West Nile virus impacts in American crow populations are associated with human land use and climate

West Nile virus (WNV) was first detected in the western hemisphere during the summer of 1999, reawakening US public awareness of the potential severity of vector-borne pathogens. Since its New World introduction, WNV has caused disease in human, avian, and mammalian communities across the continent. American crows (Corvus brachyrhynchos) are a highly susceptible WNV host and when modeled appropriately, changes in crow abundances can serve as a proxy for the spatio-temporal presence of WNV. We use the dramatic declines in abundance of this avian host to examine spatio-temporal heterogeneity in WNV intensity across the northeastern US, where WNV was first detected. Using data from the Breeding Bird Survey, we identify significant declines in crow abundance after WNV emergence that are associated with lower forest cover, more urban land use, and warmer winter temperatures. Importantly, we document continued declines as WNV was present in an area over consecutive years. Our findings support the urban-pathogen link that human WNV incidence studies have shown. For each 1% increase in urban land cover we expect an additional 5% decline in the log crow abundance beyond the decline attributed to WNV in undeveloped areas. We also demonstrate a significant relationship between above-average winter temperatures and WNV-related declines in crow abundance. The mechanisms behind these patterns remain uncertain and hypotheses requiring further research are suggested. In particular, a strong positive relationship between urban land cover and winter temperatures may confound mechanistic understanding, especially when a temperature-sensitive vector is involved.