基于泉州极大风特征对ECMWF细网格阵风预报产品的随机森林订正

作者：吴一非, 蓝秋萍, 韩赓

单位：福建省泉州市气象局, 福建 泉州 362000

关键词：泉州;变压梯度;随机森林;阵风预报订正

分类号：P732

出版年·卷·期（页码）：2026·43·第4期(97-106)

摘要：

利用2022年1月—2024年7月欧洲中期天气预报中心（ECMWF）细网格阵风预报产品和泉州265个气象观测站逐小时数据，分析泉州极大风特征；结合沿海大风影响因子的重要性排序，采用随机森林算法订正ECMWF细网格阵风产品。结果表明：泉州极大风速由沿海向内陆递减，季节变化性强，夏季多为低空急流引发的西南大风，秋冬季则为冷空气导致的东北大风；台湾海峡北部西南低空急流发生频率最高，这一现象与地形阻挡及气压场配置密切相关；冷空气大风中，气压梯度对风速影响显著，但需结合气压梯度风向判断实际风速大小，24 h变压梯度与风速增幅变化一致性高（相关系数>0.8），可作为大风预报的良好指标；随机森林订正后，泉州265个气象观测站极大风速的平均误差、平均绝对误差、均方误差、均方根误差均明显减小，预报精度显著提高，表明随机森林算法对ECMWF细网格阵风预报具备正向订正能力，可为泉州沿海大风气象服务提供有价值的参考依据。

Using the ECMWF fine-grid gust products from January 2021 to July 2024 and the hourly observational data of 265 observation stations in Quanzhou, this study analyzes the characteristics of maximum wind in Quanzhou. According to the importance of the influencing factors of coastal strong winds, the Random Forest algorithm is used to correct the ECMWF fine-grid gust products. The results show that: The average maximum wind speed in Quanzhou decreases from the coast to the inland, with obvious seasonal variations. In summer, southwesterly strong winds are mostly caused by low-level jets, while in winter, northeasterly strong winds are influenced by cold air. The occurrence frequency of the southwesterly low-level jet in the northern Taiwan Strait is the highest, which is closely related to the terrain obstruction and the pressure field. Regarding cold air strong winds, the pressure gradient has a significant impact, but the wind speed judgment needs to consider its direction; the 24-hour pressure change gradient has a high consistency with the variation of wind speed increase(correlation coefficient>0.8), making it a good indicator for strong wind forecasting. After correction by the Random Forest algorithm, the mean error, mean absolute error, mean square error, and root mean square error of the wind speeds at three representative stations along the Quanzhou coast are significantly reduced, and the forecasting accuracy is remarkably improved. This suggests that the Random Forest algorithm has a positive correction ability for ECMWF fine-grid gust forecasts and provides valuable references for meteorological services of coastal strong winds in Quanzhou.

