Decision analysis for designing marine protected areas for multiple species with uncertain fishery status

Marine protected areas (MPAs) are growing in popularity as a conservation tool, and there are increasing calls for additional MPAs. Meta-analyses indicate that most MPAs successfully meet the minimal goal of increasing biomass inside the MPA, while some do not, leaving open the important question of what makes MPAs successful. An oftenoverlooked aspect of this problem is that the success of fishery management outside MPA boundaries (i.e., whether a population is overfished) affects how well MPAs meet both conservation goals (e.g., increased biomass) and economic goals (e.g., minimal negative effects on fishery yield). Using a simple example of a system with homogeneous habitat and periodically spaced MPAs, we show that, as area in MPAs increases, (1) conservation value (biomass) may initially be zero, implying no benefit, then at some point increases monotonically; and (2) fishery yield may be zero, then increases monotonically to a maximum beyond which further increase in MPA area causes yield to decline. Importantly, the points at which these changes in slope occur vary among species and depend on management outside MPAs. Decision makers considering the effects of a potential system of MPAs on multiple species are confronted by a number of such cost–benefit curves, and it is usually impossible to maximize benefits and minimize costs for all species. Moreover, the precise shape of each curve is unknown due to uncertainty regarding the fishery status of each species. Here we describe a decision-analytic approach that incorporates existing information on fishery stock status to present decision makers with the range of likely outcomes of MPA implementation. To summarize results from many species whose overfishing status is uncertain, our decisionanalysis approach involves weighted averages over both overfishing uncertainty and species. In an example from an MPA decision process in California, USA, an optimistic projection of future fishery management success led to recommendation of fewer and smaller MPAs than that derived from a more pessimistic projection of future management success. This example illustrates how information on fishery status can be used to project potential outcomes of MPA implementation within a decision analysis framework and highlights the need for better population information.