A sequential Bayesian methodology to estimate movement and exploitation rates using electronic and conventional tag data: application to Atlantic bluefin tuna (Thunnus thynnus)

This paper presents a Bayesian methodology to estimate fishing mortality rates and transoceanic migration rates of highly migratory pelagic fishes that integrates multiple sources of tagging data and auxiliary information from prior knowledge. Exploitation rates and movement rates for Atlantic bluefin tuna (Thunnus thynnus) are estimated by fitting a spatially structured model to three types of data obtained from pop-up satellite, archival, and conventional tags for the period 1990-2006 in the western North Atlantic. A sequential Bayesian statistical approach is applied in which the key components of the model are separated and fitted sequentially to data sets pertinent to each component with the posterior probability density function (pdf) of paramenters from on analysis serving as the prior pdf for the next. The approach sequentially updates the estimates of age-specific fishing mortality rates (F) and transoceanic movements rates (T). Estimates of recent F are higher than the estimated rate of natural mortality and higher in the east than in the west. Estimates of annual T from the west to the east are higher for larger fish (6% for ages 0-3 to 16% for ages 9+). These estimates are also higher than those obtained from tagging studies before the 1990s and could be associated with changes in stock composition.