Hailstorm simulations using the Weather Research and Forecasting (WRF) model: Microphysical parameterization sensitivities and preliminary verification

The development of the Weather Research and Forecasting (WRF) model has been useful to both the operational and research atmospheric science communities, allowing forecasters and researchers to benefit from the use of a sophisticated dynamical model with multiple options for physical parameterizations. However, very little work has been documented in the formal literature on the ability of the WRF model and its microphysics parameterizations to accurately simulate hailstorms. Hailstorms, which are responsible for nearly $1.4 billion of damage to crops and property annually in the United States, remain elusive to forecast. These forecasts often rely on rules of thumb such as large convective available potential energy in the storm environment, or outdated empirical methods for nowcasting such as the examination of the vertically integrated liquid water content (VIL) from storm reports upstream of the forecast area.

This study is a first step towards utilizing the WRF model for accurate simulations of historical hailstorms across the continental United States, in an effort to determine biases in the WRF model for hail forecasting. This study will focus on the well-documented “Mayfest” hailstorm that occurred on 5 May 1995 and produced numerous hailstones up to 114 mm (4 in.) in diameter along with extreme rainfall. The resulting damage to the Fort Worth, Texas area cost over two billion dollars and over 20 fatalities were reported. This paper will present preliminary results from WRF model storm simulations, highlighting both the microphysical and large-scale differences observed in the storm structure from the use of different microphysical schemes. Additionally, direct comparisons will be made between archived National Weather Service WSR-88D radar reflectivity and WRF simulated reflectivity fields throughout the lifetime of the storm and hailfall at the ground.