Many mosquito-borne diseases are currently emerging in new locations or resurging in areas where they were previously eliminated. An important example from North America is West Nile virus, which was first detected in New York City in 1999 and then rapidly spread across the entire continent. The problem of WNV has been particularly acute in the northern Great Plains, where the long-term incidence of human disease has remained much higher than the rest of the United States. Efforts are currently underway to model and forecast interannual fluctuations in WNV cases and mosquito populations using vegetation indices and last surface temperature data from MODIS and precipitation data from TRMM. However, a major limitation of remote sensing based disease early warning systems is that these satellite-derived variables provide only indirect measurements of the proximal environmental factors influencing mosquito populations and disease risk. The recent development of a new set of daily global land surface parameters derived from the Advanced Microwave Scanning Radiometer on EOS (AMSR-E) offers new opportunities for developing improved environmental models of mosquito-borne disease risk. In particular, the new AMRS-E products provide several critical environmental variables that are directly relevant to mosquito ecology, including near-surface air temperature, soil moisture, and fractional water cover. The specific objectives of this research are to (1) develop statistical models of WNV risk and mosquito population dynamics using the AMSR-E land surface parameters, (2) compare their performance with models based on MODIS and TRMM products to quantify potential improvements in the forecasting of WNV outbreaks, and (3) generate and disseminate early warning predictions from these models and gain qualitative feedback from vector control experts and public health practitioners as to their utility. To accomplish these objectives, we will obtain historical AMSR-E land surface parameter data from 2002 to present and summarize it temporally to match the 8-day MODIS composite periods, and spatially to match the resolutions of the human case data (counties) and mosquito data (city boundaries). We will apply appropriate statistical methods to account for spatial and temporal autocorrelation, seasonality, and the effects of other environmental variables such as land cover and land use. Performance of the models based on AMSR-E variables will be quantified based on multi-model inference using the AIC and DIC statistics, and by comparing forecasts in future years with human case and mosquito surveillance data. Based on the results of these analyses, we will provide seasonal forecasts of WNV risk to the South Dakota Department of Health. The major significance of this project is the potential for incorporating novel information from a portion of the electromagnetic spectrum that is currently underutilized in mosquito-borne disease research, with a potential benefit of developing more accurate and effective early-warning systems. The state of South Dakota and the broader northern Great Plains region will benefit from access to improved WNV risk forecasts, which will aid in the development of more effective vector control and disease prevention strategies. Understanding the potential for using AMSR-E data for disease early warning will also have broader impacts, as these data could also be applied to enhance the forecasting of malaria, dengue fever, Rift Valley fever, and a variety of other mosquito-borne diseases.