Abstract: Geographic information systems (GIS) and remote sensing technologies offer new opportunities to investigate environmental correlations with mosquito vector species. Advances in modeling approaches combined with routinely collected trap data provide the basis to maximize mosquito control surveillance to predict distributions of individual and combinations of mosquito vector species. One approach to predicting the distribution of multiple mosquito vectors across different landscapes is the use of a joint species distribution modeling framework (jSDM). Rather than predicting a single vector species, this method generates model outputs that allow predictions of multiple species and their traits, including vector competency. Here, we used a jSDM approach to investigate correlations between mosquito community composition, community-weighted mean traits (CWMT), and LULC in Manatee County, Florida to identify areas predicted to have high proportions of West Nile virus vectors. Species presence/absence across 60 trap sites sampled in 2016, 2017, 2019, and 2020, served as response variables; percent land cover for developed, cropland, herbaceous wetland, and woody wetland within 5 km of trap sites served as environmental variables. Community weighted mean values were calculated from a binary matrix of WNV vector competency to predict proportions of WNV vectors across the study area. Results indicated that proportions of variance explained by percent land cover for individual species coincided with known habitat associations for most species. Maps visualizing spatial predictions of CWMT predicted mosquito communities with the highest proportions of WNV vector competent species in urbanized areas. Joint species distribution models that leverage GIS and remote sensing data provide a powerful tool that can help optimize targeted surveillance and control efforts and have the potential to provide insights that may help improve prevention and management of mosquito-borne disease.
