The objective of this study is to modify a regional climate model (RCM)—the Regional Climate Model of the National Climate Center of China—to simulate the summer monsoon rainfall over South China and the South China Sea. Such a modification is necessary because this RCM was designed for studying the climate over central and north China where precipitation processes are very different from those occurring further south so that simulations using the basic parameters of the original model give precipitation amounts much less than those observed.
Using the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis data, many sensitivity experiments have been carried out for the months of April to June 1998, which include modifying the cumulus parameterization scheme, the large-scale precipitation scheme, the radiation scheme, the surface exchange processes, and the size of the buffer zone. The design of all the experiments is either to enhance the moisture provision to the atmosphere over South China and the South China Sea, or to help the model atmosphere realize the moisture to form precipitation. It is found that indeed modifications to the various parameters can produce rainfall amounts much closer to those observed.
Based on the results of these experiments, an “optimal” design of the RCM is reached and tested for its effectiveness using the NCEP reanalyses for two wet years (1994 and 1997) and two dry years (1996 and 1999) during which rainfall in May and June over South China were above/below normal respectively. The physical processes in this design include the Kuo scheme for convective precipitation, the Pal scheme for large-scale precipitation, the Holtslag scheme for the planetary boundary layer, the radiation transfer scheme of the NCAR Community Climate Model Version 3, and the Biosphere-Atmosphere Transfer Scheme for the land surface process. The width of the buffer zone is reduced and the effective cloud droplet radius is fixed at different values over land and ocean. The neutral drag coefficient is prescribed as a function of surface wind speed, and the heat and moisture exchange coefficients are set at values larger than that of momentum. The simulations based on this design for all the five years are found to be much closer to observations than those from the control.