Impacts of Distance from the Nearest Radar, Time of Day, Resident Population, and Season on Severe Warning Performance. Part I: CONUS Perspective

Severe weather warnings notify the public about imminent threats to life and property. Over the last few years, the national false alarm ratio (FAR) for severe thunderstorm warnings has been around 0.49 and the probability of detection (POD) has been 0.78. While there has been research into how tornado warning verification has evolved by time of year and time of day, a similar study has not been performed for severe thunderstorm warnings, which constitute a large majority of warnings issued across the United States. We evaluate how warning performance varies by hour of the day, month of the year, season, population, and distance of the warning or event from the nearest radar for all National Weather Service forecast areas across the continental United States. All datasets from severe thunderstorm warnings and severe weather events (severe wind gusts, wind damage and hail) were evaluated from 2010 to 2018. This 9-year period was selected to be consistent with the most recent change in severe thunderstorm warning criteria of 25.4mm (1 inch or quarter size) hail or greater. A preliminary examination of data from the north central United States has shown that FAR (POD) is higher (lower) at night than during the day while there is little difference in FAR or POD as a function of distance from radar. By examining performance over the continental United States (CONUS), we will find if severe thunderstorm warning verification is impacted by different non-meteorological factors including distance from radar, time of day, time of year, and population. Results seek to identify future research and training in severe weather warnings to address these issues and decrease their impact on warning performance.