Interannual variation in large-scale movement of Atlantic bluefin tuna (Thunnus thynnus) determined from pop-up satellite archival tags

Data from pop-up satellite archival tags (PSATs) deployed on Atlantic bluefin tuna (Thunnus thynnus) in 1999, 2000, and 2002 were analyzed synthetically using variants of a state-space statistical model to estimate tag shedding positions, in situ position estimation errors, and movement parameters. Geographic position estimates were computed from PSAT data for 30% to 50% of the days at liberty for each tag. Such relatively low position reporting rates may present biased impressions of movement for brief times at liberty. The respective longitude and latitude errors were estimated by the state-space Kalman filter model to be approximately 0.4 degrees and 1.0 degrees for the 1999 deployments, 0.4 degrees and 2.1 degrees for the 2000 deployments, and 0.9 degrees and 2.1 degrees for the 2002 deployments. Estimated movement parameters were used in stochastic simulation models to predict the distribution of tagged fish after times at liberty that exceed observed reporting dates. The distributions predicted by parameters estimated from the 2002 deployments were more restricted than those estimated from the 1999 deployments. Atlantic bluefin dispersal patterns appear to be age- or size-dependent and linked to shifts in oceanographic conditions. Future fisheries management measures should address interannual differences in stock distribution identified by electronic tags.