Long-term changes in the timing of autumn migration in Alaska's boreal songbirds

Abstract

Alaskas boreal birds face a rapidly changing environment, but we know little about shifts in migratory timing, particularly in autumn. We used quantile regression to quantify long-term changes in autumn capture date in 21 boreal passerines using 22+year datasets from two banding stations in central Alaska. We also quantified differences between sites and explored whether select climate indices during three periods of the annual cycle (breeding, post-fledge, and migration) could predict long-term changes in median capture. Long-term changes in autumn migration were detected in 86% of taxa, 76% of which exhibited advances in capture date ([~]2-3 days/decade), particularly long-distance migrants at one field site. However, site-specific differences unexpectedly highlight the need for caution before extrapolating long-term timing patterns over broad spatial extents. Warmer conditions during the breeding period (using the AO climate index) were associated with advances in autumn capture date in the greatest number of species (9). Collectively, we hypothesize that Alaskas immense size and spatially-variable climate regions impact reproductive timing, often resulting in long-term advances (with warming) and occasionally delays (with cooling). Carry-over effects of reproductive timing may therefore influence the autumn passage of different breeding populations, causing site-specific patterns, such as a species showing long-term advances at one location, but delays at another. Finally, as part of the broader effort to anticipate and reduce declines in boreal migratory birds, our study underscores the conservation value of banding station data in quantifying avian responses to and investigating drivers associated with varied climate indices.