干旱区研究

干旱区研究 ›› 2022, Vol. 39 ›› Issue (2): 368-378.doi: 10.13866/j.azr.2022.02.04

• 天气与气候 • 上一篇    下一篇

TIGGE降水预报在中国干旱半干旱地区的适用性评估

何超禄1,2(),吕海深1,2(),朱永华1,2,李文韬1,2,谢冰绮3,徐凯莉1,2,刘名文1,2   

  1. 1.河海大学水文水资源与水利工程科学国家重点实验室,江苏 南京 210098
    2.河海大学水文水资源学院,江苏 南京 210098
    3.珠江水利委员会珠江水利科学研究院,广东 广州 510630
  • 收稿日期:2021-08-06 修回日期:2021-12-14 出版日期:2022-03-15 发布日期:2022-03-30
  • 通讯作者: 吕海深
  • 作者简介:何超禄（1996-）,男,硕士研究生,主要从事水文预报研究. E-mail: hechaolu2020@163.com
  • 基金资助:
    国家重点研发项目(2019YFC1510504)

Assessment of TIGGE precipitation forecast models in arid and semi-arid regions of China

HE Chaolu1,2(),LYU Haishen1,2(),ZHU Yonghua1,2,LI Wentao1,2,XIE Bingqi3,XU Kaili1,2,LIU Mingwen1,2   

  1. 1. State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, Jiangsu, China
    2. College of Hydrology and Water Resources, Hohai University, Nanjing 210098, Jiangsu, China
    3. Pearl River Water Resources Research Institute, Pearl River Water Resources Commission, Guangzhou 510630, Guangdong, China
  • Received:2021-08-06 Revised:2021-12-14 Online:2022-03-15 Published:2022-03-30
  • Contact: Haishen LYU

RichHTML

8

PDF (PC)

273

赞

摘要：

TIGGE降水产品作为目前中短期集合预报最权威的数据集合的重要组成部分,其在中国干旱半干旱地区的适用性如何,还需要进一步探讨。基于中国干旱半干旱地区2015—2017年实测降水数据,采用平均绝对偏差、均方根误差、TS评分等指标,从降水量预报、降水分级预报、降水探测能力和空间预报精度等多角度出发,综合评估了TIGGE数据中心的ECMWF、JMA、KMA和UKMO 4种模式在研究区的预报效果。结果表明：（1） 4种模式对小雨的预报效果均较好,在进行不同降水量级预报时,JMA对于小雨的降水预报效果最佳,而对于其他降水量级的降水预报,4种模式无明显差别;（2） 针对日降水量预报,KMA效果最差,而ECMWF最精确;（3） 在不同降水阈值下对降水的探测能力评估结果显示,ECMWF更具优势,尤其在以25 mm·d-1为阈值时优势明显;（4） 空间预报精度检验结果显示,各模式在80°~100°E,35°~45°N范围表现最佳,主要是新疆中部以及新疆、甘肃和青海三省交界处,对比各模式之间,ECMWF表现更稳定,KMA则表现较差。

关键词: TIGGE, 降水预报, 降水量级, 空间预报精度, 干旱半干旱地区, 中国

Abstract:

Short- and medium-term precipitation forecast products are important to improve the prediction period and accuracy of flood forecasts. With global climate change, the prediction of precipitation patterns becomes increasingly complex and important. To date, there have been no available reports on the applicability of TIGGE precipitation products (a significant aspect of the most authoritative data set for short-and medium-term ensemble forecasts) in the arid and semi-arid regions of China. On the basis of precipitation data measured from 2015-2017 in arid and semi-arid regions of China, the mean absolute deviation, root mean square error, TS score, and other indicators were used to analyze the precipitation forecast, precipitation classification forecast, precipitation detection ability, and spatial prediction accuracy. The prediction effects of the ECMWF, JMA, KMA, and UKMO models in the TIGGE data center were evaluated comprehensively in the study area. The results show that the four models effectively forecast light rain. The JMA model has the best forecast efficacy for light rain at different precipitation levels; for other precipitation levels, no significant difference was seen between the four models. The KMA model performs the worst for the daily precipitation forecast, whereas the ECMWF model is the most accurate. The evaluation results of precipitation detection capabilities under different precipitation thresholds show that ECMWF is more advantageous, especially when the threshold is 25 mm·d-1. The spatial accuracy test results show that each model performs better in the range of 80°-100°E and 35°-45°N, mainly in central Xinjiang and at the junction of Xinjiang, Gansu, and Qinghai provinces. Out of all the models, the ECMWF model demonstrated the best performance and the KMA model the worst.

Key words: TIGGE, precipitation forecast, precipitation level, space-prediction accuracy, arid and semi-arid regions, China