations that could lead to augmented socioeconomic value in terms of increased economic impact stemming from increased fishing effort (at a regional, local or species specific scale), or increased benefit from increased catch rates, but that there was little reason to expect stocking smaller fish (i.e. smaller than the siz e of cessation of density dependent mortality) to have much benefit. This is notable since many (especially marine) hatcheries stock fish at very small sizes. To illustrate this with a specific case study, I showed that increasing the abundance of Red Dr um by 25% in a popular fishery (Tampa Bay, Florida) could be achievable at a potentially reasonable cost (thousands to hundreds of thousands of dollars per year), provided that fish were stocked at larger sizes (>100mm). Achieving such an increase by stoc king smaller fish (~25mm total length, a size at which it is expected that stocked
156 fish must pass through nearly all of the density dependent mortality phase in the wild) could only be achieved at unreasonably high costs ($10e16). This is very important, given the state of Florida is considering moving to more intensive rearing of Red Drum that would correspond to a very small release size. This work revealed an important trade off between wild fish populations (a common conservation objective) and aggreg ate effort and catch rates (common socioeconomic objectives). A secondary trade off that this work made more explicit is between economic impact (revenue and jobs generated from increased fishing effort) and economic benefits (believed to be directly rela ted to catch rates). These trade offs should exist in nearly any open access recreational fishery. Chapter 4 The trade off between conservation and socioeconomic or current and future use is perhaps the defining element of natural resource science, and it can be very useful to describe the shape of this trade off realized under one strategy, relative to alternative management strategies (Walters and Martell 2004). For example, concave down trade offs between two objectives allow for an optimum, where th e most of each is attainable, whereas convex up shapes require that substantial amounts of one value must be forgone to achieve marginal increases in the other (Walters and Martell 2004). The work of this chapter represented the first evaluation of the tr ade off shape of enhancement, revealing that stocking will likely realize a convex up shape, which is quite inefficient. Further, I showed that uncertainties in angler effort dynamics did little to alter the shape of this trade off. Interestingly, the on ly marked change occurs if one was radically dependent on the existence of a stocking program. In such a case,
157 stocking very small numbers of small fish would cause essent ially no decrease in wild fish (because almost no stocked fish would survive) or any corresponding increase in effort related economic impact, but would benefit from the increased satisfaction stakeholders experience from the operation of a hatchery. This actually altered the shape of the trade off to concave down, a more favorable shape. However, comparisons of the trade offs realized under stock enhancement to alternative augmentative strategies, such as habitat restoration or fishing facilities improve ments (e.g. better restrooms) revealed that stock enhancement was generally less efficient. Key C ontributions To M anagement Three key areas of uncertainty relative to enhancement outcomes are (1) the relationship between density dependent mortality and f ish size, (2) how angler effort is related to catchable abundance of fish and (3) how enhancement alters Given the possibility that targeted fishing effort is de coupled fro m fish abundance, Florida (and other areas where recreational fishing is increasing in popularity) may need to follow the advice of Cox et al. (2003) and consider restricting effort if maintaining quality (in terms of catch rate) wild fisheries is desired. It is very likely stocking Red Drum in Florida will have a negative effect on the wild fish populations, though it may increase some socioeconomic measures . Stocking small Red Drum in Florida is not likely to cause any noticeable increase in overall abu ndance or socioeconomic outcomes aside from satisfying a potential direct demand for hatcheries. Probably the best scenario for stocking Red Drum in Florida would be to release larger (>150mm) fish in areas near population centers that have low natural re cruitment (either owing to recruitment overfishing or lack of essential habitat at very small sizes), and using the stocking to experimentally explore not only biological uncertainty regarding recruitment, but also responses of stakeholder beliefs and inve stment.
158 To D iscipline(s) The outcomes of any given management strategy depend on a chain of interactions that includes biological effects, how the biological effects interact with stakeholder uses of the resource, and even more broadly, how the management action directly effects opinions. An honest approach to angler effort dynamics should include admitting that currently I have little predictive ability regarding aggregate recreational effort, which has substantial implications for the motivation of many management practices, including stock enhancement. Quantitative enhancement modeling frameworks that integrate the necessary chains of biological, resource use, and stakeholder interactions are achievable and I argue are useful despite their necessary abst ractions. Such models can be tuned to most marine recreational fisheries given available data (from the National Marine Fisheries Service) and can be used to make reasonable predictions of the outcomes of enhancement prior to implementation. Stocking is not a silver bullet for fisheries management, and should be expected to realize an unadventageous trade off between wild fish populations and socioeconomic objectives under most implementations. Trade offs realized with enhancement are likely to be much strategies (e.g., habitat restoration, fishing facility improvement) because enhancement is likely to explicitly harm wild fish . While angler effort dynamics are quite important for understanding which type of socioeconomic effects (impacts, satisfaction) are realized, the deleterious effects of stocking on wild populations are largely not sensitive to these. This implies that while dynamics are uncertain, the harm enhancement causes wild fish is more so. Economic impact and socioeconomic benefits are quite different metrics that actually trade off in the context of stock enhancement. One of the elements most likely to alter trade off shapes is how the management strategy itself (not its indirect effect v ia altered natural resource levels) is viewed by stakeholders. Evaluations of such effects require explicitly an integrated framework. Future Directions Nearly all of this work has focused on equilibrium outcomes of a given management strategy, stock enh ancement. This is very useful for understanding the statistical expectation s of outcomes under average conditions, but is less useful for
159 making predictions of the range of out comes possible under unlikely but possible conditions. This means that the pre dictions of this work are best used for increasing explicit understanding of how enhancement can and cannot be expected to alter integrated socioecological systems, not for making precise quantitative predictions of what will happen in the short run (e.g., one year after commencement of stock enhancement). Additionally, while implications of this work can potentially be made relative to the effects of stock enhancement on system resilience, further work is needed to explicitly address this. Considering a response metric of system resilience, there are at least two primary pathways by which this may be done. First, a logical/conceptual evaluation of resilience may involve assessing the antecedents of a resilient system (e.g. genetically health fish populat ions, invested stakeholders) and a subsequent evaluations of how these antecedents would be affected by enhancement. Second, a quantitative approach using stochastic dynamic programming can be employed to subject underlying quantitative models of biologica l/ecological, socioeconomic, and even management components to unexpected environmental or social conditions. Either of these approaches would probably be most useful for comparing the utility of alternative management strategies at preserving the current or desired structure of a socioecological system faced with the unexpected changes that history has suggested are almost guaranteed to occur. Evaluations of how management strategies, like enhancement, affect resilience would be aided by two specific stud ies. Of primary use would be empirical explorations of how stakeholder interest, opinion and investment respond to natural resource management paradigms, strategies, and scenarios. This probably will require coupling
160 quantitative evaluations of actual re source (e.g. fish populations) and use (fishing effort) responses with possibly less quantitative evaluations of stakeholder characteristics. Stock enhancement may present an excellent natural experiment framework for such work. A second area of study th at would be very useful is to place the work described in this dissertation in a spatial context, such that larger scale (e.g. statewide) outcomes of local stock enhancement could be explored. Such spatially explicit work, especially when coupled with an understanding of dynamics of stakeholder investment, should be particularly useful to informing management paradigms and visions and is quite compatible wi th current ideas of stakeholder led and often local co management, recognized as a key component of o verall system resilience (Guiterez et al. 2011).
