Key points are not available for this paper at this time.
We have performed three-dimensional thunderstorm real simulations using the two-moment and three-moment bulk microphysics schemes in the Weather Research and Forecasting (WRF) model. We have analyzed three cases to understand the potential differences between the double-moment (Morrison-2M) and National Taiwan University triple-moment (NTU-3M) microphysics parameterizations in capturing the characteristics of lightning events over the Indian subcontinent. Despite general resemblances in these schemes, the simulations reveal distinct differences in storm structure, cloud hydrometeors formation, and precipitation. The lightning flash counts from the in situ lightning detection network (LDN) are also used to compare the simulation of storms. The Lightning Potential Index (LPI) is computed for Morrison-2M and NTU-3M microphysics schemes and compared it with the Lightning Detection Network (LDN) observation. In most cases, the Morrison-2M shows more LPI than the NTU-3M scheme. Both the schemes also differ in simulating rainfall and other thermodynamical, dynamical, and microphysical parameters in the model. Here, we have attempted to identify the basic differences between these two schemes, which may be responsible for the discrepancies in the simulations. In particular, the Morrison-2M produced much higher surface precipitation rates. The effects on the size distributions cloud hydrometeors between two microphysical schemes are important to simulate the biases in the precipitation and lightning flash counts. The inclusions of ice crystal shapes are responsible for many of the key differences between the two microphysics simulations. Different approaches in treating cloud ice, snow, and graupel may have an impact on the simulation of lightning and precipitation. Results show that the simulation of lightning events is sensitive to microphysical parameterization schemes in NWP models.
Mallick et al. (Fri,) studied this question.