中文摘要：

基于GRAPES中尺度数值天气预报模式,研究了不同复杂程度陆面过程和不同土壤初始条件对2007年中国夏季淮河流域降水数值预报的影响。重点分析了下垫面非均匀性对造成我国夏季强降水的中小尺度对流系统的启动和发展的影响,探讨了中小尺度对流发生发展中陆气相互作用的可能影响。对2007年7月8日发生在淮河流域的一次对流过程采用两组敏感性试验进行研究,结果表明对流启动对陆面过程有很强的敏感性。Noah陆面模式能合理地模拟出对流的启动,而Slab陆面模式模拟的对流会延迟1～2小时,这与Noah陆面模式能合理描述地表感、潜热通量有关。GLDAS初始土壤资料能更加合理地描述实际土壤温湿的分布状况,有利于更加准确地模拟出对流降水的启动和分布位置。另外,对2007年7月淮河流域的持续性强降水天气过程进行了对比模拟,研究表明：不同复杂程度陆面过程对连续强降水数值预报有明显的影响。使用Noah陆面模式能提高对强降水预报的能力,而且随着模式分辨率的提高,降水的TS评分也在提高,Noah的TS评分总体上高于Slab。

英文摘要：

The impacts of different land surface processes and soil initial conditions on the rainfall forecast over Huai River basin in summer of 2007 have been investigated by using a regional numerical weather predication model GRAPES（Global/Regional Assimilation and Prediction System）.This study focuses on analyzing the effects of surface heterogeneity and land-atmosphere interactions on the initiation and development of meso-scale convections which always induce the heavy rainfall during summer in China.Two numerical experiments are conducted for a convection initiation case occurring in July 8,2007 over Huai River basin.The results show that the convection initiation is obviously sensitive to the land surface process.The NOAH land surface model（LSM） applied in this study performs much better in simulating the initiation of convections than the SLAB LSM does.The initiation of convection is delayed by 1～2 hours when SLAB LSM is utilized.Such difference is strongly related to the different treatment of surface sensible and latent heat fluxes in these two LSMs,while the NOAH LSM describes the surface sensible and latent heat fluxes much well.In addition,soil initial conditions from GLDAS show better simulation of convection initiation due to the reasonable description of the surface characteristics.Furthermore,the successive 24 hour rainfall forecast is conducted for a heavy rainfall period occurred in July 2007 over Huai River basin.The results show the distinct impacts of different land surface processes on the numerical prediction of the summer continuous precipitation.Obvious improvements are found in simulations of location and intensity of heavy rainfall by the NOAH,which is close to the observation.The threat score（TS） of precipitation increases when the model resolution becomes higher and the TS in NOAH is larger than in SLAB.