Comparing Approximated Heat Stress Measures Across the United States

Abstract

Abstract Extreme heat is a great threat to human health, and a growing public health concern, with recent higher intensity and duration of heat days coupled with increasing population exposure to hot environments. Utilizing accurate weather information and measures that reflect what people experience is a key component to understanding extreme heat impacts on human health. Numerous studies have suggested various heat metrics Heat Index (HI) and Wet Bulb Globe Temperature (WBGT) have been widely used in heat exposure guidelines. However, there are few systematic comparisons of daily heat measures and weather variables such as daily relative humidity, wind speed, and solar radiation which are highly related to human body thermoregulation and physiologic impact of heat. We compared three relevant heat measures (HImax, WBGTBernard, and WBGTLiljegren), derived from three widely-used gridded weather datasets (ERA5, PRISM, and Daymet) with ground-based weather observations. The heat measures calculated from gridded weather data and station data showed fairly strong agreement (R2 0.82–0.96, Root Mean Square Error (RMSE) 1.69–5.37°C). However, the discrepancies varied according to Köppen-Geiger climates (e.g., Adjusted R2 HImax (0.61–0.96), WBGTBernard (0.64–0.94), and WBGTLiljegren (0.34–0.94)). Gridded weather datasets offer a fairly reliable approach to assessing heat exposure of meteorological variables and heat measures. However, further research and establishing local ground station networks are necessary to reduce exposure measurement error and improve accuracy to ultimately better and more robustly understand the links between humid heat and health outcomes.