中文摘要：

基于GRAPES_Meso区域集合预报系统,设计了三种集合预报扰动方案,即多初值、多初值多物理、多初值多物理多边值,并针对三种方案进行了连续一个月的批量试验,重点分析了2008年7月23日江淮暴雨过程.结果表明,对于降水预报,三种集合扰动方案均相对于控制预报均有所改善,多初值多物理与多初值多物理多边值方案对小雨、中雨预报改进效果显著,对暴雨预报略有改进;多初值方案仅能产生有限的集合离散度且难以增长,引入物理参数方案扰动及边界条件扰动能显著提高集合离散度,改善各物理量场的预报效果;通过比较,多初值多物理多边值为最优方案.该批量试验表明,模式物理过程及边界条件是影响GRAPES _Meso区域集合预报不确定性的不可忽视因素.

英文摘要：

Based on the GRAPES_Meso regional ensemble forecast system,three ensemble perturbation schemes are designed and compared,namely,multi-initial,multi-initial-physics and multi-initial-physicsboundary.Consecutive experiments for a period of one month in July 2008,during which heavy rainfall processes occurred in Jianghuai,are conducted in this paper.The results are as follows：All ensemble schemes can outperform the control forecast,the improvement for small rain forecast is obvious whereas the improvement for heavy rain forecast is slight;the multi-initial scheme can only produce limited ensemble spread;the introduction of physical parameterization perturbation and lateral boundary perturbation could significantly amplify ensemble spread and improve the ensemble forecast of each quantity.After all,the scheme of multi-initial-physics-boundary is the best.The results indicate that model physics and lateral boundary are important factors for GRAPES_Meso regional ensemble forecast uncertainty.