Abstract
Introduction Atlantic bluefin tuna (ABT) are large, migratory pelagic predators of high economic importance. ABT are currently managed as two independent stocks assigned to discrete spawning areas (Gulf of Mexico, and Mediterranean); however, stock overlap outside spawning areas makes accurate assignment of catch to stock-of-origin difficult. Methods Within this two-stock paradigm, we characterised stock-specific spatial distributions and behaviours by comparing habitat usage and vertical movement behaviours of 118 electronically tagged adult ABT spatially assigned to the GOM and Med spawning grounds. These spatial and behavioural differences were used in tests to probabilistically assign unknown individuals (which did not visit the GOM/Med spawning areas) to a stock. Results This new methodological approach using existing tag data, enables increased assignment of a track to a potential stock, to be achieved before genetic assignments. We identified certain markedly different movement patterns, range extents, depth use preferences (and associated area usage), migration directness and speeds, corresponding distance from shore, and mesopelagic-layer visitation. The probabilistic assignment approach had 97% in-bag testing accuracy, then assigned 190 individuals to a stock, doubling the number of assigned stock-of-origin tracks in our dataset, and potentially revealing novel movement and behaviour patterns among pre-spawning-age ABT. Discussion This approach can be easily adapted to other study species, more stocks, and different testing variables, hopefully serving as a useful addition to the fisheries management toolkit.