Water availability has stronger effects on West Nile virus dynamics in water-limited regions

Abstract

West Nile virus dynamics are shaped by hydrological conditions that influence mosquito habitat and pathogen transmission, but identifying causal relationships is difficult in managed landscapes where irrigation decouples local water conditions from precipitation, complicating climate-disease inference. We address this challenge using a 21-year panel of more than 19 million Culex tarsalis mosquitoes from Californias Central Valley, applying fixed-effects panel models to estimate how surface water availability affects mosquito abundance and infection rates while accounting for spatial differences and shared temporal variation. We find that wetter conditions lead to higher mosquito abundance but slightly lower infection rates, suggesting divergent responses of vector population growth and pathogen amplification. These patterns are consistent across multiple hydrological measures, including drought indices, soil moisture, surface water, and evapotranspiration. Effects are strongest in water-limited regions, where hydrological variability is greatest and buffering by snowmelt-fed river systems is weakest. Overall, hydrological conditions exert contrasting effects on key components of West Nile virus dynamics, and these relationships are strongly conditioned by human water management. Our results highlight how irrigation decouples local hydrological conditions from broader climatic variability, underscoring the need for fine-scale hydrological data and panel-based approaches to identify drivers of disease dynamics in human managed landscapes.