Assistant Professor University of California, Santa Barbara, United States
West Nile virus (WNV), transmitted between Culex mosquitos and wild birds, has been a persistent threat to human and avian health since its introduction to California in 2003. It is particularly prevalent in the agriculturally dominated Central Valley where human land use decisions—from urban development, to what crops to grow where—can strongly influence microclimatic and habitat conditions for mosquitos and their life cycle processes. Moreover, these human-dominated agricultural landscapes also influence the abundance and diversity of bird hosts of WNV, as well as the foraging activity of potential mosquito predators, like Mexican free-tailed bats. Here we use a variety of landscape-scale spatial data and remote sensing-based approaches, in combination with community science and WNV surveillance data, to map WNV risk and explore downstream applications to the investigation of the community and landscape ecological drivers of WNV in the Central Valley. First, we use high resolution land surface temperature from the ECOSTRESS sensor to model diurnal air temperature patterns and map temperature-dependent WNV risk, and explore underlying land use patterns. We then use WNV surveillance data from VectorSurv to validate this approach. Finally, we explore the use of this WNV risk mapping and surveillance data, in combination with ebird community science data and weather radar-based mapping of free-tailed bat activity to investigate the community ecological drivers of WNV risk at the landscape scale. We find that ECOSTRESS-based risk mapping of WNV can identify variation in temperature-dependent risk across different land uses and land covers, with transmission probability highest in urbanized areas and orchards, and mosquito biting rates highest in orchards and fruit crops, for example. Moreover, this temperature-dependent trait mapping approach is predictive of WNV surveillance in the field, controlling for other abiotic conditions like irrigation activity and vegetation greenness, as well as host availability and abundance. Additionally, we illustrate that free-tailed bat activity is elevated over regions of higher WNV risk, suggesting a possible role for bats in the control of WNV that we plan to further explore. Finally, we also demonstrate the potential role of bird host abundance and diversity in influencing risk at the landscape scale. Through this series of approaches and investigations, we illustrate how the integration of novel data streams—from earth observation and remote sensing, to community science and vector surveillance—can inform vector-borne disease ecology and risk across rapidly changing landscapes.
