The Sensitivity of a Supercell Cold Pool to Changes in Raindrop Breakup and Hail Fall Speed in Multiparameter WRF/DART EnKF Simulations

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Thursday, 8 January 2015: 2:45 PM
131C (Phoenix Convention Center - West and North Buildings)
Anthony E. Reinhart, Texas Tech University, Lubbock, TX; and C. C. Weiss, D. C. Dowell, and H. Morrison

An ongoing study is considering the ability of numerical weather prediction microphysical parameterizations to properly simulate supercell cold pools. Inaccuracies in these parameterizations have led typically to an overestimation of high-level clouds, precipitation amounts, and magnitude of evaporative cooling. Two-moment microphysics addresses several issues with particle size distributions, but challenges remain to further improve the effectiveness of these parameterizations. The parameterization of raindrop breakup and hail fall speeds used in schemes lead to a large difference in cold pool temperature and evolution of the supercell.

This study investigates a case from the Verification of the Origins of Rotation in Tornadoes Experiment 2 (VORTEX2) to determine the effect of changes in the aggressiveness of the drop breakup scheme and hail fall speed present in the Morrison microphysics parameterization on the supercell cold pool. WRF/DART simulations using full physics and assimilating WSR-88D and mobile radar radial velocity data onto a 1 km domain every two minutes were conducted. The EnKF technique was used in order to minimize the initial condition error and otherwise best produce the observed atmospheric state, allowing for a focus on differences that can be attributed to the changes in drop breakup and/or changes in the fall speed of hail. Each ensemble member has varying drop breakup aggressiveness and a different hail fall speed leading to a larger coverage of possible true observed values, maintaining ensemble spread. Detailed results of the effect of the changes in drop breakup and hail fall speed will be presented. Results will suggest that a range of values is necessary to create an accurate simulation for this particular case.