干旱气象 ›› 2024, Vol. 42 ›› Issue (2): 305-314.DOI: 10.11755/j.issn.1006-7639(2024)-02-0305

• 新气象数据检验与订证 • 上一篇    下一篇

基于准对称混合滑动训练法对CLDAS气温的订正检验

李奇临1(), 刘超2(), 朱君1, 刘昉1, 旷兰3   

  1. 1.中国气象局气候资源经济转化重点开放实验室,重庆市气象信息与技术保障中心,重庆 401147
    2.重庆市气象台,重庆 401147
    3.重庆市綦江区气象局,重庆 綦江 401420
  • 收稿日期:2022-07-20 修回日期:2023-05-04 出版日期:2024-04-30 发布日期:2024-05-12
  • 通讯作者: 刘超(1989—),女,高级工程师,研究方向为智能预报算法客观订正。E-mail:ajuchao@qq.com
  • 作者简介:李奇临(1987—),女,高级工程师,研究方向为气象观测数据质量控制与网格实况产品评估。E-mail:342455539@qq.com
  • 基金资助:
    重庆市气象部门业务技术攻关项目“重庆气温网格实况产品的偏差订正与评估”(YWJSGG-202110);“降水融合实况分析产品在重庆地区的订正”(YWJSGG-202307)

Correction and evaluation of CLDAS air temperature products based on quasi-symmetric hybrid sliding training method

LI Qilin1(), LIU Chao2(), ZHU Jun1, LIU Fang1, KUANG Lan3   

  1. 1. CMA Key Open Laboratory of Transforming Climate Resources to Economy, Chongqing Meteorological Information and Technical Support Center, Chongqing 401147, China
    2. Chongqing Meteorological Observatory, Chongqing 401147, China
    3. Qijiang Meteorological Bureau, Qijiang 401420, Chongqing, China
  • Received:2022-07-20 Revised:2023-05-04 Online:2024-04-30 Published:2024-05-12

摘要:

基于准对称混合滑动训练期方法,对近两年来中国气象局陆面同化分析系统(CMA Land Data Assimilation System,CLDAS)输出的日最高气温、日最低气温网格实况分析产品进行订正,以期提高该产品在重庆地区的适用性。结果表明:订正前,2021年CLDAS日最高气温产品的平均误差为0.63 ℃,平均绝对误差为1.14 ℃,订正后平均误差减小至-0.03 ℃,平均绝对误差减小至0.64 ℃,误差小于或等于1 ℃的准确率由约64%提高到约90%,明显改善了该产品在重庆西部和东南部地区的适用性;订正前,2021年CLDAS日最低气温的平均误差为-0.22 ℃,平均绝对误差为0.75 ℃,订正后平均误差减小至-0.03 ℃,平均绝对误差减小至0.55 ℃,误差小于或等于1 ℃的准确率由约91%提高到约93%,改善了该产品在重庆中部地区的适用性。

关键词: 偏差订正, 准对称混合滑动训练, CLDAS, 评估

Abstract:

Based on the quasi-symmetric mixed sliding training period method, the grid analysis products of daily maximum temperature and daily minimum temperature from Land Data Assimilation System of China Meteorological Administration (CLDAS) in recent two years are revised, in order to improve the applicability of the products in Chongqing. The results show that before revision, the average error of daily maximum temperature products in 2021 was 0.63 ℃, and the average absolute error was 1.14 ℃. After revision, the average error decreased to -0.03 ℃, the average absolute error decreased to 0.64 ℃, the accuracy of error less than or equal to 1 ℃ was improved from about 64% to 90%, which obviously improved the applicability of products in western and northeastern Chongqing. Before revision, the average error of daily minimum temperature products in 2021 was -0.22 ℃, and the average absolute error was 0.75 ℃. After revision, the average error was reduced to -0.03 ℃, the average absolute error was reduced to 0.55 ℃, and the accuracy of error less than or equal to 1 ℃ was improved from about 91% to 93%, which improved the applicability of products in central Chongqing.

Key words: deviation correction, quasi-symmetric hybrid sliding training, CLDAS, evaluation

中图分类号: