Evaluation of Hourly Snow-to-Liquid Ratio Algorithms for the U.S. Air Force

Operational weather units within the United States Air Force (USAF) disseminate watches, warnings, and advisories (WWA) for adverse weather conditions in order to minimize impacts to USAF missions, resources, and personnel. One such WWA requires forecasters to identify periods of heavy snow defined as >=2 inches of accumulation within 12 or less hours, which impacts personnel movement and airfield operations. Forecasters often use hourly snowfall rates to determine the likelihood of a heavy snow event; however, this parameter involves assumptions based on environmental factors (e.g. temperature) to compute snow-to-liquid ratios as it is not a direct model output variable. This study evaluates four different algorithm techniques for determining these ratios and verifies them using a heavy snow event that occurred over the central United States over a 24-hour period ending at 12 UTC on December 2, 2018. Given that hourly, observed snowfall accumulations for the time period are not available from observational weather datasets, this study proposes a novel approach as a proxy to perform objective verification. In particular, this technique disaggregates the daily national snowfall analysis provided by the National Weather Service (https://www.nohrsc.noaa.gov/snowfall/) into hourly components by combining information on precipitation type from standard surface observations with quantitative precipitation estimates from high-resolution gridded radar data (i.e. Multi-Radar Multi-Sensor system; https://www.nssl.noaa.gov/projects/mrms/). Standard contingency table metrics and skill scores are presented, and for comparison purposes, similar results for an alternate ground-truth source are generated using a 30-year climatology (http://www.eas.slu.edu/CIPS/slr.html).