Impact of Microphysical Parameterizations on Supercell Thunderstorm Cold Pools using EnKF Data Assimilation Experiments

Errors in high-resolution numerical simulations of supercell thunderstorms can stem from inaccuracies in initial conditions, parameterizations, and model numerics. Assimilating mesoscale data into an ensemble of parallel simulations can mitigate initial condition and, to a lesser extent, model bias errors, permitting focus on errors largely attributable to bulk microphysical parameterizations. Inaccuracies in these parameterizations have led typically to an overestimation of high-level clouds, precipitation amounts, and the magnitude of evaporative cooling. All of these effects can affect the amount of observations assimilated and impact that those newly assimilated observations have on the EnKF analysis.

This ongoing study considers three target storms from the Verification of the Origin of Rotation in Tornadoes Experiment (VORTEX2) and uses three different microphysical parameterizations (LFO, Milbrandt-Yao, and Morrison) to determine which of these popular schemes produce the most realistic cold pools. Data Assimilation Research Testbed (DART) software, using an EnKF technique coupled with the Weather Research and Forecasting (WRF) model, is used to develop a mesoscale background and then scale down to a domain with 1 km horizontal grid spacing. Mobile radar radial velocity data and WSR-88D data are assimilated into this 1 km domain every two minutes. The impact of the different microphysical parameterization on the model analysis and on the assimilation of new observations is being investigated. Updating number concentration and mixing ratio from multi-moment microphysical parameterizations result in a more realistic analysis. However, we find a finely tuned single moment microphysical parameterization can rival the sophisticated multi-moment schemes.