中文摘要：

积云对流和云微物理是数值天气预报模式中最为重要的两类湿物理过程,它们共同影响云和降水的预报性能。通过采用CMAP降水资料和MODIS、MLS及Cloud Sat卫星云观测资料对全球中期数值预报模式GRAPES_GFS中这两类湿物理过程参数化方案的不同组合所预报的降水场和云宏微观场进行诊断和评估,以揭示其对云和降水的预报性能。结果表明：（1）云微物理方案是中高纬度地区总降水预报差异的主因,三种云微物理方案预报的降水强度为SI〉NCEP3〉NCEP5。赤道及低纬地区降水差异主要是由积云对流方案引起的,KF_SI组合与CMAP降水最为一致。（2）SI方案和NCEP3方案在中纬度地区格点降水要显著多于混合相云NCEP5方案;与SAS方案和KF方案相比,BM方案使与其组合的云方案产生的格点降水明显偏少。（3）BM方案产生的对流降水要明显多于SAS方案和KF方案,中高纬地区SAS方案和KF方案预报的对流降水基本一致,在低纬地区SAS方案对流降水最少。（4）NCEP5方案预报的云顶温度与MODIS观测吻合较好,NCEP3方案和SI方案预报的云顶温度要较实况偏暖。三种对流方案预报的云顶温度冷暖关系为SAS，

英文摘要：

Cumulus convection schemes and cloud microphysical schemes are the most important moist physical processes in numerical weather model,which have significant effects on cloud and precipitation forecast performance.Cloud and precipitation forecast performances by different combinations between the above precipitation processes parameterization schemes in GRAPES_GFS model have been diagnosed and validated using CMAP precipitation observation data and MODIS,MLS and Cloud Sat satellite data.The results show that：（1）The distinction of cloud microphysical schemes is a predominant reason for precipitation forecast difference at mid-high latitude areas.There is a sequence of SINCEP3NCEP5 according to the precipitation intensity of order.While the distinction of cumulus convection schemes is the main reason for those at low latitude areas.（2）There is more grid-scale precipitation of SI and NCEP3 schemes than that of NCEP5 scheme at middle latitude areas.Compared with SAS and KF scheme,BM scheme is prone to causing more grid-scale precipitation with its combinations partner cloud scheme.（3）There is obviously more convection precipitation of BM scheme than those of SAS and KF schemes.The convection precipitation magnitude of SAS scheme and KF scheme is very approximate at mid-high latitude areas,nevertheless,that of SAS scheme is the least among the three convection schemes at low latitude areas.（4）The cloud top temperature of NCEP5 is in good agreement with MODIS satellite data,while those of NCEP3 scheme and SI scheme are warmer than MODIS.There is a sequence of SASBMKF from cold to warm for their cloud top temperature.（5）The integration cloud water content of NCEP5 is closest to MODIS observation and there are significantly less for the two simple ice microphysical schemes,especially for SI scheme.KF scheme has an obvious better performance for integration cloud water prediction than those of SAS and BM scheme.（6）There are all negative biases of cirrus cloud compared with MLS observation for thre