参考文献：

[1] 曾瑾瑜,韩美,吴幸毓,等.WRF、EC和T639模式在福建沿海冬半年大风预报中的检验与应用[J].海洋科学, 2015, 39(7):75-85.Zeng Jinyu, Han Mei, Wu Xingyu, et al.Test and application of WRF, EC and T639 models in strong wind forecasting in Fujian coast in winter half year[J].Marine Sciences, 2015, 39(7):75-85. [2] 胡海川,黄彬,魏晓琳.我国近海洋面10 m风速集合预报客观订正方法[J].气象, 2017, 43(7):856-862.Hu Haichuan, Huang Bin, Wei Xiaolin.Objective correction method of ensemble forecast of 10 m winds on Chinese offshore[J].Meteorological Monthly, 2017, 43(7):856-862. [3] 吴俞,冯文,李勋,等.ECMWF细网格10 m风场产品在南海海域的预报检验[J].西南师范大学学报(自然科学版), 2015, 40(9):204-212.Wu Yu, Feng Wen, Li Xun, et al.On forecast test of ECMWF 10 m wind field products in South China Sea[J].Journal of Southwest China Normal University(Natural Science Edition), 2015, 40(9):204-212. [4] 罗江珊,杨凡,毕玮,等.EC细网格10 m风场产品在青岛港区的预报检验和随机森林订正[J].海洋预报, 2024, 41(3):110-119.Luo Jiangshan, Yang Fan, Bi Wei, et al.Verification and random forest correction of the EC fine-grid model 10 m wind in Qingdao port[J].Marine Forecasts, 2024, 41(3):110-119. [5] Rigatti S J.Random forest[J].Journal of Insurance Medicine, 2017,47(1):31-39, doi:10.17849/insm-47-01-31-39.1. [6] 许立兵,孔扬,周峥,等.基于机器学习的风场预报订正方法研究[J].陕西气象, 2023(1):15-20.Xu Libing, Kong Yang, Zhou Zheng, et al.Research on prediction and correction method of wind field based on machine learning[J].Journal of Shaanxi Meteorology, 2023(1):15-20. [7] Shi Kunpeng, Qiao Ying, Zhao Wei, et al.An improved random forest model of short-term wind-power forecasting to enhance accuracy, efficiency, and robustness[J].Wind Energy, 2018, 21(12):1383-1394, doi:10.1002/we.2261. [8] 曹钢锋,张善君,朱官忠,等.山东天气分析与预报[M].北京:气象出版社, 1988:245-249.Cao Gangfeng, Zhang Shanjun, Zhu Guanzhong, et al.Analysis and forecast on Shandong weather[M].Beijing:Meteorology Press, 1988:245-249. [9] 王雷.一次冷空气强风的成因分析[J].海洋预报, 2005, 22(4):96-101.Wang Lei.The reason analysis of a case of strong cold air gale[J].Marine Forecasts, 2005, 22(4):96-101. [10] 陈瑞闪.试论等变压梯度对福建沿海大风的影响[J].海洋通报,1987, 6(1):16-21.Chen Ruishan.Preliminary discussion of the effects of isallobaric gradient on the coastal gale of Fujian province[J].Marine Science Bulletin, 1987, 6(1):16-21. [11] 杜雅菡,张军.2019-2021年南京市区域站气压资料质量评估与分析[J].现代农业科技, 2023(16):172-175.Du Yahan, Zhang Jun.Quality assessment and analysis of barometric data from regional stations in Nanjing from 2019 to2021[J].Modern Agricultural Science and Technology, 2023(16):172-175. [12] Wang Dongqian, Zhang Yaocun, Huang Anning.Climatic features of the south-westerly low-level jet over southeast China and its association with precipitation over east China[J].Asia-Pacific Journal of Atmospheric Sciences, 2013, 49(3):259-270, doi:10.1007/s13143-013-0025-y. [13] Liu Hongbo, He Mingyang, Wang Bin, et al.Advances in lowlevel jet research and future prospects[J].Journal of Meteorological Research, 2014, 28(1):57-75, doi:10.1007/s13351-014-3166-8. [14] Chen Yileng, Tu Chuanchi, Hsiao F, et al.An overview of lowlevel jets(LLJs)and their roles in heavy rainfall over the Taiwan area during the early summer rainy season[J].Meteorology, 2022,1(1):64-112, doi:10.3390/meteorology1010006. [15] 朱乾根,林锦瑞,寿绍文,等.天气学原理和方法[M].4版.北京:气象出版社, 2007:267-268.Zhu Qiangen, Lin Jinrui, Shou Shaowen, et al.Principles and methods of synoptic meteorology[M].4th ed.Beijing:Meteorology Press, 2007:267-268. [16] 邱王泽禾,章蓝文.椒江沿海大风特征分析及精细化预报方法研究[J].海洋预报, 2024, 41(3):98-109.Qiu Wangzehe, Zhang Lanwen.Research on characteristic analysis and forecast method of gales in the Jiaojiang coastal area[J].Marine Forecasts, 2024, 41(3):98-109. [17] 韩永清,张少林,郭俊建.黄渤海一次强风天气成因分析[J].气象科技, 2015, 43(2):283-288.Han Yongqing, Zhang Shaolin, Guo Junjian.Causal analysis of a strong wind process over Yellow and Bohai Seas[J].Meteorological Science and Technology, 2015, 43(2):283-288.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】