Fishery Management Problems Research Articles

Deep reinforcement learning for conservation decisions

Abstract Can machine learning help us make better decisions about a changing planet? In this paper, we illustrate and discuss the potential of a promising corner of machine learning known as deep reinforcement learning (RL) to help tackle the most challenging conservation decision problems. We provide a conceptual and technical introduction to deep RL as well as annotated code so that researchers can adopt, evaluate and extend these approaches. RL explicitly focuses on designing an agent who interacts with an environment that is dynamic and uncertain. Deep RL is the subfield of RL that incorporates deep neural networks into the agent. We train deep RL agents to solve sequential decision‐making problems in setting fisheries quotas and managing ecological tipping points. We show that a deep RL agent is able to learn a nearly optimal solution for the fisheries management problem. For the tipping point problem, we show that a deep RL agent can outperform a sensible rule‐of‐thumb strategy. Our results demonstrate that deep RL has the potential to solve challenging decision problems in conservation. While this potential may be compelling, the challenges involved in successfully deploying RL‐based management to realistic scenarios are formidable—the required expertise and computational cost may place these applications beyond the reach of all but large, international technology firms. Ecologists must establish a better understanding of how these algorithms work and fail if we are to realize this potential and avoid the pitfalls such a transition would bring. We ultimately set forth a research framework based on well‐posed, public challenges so that ecologists and computer scientists can collaborate towards solving hard decision‐making problems in conservation.

Assessing the Uncertainty of Total Seabird Bycatch Estimates Synthesized from Multiple Sources with a Scenario Analysis from the Western and Central Pacific

Each year, billions of seabirds undertake migrations, connecting remote regions of the world, potentially synchronizing population fluctuations among distant areas. This connectedness has implications for the uncertainty calculations of the total seabird bycatch estimate at a regional/global scale. Globally, fisheries bycatch poses a major problem in fishery management, and estimating the uncertainty associated with a regional/global seabird bycatch estimate is important because it characterizes the accuracy and reliability of the fisheries’ impact on the seabird populations. In this study, we evaluate different assumptions underlying the estimation of the variability of the total seabird bycatch at a regional/global scale based on local assessment reports. In addition to theoretical analysis, we also simulate multiple spatially distant separately managed areas with relatively low levels of observer coverage, based on bycatch data from the Western and Central Pacific Fisheries Commission convention area. The results show that assuming a completely synchronized variation produced the most conservative uncertainty estimate and it also missed an opportunity to improve the precision. Simplified correlation structures also failed to capture the complex dynamics of bycatch rates among spatially distant areas. It is recommended to empirically estimate the correlation of bycatch rates between each pair of sources based on bycatch rate time series.

Estimation of conservation unit and population contribution to Chinook salmon mixed-stock fisheries in British Columbia, Canada, using direct DNA sequencing for single nucleotide polymorphisms

Determination of population structure and stock identification is a general problem in fisheries assessment and management. Pacific salmon fishery management regimes are evolving to require higher resolution of stock composition on increasingly smaller reporting units. For Chinook salmon (Oncorhynchus tshawytscha), a stock identification baseline comprised of some 125 198 individuals from 369 populations ranging from Russia to California was employed for genetic stock identification (GSI). GSI analysis based on variation at up to 547 single nucleotide polymorphisms (SNPs) was demonstrated to provide accurate estimates of stock composition for 68 conservation units (CUs) in British Columbia, 23 reporting groups in the United States, and one reporting group in Russia. In many instances, accurate population-specific estimates of stock composition within a CU were possible in fishery samples, as well as identifying individuals to some specific populations. A genetics-based assessment system provides an opportunity for conservation-based management of Canadian Chinook salmon.

