A comparison of initial statistical catch-at-age and catch-at-length assessments of Eastern Atlantic Bluefin tuna.

A concern associated with existing Atlantic bluefin tuna age-based assessments using VPA is that the catch-at-age data inputs are obtained by the cohort-slicing method, which is approximate and might introduce appreciable bias into the results. Current custom in such circumstances is rather to fit the assessment model directly to the basic catch-at-length data available, under the assumption of invariance of the distributions of length-at-age of the fish over time, with statistical models used to formulate the likelihoods maximised in the model fitting process. Initial results are presented for a process of comparing the 2010 ICCAT SCRS VPA assessment of the eastern plus Mediterranean stock with first a statistical catch-at-age assessment approach which also uses the same cohort-sliced catch-at-age inputs, and then a statistical catch-at-length method which fits instead to catch-at-length distributions. Spawning biomass estimates for both the statistical catch-at-age and -at-length analyses are appreciably larger than for the corresponding VPA assessment, which suggests that the specifications of the relationships between the fishing mortality on the plus group and that on immediately younger ages in the VPA merit reconsideration.