Spatiotemporal dynamics of the dolphin-associated purse-seine fishery for yellowfin tuna (Thunnus albacares) in the eastern Pacific Ocean

Abstract

Abstract Spatiotemporal models of catch rate data are used in fisheries science to understand the spatiotemporal dynamics of a fishery and, more importantly, to produce a standardized index of relative abundance for stock assessment models. Here, we apply a spatiotemporal delta-generalized linear mixed model to catch rate data collected from the dolphin-associated purse-seine fishery for yellowfin tuna (Thunnus albacares) in the eastern Pacific Ocean during 1975–2016. Results show that the density of yellowfin tuna had both pronounced spatial variation across the eastern Pacific Ocean and spatiotemporal variation at time scales from quarterly to decadal. In comparison to the index of relative abundance standardized via the spatiotemporal approach, that computed via the nominal approach, which is currently used in the stock assessment, under- and over-estimates initial and terminal abundances, respectively, by as much as 50% or more. The standardized index represents an improvement over the nominal index because the spatiotemporal approach imputes population density in unfished areas to obtain an area-weighted index and accounts for vessel-specific differences in fishing efficiency. The importance of area-weighting for index standardization is highlighted by the fact that ignoring it has a greater effect on the abundance index than does ignoring differences in fishing efficiency among vessels. The stock assessment model that includes standardized indices has significantly larger log-likelihood and reduced pattern in residuals than the one that includes nominal indices, suggesting that replacing nominal indices with indices from a spatiotemporal model can improve the stock assessment for yellowfin tuna in the eastern Pacific Ocean.