161 LIST OF REFERENCES Aas, O. 2008. Global Challen g es in Rec r eational F isheries. Ames, I o w a: Blackwell Publishing. Adams, D. H., and D. M. T remain. 2000. Association of la r ge ju v enile Red Drum , Sciaenops oce llatus, with an estuarine creek on the Atlantic coast of Florida. E n vi r onmental Biology of F ishes, 58: 183 194. Adkinson, M. D. 2009. Drawbacks of complex models in frequentist and Bayesian approaches to natural resource management. Ecological Applicati ons 19: 198 205. Allen M.S., R.N.M. Ahrens, M.J. Hansen and R. Arlinghaus. 2013. Dynamic angling effort influences the value of minimum length limits to prevent recruitment overfishing. Fisheries Management and Ecology 20: 247 257. American Sportfishing Association ASA. 2001. Data and statistics, saltwater. A v ailable from http://www.asafishing.org Araki H., B. Cooper, and M. S. Blouin, M. S. 2007. Genetic effects of captive breeding cause a rapid, cumulative fit ness decline in the wild. Science 318: 100 103. Araki, H., B.A. Berejikian, M.J. Ford, and M.S. Blouin. 2008. Fitness of hatchery reared salmonids in the wild. Evolutionary Applications 1: 342 355. Ar b uckle, M. 2000. Fisheries management under ITQs: In n o v ations in N e w Zealand s Southern Scallop Fisher y . P aper presented at the 2000 Confe r ence of the International Institute of F isheries Economics and T r ade IIFET. A v ailable from http://oregonstate. du/Dept/IIFET/ 2000/papers/arbuckle.pdf. Arlinghuas R. 2006. On the apparently striking disconnect between motivation and satisfaction in recreational fishing: the case of catch orientation of German anglers. North American Journal of Fisheries Management 26: 592 605. Arlinghaus, R., and T. Mehner 2005. Determinants of management preferences of recreational anglers in Germany: habitat management versus fish stocking. Limnologica Ecology and Man agement of Inland Waters, 35: 2 17. Arlinghaus, R. and S. J. Cooke. 2009. R ecreational fisheries: socioeconomic importance, conservation issues and management challenges. Pages 39 58 in B. Dickson, J. Hutton, and W. A. Adams editors. Recreational Hunting, Conservation and Rural Livelihoods: Science and Practice. Oxford: Blackwell Publishing. Arlinghaus, R., S.J. Cooke and W. Potts. 2013. Towards resilient recreational fisheries on a global scale through improved understanding of fish and fisher behavior. Fisheries Management and Ecology 20: 91 98
162 Arnason, R. 2001. The economics o f ocean r an c hing: e xperiences, outlook and theory. F A O Fisheries T echnical P aper 413. ASA. 2001. American Sportfishing Association. Data and Statistics, Saltwater http://www.asafishing.org Askey, P.J., E. A. Park inson, and J. R. Post. 2013. Linking fish and angler dynamics to assess stocking strategies for hatchery dependent, open access recreational fisheries. North American Journal of Fisheries Management 33: 557 568. Bacheler, N. M., L. M. P aramore, J. A. Bu c k el, and F . S. Scharf. 2008 . Recruitment of ju v enile Red Drum in North Carolina: Spatiotemporal patterns of year class strength and v alidation of a seine sur ve y. North American J ournal of F isheries Man a g ement 28: 1086 1098. Baer, J., K. Blasel, and M. Die kmann. 2007. Benefits of repeated stocking with adult, hatchery reared br o wn trout, Salmo trutta, to recreational fisheries? F isheries Man a g ement and Ecology, 14: 51 59. Baer, J. and A. Brinker. 20 10 . The response of a brown trout stocks and perception of anglers to cessation of brown trout stocking. Fisheries Management and Ecology 17: 455 464. Beard, T . D. J r ., S. P . Cox, and S. R. Carpenter. 1993. Impacts of daily bag limit reduction on angler e f fort in W isconsin w all e ye la k es. North American Journal of Fisheries Management 23: 1283 1293. Beardmore B., W. Haider, L.M. Hunt and Robert Arlinghaus. 2011. The Importance of Trip Context for Determining Primary Angler Motivations: Are More Specialized Anglers More Catch Oriented than Previously Believed? Nor th American Journal of Fisheries Management 31: 861 879. Beck, J. L., and J. R. Roo k er. 2007. The impact of diet on the surv i v al skills of hatchery Red Drum Sciaenops ocellatus. P r oceedings of the Annual Gulf and Carib b ean F isheries Institute 60: 368 371. Bell, J. D., K. M. Leber, H. L. Blan k enship, N. R. Lonera g an, and R. Masuda . 2008. A new era for restocking, stock enhancement and sea ranching of coastal fisheries resources. R e views in F isheries Science 16: 1 9. Beverton, R.J.H. and S.J. Holt, S.J. 1957. On the Dynamics of Exploited Fish Populations. London: HMSO. Blankenship H. L., and K. M. Leber. 1995. A responsible approach to marine stock enhancement. American Fisheries Society Symposium 15: 67 175. Boyd I.L. 2012. The art of ecological modeling. Science 337:306
