The paper presented in this seminar is co-authored by Ian Sue Wing and Shouro Dasgupta.

Estimates of flu-associated deaths in the United States range from around 3,000 to a high of about 49,000 with direct medical costs estimated at over $10 billion. Previous empirical studies on the impact of climate exposure and influenza related mortality have observed that environmental factors are among the key determinants. Despite these hypothesized mechanisms, the impacts of humidity on mortality have not always been well established in the epidemiological literature. Using 3-hourly temporal resolution climatic data and weekly influenza mortality data, this paper utilizes Generalized Additive Models (GAM)
with smoothing functions and fixed effects panel regressions to investigate the relationship between climatic variables and influenza mortality rates in 122 cities in the US during 1970 – 2010. We find non-linear effect of temperature and specific humidity on influenza related mortality rates. Furthermore, we provide empirical evidence for epidemiological theories that lower temperatures and humidity levels have greater positive impact on influenza related mortality.