Mosquito population dynamics are shaped by interactions among larval density, temperature, and relative humidity.

Understanding how variation in key abiotic and biotic factors interact at ecologically relevant spatial scales is crucial for predicting population dynamics, distributions, and abundances. This is especially true for vectors that transmit human pathogens. However, studies investigating the effects of environmental variation on vectors have typically investigated environmental factors in isolation or in laboratory experiments that examine constant environmental conditions that often do not occur in the field. To address these limitations, we conducted a semi-field experiment in Athens, Georgia, using the invasive Asian tiger mosquito (Aedes albopictus). We selected nine sites that varied in impervious surface and vegetation cover to explore effects of natural variation in microclimate (specifically, temperature, and relative humidity) on mosquitoes. We manipulated conspecific larval density at each site and repeated the experiment in summer and fall to further increase the variability in temperature and relative humidity. We then evaluated the original design features (land cover, larval density, and season), and their interactions, on the mean proportion of females emerging, juvenile development time, size upon emergence, and estimated per capita population growth (i.e., fitness). We found significant effects of larval density, land cover, and season on all response variables, including a non-intuitive decrease in development time with increasing larval density in fall. We repeated these analyses using the hypothesized microclimate drivers of these effects: temperature and relative humidity. In general, the model using the microclimate variables outperformed the model using land use and season, highlighting the roles of temperature and relative humidity and their interactions with density on mosquito traits and dynamics. Our study demonstrates that ignoring the interaction between variation in biotic (e.g., intraspecific competition) and abiotic (e.g., temperature and relative humidity) variables could reduce the accuracy and precision of models used to predict mosquito population and pathogen transmission dynamics, especially those inferring dynamics at finer spatial scales across which transmission and control occur.