Hierarchical habitat selection by a predatory fish in a patchy seascape.

Faunal habitat selection, or the disproportionate use of available resources, is closely linked to habitat composition and configuration across a seascape. However, the drivers of habitat selection operate across multiple scales and require a hierarchical approach to study. This study combines acoustic telemetry, field survey data, remote sensing, and machine learning to investigate the multi-scale (seascape and patch) habitat selection of spotted seatrout (Cynoscion nebulosus) in Florida Bay, Everglades National Park, USA. Spotted seatrout responded to both scales, as there were three patch-scale (Halodule cover, standard deviation of submerged aquatic vegetation (SAV) cover, and SAV species richness) and one seascape-scale (patch density) predictor in the top four. However, responses were scale-specific, exhibiting logistic responses to seascape-level variables and optimal (specific-range) responses to patch-level characteristics. This study highlights the importance of investigating habitat selection across multiple scales as climate change alters not only species ranges, but local seascapes as well.