In Memoriam

He liked fish. And he liked people who liked fish. In the span of a week, Michael J. Maceina, Professor Emeritus at Auburn University, went from fishing with his family near Mobile Bay to passing peacefully at home in his own bed on June 10, 2020. During his brief battle with prostate cancer he did not stop fishing, he did not stop hunting, and he did not stop spending quality time with family and friends. Maceina earned his fisheries degrees at the University of Florida (BS, MS) and Texas A&M University (PhD). He worked as a senior environmental scientist with the South Florida Water Management District for several years before joining the Auburn University faculty in 1990, where he worked until he retired in 2015. His population dynamics graduate course was legendary. Maceina was an intense and passionate instructor. He never settled for anything less than your best and he expected you to bring your A-game every day. He certainly always brought his. His students benefitted from his energy and drive to solve fishery problems with rigorous data analysis and a critical eye for investigating data trends and patterns. When he wasn’t fishing or hunting with family and friends, Maceina was a dynamic researcher and prolific author, penning 127 peer-reviewed manuscripts and book chapters. His nickname “Iron Mike” may have come from his dealings with reviewers and editors. He also answered to “Dr. N,” a nod towards his advocacy for large sample sizes and resultant statistical power. Mike and his collaborators made seminal contributions in their studies of relationships between trophic states and fisheries, harvest regulations, age and growth of fish, fish habitat requirements, vegetation management, and environmental drivers of fish recruitment. Mike developed the methodology to use residuals from catch-curves to index year class strength and that influential 1997 paper has been widely adopted by freshwater and marine biologists. Yield models have been the bread-and-butter of marine stock assessments for many years, but were less commonly employed in studies of freshwater fisheries before 2000. Maceina and his student coauthor helped change that when they developed Windows-based FAST/FAMS software. Together, they hosted dozens of workshops to teach several generations of freshwater biologists how to use sophisticated yield models to forecast the effects of regulation changes on fish populations. Given his penchant for studying fisheries management problems, it is not surprising that he coauthored the Best Paper of 2016 in the North American Journal of Fisheries Management (10 years earlier, he coauthored a paper that was a finalist for that prestigious award). Mike was inducted into the Fisheries Management Section’s Hall of Excellence in 2017. He was an AFS Life Member and an active participant at state, regional, and AFS Annual Meetings and symposia for 35 years. He was also active in the North American Lake Management Society and the Aquatic Plant Management Society. Maceina always felt a connection to southeastern U.S. ecosystems and the management of aquatic vegetation and nutrients in lakes and reservoirs. Your endeavors were nearly always successful if Maceina was your co-principal investigator or coauthor. If he was your friend, well, you could find no better. If he was your advisor you counted yourself lucky—perhaps not at the exact moment when he was delivering a well-timed “attitude adjustment,” but later you certainly did. Maceina mentored 28 graduate students who fanned out across the USA and the world to continue his work. That achievement is perhaps his greatest legacy as a fisheries scientist. His son Vaughn Trout Maceina, his daughter-in-law Claire, and their two children survive him. His wife Nellie passed away before he did. A memorial endowment in the School of Fisheries, Aquaculture, and Aquatic Sciences at Auburn University has been established. Contributions can be made out and sent to the AU Foundation in Mike’s name at College of Agriculture Development Office, 100 Comer Hall, Auburn, AL 36849-5401.

Me and My Behavior: An Experiment on Individual Characteristics and Compliance Behavior in Recreational Fishing

Non-compliance with regulations is a complex problem in recreational fisheries management, having the potential to evoke uncertainty for conservation and socio-ecological outcomes and to undermine management efforts. While we know that in fisheries people make trade-offs between following or breaking rules, it is of interest to determine how people respond to different management incentives to curtail non-compliance. The overall aim of this study is to examine what individual psycho-social characteristics are associated with responses to instrumental and normative management incentives in a recreational fisheries context through the use of an economic experiment. We examined five psycho-social characteristics, three of which (expectation of behavior of others, social norms, and risk preferences) have separately been explored within the fisheries compliance literature, while two factors (ecological values and personality types) have yet to be explored. While information about these two latter characteristics is limited within the fisheries compliance literature, our results suggest that they are relevant predictors for certain compliance groups across compliance incentives. The findings underline that there is significant heterogeneity in the associations between psycho-social make-up and compliance behaviors. Knowledge of this behavioral relationship can progress fisheries management toward increased innovation by encouraging the management of the individual fisher rather than the average fisher.

Accurate estimation of conservation unit contribution to coho salmon mixed-stock fisheries in British Columbia, Canada, using direct DNA sequencing for single nucleotide polymorphisms

Determination of population structure and stock identification is a ubiquitous problem in fisheries assessment and management. Pacific salmon fishery management regimes are evolving to require higher resolution of stock composition on increasingly smaller reporting units. For coho salmon (Oncorhynchus kisutch), a stock identification baseline composed of some 57 982 individuals from 332 populations ranging from southeast Russia to California was employed for genetic stock identification (GSI). GSI analysis based upon variation at up to 480 single nucleotide polymorphisms (SNPs) was demonstrated to provide accurate estimates of stock composition for 37 conservation units (CU) in British Columbia, 13 reporting groups in the United States, and one reporting group in Russia. In many instances, accurate population-specific estimates of stock composition within a CU were possible in fishery samples, as well as identifying individuals to some specific populations. A genetics-based assessment system provides an opportunity for conservation-based management of Canadian coho salmon.

