West Nile virus (WNV) is an invasive species in North America.  Since its introduction to New York City in 1999, WNV has spread rapidly across the continental United States (US) and has become a major public health concern throughout North America.  Therefore, it is crucial to predict human WNV risks for cost-effective controls of the disease and optimal allocations of limited resources.  Droughts of previous years induce mosquito outbreaks by reducing their predators or competitors.  Consequently, droughts increase human WNV risks 1-2 years post-drought.  The state of Mississippi had the highest human incidence rate during the 2002 WNV outbreak in the US.  We analyzed data on human WNV incidences in the 82 counties of Mississippi in 2002, using standard morbidity ratio (SMR) and Bayesian hierarchical models, to estimate relationships between precipitation and human WNV risks.  We also evaluated spatial autocorrelations of human WNV risks with conditional autocorrelative (CAR) models. 

Results/Conclusions

The human WNV risk of 2002 was inversely correlated with annual precipitation during 2001.  Annual precipitation was a predictor of spatial variation of WNV risk.  Therefore, precipitation patterns and ultimately, climate change, can influence WNV risk by altering trophic interactions between mosquitoes and their predators.