Macroscale population estimates for a migrant songbird using weather radar

Abstract

Changes in migratory bird populations can signal environmental change occurring on the wintering grounds, the breeding grounds, or along the migratory route between them. Aerial insectivores (swifts, swallows, etc.) additionally serve as an indicator of the abundance of airborne arthropods across these domains. The Purple Martin (Progne subis) is one such species and aspects of its life history make it ideal for remote, large-scale population studies. During the summer months, Purple Martins form dense roosting colonies across eastern North America. As they leave these sites on diurnal foraging trips in the lower atmosphere, their exodus is routinely captured by NEXRAD weather radar facilities across the country. Data collected by these facilities are publicly available through the National Weather Service and represent a free, large-scale archive of ecological and behavioral information. I present a method of translating radar reflectivity into a biological density product to generate raw and modeled estimates of Martin populations to examine abundance, phenology, and habitat use across the eastern U.S. Results suggest that while radar-derived estimates are variable within individual sites, aggregate population measures across roosts illustrate trends in abundance, phenology, and habitat use at continental spatial scales, and decadal temporal scales. These data can help inform and validate other citizen science efforts (e.g. Breeding Bird Survey, eBird, etc.), lay the groundwork for large-scale, long-term remote monitoring of these populations, and have broad applicability to other colonially roosting bird and bat species.