An integrative machine learning approach to understanding South Pacific Ocean albacore tuna habitat features

This study employs a random forest model combined with interpretable machine learning techniques to analyze the habitat preferences of South Pacific albacore tuna, incorporating a broad range of marine environmental variables. Among these, several factors derived from mesoscale eddy structures, including eddy polarity, eddy radius, and eddy kinetic energy, are integrated to further enhance the characterization of mesoscale eddy features. Interpretable methods were applied to provide intuitive visualizations of albacore tuna habitat preferences, with a focus on the most influential factors, including seawater temperature, dissolved oxygen concentration, and normalized mesoscale eddy radius. Seawater temperature and oxygen concentration are directly linked to the physiological needs of albacore tuna, while mesoscale eddy characteristics influence foraging and behavior by altering water column properties. This study provides a comprehensive perspective on the characteristics of albacore tuna habitat and the mechanisms driving its oceanographic variables, providing valuable insights for developing location-based, practical science-based management strategies for fishery resources.