Good Nus and Bad Nus: An Investigation of the Raindrop Shape Parameter in a Deep Convective Simulation

While double moment microphysical parameterizations require fewer a priori parameter selections than single moment schemes, they still need some prior-set description of the hydrometeor size distributions. In most microphysical parameterizations, the gamma probability distribution is used with some selected shape parameter (nu), which controls the width of the drop size distribution (DSD). It is anticipated that a fixed shape parameter might exert a significant impact on cloud microphysical and dynamical processes, and yet the sensitivity of three-dimensional simulations of deep convective storms to this parameter, when applied to rain, has not been fully explored. In order to address this, a suite of simulations of an idealized supercell using the Regional Atmospheric Modeling System (RAMS) is conducted, with changes to the rain shape parameter run along with changes to the collection kernel. This allows us to understand the sensitivity of the model solution to both of these factors and interpret how important the selection of the rain shape parameter is when simulating deep convection. The results indicate that the solution is much more sensitive to the selection of the rain shape parameter than to the collection kernel when examining surface rainfall. The differences in the surface rainfall are primarily driven by three factors as the DSD narrows: (1) an increase in below-cloud evaporation, (2) decreased rain production due to decreases in ice melting, and (3) slower raindrop fall speeds, leading to longer residence time. Care should be taken when selecting this a priori parameter.