Bird community composition linked to human West Nile virus cases along the Colorado front range.

In the present study, we examined whether bird community composition can predict the annual number of human West Nile virus (WNV) cases on a per county basis in the Colorado Front Range, a region that experienced high numbers of human cases during the early part of the North American epidemic. We analyzed data sets pertaining to birds and human WNV cases from multiple existing databases between the years 2002 and 2008. Based on previous studies that used amplification fractions to compare the relative competence of different bird species, ten bird species that are common in Colorado were selected and categorized as high amplification birds, such as the American Robin (Turdus migratorius), or low amplification birds, such as the American Crow (Corvus brachyrhynchos). A general linear modeling analysis was used with an information theoretic (AIC) model sorting approach to examine which of the models best predicted the number of human WNV cases per county. Candidate models included year as a covariate and one of several bird community descriptors (e.g., richness, diversity, total bird abundance, high amplification abundance, or low amplification abundance). Results demonstrated that high amplification birds were a significant predictor of human WNV cases between 2002 and 2008. Our results suggest that a small subset of the bird community with high amplification fractions may drive the dynamics of human disease risk for West Nile. This study has implications for surveillance of West Nile and may offer insight into disease risk associated with other vector-borne zoonotic diseases.