163 Brennan, N. P ., M. C. Dar c y, and K. M. Leber. 2006. Predator free enclosures impr ov e post release surv i v al of stoc k ed common snook. J ournal of Experimental Ma r ine Biology and Ecology 335: 302 311. Brock W.A. and S.R. Carpenter. 2007. P anaceas and diversification of environmental policy. Proceedings of the National Academy of Sciences 104:15206 15211. Bureau of Labor Statistics BLS. 2013. Florida unemployment Series ID LASST12000003. http://data.bls.gov/pdq/SurveyOutputServlet Bureau of Economic and Business Research BEBR. 2011. Population Estimates for Florida. University of Florida. Camp E.V., K. Lorenzen, R.N.M. Ahrens, L. Barbieri and K.M. Leber. 2013. Understanding soci oeconomic and ecological trade offs in the enhancement of recreational fisheries: an integrated review of potential Florida Red Drum enhancement. Reviews in Fisheries Science 21: 388 402 Carpenter S.R., A. Munoz Del Rio, S. Newman, P.W. Rasmussen and B.M. Johnson. 1994. Interactions of anglers and wallyers in Escanaba Lake, Wisconsin. Ecological Applications 4: 822 832. Carpenter, S. R., and W. A. Brock. 2004. Spatial complexity, resilience and policy diversity: Fishing on lake rich landscapes. Ecology a nd Society 9. Chase Dunn C., and T.D. Hall. 1997. Rise and demise: comparing world systems. Westview Press, Boulder, Colorado, USA. Cheung, W.W. and U.R. Sumaila. 2008. Trade offs between conservation and socio economic objectives in managing a tropical m arine ecosystem. Ecological Economics 66: 193 210. Coggins, J r . L. G., M. J. Catalano, M. S. Allen, W . E. Pine III, and C. J. W alters. 2007. E f fects of cryptic mortality and the hidden cost of using length limits in fisheries management. Fish and Fisheries 8: 196 210. Cooke, S. J., and I. G. Cowx. 2006. Contrasting recreational and commercial fishing: searching for common issues to promote unified conservation of fisheries resources and aquatic environments. Biological Conservation 128: 93 108. Cowx, I .G. 1994. Stocking strategies. Fisheries Management and Ecology 1: 15 30. C o wx, I. G. 2002. Recreational fisheries . Pages 367 390 in P. Hart, and J. R e ynolds editors. Handbook of F ish Biol o gy and F isheries, V ol. II. Oxford, UK: Blackwell Science.
164 Cowx I.G ., R. Arglinghaus and S.J. Cooke. 2010. Harmonizing recreational fisheries and conservation objectives for aquatic biodiversity in inland waters. Journal of Fish Biology 76: 2194 2215. Cox, S. P ., C. J. W alters, and J. R. Post. 2003. A model based ev alu ations of act i v e management of recreational fishing e f fort. North American Journal of Fisheries Management, 23: 1294 1302. Cox, S. P., T. D. Beard, and C. Walters. 2002. Harvest control in open access sport fisheries: hot rod or asleep at the reel? Bulleti n of Marine Science 70: 749 761. Cox S.P and C. Walters. 2002. Modeling exploitation in recreational fisheries and implications for effort management on British Columbia rainbow trout lakes. North American Journal of Fisheries Management 22: 21 34. Denem ark, R. A., J. Friedman, B. K. Gills, and G. Modelski, editors. 2000. World system history: the social science of long term change. Routledge, London, UK. Ditton, R. B., and A. J. Fedler. 1989. Importance of fish consumption to sport fishermen: A reply to Matlock et al. 1988. F isheries 14: 4 6. Durbin J. and G.S. Watson. 1950. Testing for Serial Correlation in Least Squares Regression I. Biometrika 37: 409 428. Economic Development and Research EDR. 2013. County population estimates. Based on the result s from the Florida Demographic Estimating Conference, February 2013 and UF, BEBR, Florida Population Studies, Volume 46, Bulletin 165, March 2013. http://edr.s tate.fl.us/Content/population demographics/data/CountyPopulation.pdf Engstrom Fisheries 6: 11 16. Evans M.R., K.J. Norris and T.G. Benton. 2012. Predictive ecology: system appro aches. Philosophical Transactions of the Royal Society of Britian B Series 367: 163 169. Fedler, A. J., and R. B. Ditton. 1994. Understanding angler mot i v ations in fisheries management. F isheries 19: 6 13. Fenichel E.P., J.K. Abbott and B. Huang. 2013. Modeling angler behavior as a part of the management system: synthesizing a multi disciplinary literature. Fish and Fisheries 14: 137 157. Figge, F. 2004. Bio folio: Applying portfolio theory to biodiversity. Biodiversity Conservation 13: 827 8 49.
