A three-dimensional, time-dependent, non-hydrostatic model is used to simulate the microphysical process in a convective precipitation system developed in the northern Taiwan. This simulation uses the hail parameterization model (HPM) version of the Wisconsin Dynamical/Microphysical Model (WISCDYMM), in which the precipitation hydrometeors are assumed to follow exponential size distribution, while cloud water and cloud ice are assumed to be monodispersed. The simulation is carried out for 120 minutes. The results exhibit key dynamic and thermodynamic features characteristic of the observed system, including an abrupt rainfall, a cold pool and split cell developing as system propagates. However, the orographic effect is not simulated here.

Integrated hydrometeor mass for the entire domain shows that about 50% hydrometeor mass is in ice phase, with graupel and hail the most predominant type. The other 50% are contributed almost entirely by rain water. The total mass of rain, graupel and hail increase rapidly as system intensifies, while the total mass of ice and snow increase much more slowly and remain fairly steady as shown by time evolution of the hydrometeors. After 40 minutes, the percentage mass of each hydrometeor type in the domain seems to reach a quasi-steady state even though the total mass keeps growing. Production and depletion curves of microphysical processes for each hydrometeor type indicate that rain primarily comes from the shedding of liquid layer and melting of graupel and hail. Snow is primarily initiated by Bergeron-Findeisen process and accretion of rain droplet by cloud ice, and grows predominately from the Bergeron-Findeisen process. Hail and graupel are primarily initiated through rain-snow collision and deposition of water vapor, and grow mainly by accretion by rain and cloud water. The main depletion mechanism of rain is the accretion by graupel and hail for rain. Melting and shedding to form rain drops is the largest depletion process of graupel and hail.