| Summary: | Dissertation (Ph.D.) University of Alaska Fairbanks, 2012 A multispecies age-structured assessment model (MSASA) for the Gulf of Alaska (GOA) is developed to examine the effects of integrating predation mortality into stock assessment efforts. Age-specific predation mortality is modeled as a flexible function of predator and prey abundances, constructed from species-preference and size-preference parameters and fitted to stomach-content data. Modeled species include arrowtooth flounder ( Atheresthes stomias), Pacific cod (Gadus macrocephalus), walleye pollock (Theragra chalcogramma), Pacific halibut (Hippoglossus stenolepis) and Steller sea lion ( Eumatopias jubatus). Recruitment, residual natural mortality, full-recruitment fishing mortality, and fishery/survey selectivities are estimated for pollock, cod, and flounder; abundances for apex predators sea lions and halibut are input. Estimated trophic structures and predation links show significant changes as a result of the inclusion of higher trophic level predators, and model results are highly sensitive to assumptions regarding sea lion diet. Simulation exercises suggest that model performance degrades more due to model misspecification and data scarcity than assumptions regarding data weighting and variance. Estimates of predation mortality work in tandem with survey data, constraining predation estimates in the face of incomplete diet data and potentially improving estimates of cohort structure. Exploration of predator functional responses (PFR) shows the default GOA MSASA Holling Type II PFR to be more flexible than initially thought, and that explicitly modeling predator competition for the same prey can improve model fit to stomach-content data. Median parameter estimates and their respective variances from the fitted MSASA model are used to construct management strategy simulations. Reducing fishing pressure on pollock during periods of high predator biomass is less effective at preserving pollock stocks than raising fishing pressure on flounder, and ...
|
|---|