Assessing the Sustainable Fishery-based Industry in Malaysia Using the Analytic Hierarchy Process (AHP)

The concept of sustainability in the fishing-based industry is to reach and maintain the fishery at the maximum sustainable yield. This principle is important to ensure that the population of fish is maintained without compromising the ability of the future generations to meet their own needs. Inefficient sustainable management could cause the depletion of natural resources, pollutions and natural habitat loss in the future. Therefore, a good sustainable management for fishery should be emphasized so that future generations can enjoy what we have today. However, the fishery management is characterized by multiple conflicting attributes which are complicated to be inspected. In this paper, we would like to investigate the fishery management problem analytically by using the Analytic Hierarchy Process (AHP) method. Considering four criteria and thirteen sub-criteria influencing to the sustainability in the fishery-based industry, the AHP method will analyse and prioritize each criterion and sub-criterion. The BPMSG AHP Online System is used as the tool to process the collected data and the final results show that the protected species interaction (PSI) is the most important sub-criteria towards achieving the sustainable fisheries. This followed by the sub-criteria of profit (14.2%), employment (10.9%) and overfishing (8.90%). These four top sub-criteria must be monitored regularly as it can influence the sustainability of fishery-based industry. The other sub-criteria have their own priorities which should not be neglected as well.

The mitigation hierarchy for sharks: A risk‐based framework for reconciling trade‐offs between shark conservation and fisheries objectives

AbstractSharks and their cartilaginous relatives are one of the world's most threatened species groups. The primary cause is overfishing in targeted and bycatch fisheries. Reductions in fishing mortality are needed to halt shark population declines. However, this requires complex fisheries management decisions, which often entail trade‐offs between conservation objectives and fisheries objectives. We propose the mitigation hierarchy (MH)—a step‐wise precautionary approach for minimizing the impacts of human activity on biodiversity—as a novel framework for supporting these management decisions. We outline a holistic conceptual model for risks to sharks in fisheries, which includes biophysical, operational and socioeconomic considerations. We then demonstrate how this model, in conjunction with the MH, can support risk‐based least cost shark conservation. Through providing examples from real‐world fishery management problems, we illustrate how the MH can be applied to a range of species, fisheries and contexts, and explore some of the opportunities and challenges hereto. Finally, we outline next steps for research and implementation. This is important in the context of increasing international regulation of shark fishing and trade, which must lead to reductions in shark mortality, while managing trade‐offs between conservation objectives and the socioeconomic value of fisheries.

Applying science to pressing conservation needs for penguins.

More than half of the world's 18 penguin species are declining. We, the Steering Committee of the International Union for Conservation of Nature Species Survival Commission Penguin Specialist Group, determined that the penguin species in most critical need of conservation action are African penguin (Spheniscus demersus), Galápagos penguin (Spheniscus mendiculus), and Yellow‐eyed penguin (Megadyptes antipodes). Due to small or rapidly declining populations, these species require immediate scientific collaboration and policy intervention. We also used a pairwise‐ranking approach to prioritize research and conservation needs for all penguins. Among the 12 cross‐taxa research areas we identified, we ranked quantifying population trends, estimating demographic rates, forecasting environmental patterns of change, and improving the knowledge of fisheries interactions as the highest priorities. The highest ranked conservation needs were to enhance marine spatial planning, improve stakeholder engagement, and develop disaster‐management and species‐specific action plans. We concurred that, to improve the translation of science into effective conservation for penguins, the scientific community and funding bodies must recognize the importance of and support long‐term research; research on and conservation of penguins must expand its focus to include the nonbreeding season and juvenile stage; marine reserves must be designed at ecologically appropriate spatial and temporal scales; and communication between scientists and decision makers must be improved with the help of individual scientists and interdisciplinary working groups.

Moving Forward: A Simulation-Based Approach for Solving Dynamic Resource Management Problems

Standard dynamic resource optimization approaches, such as value function iteration, are challenged by problems involving complex uncertainty and a large state space. We extend a solution technique to address these limitations called approximate dynamic programming (ADP). ADP recently emerged in the macroeconomics literature and is novel to bioeconomics. We demonstrate ADP in solving a simple fishery management model under uncertainty to show: the mechanics of ADP in simplest form; the accuracy of ADP; the value of a nonparametric extension; and readily adaptable, non-specialized code. We then demonstrate ADP’s capacity to handle rich bioeconomic problems by solving the fishery management problem subject to four autocorrelated shock processes (governing economic returns and biological dynamics) which entails four sources of stochasticity and five continuous state variables. We find that accounting for multiple autocorrelation has important impacts on harvest policy and generates gains that depend crucially on the structure of harvest cost.

