Modeling Aedes aegypti Dynamics with Climate Factors

We developed statistical models to identify the most important lagged local climate variables influencing Ae. aegypti population dynamics and to test whether significant climate factors varied between the study localities. We modeled log₁₀-transformed ovitrap data from the CA, PA, and both localities combined (eggs/ovitrap/week) as a function of climate using a general linear model. We identified the most important lags to test in the model by assessing significant correlations between ovitrap data and climate variables at lags from 0 to 19 weeks, a similar time frame as the lags tested in a recent study of dengue and climate in the same region (See raw climate and ovitrap data in Fig. S1) [22] (link). We used these parameters to derive a best-fit model using glmulti in R [38] . A dummy variable for study locality was included in the best-fit model for both localities combined to capture confounding factors (e.g., socioeconomic differences, microclimate variability).