基于CRA技术的华南前汛期强降水EC模式预报误差分析

利用CRA空间检验技术对ECMWF模式36 h时效预报的2016—2018年华南前汛期(4—6月)69个降水目标进行了检验及误差统计分析，并将预报落区偏差相似个例的环流形势及天气尺度影响系统进行了分析。(1) 87% 的强降水目标存在明显落区预报偏差，最大偏差为2.75 °。偏差以经向偏差为主，其中偏北的目标多于偏南的目标，平均偏北0.6 °；无系统性纬向偏差。(2) 模式预报的降水面积较实况偏大的个例多。(3) 不同降水落区预报偏差类型月份分布、对应的环流特征与天气尺度影响系统具有一定的差异性。4月各偏差类型出现的频次相当，5月以西北型个例为主，6月东北型个例最多。西北型个例天气尺度影响系统以长波槽或东北冷涡、冷式切变线为主，西南型、东北型个例主要受南支波动与中纬度短波槽影响，低层低涡、冷、暖式切变线等出现的频次差不多。通过降水预报落区偏差较大和较小的个例对比分析，表明模式强降水落区预报偏差可能与对流组织化发展程度以及暖区是否存在有利于对流发展条件等有关。 

FORECAST ERROR ANALYSIS OF EC MODEL FOR HEAVY RAINFALL DURING ANNUALLY FIRST RAINY SEASON IN SOUTH CHINA BASED ON CRA METHOD

The present study verified the ECMWF model 36-hour forecast of 69 precipitation events from April to June in 2016—2018 during the annually first rainy season in south China and the forecast error was statistically analyzed by using the contiguous rain area (CRA) spatial verification method. Furthermore, analysis of the circulation situation and synoptic scale system of the cases with similar forecast rainfall location error were carried out. The results showed that: (1) 87% forecasts of the heavy rainfall events have obvious error in rainfall location and the maximum error is 2.75° deviation from the observation. The deviation is dominantly meridional ones, in which cases with north rainfall error are more than the south one, with an average of 0.6° northward deviation to the observation. No obvious zonal deviation is seen. (2) There are many cases where the precipitation area predicted by the model is larger than the observed. (3) The monthly distribution of forecast deviation types in different precipitation areas, the corresponding circulation characteristics and the synoptic influence system show certain differences. In April, the cases of different types show uniform distribution. However, the northwest and northeast types are the main types in May and June, respectively. The synoptic system of the northwest type is dominated by long-wave trough or the northeast cold vortex and cold shear line. The northeast and southwest types are mainly affected by the south branch trough and mid-latitude shortwave trough, while the effect by the vortex, cold and warm shear line are almost the same. Finally, comparative analysis of two cases with large and small rainfall location error indicate that predictability of heavy rainfall depends on the development of organization of convective systems and initiation of convective systems in the warm area, etc.