| 摘要: |
| 为了解ECMWF高分辨率数值预报模式（以下简称“ECMWF”）对广安地区暴雨的预报性能，提高预报质量。利用ECMWF模式资料、广安地区4个国家观测站和160个区域自动站的逐日降水资料，对2015～2017年广安地区出现的34次中雨及以上量级降水（其中，14次暴雨）天气过程进行分析。结果表明：直接以ECMWF暴雨预报结果进行暴雨预报，易出现漏报，评分质量较低。当ECMWF预报满足：中、短期均预报降水R≥50mm；中期预报R≥25mm、短期预报R≥50mm；中期预报R≥50mm、短期预报R≥25mm；中、短期均预报R≥35mm四个订正条件之一时，进行订正预报暴雨，评分高，空、漏报率低，命中率高，可有效提高暴雨预报质量。同时，ECMWF在暴雨的持续时间、强降水时段、落区和面雨量预报中均有良好的预报指导意义。 |
| 关键词: 高分辨率数值模式;暴雨;检验 |
| DOI： |
| 投稿时间：2018-04-28修订日期：2018-08-10 |
| 基金项目:广安市气象局2017年度科学技术研究课题“ ECMWF高分辨率温度、降水产品在广安的释用性能检验(2017-5)” |
|
| Verification and Analysis on the Rainstorm Forecast Performance of ECMWF High Resolution Numerical Model in Guang''an |
| WANG Ping |
| (Guang''an Meteorological Bureau) |
| Abstract: |
| To better understand and improve the forecast performance of the rainstorm in Guang''an city, the authors explored the weather circumstance of 34 precipitation processes (including 14 rainstorm cases) with moderate precipitation and beyond occurred in Guang''an city during period 2015-2017 by employing the precipitation data from the ECMWF high-resolution numerical forecasting model (hereinafter referred to as ECMWF), 4 national observation stations in Guang''an, and 160 regional auto stations. The result shows that a direct forecasting of rainstorm based on the ECMWF rainstorm forecast output would result in a missed report and a low score. However, when the ECMWF forecast satisfies one of the following four correction conditions: medium and short-range forecast R ≥ 50mm, medium-range forecast R ≥ 25mm and short-range forecast R ≥ 50mm, medium-range forecast R ≥ 50mm and short-range forecast torrential rain R ≥ 25mm, and both the medium and short-range forecast R ≥ 35mm, the forecasting of torrential rains will be carried out with a high score, and a low rate of missing reports, and a high hit rate, which can effectively improve the forecast quality. And it is also suggested that the ECMWF shows a good forecasting skill in the forecasting of rainstorm duration, strong precipitation period, falling area of rainstorm, and area rainfall amount. |
| Key words: High-resolution numerical model;rainstorm;verification |
