Mosquito-based surveillance is a practical way to estimate the risk of transmission of West Nile virus (WNV) to people. Variations in temperature and precipitation play a role in driving mosquito infection rates and transmission of WNV, motivating efforts to predict infection rates based on prior weather conditions. Weather conditions and sequential patterns of meteorological events can have particularly important, but regionally distinctive, consequences for WNV transmission, with high temperatures and low precipitation often increasing WNV mosquito infection. Predictive models that incorporate weather can thus be used to provide early indications of the risk of WNV infection. The purpose of this study was first, to assess the ability of a previously published model of WNV mosquito infection to predict infection for an area within the region for which it was developed, and second, to improve the predictive ability of this model by incorporating new weather factors that may affect mosquito development. The legacy model captured the primary trends in mosquito infection, but it was improved considerably when calibrated with local mosquito infection rates. The use of interaction terms between precipitation and temperature improved model performance. Specifically, temperature had a stronger influence than rainfall, so that lower than average temperature greatly reduced the effect of low rainfall on increased infection rates. When rainfall was lower, high temperature had an even stronger positive impact on infection rates. The final model is practical, stable, and operationally valid for predicting West Nile virus infection rates in future weeks when calibrated with local data.