Evaluating the performance of data-moderate and catch-only assessment methods for U.S. west coast groundfish

Abstract

The estimation of sustainable harvest limits for stocks that have never been assessed and have limited data available can be challenging. Harvest limits for previously un-assessed U.S. west coast groundfish have been set using catch-only methods (Depletion Corrected Average Catch [DCAC] and Depletion-Based Stock Reduction Analysis [DB-SRA]) for data-limited stocks, as well as catch and index based methods (Extended Depletion-Based Stock Reduction Analysis [XDB-SRA] and Extended Simple Stock Synthesis [XSSS]) for data-moderate stocks. To account for uncertainty and to prevent overfishing the harvest levels for U.S. west coast groundfish are reduced based upon the estimation method and the amount of data used. A management strategy evaluation was applied to evaluate the performance of each estimation method to provide benchmark harvest levels for two life-history types (U.S. west coast flatfish and rockfish) under varying misspecifications of the parameter distributions used by these methods. Both, data-moderate and catch-only (data-limited) methods, resulted in overfishing>0.50 (except XSSS in select scenarios) when the simulated stock was below the relative biomass target for both life-histories. Each of the data-moderate methods (XDB-SRA and XSSS) that applied biomass index data failed to estimate the correct stock status in the first assessment when an overly optimistic prior distribution about the stock status was assumed. However, during the projection period as the biomass index lengthened, the estimates of current stock size improved for both of these estimation methods, reducing the probability of overfishing to <0.50 (except XSSS for one scenario). The ability to incorporate index data by the data-moderate methods resulted in improved estimates, as the data became more informative, for stock status and the subsequent harvest limits, that when reduced to account for uncertainty resulted in population stock sizes that either remain stable or rebuild toward the biomass targets for both life-histories. A notable exception was the performance of XDB-SRA for the flatfish life-history when the stock was at the target level and the prior was assumed correctly. In this instance the index data were non-informative, resulting in overfishing due to overly optimistic estimates of relative population size.