BROADENING CLIMATE PERSPECTIVES IN VECTOR ECOLOGY BY EXAMINING MULTIDIMENSIONAL DRIVERS ACROSS SCALES

Abstract

Mosquitoes transmit parasites and arboviruses responsible for hundreds of thousands of deaths and hundreds of millions of infections annually, and even nonfatal infections can have substantial health, social, and economic impacts. These burdens arise in part because mosquito population dynamics are tightly coupled with climate, which shapes where and when transmission can occur. However, most climate-driven projections of mosquito and disease risk rely primarily on temperature and rainfall, overlooking other aspects of atmospheric moisture, hydrometeorology, and wind, and they rarely account for how urbanization reshapes local climates. This dissertation reconceptualizes climatic suitability for mosquito vectors as a multidimensional product of thermal, hydric, hydrometeorological, and aerodynamic conditions interacting with urban landscapes and human populations. It combines regional-scale epidemiological modeling, local-scale entomological surveillance, and continental-scale trait-based demography to ask how adding humidity, drought, soil moisture, solar radiation, and wind to temperature-based frameworks alters our understanding of when and where mosquito populations can grow and intersect with people.At the regional scale, I analyzed West Nile virus (WNV) risk in Louisiana and South Dakota using county-level weekly human case data from 2004 to 2022 linked to daily meteorological data. Logistic distributed-lag models showed that wind speed provided independent, time-lagged information about WNV risk and substantially improved model fit beyond temperature, precipitation, and humidity. Higher-than-normal wind speeds over preceding weeks to months were consistently associated with reduced odds of WNV cases, consistent with wind-driven suppression of mosquito flight and vector–host contact. Temperature anomalies had opposite effects in the two states: warmer anomalies increased risk in cooler South Dakota, whereas cooler anomalies increased risk in warmer Louisiana, highlighting unimodal temperature–transmission relationships and the dependence of climate–disease associations on baseline climate. Effects of vapor pressure deficit and precipitation further indicated that drought–wet sequences and atmospheric dryness, rather than rainfall totals alone, modulated WNV amplification. At the local scale, I examined urban mosquito abundance in Norman, Oklahoma, using BG-Sentinel traps over two seasons (2023–2024) at four sites, coupled with high-frequency meteorological and soil moisture observations. Negative binomial generalized additive models with linearly distributed lags revealed species-specific hydrometeorological niches. Aedes albopictus abundance was best explained by shallow soil moisture and a standardized precipitation–evapotranspiration index, with short-lag increases following dry conditions and longer-lag increases following wetter antecedent soils. Aedes trivittatus responded to a similar wet–dry sequence, but with greater contributions from mean temperature and accumulated solar radiation. For the Culex pipiens/quinquefasciatus complex, maximum temperature and vapor pressure deficit provided comparable explanatory power, indicating that apparent heat effects partly encoded atmospheric moisture constraints. These results demonstrated that soil moisture, drought indices, and solar radiation captured ecological processes omitted from temperature–rainfall models and provided actionable indicators for timing urban mosquito control. At the continental scale, I developed a trait-based demographic model for the invasive urban malaria vector, Anopheles stephensi, that jointly accounted for temperature and relative humidity. Laboratory-derived temperature–humidity performance surfaces for development, survival, fecundity, and mortality were combined with downscaled CMIP6 climate projections to estimate intrinsic population growth rates across Africa and South Asia under a temperature-only formulation and a temperature–humidity formulation. Humid, moderately warm environments supported substantially higher growth rates than a temperature-only view suggested, while hot, dry lowlands that appeared suitable under temperature alone became much less favorable once desiccation was considered. Overlaying suitability with 1-km urban population projections yielded an exposure index that showed how humidity reduced projected exposure in some large but arid cities and elevated it in smaller, humid upland centers. Across these scales, the dissertation demonstrates that humidity, hydrometeorological variability, and wind are fundamental climatic dimensions for mosquito vectors and that temperature and rainfall only models systematically mischaracterize climatic windows for vector persistence in an urbanizing, warming world. By explicitly incorporating these additional dimensions, the work refines estimates of when and where transmission is most likely to occur, reveals context-dependent climate–disease relationships that are obscured in temperature-only models, and identifies meteorological indicators that can improve early warning systems and more targeted vector control.