Seasonally frugivorous forest birds and window collision fatalities: novel integration of bird counts in fall improves assessment of species vulnerability to collisions

Bird-window collisions are a significant and growing threat to birds, but the issue is still understudied in many geographical areas and stages of the avian annual life cycle. The mountainous topography and numerous distinct biogeoclimatic zones along the Pacific coast of Canada and the United States may result in regional and seasonal differences in collision mortality and species vulnerability to collisions. We surveyed daily for evidence of bird-window collisions over six 21-day periods in fall, early winter, and late winter between 2019 and 2022 at a university campus in southwestern British Columbia, Canada, and assessed individual species' vulnerability to collisions by examining whether species-specific collision rates were disproportionate to their local abundance. We accounted for poor detectability of some species in fall, by integrating point count data from our study site with mist net capture data from a nearby banding station to improve abundance estimates. Collision mortality peaked in fall, but early winter collision mortality was significantly higher than in the later winter months, potentially due to movements of altitudinal migrants into our low-elevation study area in early winter. We estimated that an average of 885-1,342 (median = 1,095) birds are killed at 51 buildings campus-wide each year between September 15 and February 10, the peak fall migration wintering period. Forest birds, particularly species that switch to highly frugivorous diets in fall and winter, were most vulnerable to collisions across the seasons studied. Non-breeding season mortality due to collisions may be substantial for these species, particularly when considering cumulative mortality across the entire non-breeding period. The potential role of collision mortality in species declines should be further explored by assessing collision frequency and species vulnerability across life cycle stages in other geographical locations, and through improved data on migratory connectivity and linkages between declining breeding populations and non-breeding season mortality.