Applying mixed-effects growth models to back-calculated size-at-age data for Atlantic bluefin tuna (Thunnus thynnus)

We fit growth models to back-calculated size-at-age data for Atlantic bluefin tuna (Thunnus thynnus) captured in the western Atlantic management area to inform alternative growth scenarios within the current management framework of the species. The Modified-Fry function was the best-performing of three back-calculation functions based on a leave-one-out cross-validation for within-cohort comparisons. We fit multiple growth models to the back-calculated growth trajectories, and demonstrated that growth parameters were highly sensitive to whether the lack of independence between back-calculated lengths from each individual was accounted for by the model. Non-linear mixed-effects modelling provides a suitable approach for accounting for this lack of independence and the autocorrelation between back-calculated lengths from the same individual. We further demonstrated the utility of mixed-effects models for predicting future growth of individuals using a forecasting test. We used this modelling approach to demonstrate that male bluefin tuna had greater estimated asymptotic length than females, and found that there was no significant difference in growth parameters between individuals assigned to genetically distinct spawning stocks.