Physical Oceanographic Conditions Drive Patterns of Seasonal Core Habitat Among Marine Top Predators in the San Juan Archipelago

Abstract

Species distribution models (SDMs) are valuable tools for characterizing ecological-environmental interactions and identifying regions of core habitat zones of high conservation importance. This is particularly true for marine top predators, organisms at or near the top of their respective food chains, which can be challenging to observe and difficult to manage owing to their relatively low population sizes and highly migratory natures. In this study, I construct species distribution models for the four most abundant species from four predominant families of marine predator across the San Juan Archipelago (SJA). Using a combination of logistic regression analysis and generalized additive models, I describe how nine environmental variables influence patterns of distribution among marine mammals and seabirds respectively and identify regions of core habitat where predators are predicted to be most abundant. Water depth and channel width were identified as highly significant explanatory variables in three of my four SDMs, highlighting the importance of physical habitat structure in the SJA, where strong tidal currents interact with topographical features to create foraging opportunities for piscivorous marine predators. Core habitat zones for multiple predator species occurred almost exclusively within 1-km from shore, suggesting that nearshore areas are critical to ecosystem function from a top-down perspective in this region. Conservation efforts to preserve marine predator populations should focus on restoring altered shorelines to create optimum habitat for marine predators and the forage fish on which they depend. Future shoreline degradation might be prevented by restricting private development of shorelines, with the added benefit of allowing more equitable access to natural spaces where the public can connect with and observe the marine top predators which share their unique ecosystem.