| 摘要: |
| 利用CMORPH融合降水产品、地面常规雨量资料和ECMWF集合预报产品,在对CMORPH资料应用于淮河流域实况面雨量监测的合理性进行充分探讨的基础上,检验评估ECMWF集合预报的性能。结果表明：两种实况面雨量计算结果相关程度高,CMORPH融合降水产品可用于淮河流域面雨量监测；ECMWF确定性预报晴雨预报正确率随着预报时效的延长呈下降趋势,各等级TS评分平原高于山区,大暴雨以下量级降水预报表现出不同程度的高估；除沂蒙山区,其他地区各累积时段集合预报性能随着预报时效的延长逐渐降低,汛期山区整体预报性能好于平原地区,非汛期反之；集合预报相对于气候均值预报的优势在0-96小时更为明显。 |
| 关键词: 淮河流域;面雨量预报;检验评估;预报性能 |
| DOI： |
| 投稿时间：2018-09-26修订日期：2018-11-27 |
| 基金项目:(HRM201409)。 |
|
| Performance evaluation of areal precipitation forecasting for Huaihe River Basin based on ECMWF Ensemble System |
| LUO Jing,YE Jinyin,WANG Xi,LIU Jing |
| (Anhui Meteorological Observatory;Anhui Branch of Meteorological Cadre Training College of China Meteorological Administration;Taizhou Meteorological Bureau;Bengbu Meteorological Bureau) |
| Abstract: |
| Using CMORPH fusion precipitation products, conventional observation precipitation data and ECMWF ensemble forecast products, whether CMORPH data can be used in monitoring the areal precipitation in Huaihe River Basin are discussed before testing and evaluating the performance of ECMWF ensemble forecast products. The results show that: the correlation of two kinds of areal precipitation is significant, so CMORPH fusion precipitation products can be used to monitor areal precipitation for Huaihe River Basin；the accuracy of clear or rain forecast of ECMWF deterministic forecast decreases with the increase of forecast period, Treat Score of each precipitation grade is higher in plain than in mountain area,there are different degrees of overestimation of precipitation below heavy rain；in addition to the Yimeng mountain area, the performance of ensemble forecast gradually decreases with the increase of the forecast period in other areas for every accumulative period,the overall forecast performance of the mountain areas in flood season is better than that in the plain, and the non flood season is the opposite；the advantage of ensemble forecast relative to climate mean forecast is more obvious in 0-96 hours. |
| Key words: Huaihe River Basin;precipitation forecast;test and evaluate;forecast performance |
