WORDS)Temperature is a key driver of mosquito-borne disease because it affects the physiology and life history traits of mosquitoes and the pathogens they transmit. These trait thermal responses are typically nonlinear and vary by species. However, the impact of temperature on many important temperate mosquito-borne diseases has never been quantified, and it is unclear if thermal responses of temperate mosquito-borne diseases differ systematically from the patterns found in better-studied tropical systems. These environment-transmission relationships are critical for understanding current seasonal and geographic distributions of disease and future shifts due to climate change. We used a mechanistic, trait-based approach to characterize temperature-dependent transmission of 10 vector-pathogen pairs in an overlapping network of mosquitoes (Culex pipiens, Cx. quinquefascsiatus, Cx. tarsalis, Aedes triseriatus, and others) and viruses (West Nile virus [WNV], Eastern and Western Equine Encephalitis viruses [EEEV and WEEV], St. Louis Encephalitis virus [SLEV], Sindbis virus [SINV], and Rift Valley Fever virus[RVFV]), most of which have predominantly temperate geographic distributions worldwide. We found that several of these temperate pathogens have lower thermal limits (8.6-19.0ºC) and optima (22.7-26.0ºC) that are cooler than those of tropical pathogens (lower thermal limits:16.2-22.8ºC; optima: 25.4-29.1ºC), as expected based on geography. However, their upper thermal limits (31.9-37.8ºC) were very similar to those of tropical pathogens (31.4-34.5ºC).Together, these patterns led to wider thermal breadths (NUMºC) for several of the temperate viruses compared to most tropical pathogens (NUMºC). Among temperate viruses, thermal limits for transmission varied more and were more uncertain than the optima. While the qualitative shapes of trait thermal responses generally matched prior studies, lifespan declined monotonically at temperatures above 14ºC for the temperate Culex vectors, in contrast to 4 unimodal responses observed within this temperature range for tropical species. We validated the mechanistic models with independent data on human disease cases. The incidence of WNV across US counties responded unimodally to temperature and peaked at 23.9ºC, closely matching model predictions (optima for WNV in North American vectors ranged from 23.9-25.2ºC).Additionally, data on the month of onset of cases of WNV, SLEV, and EEEV suggested that temperature is important for determining the seasonality of transmission. The thermal responses for parasite development rate, vector competence, biting rate, and lifespan, all of which varied among species and determined thermal limits in models, are a critical need for future empirical work. Because temperature effects on vector-borne disease transmission are pervasive and nonlinear, and climate regimes are changing rapidly, trait-based models are a key tool for predicting how temperature will affect mosquito-borne disease dynamics. 5
