Modeling the importance of subsurface environmental variables in driving swordfish (Xiphias gladius) catchability in the Western Indian Ocean.

Many oceanic species in pelagic habitats move vertically through the water column, highlighting the ecological importance of that spatial environment for modeling habitats of marine species. The role and importance of multiple oceanic subsurface environmental variables in modeling the habitat suitability of swordfish (Xiphias gladius), a highly migratory large pelagic fish, is poorly understood. In this study, we analyzed adult swordfish data from the 2017–2019 Chinese Indian Ocean tuna longline fishery observer. We used the maximum entropy model (MaxEnt) and random forest model (RF) to compare modeling schemes that included multiple subsurface environmental datasets. The area under receiver operating characteristic curve (AUC) from training and test sets was evaluated to investigate whether the inclusion of subsurface variables could enhance model performance and affect the simulation results. This analysis showed that model performance was significantly enhanced after addition of subsurface environmental variables, and the best model fit was achieved at 200–300 m depth. Sea water temperature, dissolved oxygen, net primary production, and ocean mixed layer depth were the critical environmental factors constituting catchability for swordfish in the Western Indian Ocean. As the depth increased, dissolved oxygen became the most important environmental factor, replacing surface temperature. Compared with the surface model, the location and extent of areas of high catchability in certain months changed significantly after the addition of subsurface variables. The results of this study provide evidence for a better understanding of the selection of critical environmental variables and improvement of model performance in 3D habitat modeling of pelagic fish. [ABSTRACT FROM AUTHOR]