Window related avian mortality at a migration corridor.

Abstract

Various man-made structures act as barriers to the movement of migrating birds. In the United States alone hundreds of millions of birds are estimated to die annually from window collisions. This risk increases when urban environments exist along important migratory corridors. Although all birds are susceptible to collisions with windows, nocturnally migrating songbirds make up the largest portion of observed window fatalities. Minnesota Point located on the western tip of Lake Superior is a valuable study location for quantifying collision fatalities because birds rely on this location to rest and forage during migration and because it is a highly developed landscape. To estimate mortality rates, a total of 42 residential homes were searched periodically during five migratory periods between 2006 and 2009 for avian window kills. To increase the accuracy of this estimate, an experiment was incorporated into the study to account for fatalities undetected by searchers due to removal by scavengers. A total of 40 identifiable species and 108 individual birds were recorded as window collision fatalities and of these, 90% were short and long distance migrants. During the spring and fall of 2009, bird density estimates were calculated to compare the abundance of species flying through the area to the abundance of observed species fatalities. The density of migrants utilizing the area was greater than that of resident species and was reflected by the higher mortality rates observed for migrants. The probability of a carcass being scavenged within 6 d averaged 79%. Variation in scavenging rate was best explained by the location of the house and the size of the carcass used. With adjustments made for scavenging, an annual estimate of avian window fatalities for houses on Minnesota Point during peak migration is 1333 ± 73 birds per year. Avian window collisions contribute significantly to annual mortality rates and are potentially avoidable. Deriving accurate estimates of mortality is vital to predicting long term population effects, especially for species susceptible to window collisions.