Mosquito surveillance and pesticide treatment data can be combined in statistical models to provide insight into drivers of mosquito population dynamics. In cooperation with the county-based public health authority, multiple municipalities in Tarrant County, Texas, supplied surveillance and pesticide treatment data available from the 2014 mosquito season for analysis. With these data, general linear mixed modeling was used to model population dynamics of Culex quinquefasciatus, the primary vector for West Nile virus. Temporally lagged pesticide treatment information, weather data, and habitat variables were used as predictors of log + 1 transformed mosquito count data, and Akaike information criteria corrected for small sample sizes (AICc)-based model selection and multimodel averaging was used to produce a final model of mosquito abundance. The model revealed that mosquito counts were driven mainly by seasonally fluctuating temperature, precipitation, human population density, and treatment. In particular, interactions between temperature and treatment, and precipitation and human population density significantly contributed to the interpretation of the effects of the nonweather variables.
Keywords: model; mosquito control; surveillance.