Simulation-Based Evaluation of Assessment Approaches for Intermixing Fish Populations

Fish from distinct spawning populations often occupy the same location when harvested, and performance of assessment approaches in light of mixing requires further study. We used MSE-type simulations to explore the performance of three different statistical catch-at-age approaches (separate, pooled, overlap) under differing productivity, mixing, and harvest levels. The “separate” approach ignored intermixing and treated mixed populations in discrete regions as stocks; the “pooled” approach lumped all populations together and assessed them as a single stock: the “overlap” approach estimated population dynamics of individual spawning stocks when supplied with correct movement rates.  Simulations were based on populations of economically valuable walleye, lake trout, and lake whitefish in the Laurentian Great Lakes region of North America, species for which distinct spawning populations intermix during harvest.  Bias in spawning stock biomass (SSB) estimates increased as intermixing increased for the separate assessments, whereas pooled and overlap assessments became nearly unbiased.  The overlap method frequently failed to provide converged estimates for high movement rates, but when it converged it provided the best fishery management results in terms of sustaining high relatively stable yields and avoiding depletion of SSB.  Differences in results among the species will be highlighted during the presentation.