Numerical Simulation of a Torrential Rain Event in the Northeast of Huaihe Basin. Part I: Model Verification and Analysis of MCSs

A torrential rain event accompanying Typhoon Prapiroon occurred in 2000, with 24-h rainfall amount reaching 800mm near the typhoon center. This event is simulated by the nonhydrostatic mesoscale model ARPS (V5.2), with thriple one-way nested-grids. Grid spacings of 45, 15, and 5 km are chosen for the three nested domains. The corresponding grid sizes are 75×75, 140×140, and 180×180, respectively. The NCEP/NCAR reanalysis data, radar echoes, and GMS-5 satellite images are assimilated with the ARPS model initially using a 3-D data assimilation system--ADAS. The simple ice phase scheme and the Kain-Fritsch cumulus parameterization scheme are used. There are 35 layers in the vertical, with a vertical grid spacing of about 625 m. The integration is performed up to 48 h from 0800 BT 29 to 0800 BT 31 August 2000. Compared with radar echoes, GMS-5 satellite images, and intensive surface observations, the results show that the heavy rain area down between the 500-hPa trough and the subtropical high in the left-front of Prapiroon is well simulated by the model ARPS, and the simulated rainfall centers are consistent with observations. A comparison of the radar echoes with these retrieved from the simulated hydrometeors reveals that there are meso-β scale convective systems that exhibit distinctive characteristics, and there are four convective belts converging in the vicinity of Xiangshui, where the maximum rainfall is observed. A further comparision of skew T-lgp diagrams from simulated and observed data demonstrates significant instability in this torrential rain process. The persistent vertical wind shear provides kinetic energy for the development of the MCSs, hence promoting the baroclinic development of convective cells, and the concentration of heavy rain at the specific location. The consistency between model results and observations encourages a further study of the torrential rain event using the simulation data.