Incorporating drivers of global change throughout the annual cycle in species distribution models for migratory birds: a gap in ecological forecasting

Abstract

Understanding the consequences of global change for migratory birds is complex as individuals are exposed to diverse conditions and experiences that interact across their annual cycle. Species distribution models (SDMs) can serve as a powerful tool that help us understand how species distributions respond to global change. However, SDMs applied to migratory birds may fail to capture the effects of seasonal variability on species distributional changes, likely due to a lack of appropriate modeling frameworks and limited data availability that hamper the inclusion of events and conditions throughout the annual cycle. Here, we review patterns in the migratory bird SDM literature over the last two decades using a vote counting approach, and provide a framework for migratory bird SDMs moving forward. We found evidence that species distribution models applied to migratory birds (1) typically incorporate data from only one season of the full annual cycle and do not account for seasonal interactions, (2) are focused on terrestrial species in North America and Europe, (3) tend to model the distributions of obligate migratory species, especially songbirds and waterfowl, and (4) largely lack biologically relevant threat layers. To improve our ability to forecast how species cope with global change, we recommend a Bayesian modeling framework where existing knowledge about a species’ migratory connectivity, threats, and/or other biologically relevant factors can be specified via model priors. Full annual cycle species distribution models are important tools for improving forecasts of migratory bird distributions in response to global change.