Method evaluation and risk assessment: A framework for evaluating management strategies for data‐limited fisheries

Abstract

Fisheries managers are in need of quantitative tools to inform decisions regarding selection of robust management practices, prioritising research gaps and stocks to focus on, particularly where there are limited resources or data. To support these decisions, the use of Management Strategy Evaluation (MSE), that is, closed loop simulationtesting of management procedures, is widely regarded as best practice. However, applying MSE is time-and computationally intensive, and requires highly skilled expertise and processes for stakeholder input and peer review. For data-and capacitylimited fisheries, MSE may be particularly challenging to implement. Yet, these are the contexts where it is most critical to test assumptions, evaluate the implications of all sources of uncertainty and identify the most informative data sources. To facilitate wider use of MSE, the Method Evaluation and Risk Assessment (MERA) framework was developed as an accessible online interface, with quick processing time, focused on generic data-limited management procedures, but allowing progression to tailored and more data-rich methods. The framework links a quantitative questionnaire and data input standard to a flexible operating model with optional customisation via command line access to the back-end open-source R libraries. Here, we illustrate a case study application of MERA for the bocinegro (Pagrus pagrus, Sparidae) fishery in the Gulf of Cadiz, where in conjunction with fishery stakeholders, a custom management procedure was developed and tested and key research gaps and data collection priorities were identified. We discuss implications for wider use of MSE in various contexts, including eco-certification and fishery improvement projects.

Rejection sampling and agent-based models for data limited fisheries

“…The conditioning of any operating model relies on the integration of basic fishery statistics of catch, effort and abundance trends. In data-limited fisheries, methods have been developed to infer trends in these statistics for the fishery (Goethel et al, 2019;Dowling et al, 2022;Carruthers et al, 2023). This agentbased/rejection approach is an addition to these methods.…”

“…Rejection/agent-based methods of conditioning and MSEs can be used as an alternative to questionnaires, expert opinion (Dowling et al, 2022;Carruthers et al, 2023). However, ultimately all data-limited methods including rejection/agentbased are based on priors of individuals familiar with the fishery.…”

“…This example demonstrates that the agent-based model can be conditioned with limited data. However, the actual computing time and the basic knowledge of the model required by users make it, perhaps, more difficult to implement than other approaches (Dowling et al, 2022;Carruthers et al, 2023).…”

“…Data-limited fisheries are ubiquitous and often are important components of the fisheries economy (Prince and Hordyk, 2018). Thus, recent research has addressed methods which can be used in data-poor fisheries (Carruthers et al, 2014;Dowling et al, 2022;Carruthers et al, 2023). While inadequacies of management advice resulting from the lack of data are not solved by modeling alone, data-limited MSEs can be designed to appropriately use the data that are available and to organize the results so that they can be cogently provided in terms that are useful for management.…”

“…The approaches are: 1) rejection sampling methods for conditioning operating models for data-limited MSEs and 2) the incorporation of more complex socio-economic dynamics into that conditioning and also into the MSE projections using agent-based models that depict fishery socio-economic behavior of individuals and/or vessels on a daily basis. These methods may be viewed as adding to the continuum of modeling approaches and procedures for providing MSEs for datalimited fisheries (Dowling et al, 2022;Carruthers et al, 2023).…”