Comparison of Bulk Multi-Category Rimed-Ice Microphysics Parameterization using an Idealized Supercell Storm

Ice microphysical processes are important in simulating realistic convective storms, especially in high resolution models with deep convection. They change ambient thermodynamics which modify storm dynamics, that in turn modify microphysical processes. Among a wide range of ice-phase particles, dense rimed ice has received much attention lately due to large latent heating/cooling associated with its phase change. Currently, many flavors of multi-category, multi-moment bulk microphysics schemes (BMPs) are available for storm-scale modeling. Surface distributions of rain, hail, and wind are found to be sensitive to representation of rimed-ice processes within each scheme. In this study, we compare three complex microphysics schemes in WRF, focusing on differences in rimed-ice parameterizations and their impact on surface rain and ice. They are the Milbrandt-Yau (MY), NSSL, and the new Predicted Particle Properties (P3) scheme and contain two categories with rimed ice in idealized supercell storm simulations. A polarimetric radar data simulator is used to evaluate their ability to reproduce well-known polarimetric signatures. Both the MY and NSSL schemes replicate a Z_DR arc, while the P3 scheme is unable to replicate this signature despite properly simulating ice size sorting. Only the NSSL scheme simulates the correct location of the hail signature in the FFD. The P3 scheme produces large rimed ice comparable to the MY and NSSL schemes but melts quickly below the melting layer. The MY scheme simulates hail size properly compared to previous iterations, while the P3 scheme reasonably parameterizes the production and sedimentation of ice.