Assessment and Evaluation of Heat Index and Wet Bulb Globe Temperature in North Carolina

Extreme heat days are among the highest impact weather events as they are the leading cause of weather-related mortality in the United States. As a result, a comprehensive assessment and evaluation of heat stress metrics is a necessary step in building a Weather Ready Nation, especially in light of future warming posed by climate change. Currently, the National Weather Service (NWS) issues operational heat alerts (heat advisories and excessive heat warnings) based on heat index (HI). These HI-based alerts only incorporate temperature and humidity, but vary by geographic region and Weather Forecast Office (WFO). Wet Bulb Globe Temperature (WBGT), an alternative heat stress metric that includes the effects of wind speed and solar radiation, has been gaining attention due to its ability to better represent the impacts of heat on the human body. While no formal heat warnings are issued based on WBGT, the National Digital Forecast Database has operationalized a WBGT product to supplement NWS heat forecasting efforts. In this work, we conduct a comprehensive assessment of HI and WBGT during the heat seasons (May-September) of 2020-2023. Our specific objectives are to 1) determine how well gridded dataset derived HI and WBGT compare to HI and WBGT observations in North Carolina; 2) evaluate the accuracy of operational NWS HI-based heat alerts and compare them to WBGT thresholds; and 3) determine which weather patterns contribute to potentially dangerous HI and WBGT levels in North Carolina.

To accomplish these goals, we first pair hourly HI and WBGT observations from the 44 stations of North Carolina’s Environment and Climate Observing Network (ECONET) to the nearest grid cells from the ERA5-Land dataset. Once the data has been paired, standard evaluation metrics such as correlation, root mean square error, and mean bias are computed to assess how well the ERA5-Land data represents heat observations. To evaluate the accuracy of NWS HI-based heat alerts, we gather archived NWS heat alerts from the Iowa Environmental Mesonet and identify potentially dangerous HI episodes from ECONET observations. Using this data, we then calculate the probability of alert issuance, false alarm ratio, and critical success index of the NWS heat alerts for each station. The observed potentially dangerous HI events are then compared against categorical WBGT threat levels (elevated, moderate, high, and extreme) to determine whether these metrics identify similar or different potentially dangerous heat conditions. To investigate the influence of weather patterns on extreme heat events, ERA5-Land data such as precipitation, MSLP, and wind speed are compared to potentially dangerous HI and WBGT events. Preliminary results suggest the probability of heat alert issuance is higher in the Coastal Plains, but false alarm rates are highest in the immediate coast and central NC. The mountains also experience more days per year over elevated WBGT criteria, but statistical evaluation suggests no discernable spatial patterns in WBGT biases. Understanding the implications of WBGT can help improve heat alerts that can better inform North Carolina communities.