Data Needs for Spatially Explicit Stock Assessments: A Simulation Study Using Stock Synthesis

LaTreese Denson , Fisheries and Wildlife, Oregon State University, Corvallis, OR
David Sampson , Coastal Oregon Marine Experiment Station, Oregon State University, Newport, OR
Environmentally driven dispersal of larvae can result in spatial settlement patterns of recruits. If there is limited movement of post-settlement fish and non-uniform spatial harvesting, important spatial age-structure can develop within a stock. Stocks of nearshore rockfish (Sebastes spp.) often have regionally variable age-compositions, indicative of different histories of recruitment or exploitation. To assess these stocks, data typically are aggregated across subpopulations. A variety of assessment data and model configurations (MCs) can be used to estimate stock status. Using an inappropriate MC can introduce bias and imprecision into assessment estimates. This study uses an operating model (OM) and simulated data in Stock Synthesis to assess the bias and imprecision in estimates of virgin and current spawning biomass and depletion. The experimental design includes the following factors: (a) use of survey data and environmental indices in the MC; (b) using disaggregated (2 regions as generated by the OM) or aggregated data (1 region); and (c) incorporating different patterns in the OM’s regional exploitation and environmentally driven recruitment distribution. Results indicate that the model configuration has greater influence on estimates than the environmental or exploitation histories. MCs that use more of the available disaggregated data generally are less biased and more precise.