“23·7”华北特大暴雨数值预报检验评估

Verification and Assessment of "23·7" Severe Rainstorm Numerical Prediction in North China

  • 摘要: 针对“23·7”华北特大暴雨过程, 采用天气学检验及TS(threat score)评分和MODE(method for object-based diagnostic evaluation)方法对中国气象局高分辨率全球同化预报系统(CMA-GFS)、较低分辨率全球集合预报系统(CMA-EPS)、欧洲中期数值预报中心集合预报系统(EC-EPS)和业务预报模式(EC-HR)、美国环境预报中心全球预报系统(NCEP-GFS)等全球模式和中国气象局区域台风数值预报系统(CMA-TYM)、中尺度天气数值预报系统(CMA-MESO)和区域数值预报系统(CMA-BJ)等进行中短期预报效果检验评估。结果表明:EC-EPS提前14 d预报京津冀一带有过程累积降水量超过100 mm强降水的发生概率, CMA-EPS可提前12 d报出, 但预报欠稳定且落区偏东偏南。EC-HR对100 mm以上过程累积降水量及2 d以上暴雨日的位置预报提前时效均达8 d左右, CMA-GFS的过程累积降水量预报显著偏小、强降水落区明显偏东, 可用预报时效短;NCEP-GFS预报性能介于二者之间。各模式均可提前36 h预报强降水落区和强度的变化趋势, 中尺度模式可更加精细地刻画其形态和位置分布, 尤以CMA-BJ为佳, 但其预报偏强, 其余模式不同程度偏弱, 其中CMA-GFS显著偏弱。EC-HR提前8 d预报关键影响系统发生发展, 但低层倒槽位置偏西偏北, 低空急流偏弱, 低估了地形对强降水的增幅作用, 是太行山东麓降水量预报偏弱的重要原因之一。整体上, EC-EPS、EC-HR的提前时效和稳定性, 以及CMA-BJ的落区形态和强度预报等对预报业务有较高参考价值。

     

    Abstract: During the severe rainstorm in North China from 31 July to 1 August in 2023, CMA-GFS, CMA-EPS, EC-EPS, EC-HR, NCEP-GFS, CMA-TYM, CMA-MESO, and CMA-BJ are tested and evaluated using synoptic verification, threat score (TS), and MODE (method for object-based diagnostic evaluatin). The persistence and intensity of long-term heavy rainfall, as well as the area and intensity of short-term heavy rainfall, are tested and analyzed for their effectiveness over time. Results indicate that the cumulative precipitation predicted by EC-EPS may exceed 100 mm for 14 days in advance, but there is no prediction ability for extreme heavy precipitation above 600 mm. EC-HR forecast for the location of precipitation is generally accurate up to 8 days in advance. In the short term, the daily precipitation intensity forecast by CMA-BJ closely matches the actual situation, indicating its significance in predicting precipitation extremes. The average and maximum precipitation values of CMA-GFS, EC-HR, and NCEP-GFS in the areas with concentrated heavy precipitation are lower than actual values. CMA-GFS doesn't perform very well while EC-HR is closer to the actual situation. CMA-GFS, EC-HR, and NCEP-GFS models all provide inadequate forecasts for the persistence of heavy rainfall. However, EC-HR has a relative advantage in predicting persistent precipitation 8 days in advance. TS of CMA-BJ is highest for precipitation forecasts above 50 mm and 100 mm. EC-HR and CMA-TYM precipitation forecasts above 50 mm are relatively stable. From the daily MODE results of the precipitation concentration period from 29 July to 31 July, it is evident that EC-HR exhibits a northward predictive characteristic, while the prediction of CMA-BJ is slightly southward. The forecasting ability of CMA-GFS is insufficient, and forecasts from NCEP-GFS and CMA-MESO are not stable. The high-pressure system in North China has a significant impact on the precipitation. EC-HR model forecasts the formation and reinforcement of a 500 hPa high-pressure system 3 to 4 days earlier than CMA-GFS and NCEP-GFS models. It also surpasses both in predicting the precise location and strength of intense precipitation. Additionally, EC-HR model predicts the emergence of a 925 hPa low-pressure trough and a low-level jet 7 days in advance. However, it underestimates the intensity of the trough and jet system, with the actual location being to the west to north. CMA-GFS and NCEP-GFS underestimate the impact of the Taihang Mountains on easterly winds, leading to significantly lower precipitation forecasts. The analysis of the deviation in the 36 h precipitation forecast for 30 July also shows that EC-HR has weak predictions for low-level wind fields, trough positions, and convective precipitation, resulting in a weak intensity of heavy precipitation and a west-north precipitation area.

     

/

返回文章
返回