Non-random fishery data can validate research survey observations of Pacific cod ( Gadus macrocephalus ) size in the Bering Sea

Research surveys provide the foundation for sound, effective management of fishery resources and are integral to observing fish population trends. However, bias in sampling gear and operating hours may confound observed shifts in species distributions over space and time. Within North Pacific waters off Alaska, Pacific cod ( Gadus macrocephalus ) support a large, commercial fishery and is an example of a species that experiences spatial and temporal shifts. Seasonal migratory shifts are difficult to incorporate into stock assessment models and are further complicated if research surveys do not effectively sample the underlying size distributions. In the Bering Sea, differences in median fish size between the winter Pacific cod bottom-trawl fishery (directed) and the summer research survey have been observed since the onset of the surveys in the 1980s (66 and 41 cm, respectively). Because of this, it has been suggested that large Pacific cod may not be available to the summer research survey for varying reasons. In this study, we compared standardized observations of mature Pacific cod length distributions from a summer multi-species fishery with the National Marine Fisheries Service summer research survey from 2009 to 2018 in the Bering Sea. Controlling for spatiotemporal effects, there was no difference detected in Pacific cod length distributions between the fishery and survey, suggesting that the survey accurately captures the entire size distribution of Pacific cod in the summer. Although standardized research surveys are considered to be representative samples of the entire population, using non-random fishery observations, where the fishery and survey spatiotemporally overlap, can validate survey observations and inform selectivity relationships in stock assessment models.