Recent studies have demonstrated that data from marine reserves can benefit fisheries management. Marine reserves may improve assessments by acting as a reference area when protected populations approach unfished conditions. This forms the theoretical basis for the recent development of assessment techniques that utilize data from inside marine reserves to assess fished populations out of reserves, and that better inform the selection of management control rules.
In this dissertation I examine how no-take marine reserves impact our ability to assess the status of data-poor fisheries. In my second chapter I review the evolution in scientific thinking on how reserves have been integrated into fisheries management, and describe the emerging research on how reserves may be used as reference areas for the assessment and management of fish stocks. I also examine how the characteristics of marine reserves designed for use as reference areas compare with those used to meet the more traditional goals of conservation or fishery enhancement, and suggest some avenues of future research in this vein.
In my third chapter, I demonstrate how a recently developed data-poor stock assessment method, the Length-based Spawning Potential Ratio estimator relies heavily on correctly assuming biological parameters such as growth and natural mortality, and show how this method can be extended to include information from MPAs to estimate these parameters when this biological information is unavailable.
In my fourth chapter, I compare the performance of a suite of MPA-based data-poor methods, both in the short and long term under a range of different kinds of uncertainty. The results indicate that all assessment methods are sensitive to the time since MPA creation, historical fishing pressure, and movement, but that the methods that rely on length data are more robust to these conditions than those that rely on CPUE data. When paired with a control rule, all of the assessment methods performed reasonably well, suggesting that MPA-based assessment techniques may provide a viable option for the management of sedentary data-poor stocks.
