中国农业气象 ›› 2023, Vol. 44 ›› Issue (10): 953-963.doi: 10.3969/j.issn.1000-6362.2023.10.008

• 农业气象信息技术 栏目 • 上一篇    下一篇

川渝地区汛期逐小时降水融合产品误差分析

旷兰,田茂举,李强,庞玥,吉莉,刘祥   

  1. 1.重庆市綦江区气象局,綦江 401420;2.重庆市气象台,重庆 401147;3.重庆市北碚区气象局,北碚 400700;4.重庆市气象服务中心,重庆 401147
  • 收稿日期:2022-10-25 出版日期:2023-10-20 发布日期:2023-10-11
  • 通讯作者: 田茂举,高级工程师,主要从事GIS和天气预报研究。 E-mail:147758044@qq.com
  • 作者简介:旷兰,E-mail:616032331@qq.com
  • 基金资助:
    重庆市自然科学基金面上项目(cstc2021jcyj-msxmX0057);重庆市气象部门业务技术攻关项目(YWJSGG- 202012;YWJSGG-202128);重庆市气象部门青年基金项目(QNJJ-202010)

Error Evaluation of Hourly Precipitation Fusion Products during Flood Season in Sichuan and Chongqing

KUANG Lan, TIAN Mao-ju, LI Qiang, PANG Yue, JI Li, LIU Xiang   

  1. 1. Qijiang District Meteorological Bureau of Chongqing, Qijiang 401420, China; 2. Chongqing Meteorological Observatory, Chongqing 401147; 3. Beibei District Meteorological Bureau of Chongqing, Beibei 400700; 4. Chongqing Meteorological Service Center, Chongqing 401147
  • Received:2022-10-25 Online:2023-10-20 Published:2023-10-11

摘要: 利用2021年汛期三套降水融合产品CMPAS、GSMaP和IMERG,经质量控制后的川渝地区190个国家气象站逐小时降水数据,结合川渝地区的不同地形区间和不同小时雨强对三套融合产品进行误差分析。结果表明:(1)2021年汛期总降水量空间分布上,CMPAS产品与站点实测最为接近,其次依次为GSMaP、IMERG产品。(2)针对不同地形区、不同时段,CMPAS产品的相关性、命中率、关键成功率最优,IMERG次之;CMPAS产品的偏差、均方根误差、误报率最小,GSMaP产品的误报率最大,IMERG的偏差、均方根误差最大;各融合产品在8月和9月的命中率、关键成功率最优,在20:00−次日2:00、2:00−8:00的相关性、命中率、关键成功率最优,误差和误报率最小。(3)针对不同小时雨强,CMPAS明显优于IMERG和GSMaP产品;其中CMPAS产品5月和9月的相关性、命中率、关键成功率最优;当小时雨强增大时,各融合产品的均方根误差逐渐增大。研究结果说明汛期CMPAS产品在川渝地区的精度明显优于IMERG和GSMaP产品,可以为地面实测数据缺乏的地区提供有效的降水数据补充。

关键词: 川渝地区, 降水融合产品, CMPAS, GSMaP, IMERG, 误差分析

Abstract: Three sets of precipitation fusion products CMPAS, GSMaP and IMERG in the flood season of 2021, and the hourly precipitation data of 190 national meteorological stations in Sichuan and Chongqing region after quality control were used to test the error analysis of the three sets of products in combination with different terrain intervals and different hourly rainfall intensity in Sichuan and Chongqing region, so as to provide data support for flood season hydrometeorological disaster prevention and reduction. Some results in this study showed that: (1) in terms of the spatial distribution of total rainfall in 2021 flood season, CMPAS products were the closest to the site measured precipitation, followed by GSMaP and IMERG products. (2) According to different topographic areas and different time periods, CMPAS products had the best correlation, hit rate and key success rate, followed by IMERG. The deviation, root mean square error and false positive rate of CMPAS products were the smallest, the false positive rate of GSMaP products was the largest, and the deviation and root mean square error of IMERG products were the largest. The hit rate and critical success rate of all fusion products were the best in August and September, and the correlation, hit rate and critical success rate were the best in 20:00−next 2:00 and 2:00−8:00, while the error and false positive rate were the least. (3) CMPAS was obviously better than IMERG and GSMaP for different hourly rainfall intensity. The correlation, hit rate and critical success rate of CMPAS products in May and September were the best. As the hourly rain intensity increased, the root mean square error of each fusion product increased gradually. The above results indicate that the accuracy of CMPAS products in flood season was obviously better than that of IMERG and GSMaP products in Sichuan and Chongqing, which could provide effective precipitation data supplement for areas lacking ground measured data.

Key words: Sichuan and Chongqing region, Precipitation fusion product, CMPAS, GSMaP, IMERG, Error analysis