Modeled and Observed Variations in Storm Divergence and Stratiform Rain Production in Southeastern Texas

Storm divergence profiles observed by an S-band Doppler radar are compared to ensemble simulations of 10 disparate precipitating systems occurring in warm-season, weakly baroclinic, and strongly baroclinic environments in southeastern Texas. Eight triply nested mesoscale model simulations are conducted for each case using single- and double-moment microphysics with four convective treatments (i.e., two convective parameterizations and explicit versus parameterized convection at 9 km). Observed and simulated radar reflectivities are objectively separated into convective, stratiform, and nonprecipitating anvil columns and comparisons are made between ensemble mean echo coverages and levels of nondivergence (LNDs). In both the model and observations, storms occurring in less baroclinic environments have more convective rain area, less stratiform rain area, and more elevated divergence profiles. The model ensemble and observations agree best for well-organized leading-line trailing-stratiform systems. Excessive convective area fractions are simulated for several less-organized cases, especially those whose ensemble mean LNDs are about 1-2 km more elevated than observed. Simulations parameterizing convection on the intermediate grid produced less-elevated divergence profiles with smaller magnitudes compared to their explicit counterparts. In one warm-season case, utilizing double-moment microphysics when parameterizing convection on both outer grids generated lower LNDs associated with variations in convective intensity and depth, detraining less ice to anvil and stratiform regions at midlevels relative to its single-moment counterpart. Similarly, mesoscale convective vortex simulations employing an ensemble-based versus a single-closure convective parameterization on both outer grids produced the least-elevated heating structures (closer to observed), resulting in the weakest midlevel vortices.