A mixture model interpretation of stock of origin data for Atlantic bluefin tuna

Stock of origin data provide uncertain inferences for the stock composition of catches even in natal spawning areas. When conditioning operating models that account for stock mixing, these data must be interpreted in as unbiased a manner as possible in order to prevent overestimation of the extent of East-West mixing which leads to an overestimation of Western stock size because of the imbalance between the sizes of the two stocks. Here we provide a data summary of the distribution of current genetics and otolith microchemistry data, and identify an approach using mixture distributions to derive less biased inferences (compared to approaches suggested earlier) from stock of origin data for conditioning operating models. The estimated mixture distributions provide a reasonable fit to assignment data and remove fewer raw data than approaches that discard data based on somewhat ad-hoc rules. It is recommended that operating models be modified to be based on a simpler 7 area model.