Using full annual cycle species distribution models to predict abundance change in migratory birds

Natural and anthropogenic stressors of global change are restructuring distributions of flora and fauna worldwide. Species distribution models allow ecologists to understand and predict changes in species distributions based on environmental conditions. To date, species distribution models for migratory birds are based on conditions from only the breeding season, overlooking conditions from the rest of their annual cycles. Here, with data from the well-studied American Redstart ( Setophaga ruticilla ), a long-distance migratory bird, we examine the importance of nonbreeding conditions in models of breeding abundance. We compared the accuracy of estimated breeding redstart abundance from models based on: (1) breeding conditions only, (2) breeding and nonbreeding conditions, and (3) breeding and nonbreeding conditions with an interaction between nonbreeding rainfall and migration distance. We then assessed whether conditioning on counts of other migratory species improved predictions of current redstart breeding abundance by capturing additional variation in nonbreeding conditions. Finally, we forecasted future breeding abundance based on the full annual cycle model that incorporates the seasonal interaction. Overall, all three models performed similarly in predicting current breeding redstart abundance. Conditioning predictions on counts of other migratory species improved predictions of current breeding redstart abundance. The full annual cycle model with an interaction forecasts range-wide population change that differs from predictions from the breeding-only model in key breeding redstart regions. Although the three models performed similarly well, the models that incorporated nonbreeding conditions identified relationships between nonbreeding conditions and breeding abundance. Continued, iterative revisions to these quantitative tools are critical for reconstructing past, estimating current, and forecasting future species distributions under global change.