165 Figuei ra, W. and F.C. Coleman. 2010. Comparing landings of United States recreational fishery sec tors. Bulletin of Marine Science. 86:499 514. Finn K.L. and D.K. Loomis. 2001. The Importance of Catch Motives to Recreational Anglers: The Effects of Catch Satia tion and Deprivation, Human Dimensions of Wildlife: An International Journal 6: 173 187 Fisher, M. R., and R. B. Ditton. 1994. A social and economic characterization of the U.S. Gulf of Mexico recreational shark fishery. Marine Fisheries R e v iew 55: 21 27. FL DOT. 2013. Annual visitors to Florida, Air, non air. Florida Department of Transportation office of policy planning. http://www.floridatransportationindicators.org/de tail.php?chart=11b Flecker A. S and C. R. Townsend. 1994. Community wide consequences of trout introduction in New Zealand streams. Ecological Applications 4: 798 807. Fraser, D.J. 2008. How well can captive breeding programs conserve biodiversity? A revi ew of salmonids. Evolutionary Applications 1: 535 586. Gale, R.P. 1987. Resource miracles and rising expectations: a challenge to fishery managers. Fisheries 12: 8 13. Garcia Asorey M.I., G. Escati Penaluza, A.M. Parma and M.A. Pascual. 2011. Conflicting objectives in trophy trout recreational fisheries: evaluating trade offs sing an individual based model. Canadian Journal of fisheries and Aquatic Sciences 68:1892 1904. Gaydon, D.S., H. Meinke, D. Rodriguez and D.J. McGrath. 2012. Comparing water option s for irrigation farmers using modern portfolio theory. Agricultural Water Management 115: 1 9. Gold, J. R. 2008. Stock structure and e f fect i v e size of Red Drum Sciaenops ocellatus in the northern Gulf of M e xico and implications relat i v e to stock enhanceme nt and recruitment. Pages 353 370 i n K. M. Leber, S. Kitada, H. L. Blankenship and T. Sv a sand , editors. Sto c k Enhancement and Sea Ran c hing: D e velopments, Pitfalls and Opportunities, 2nd Ed. Oxford, UK: Blackwell Publishing Ltd. Gold, J. R., L. R. Richardson, and T . F . T urner. 1999. T emporal and spatial divergence of mitochondrial DNA haplotyp e frequencies in Red Drum Sciaenops ocellatus from coastal r e gions of the western Atlantic Ocean and Gulf of M e xico. Ma r ine Biology 133: 593 602. Gotts N.M. 2007. Resilience, pararchy and world systems analysi s. Ecology and Society 121:24. Gregory, R.S., and R. L. Keeney. 2002. Making smarter environmental management decisions. Journal of the American Water Resources Association 38: 1601 1612.
166 Gunderson L.H. and C.S. Holling. 2002. Panarchy: Understanding transformations in human and natural systems. Is land Press, Washington, D.C. USA Haddon, M. 2010. Modelling and quantitative methods in fisheries. Taylor and Francis US. Halpern, B.S., C. White, S.E. Lester, C. Costello and S.D. Gaines. 2011. Using portfolio theory to assess tradeoffs between return fro m natural capital and social equity across space. Biological Conservation 1445: 1499 1507. Halverson M.A. 2008. Stocking trends: A quantitative review of governmental fish stocking in the United States, 1931 to 2004. Fisheries 33:69 75 Hazlerigg, C. R. E., K. Lorenzen, P. Thorbek, J. R. Wheeler, and C. R. Tyler. 2012. Density dependent processes in the life history of fishes: evidence from laboratory populations of zebrafish Danio rerio. PLOS One 7 e37550. doi: 10.1371/z0037550. Ham, K. D. and T. N. Pe arsons. 2001. A Practical Approach for Containing Ecological Risks Associated with Fish Stocking Programs Fisheries 26: 15 23. Harter, R. 2011. 2010 Florida motor gasoline and diesel fuel report. Florida Department of Agriculture and Consumer Services. http://www.freshfromflorida.com/offices/energy/docs/2010_Gasoline_Diesel_Rep ort.pdf Hilborn, R. 1998. The economic performance of marine stock enhancement . Bulletin of Ma r ine Science 62: 661 674. Hilborn R. 2007. Defining success in fisheries and conflicts in objectives. Marine Policy 31:153 158. Holling, C. S. 1978. Adaptive environmental assessment and management. Blackburn press, New Jersey. Holling, C.S. 1998. Two cultures of ecology. Conservation Ecology 22: 4. Holling, C.S., and G.K. Meffe. 1996. Command and control and the pathology of natural resource management. Conservation Biology 102: 328 337. Holling, C. S., F. Berkes, and C. Folke. 1998. Science , sustainability and resource management. Pages 342 362 in F. Berkes and C. Folke, editors. Linking social and ecological systems: management practices and social mechanisms for building resilience. Cambridge University Press, Cambridge U.K. Holling, C. S. 2004. From complex regions to complex worlds. Ecology and Society 91:11. http://www.ecologyandsociety.org/vol9/iss1/art11
