Combined use of trawl fishery and research vessel survey data in a multivariate autoregressive state-space (MARSS) model to improve the accuracy of abundance index estimates

Abstract

Fishery-dependent catch per unit effort (CPUE) data have been used as an abundance index (AI) in fish stock assessments. However, fishery-dependent CPUE data are influenced not only by changes in fish abundance but also by other factors, such as the choice or restrictions of fishing grounds to operate. Accordingly, bias may arise in AIs due to a lack of data from unfished or rarely fished areas. To improve the accuracy of AI estimates, spatially arranged CPUE datasets from both trawl fisheries and research vessel surveys in the East China Sea were concurrently analyzed in the present study using a multivariate autoregressive state-space (MARSS) model. Survey datasets complemented information on stock status in the fishing areas where fishery-dependent datasets were limited. As a result, the combined use of datasets from both sources effectively improved the accuracy of estimates of AIs and the spatial distribution of the population density of each fish species.