Modeling the time-varying density distribution of highly migratory species: Atlantic blue marlin as an example

Blue marlin (Makaira nigricans) are among many species of tunas and billfishes that require pelagic longline catch-effort statistics for stock assessments. Major controversies about stock status have arisen because of issues concerning habitat influences on catch rates, but models to describe how the species is distributed throughout its entire habitat are lacking. A habitat suitability model (HSM) of the size and spatial distribution of blue marlin habitat by month using habitat weighted-volume (H) is presented. It is estimated from oceanographic data partitioned by 1° of latitude and 1° of longitude in 50 layers from the surface to a depth of 1200 m using dissolved oxygen tolerances and temperature preferences compiled from electronic tagging. The physical habitat is an amorphous 3-dimensional space whose boundaries are constantly changing with seasonal and longer-term variations in climate. Fluctuations in habitat volume likely contribute to fluctuations in CPUE that are independent of population abundance and add unrecognized uncertainty to abundance indices used to estimate population benchmarks. The results highlight the need to expand stock assessments to include seasonal and annual climatology. The HSM-based habitat weighted volume model offers a way to validate analytical methods for using longline CPUE to monitor population health.