167 Holling, C. S., L. H. Gunderson, and G. D. Peterson. 2002. Sustainab ility and panarchies. Pages 63 102 in L. H. Gunderson and C. S. Holling, editors. Panarchy: understanding transformations in human and natural systems. Island Press, Washington, D.C., USA. Hunt, L. M. 2005. Recreational fishing site choice models: Insights and future opportunities. Human Dimensions of W ildlife 10: 153 172. Ihde T.F., M.J. Wilberg, D.A. L oewensteiner, D.H. Secor and T.J. Miller. 2011. The increasing importance of marine recreational fishing in the US: Challenges for management. Fisheries Res earch 108: 268 276. Johannes R.E. 1981. Words of the Lagoon: fishing and marine lore in the Palau district of Micronesia. University of California Press, Berkely CA, USA. Johnson B.M. and S.R. Carpenter. 1994. Functional and numerical responses: a framework for fish angler interactions. Ecological Applications 4: 808 821. Johnson, B. M., and P. J. Martinez. 2000. Trophic economics of lake trout management in reservoirs of differing productivity. North American Journal of Fisheries Management 20: 127 143 Johnston F.D., R. Arlinghaus and U. Dieckmann. 2010. Diversity and complexity of angler behavior drive socially optimal input and output regulations in a bioeconomic recrea tional fisheries model. Canadian Journal of Fisheries and Aquatic Sciences 67:1507 1531. Kennedy, G. J. A. and C. D. Strange. 1986. The effects of intra and inter specific competition on the survival and growth of stocked juvenile Atlantic Salmon Salmo sa lar L., and resident trout Salmo trutta L., in an upland stream. Journal of Fish Biology 28: 479 489. Kery, M., 2010. Introduction to WinBUGS for ecologists. Elsevier Publishers, Amsterdam. Kiker G.A., T.S. Brigdges, A. Varghese, T.P. Seager and I. Linkov . 2005. Applications of multi criteria decision analysis in environmental decision making. Integrated Environmental Assessment and Management 1:95 108. Koe h n J. 2010. Conservation and utilization: harnessing forces for better outcomes. Ecological Mana gement and Restoration 11: 86 87 Leber, K. M. 2002. Ad v ances in marine stock enhancement: Shifting emphasis to theory and accountabilit y . Pages 79 90 in R. R. Stickn e y , and J. P . Mc V e y, editors. Responsible Marine Aquacultu r e, N e w Y ork: CABI Publishing.
168 Le ber, K. M., H. L. Blan k enship, S. M. Arce, and N. P . Brennan. 1997. Influence of release season on size dependent surv i v al of cultured striped mullet, Mugil cephalus, in a H a w aiian estuary. F isheries Bulletin 95: 267 279. Leber, K. M., N. P . Brennan, and S . M. Arce. 1998. Recruitment patterns of ju v enile, cultured P acific threadfin, P olydactylus s e xfilis Polynemidae, released along sandy marine shores in H a w aii. Bulletins Ma r ine Science 62: 389 408. Leber, K. M., R. N. Cantrell, and P . S. Leung. 2005. Optim izing cost e f fect i v eness of size at release in stock enhancement programs. North American Journal of Fisheries Management 25: 1596 1608. Lester, S.E, C. Costello, B.S. Halpern, S.D. Gaines, C. White and J.A. Barth. 2013. Evaluating tradeoffs among ecosyste m services to inform marine spatial planning. Marine Policy 38: 80 89. L e vin, P . S., and G. W . Stunz. 2005. Habitat triage for e xploited fishes: Can I identify essential Estuarin e , Coastal and Shelf Science 64: 70 78. Lewin W., R. Arlinghaus and T. Mehner. 2006. Documented and potential biological impacts of recreational fisheries: insights for management and conservation. Reviews in Fisheries 14: 305 367. Loomis J. and P. Fix. 1998. Testing the importance of fish stocking as a det erminant of the demand for fishing licenses and fishing effort in Colorado. Human Dimensions of Wildlife: An International Journal 3: 46 61. Lorenzen, K. 1996. The relationship between body weight and natural mortality in fish: a comparison of natural ecos ystems and aquaculture. Journal of Fish Biology 49: 627 647 Lorenzen, K. 2005. Population dynamics and potential of fisheries stock en hancement: Practical theory for assessment and poli c y analysis . P hilosophical T ransactions of the R oyal Soc iety of London B 360: 171 189. Lorenzen , K. 2008. Understanding and managing enhancement fisheries systems. Reviews in fisheries Science 16: 10:23. Lorenzen , K., K.M. Leber and H.L. Blankenship. 2010. Responsible approach to marine stock enhancement: an update. Reviews in Fisheries Science 18: 198 210 Lorenzen, K., M.C.M. B e v eridge, and M. Mangel. 2012. Cultured fish: Integrat i v e biology and management of domestication and interactions with wild fish . B iological Rev iews of the Cambridge Philosophical Society 87: 639 660 . Ludwig, D. 2001. The era of management is over. Ecosystems 4: 758 764.