Fishery management under poorly known dynamics

Adapting fishing regulations in a highly uncertain environment remains a complex challenge for managers who have to deal with non-linear dynamics of fish population and harvest levels. In this research, a recent method of stochastic control is adapted to a general fishery management problem under multiples sources of uncertainty related to the dynamics of the fish population and the effect of fishing on its growth. The question is about adjusting permanently the management rule or to hold a fixed policy thus avoiding additional noise. The mathematical problem developed here, though oversimplified, represents an original approach to the fishery management issue inspired by the monetary policy challenge of a central bank (Brainard principle). It assumes that Control Variation Increases the level of Uncertainty (namely CVIU approach) under particular conditions, resulting in preferable inaction regions for managers. We specify these conditions to show that the management of a poorly known fishery is still possible by using a CVIU approach.

Decomposition methods with deep corrections for reinforcement learning

Decomposition methods have been proposed to approximate solutions to large sequential decision making problems. In contexts where an agent interacts with multiple entities, utility decomposition can be used to separate the global objective into local tasks considering each individual entity independently. An arbitrator is then responsible for combining the individual utilities and selecting an action in real time to solve the global problem. Although these techniques can perform well empirically, they rely on strong assumptions of independence between the local tasks and sacrifice the optimality of the global solution. This paper proposes an approach that improves upon such approximate solutions by learning a correction term represented by a neural network. We demonstrate this approach on a fisheries management problem where multiple boats must coordinate to maximize their catch over time as well as on a pedestrian avoidance problem for autonomous driving. In each problem, decomposition methods can scale to multiple boats or pedestrians by using strategies involving one entity. We verify empirically that the proposed correction method significantly improves the decomposition method and outperforms a policy trained on the full scale problem without utility decomposition.

The influence of nudges on compliance behaviour in recreational fisheries: a laboratory experiment

Abstract Non-compliance is a tenacious problem in recreational fisheries management, posing a risk to marine conservation and socio-ecological systems by, for example, undermining management efforts and creating conflict between resource user groups. In fisheries management, deterrence-based approaches have traditionally been used to tackle non-compliance. However, enforcement is often limited in recreational fisheries and an alternative approach is needed to improve compliance. In this paper, we explore the lessons from behavioural economics and apply nudge theory as the basis of alternative management approaches. Nudge theory argues that through positive reinforcement or indirect suggestion, voluntary compliance can be achieved. We test the influence of a nudge, based on a descriptive social norm, through an economic laboratory experiment in a recreational fisheries context. Our results show that the presence of this nudge can increase compliance behaviour by 10%. We find that a nudge was more effective when deterrence is low, but its effects become weaker when deterrence is already high. We also find heterogeneity across individual’s responses to the nudge and risk preferences significantly related to compliance behaviour. Nudges based on social norms have the potential to complement traditional deterrence methods and could prove successful as a cost-effective compliance tool in the marine environment.

Exploring the legal status and key features of ecosystem‐based fisheries management in international fisheries law

Ecosystem‐based fisheries management (EBFM) has been considered to be a solution to the multifarious problems of fisheries management in areas within and beyond national jurisdictions. However, the literature has introduced different versions of EBFM and there are controversies among commentators concerning the legal status of EBFM in international fisheries law. This article seeks to examine the legal status of EBFM. It also explores the essential features that an EBFM model should incorporate to function effectively. The article argues that the implementation of EBFM has been gaining ground as a legal obligation in international fisheries law.

From Infinite to Finite Programs: Explicit Error Bounds with Applications to Approximate Dynamic Programming

We consider linear programming (LP) problems in infinite dimensional spaces that are in general computationally intractable. Under suitable assumptions, we develop an approximation bridge from the infinite dimensional LP to tractable finite convex programs in which the performance of the approximation is quantified explicitly. To this end, we adopt the recent developments in two areas of randomized optimization and first-order methods, leading to a priori as well as a posteriori performance guarantees. We illustrate the generality and implications of our theoretical results in the special case of the long-run average cost and discounted cost optimal control problems in the context of Markov decision processes on Borel spaces. The applicability of the theoretical results is demonstrated through a fisheries management problem.

