While emerging infectious diseases can exert considerable long-term effects on wildlife populations, impact assessment is difficult when contemporaneous surveillance requires data for pathogens, vectors, and hosts. West Nile virus (WNV) noticeably affected many North American bird species after its arrival in 1999, but community-level responses have not been broadly assessed in over 10 years. We used mosquito surveillance data from Pennsylvania, USA to model local patterns of WNV occurrence based on environmental trapping conditions, thereby allowing implementation of predicted WNV occurrence as a predictor of bird abundance along Breeding Bird Survey routes. Occurrence varied substantially in mosquito samples across space and time, and was correlated with environmental factors such as temperature, elevation, and land cover. Correlations between annual WNV indices and bird counts revealed impacts for at least 20% of species (31/155), although the specific mechanisms of impacts require further investigation. Additionally, comparable models examining potential effects of heat stress produced a dissimilar and reduced cohort of affected species, supporting the distinctness of our WNV findings from effects of broader environmental factors. Many affected species also did not exhibit overall declines in Pennsylvania, suggesting that while WNV may not prevailingly drive current trends, it could exacerbate future declines in tandem with other stressors. Our findings indicate that WNV continues to influence bird populations after 25 years, with implications for both ecosystems and human settlements where the virus occurs. Our approach also shows promise for assessing other localized health threats, demonstrating how community-sourced data can be leveraged toward proactive conservation.
Geary et al. (2026) studied this question.