169 Markowitz, H. 1959. Portfolio Selection: Efficient Diversification of Investments. New York: John Wiley and Sons. Martin, J., Runge, M. C., Nichols, J. D., Lubow, B. C., and Kendall, W. L. 2009. Structured decision making as a conceptual framework to identify thresholds for conservation and management. Ecological Applications 19: 1079 1090. McEacheron L. W. and K. Daniels. 1995. Red Drum in Texas: A success story in partnership and com mitment. Fisheries 20: 6 8. McEacheron, L. W., R. C. Colura, B. W. Bumguardner, and R. Wood. 1998. Survival of stocked Red Drum in Texas. Bulletin of Marine Science 62: 359 368. McNie E.C. 2007. Reconciling the supply of scientific information with user de mands: an analysis of the problem and review of the literature. Environmental Science and Policy 10: 17 38. Melnychuk, M. C., V. Christensen, and C. J. Walters. 2013. Meso scale movement and mortality patterns of juvenile Coho salmon and steelhead trout mi grating through a coastal fjord. Environmental Biology of Fishes 96: 325 339. Miller, T . J., J. A. Blair, R. M. Ihde, D. H. Secor, and M. J. W ilbe r g. 2010. FishSmart: An inn o v at i v e role for science in sta k eholder centered approaches to fisheries man agement. F isheries 35: 424 433. Molony, B. W., R. Lenanton, G. Jackson, and J. Norriss. 2003. Stock enhancement as a fisheries management tool. Reviews in Fish Biology and Fisheries 13: 409 432. Morgan M . G . , and H. Dowlatabadi . 1996. Learning from integra ted assessment of climate change. Climate Change 34: 337 68. Muller R.G and R.G. Taylor. 2012. A stock assessment update of common snook, Centropomous undecimalis, in Florida through 2007. Report to the Florida Fish and Wildlife Conservation Commission fr om the Fish and Wildlife Research Institute, St. Petersburg. FWRI In House Report 2012 002. Muone k e, M. I., and W . M. Childress. 1994. Hooking mortality: a r e vi e w for recreational fisheries. R e views in F isheries Science 2: 123 156. Murphy, M. D. and R. G. Taylor. 1990. Reproduction, growth, and mortality of Red Drum Sciaenops ocellatus , in Florida waters. Fishery Bulletin 88: 531 542. Murphy, M. D., and R. E. Crabtree. 2001. Changes in the age structure of nearshore adult Red Drum o f f west central Florida r elated to recruitment and fishing mortalit y . North American Journal of Fisheries Management 21: 671 678. Murphy M. D., R.G. Muller and K. Guindon. 2008. A stock assessment for Pompano, Trachinotus carolinus, in Florida through 2007. Report to the Florida F ish and
170 Wildlife Conservation Commission from the Fish and Wildlife Research Institute, St. Petersburg. FWRI In House Report 2008 004. Murphy, M. D. and J. Muyandorero. 2009. A sto c k assessment of Red Drum , Sciaenops ocellatus in Florida: Status of the sto c ks th r ough 2007. Report to the Florida Fish and W ildlife Conser v ation Commission from the Fish and W ildlife Research Institute, St. Peters b u r g. FWRI In House Report 2008 008. Murphy M. D., D. Chagaris and D. Adis. 2011. An assessment of the status of sp otted seatrout in Florida waters. Report to the Florida Fish and Wildlife Conservation Commission from the Fish and Wildlife Research Institute, St. Petersburg. FWRI In House Report 2011 002. Naish, K. A., Taylor III, J. E., Levin, P. S., Quinn, T. P., Wi nton, J. R., Huppert, D., and Hilborn, R. 2007. An evaluation of the effects of conservation and fishery enhancement hatcheries on wild populations of salmon. Advances in Marine Biology 53: 61 194. National Marine Fisheries Service NMFS. 2010. Fisheries Ec onomics of the United States, 2009. U.S. Dept. Commerce, NOAA Tech. Memo. NMFS F/SPO 118, 172p. Available at: https://www.st.nmfs.noaa.gov/st5/publication/index.html . National Marine F isheries Service NMFS. 2013. Personal communication from the National Marine Fisheries Service, Fisheries Statistics Division . April 20 2013. National Oceanic and Atmospheric Administration NOAA. 2013. Hurricane research division. http://www.aoml.noaa.gov/hrd/tcfaq/E11.html National Oceanic and Atmospheric Administration NOAA. 2013. National climatic data center. http://www.ncdc.noaa.gov/ and http://climatecenter.fsu.edu/products services/data/statewide averages/precipitation Nicholson E. and H.P. Possingham. 2006. Objectives for multiple species conserva tion planning. Conservation Biology 20: 871 881. Oa k erson, R. 1992. Analyzing the commons: A fram e w ork. Pages 41 59 in D. W. Bromley and R. Feeny, editors. Making the Commons W ork: Theor y , P r actice and P olicy . San Francisco, CA: ICS. Oh, C. O., and R. B. D itton. 2006. Specialization di f ferences in for Red Drum Sciaenops ocellatus har v est r e gulations. P r oceedings of the Gulf and Carib b ean F isheries Inst itute 57: 869 880. Ostrom, E. 1990. G o verning the Commons: The Evolution of Institutio ns for Collective Action. Cambridge: Cambridge Un i v ersity Press . Ostrom, E., J. Burger, C. B. Field, R. B. Norgaard, and D. Policansky. 1999. R e visiting the commons: Local lessons, global challenges. Science 284: 278 282.
171 Ostrom, E. 2009. A general framew ork for analyzing sustainability of social ecological systems. Science 325: 419 422. Pareto, V. 1971. Manual of Political Economy. New York: A. M. Kelley. Pereira D.L and M.J. Hansen. 2003. A Perspective on Challenges to Recreational Fisheries Management : Summary of the Symposium on Active Management of Recreational Fisheries, North American Journal of Fisheries Management, 23: 1276 1282. Peters, K. M., and R. G. McMichael. 1987. Early life history of Sciaenops ocellatus Pisces: Sciaenidae in Tampa Bay, F lorida. Estuaries 10: 92 107. Petrosky, C. E. and T. C. Bjornn. 1988. Response of wild rainbow Salmo gairdweri and cutthroat trout S. clarkito stocked rainbow trout in fertile and infertile streams. Canadian Journal of Fisheries and Aquatic Sciences 45: 20 87 2105. Pinheiro, J. C. and D. M. Bates. 2000. Mixed Effects Models in S and S PLUS, Springer, New York. Pin k erton, E. 1994. Economic and management benefits from the coo r dination of capture and culture fisheries: The case of Prince W illiam Sound pink sal mon. North American Journal of Fisheries Management 14: 262 277. Pinkerton E. 2009. Coastal marine systems: conserving fish and sustaining community livelihoods with co management. Pages 2 41 257 in F. S. Chapin III, G. P. Kofinas and C. Folke, editors. Pri nciples of Ecosystem Stewardships. Springer Science and Buisness Media, New York, NY, USA. Possingham, H. P., S. J. Andelman, B. R. Noon, S. Trombulak, and H. R. Pulliam. 2001. Making smart conservation decisions. Pages 225 244 in G. Orians and M. SoulÃ©, e ditors. Research priorities for conservation biology. Island Press, Washington , D.C. Post J.R., M. Sullivan, S. Cox, N.P. Lester, C.J. Walters, E.A. Parkinson, A.J. Paul, L. colla pse? Fisheries 27: 6 17. Post J.R., L. Persson, E.A. Parkinson and T. van Kooten. 2008. Angler numerical response across landscapes and the collapse of freshwater fisheries. Ecological Applications 18: 1038 1049. Propst, D. B. and Gavrilis, D. G. 1987. Ro le of economic impact assessment procedures in recreational fisheries management. Transactions of the American Fisheries Society 116: 450 460.