Stabilization of an uncertain simple fishery management game

This note analyzes in a fishery management problem the effects of relaxing one of the usual assumptions in the literature of dynamic games. Specifically, the assumption that players restrict to strategies that stabilize the system. Previous works in the literature have shown that feedback Nash equilibria can exist in which a player can improve unilaterally by choosing a feedback control for which the closed-loop system is unstable. This paper considers in some more detail the implication this setting has in the framework of a simple fishery management. It is shown that if the fishermen are not short-sighted in their valuation of future profits, the considered approach implies a division of them into two groups. One group going for maximal profits and the other group taking care of the imposed stability constraint.To see how noise might impact these results we additionally consider a framework where fishermen take into account the possibility that the fish growth might be corrupted by external factors. We consider a deterministic approach of dealing with noise in this set-up.The model predicts that the more people are involved in the group taking care of the stabilization constraint, the less active they get. Furthermore it predicts that the natural reaction of any of these persons is to increase his activities if he expects more noise. But that this activity is reduced, and partly replaced by more active control policies by group members, if the size of the group increases. Activity of fishermen going for maximal profits is not affected by noise expectations.

Stabilization of an Uncertain Simple Fishery Management Game

This note analyzes in a fishery management problem the effects of relaxing one of the usual assumptions in the literature of dynamic games. Specifically, the assumption that players restrict to strategies that stabilize the system. Previous works in the literature have shown that feedback Nash equilibria can exist in which a player can improve unilaterally by choosing a feedback control for which the closed-loop system is unstable. This paper considers in some more detail the implication this setting has in the framework of a simple fishery management. It is shown that if the fishermen are not short-sighted in their valuation of future profits, the considered approach implies a division of them into two groups. One group going for maximal profits and the other group taking care of the imposed stability constraint. To see how noise might impact these results we additionally consider a framework where fishermen take into account the possibility that the fish growth might be corrupted by external factors. We consider a deterministic approach of dealing with noise in this set-up. The model predicts that the more people are involved in the group taking care of the stabilization constraint, the less active they get. Furthermore it predicts that the natural reaction of any of these persons is to increase his activities if he expects more noise. But that this activity is reduced, and partly replaced by more active control policies by group members, if the size of the group increases. Activity of fishermen going for maximal profits is not affected by noise expectations.

Optimal management of a stochastically varying population when policy adjustment is costly

Ecological systems are dynamic and policies to manage them need to respond to that variation. However, policy adjustments will sometimes be costly, which means that fine-tuning a policy to track variability in the environment very tightly will only sometimes be worthwhile. We use a classic fisheries management problem, how to manage a stochastically varying population using annually varying quotas in order to maximize profit, to examine how costs of policy adjustment change optimal management recommendations. Costs of policy adjustment (changes in fishing quotas through time) could take different forms. For example, these costs may respond to the size of the change being implemented, or there could be a fixed cost any time a quota change is made. We show how different forms of policy costs have contrasting implications for optimal policies. Though it is frequently assumed that costs to adjusting policies will dampen variation in the policy, we show that certain cost structures can actually increase variation through time. We further show that failing to account for adjustment costs has a consistently worse economic impact than would assuming these costs are present when they are not.

Accelerating viability kernel computation with CUDA architecture: application to bycatch fishery management

Computing a viability kernel consumes time and memory resources which increase exponentially with the dimension of the problem. This curse of dimensionality strongly limits the applicability of this approach, otherwise promising. We report here an attempt to tackle this problem with Graphics Processing Units (GPU). We design and implement a version of the viability kernel algorithm suitable for General Purpose GPU (GPGPU) computing using Nvidia’s architecture, CUDA (Computing Unified Device Architecture). Different parts of the algorithm are parallelized on the GPU device and we test the algorithm on a dynamical system of theoretical population growth. We study computing time gains as a function of the number of dimensions and the accuracy of the grid covering the state space. The speed factor reaches up to 20 with the GPU version compared to the Central Processing Unit (CPU) version, making the approach more applicable to problems in 4 to 7 dimensions. We use the GPU version of the algorithm to compute viability kernel of bycatch fishery management problems up to 6 dimensions.

Conditions for the Solvability of the Linear Programming Formulation for Constrained Discounted Markov Decision Processes

We consider a discrete-time constrained discounted Markov decision process (MDP) with Borel state and action spaces, compact action sets, and lower semi-continuous cost functions. We introduce a set of hypotheses related to a positive weight function which allow us to consider cost functions that might not be bounded below by a constant, and which imply the solvability of the linear programming formulation of the constrained MDP. In particular, we establish the existence of a constrained optimal stationary policy. Our results are illustrated with an application to a fishery management problem.