Predicting the occurrence of Atlantic bluefin tuna (Thunnus thynnus) larvae in the northern Gulf of Mexico: building a classification model from archival data

Although bluefin tuna are found throughout the Atlantic Ocean, spawning in the western Atlantic has been recorded predominantly in the Gulf of Mexico (GOM) in spring. Larval bluefin tuna abundances from the northern GOM are formulated into an index used to tune the adult stock assessment, and the variability of this index is currently high. This study investigated whether some of the variability in larval bluefin tuna abundances was related to environmental conditions, by defining associations between larval bluefin tuna catch locations, and a suite of environmental variables. We hypothesized that certain habitat types, as defined by environmental variables, would be more likely to contain bluefin tuna larvae. Favorable habitat for bluefin tuna larvae was defined using a classification tree approach. Habitat within the Loop Current was generally less favorable, as were warm-core rings, and cooler waters on the continental shelf. The location and size of favorable habitat was highly variable among years, which was reflected in the locations of larval bluefin tuna catches. The model successfully placed bluefin tuna larvae in favorable habitat with nearly 90% accuracy, but many negative stations were also located within theoretically favorable habitat. The probability of collecting larval bluefin tuna in favorable habitat was nearly twice the probability of collecting bluefin tuna larvae across all habitats (35.5 versus 21.0%). This model is a useful addition to knowledge of larval bluefin tuna distributions; however, the incorporation of variables describing finer-scale features, such as thermal fronts, may significantly improve the model's predictive power.