172 Purtlebaugh, C. H., and M. S. Allen. 2010. Relat i v e a b undance, gr o wth and mortality of fi v e age 0 estuarine fishe s in relation to discha r ge of the Su w annee R i v e r , Florida. T r ansactions of the American F isheries Society 139: 1233 1246. R e yier, E. A., R. H. L o wers, D. M. Scheidt, and D. H. Adams. 2011. M ov ement patterns of adult Red Drum Sciaenops ocellatus, in shall o w Florida lagoons as inferred through autonomous acoustic telemetr y . Environmental Biology of Fishes 90: 343 360. Rogers, M. W., M. S. Allen, P. Brown, T. Hunt, W. Fulton, and B. A. Ingram. 2010. A simulation model to explore the relative value of stock enh ancement versus harvest regulations for fishery sustainability. Ecological Modelling 221: 919 926. Roo k er, T . R., S. A. Holt, G. J. Holt, and L. A. Fuiman. 1999. Spatial and temporal v ariability in gr o wth, mortality and recruitment potential of post settle ment Red Drum , Sciaenops ocellatus, in a subtropical estuary. Fishery Bulletin N O AA, 97: 581 590. Roo k er, J. R., G. W . Stunz, S. A. Holt, and T . J. Minello. 2010. Population connectivity of Red Drum in the Northern Gulf of Mexico. Marine Ecology Progress S eries 407: 187 196. Sanchirico, J.N., M.D. Smith and D.W. Lipton. 2008. An empirical approach to ecosystem based fishery management. Ecological Economics 64: 586 596. Scharf, F . S. 2000. Patterns in abundance, growth and mortality of juvenile Red Drum acro ss estuaries in Texas coast with implications for recruitment and stock enhancement. Transactions of the American Fisheries Society, 129 : 1207 1222. Schindler, D.E., R. Hilborn, B. Chasco, C.P. Boatright, T.P. Quinn, L.A. Rogers and M.S. Webster. 2010. Pop ulation diversity and the portfolio effect in an exploited species. Nature 465: 609 613. Schuhmann P.W. 1998. Deriving species specific benefits measures for expected catch improvements in a random utility framework. Marine Resource Economics 13: 1 21. Sc huhmann P.W. and K.A. Schwabe. 2004. An analysis of congestion meaures and heterogeneous angler preferences in a random utility model of recreational fishing. Environmental and Resource Economics 27: 429 450 Serafy, J. E., J. S. Ault, T . R. Capo, and D. R. Schultz. 1999. Red Drum , Sciaenops ocellatus L., stock enhancement in Biscayne Ba y , FL, USA: Assessment of releasing unmar k ed early ju v eniles. Aquaculture Research 30: 737 750. Shea, K. 1998. Management of populations in conservation, harvesting and control. Trends in Acology and Evolution 3 9: 371 375.
173 Shea, K., Amarasekare, P., Mangel, M., Moore, J., Murdoch, W. W., Noonburg, E., Parma, A., Pascual, M. A., Possingham, H. P., Wilcox, C. and Yu, D. 1998. Management of populations in conservation, harve sting and control. Tre nds in Ecology and Evolution 13: 371 375. Sher w ood, E. T, D. J. Murie, and D. C. P ar k yn. 2004. Post release rate of loss of ju v enile Red Drum stoc k ed out of season in the Chassah o witzka National W ildlife Refuge, Florida. North America n Journal of Fisheries Management, 24: 1469 1479. Starfield A.M. 1997. A pragmatic approach to modeling for wildlife management. Journal of Wildlife Management 61:261 270. St e w art, C. B., and F . S. Scharf. 2008. Estuarine recruitment, gr o wth, and first year surv i v al of age 0 ju v enile Red Drum Sciaenops ocellatus in North Carolina. Transactions of the American Fisheries Society 137: 1089 1103 . Stunz, G. W ., and T . J. Minello. 2001. Habitat related predation on ju v enile wild caught and hatchery reared Red Drum Sciaenops ocellatus Linnaeus. Journal of Experimental Marine Biology and Ecology , 260: 13 25. Sutton S.G. and R.B. Ditton. 2005. The substitutability of one type of fishing for another. North American Journal of Fisheries Management 25: 546 546. Sy lvia, G., and D. Cai. 2005. Generating Policy Information for Fisheries Management: A Comparison of Alternative Approaches. Marine Resource Economics 101: 77 91. Taylor M. D., P. J. Palmer, D. S. Fielder and I. M. Suthers. 2005. Responsible estuarine fin fish stock enhancement: an Australian perspective. Journal of Fish Biology 67: 299 331. Thompson Rueters/University of Michigan. 2013. Estimated Consumer Sentiment Index. http://thomsonreuters.com/products_services/financial/financial_products/a z/umichigan_surveys_of_consumers/ T ringali, M. D. and 12 other authors. 2008a. Genetic policy for the r elease of finfishes in Florida. Flor ida Fish and W ildlife Conser v ation Commission, March 2007. Florida Fish and W ildlife Research Institute Publication No. IHR 2007 1. T ringali, M. D., K. M. Leber, W . G. Halstead, R. McMichael, J. W inner, R. Cody, C. Y oung, C. Neidig, H. W olfe, A. F orstchen, and L. Barbieri. 2008 b . Marine stock enhancement in Florida: A multidisciplinar y , stakeholder supported, accountability based approach. R e v iews in F isheries Science. 16: 51 57.
