干旱气象 ›› 2020, Vol. 38 ›› Issue (6): 1023-1030.

• 业务技术应用 • 上一篇    下一篇

两种订正方法在河南高速公路沿线气温精细化预报中的应用

肖瑶1,2,杜明哲1,2,徐鹏1,2   

  1. 1.中国气象局河南省农业气象保障与应用技术重点实验室,河南郑州450003;
    2.河南省气象服务中心,河南郑州450003
  • 出版日期:2020-12-31 发布日期:2021-01-06
  • 通讯作者: 杜明哲(1963— ),男,河南邓州人,本科,主要从事气象灾害影响及风险评估和公共气象服务方面的研究. E-mail: dmz838@126.com。
  • 作者简介:肖瑶(1990— ),女,黑龙江佳木斯人,硕士,主要从事专业气象预报方面的研究. E-mail: xiaoyao_jms@163.com。

Application of Two Correction Methods in Refined Forecast of Maximum and Minimum Temperature Along the Henan Expressway

XIAO Yao1,2, DU Mingzhe1,2, XU Peng1,2   

  1. 1. Henan Key Laboratory of Agrometeorological Safeguard and Application Technique, CMA, Zhengzhou 450003, China;
    2. Henan Meteorological Service Center, Zhengzhou 450003, China
  • Online:2020-12-31 Published:2021-01-06

摘要: 应用滑动平均、类卡尔曼滤波递减平均订正法对中国气象局公共气象服务中心提供的河南省高速公路交通沿线逐3 h最高、最低气温精细化指导预报产品进行误差订正。结果表明: 24 h预报时效内两种订正方法均表现为正的订正效果,最高气温平均预报准确率由64.2%提高到71.8%~75.6%,最低气温的平均预报准确率由47.2%提高到58.7%~61.6%。其中类卡尔曼滤波递减平均订正法的订正效果优于滑动平均法,改进的w(i)递减平均方案在各预报时效均为几种订正方案中最优,最高、最低气温均方根误差空间大值中心明显减小。对7 d滑动平均和w(i)递减平均两种订正方案进行集成试验,其订正结果与单一的w(i)递减平均方案订正效果相差不大,部分时段w(i)递减平均方案表现出更好的订正能力。

关键词: 高速公路交通精细化预报, 最高、最低气温, 误差订正

Abstract: The errors of maximum and minimum temperature every three hours refined forecast guidance product along the Henan expressway provided by the CMA Public Meteorological Service Center were corrected  by using  moving average method and Kalman Filter-typed decaying averaging methods. The results show that two correction methods had positive correction effects within 24-hour forecast leading time and the average forecast accuracy of maximum temperature increased from 64.2% to 71.8%-75.6%. The forecast accuracy of minimum temperature increased from 47.2% to 58.7-61.6%. The Kalman Filter-typed decaying averaging correction method showed better correction effect than moving average method. The improved w(i) decaying averaging scheme  was  the best of the several correction schemes in each forecast leading time, and it also made the large value center of root mean square error  of  maximum and minimum temperature  decrease. The integration test of the two correction schemes including 7-day moving average and w(i) decaying averaging had a similar correction effect compared with the single w(i) decaying averaging method and the w(i) decaying averaging scheme showed better correction ability sometimes.

Key words: refined forecast, maximum and minimum temperature, error correction