174 United States Energy Information Administration US EIA. 2010. Real gross domestic product by state, Appendix D. State energy data 2010: Consumption. http://www.eia.gov/state/seds/sep_use/notes/use_GDP.pdf van Poorten B. T., R. Arlinghaus, K. Daedlo w and S. S. Haertel Borer. 2011. Socioal ecological interactions, management panaceas, and the future of wild fish populations. Proceedings of the National Academy of Science 108: 12544 12559. Walters, C. J. 1986. Adaptive management of renewable resour ces. New York, NY: Macmillan Publishing Company. Walters C.J. and Holling, C. S. 1990. Large scale management experiments and learning by doing. Ecology 71: 2060 2068. Walters, C. and Korman, J. 1999. Linking recruitment to trophic factors: revisiting the Beverton Holt recruitment model from a life history and multispecies perspective. Reviews in Fish Biology and Fisheries 9: 187 202. Walters, C. J., and F. Juanes. 1993. Recruitment limitation as a consequence of optimal risk taking beh a vi our by ju v enile fi sh. Canadian J ournal of F isheries and Aquatic Science. 50: 2058 2070. Walters C.J. and R. Ahrens. 2009. Oceans and estuaries: managing the commons. Pages 221 240 in F. S. Chapin III, G. P. Kofinas and C. Folke, editors. Principles of Ecosystem Stewardshi ps. Springer Science and Buisness Media, New York, NY, USA. Walters,C.J and S. J. D. Martell . 2004. Fisheries Ecology and Management. Princeton University Press, Princeton, New Jersey. Wallerstein, I. M. 1993. World system vs. worldsystems, a critique. Pages 292 296 in A. G. Frank and B. K. Gills, editors. The world system: five hundred years or five thousand. Routledge, London, UK. Washington, P. M., and A. M. Koziol. 1993. Overview of the interactions and environmental impacts of hatchery practices on natural and artificial stocks of salmonids. Fisheries Research 18: 105 122. Weithman, A.S. 1999. Socioeconomic benefits of fisheries. Pages 193 213 in C. C. Kohler and D. D. Hubert, editors. Inland Fisheries Management in North America Bethesda, MD: Amer ican Fisheries Society. W illis, S. A., W . W . F alls, C. W . Dennis, D. E. Roberts, and P . G. Whitchurch. 1995. Assessment of season of release and size at release on recapture rates of hatchery reared Red Drum . Pages 354 365 in H. L. Schramm, and R. G. Pi per, editors. Uses and E f fects of Cultu r ed F ishes in Aquatic Ecosystems American Fisheries Society Symposium 15, Bethesda, MD.
175 Young, C. Project Tampa Bay fish production costs. Unpublished data.
176 BIOGRAPHICAL SKETCH Edward Camp was born in Malden, Mass achusetts to Peter and Marcia Camp. He was homeschooled by his parents throughout grade school and was encouraged to explore and study the natural world. In 2001 he graduated from Whitinsville Christian High School and entered Gordon College in Wenham Ma ssachusetts. At Gordon College , Edward pursued an interdisciplinary academic track by studying communications, ecology, economics, sociology and political sciences, and studied abroad at Washington, USA, New Zealand, and Western Samoa. In May 2005 Edward graduated with a degree in Environmental studies, and following suggestions by his advisor, Dr. Dorothy Boorse, sought to gain field experience studying fish and fisheries. In May 2005, Edward moved to Logan, Utah to begin work as a research assistant st udying Bonneville Cutthroat trout. Over the next several years Edward worked as a research assistant in Utah, Wyoming, Colorado, New York and Florida and researched a variety of ecological and management issues. Over these years Edward hiked, ran, skied and fished some amazing parts of this country. Edward began work on is Masters degree in 2008 under the guidance of Dr. Tom Frazer and Dr. Bill Pine, studying relationships between changing aquatic vegetative habitats and small bodied fish communities in During this time Edward learned to program in R and how to write scientific papers, practiced catching redfish, and met his future wife, Genevieve. He graduated in 2010 a nd moved to Ithaca, New York, where he worked for his friend Dr. Cliff Kraft, where he honed his thinking skills and learned how to hunt.
177 In August 2011, Edward moved back to Gainesville to begin his Doctors of Philosophy with Dr. Kai Lorenzen, where he evaluated the integr ated socioecological implications of stock enhancement of red drum fishery . During this time Edward honed many quantitative and integrative research skills, married Genevieve, and became a pretty good fisherman. The time working on his disserta tion were markedly improved by a group of excellent researchers and friends , including Dan Gwinn, Bryan Matthias , Rob Ahrens, Mike Allen and Carl Walters. Upon the completion of his Doctors of Philosophy degree, Edward plans to continue working with Dr. Kai Lorenzen in a post doctorate position. He hopes to eventually work at the intersection of management agencies and academia, where he believes he will be best able to conduct novel work with exceptional application. Edward hope this work will in some outcomes that yield acceptable current use of natural resources while preserving or improving the resilience of socioecological systems that is necessary for continued use in the future.