A method for estimating surface wind speed during typhoon events using GNSS-derived PWV and its accuracy validation

doi:10.11947/j.AGCS.2026.20260055

Abstract

Abstract:

Surface wind speed (WS) is a key indicator for evaluating typhoon intensity and assessing potential wind hazards. However, the sparse distribution of automatic weather stations limits high-resolution monitoring of surface wind fields during typhoon events. In this study, three PWV-based WS estimation models (LPWV-WS, PPWV-WS, and PWV-WS3) were developed based on the correlation between precipitable water vapor (PWV) and WS during typhoon periods, together with the temporal persistence of wind speed. The models were evaluated using PWV observations from 97 global navigation satellite system (GNSS) stations, WS observations from 24 automatic weather stations, and ERA5 wind field data collected over Kyushu, Japan, during typhoon events from 2019 to 2023. The results show that PWV is strongly correlated with WS during typhoon periods, whereas the correlation is weak during non-typhoon periods. A weighted K-fold cross-validation based on typhoon samples demonstrates that the PWV-WS3 model achieves the best accuracy and stability, with mean absolute error (MAE), root mean square error (RMSE), and standard deviation (STD) of 0.89 m/s, 1.25 m/s, and 1.24 m/s, respectively. The corresponding errors for the derived wind level (WL) are 0.44, 0.71, and 0.70, respectively. A case study of super Typhoon Khanun (2023) further demonstrates that GNSS-PWV-enhanced wind fields provide a more detailed representation of high-wind-speed regions and their spatial evolution during typhoon movement. These results indicate that GNSS-derived PWV can effectively complement conventional surface wind observations and provide a new approach for high-resolution surface wind field reconstruction and typhoon wind hazard monitoring.

Key words: global navigation satellite system, precipitable water vapor, typhoon, wind speed and level, cross-validation

CLC Number:

P228

Qimin He, Kangming Song, Kefei Zhang, Chao Hu, Biqing Gao. A method for estimating surface wind speed during typhoon events using GNSS-derived PWV and its accuracy validation[J]. Acta Geodaetica et Cartographica Sinica, 2026, 55(8): 1357-1368.

Add to citation manager EndNote|Reference Manager|ProCite|BibTeX|RefWorks

URL: http://xb.chinasmp.com/EN/10.11947/j.AGCS.2026.20260055

http://xb.chinasmp.com/EN/Y2026/V55/I8/1357

Figures/Tables 12

Tab. 1

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Tab. 2

Tab. 3

Tab. 4

Fig. 6

Fig. 7

Tab. 5

References 33

[1]	闵隽玉, 谈哲敏. 基于潜空间扩散模型的台风路径与强度预报模型研究[J]. 南京大学学报(自然科学), 2025, 61(3): 373-386.
	Min Junyu, Tan Zhemin. On typhoon forecasting models for track and intensity based on latent diffusion model[J]. Journal of Nanjing University (Natural Science), 2025, 61(3): 373-386.
[2]	王充. 基于遥感图像和深度学习技术的台风信息提取和预报[D]. 青岛: 中国科学院大学(中国科学院海洋研究所), 2024.
	Wang Chong. Typhoon information extraction and forecasting based on remote sensing images and deep learning technology[D]. Qingdao: Institute of Oceanology, Chinese Academy of Sciences, 2024.
[3]	Pan Honglin, Yan Junjie, Kong Tuanjie, et al. Climatology of low-level aerosol extinction coefficient of different aerosol types and their association with the meteorology during 2007—2021: insights from the CALIPSO observation over China[J]. Theoretical and Applied Climatology, 2024, 156(1): 55.
[4]	刘安琪, 窦芳丽, 刘可, 等. 基于多源数据和三重匹配方法的风云三号风场测量雷达海面风场精度分析[J]. 海洋气象学报, 2025, 45(3): 62-71.
	Liu Anqi, Dou Fangli, Liu Ke, et al. Evaluation of FY-3E WindRAD sea surface wind field products based on multi-source data and triple collocation method[J]. Journal of Shandong Meteorology, 2025, 45(3): 62-71.
[5]	周鑫. 基于多普勒天气雷达的地面风场反演及大风预警方法研究[D]. 成都: 成都理工大学, 2020.
	Zhou Xin. Study on wind field retrieval and gale warning near the surface based on Doppler weather radar[D]. Chengdu: Chengdu University of Technology, 2020.
[6]	梁旭东, 周海光, 罗义. 多普勒天气雷达反演风场综述[J]. 气象学报, 2025, 83(3): 855-872.
	Liang Xudong, Zhou Haiguang, Luo Yi. Doppler radar wind field retrieval techniques: a review[J]. Acta Meteorologica Sinica, 2025, 83(3): 855-872.
[7]	曾加东, 宋子函, 张志田. 台风风特性现场实测与数值模拟研究综述[J]. 科学技术与工程, 2023, 23(28): 11937-11946.
	Zeng Jiadong, Song Zihan, Zhang Zhitian. Review of field measurements and numerical simulations of typhoon wind characteristics[J]. Science Technology and Engineering, 2023, 23(28): 11937-11946.
[8]	Xu Hongxiong, Zhao Yang, Zhao Dajun, et al. Exploring the typhoon intensity forecasting through integrating AI weather forecasting with regional numerical weather model[J]. npj Climate and Atmospheric Science, 2025, 8: 38.
[9]	张克非, 李浩博, 王晓明, 等. 地基GNSS大气水汽探测遥感研究进展和展望[J]. 测绘学报, 2022, 51(7): 1172-1191. DOI: . doi: 10.11947/j.AGCS.2022.20220149
	Zhang Kefei, Li Haobo, Wang Xiaoming, et al. Recent progresses and future prospectives of ground-based GNSS water vapor soundina[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(7): 1172-1191. DOI: . doi: 10.11947/j.AGCS.2022.20220149
[10]	姚宜斌, 赵庆志. GNSS对流层水汽监测研究进展与展望[J]. 测绘学报, 2022, 51(6): 935-952. DOI: . doi: 10.11947/j.AGCS.2022.20220039
	Yao Yibin, Zhao Qingzhi. Research progress and prospect of monitoring tropospheric water vapor by GNSS technique[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(6): 935-952. DOI: . doi: 10.11947/j.AGCS.2022.20220039
[11]	Won J, Kim D. Analysis of temporal and spatial variation of precipitable water vapor according to path of typhoon EWINIAR using GPS permanent stations[J]. Journal of Positioning, Navigation, and Timing, 2015, 4(2): 87-95.
[12]	He Qimin, Zhang Kefei, Wu Suqin, et al. An investigation of atmospheric temperature and pressure using an improved spatio-temporal Kriging model for sensing GNSS-derived precipitable water vapor[J]. Spatial Statistics, 2022, 51: 100664.
[13]	何琦敏, 张克非, 李黎, 等. 基于PWV到达时间差估计台风运动状态的四参数模型[J]. 测绘学报, 2024, 53(11): 2125-2137. DOI: . doi: 10.11947/j.AGCS.2024.20230473
	He Qimin, Zhang Kefei, Li Li, et al. A four-parameter model for estimating typhoon motion states based on time difference of PWV arrival[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(11): 2125-2137. DOI: . doi: 10.11947/j.AGCS.2024.20230473
[14]	He Qimin, Zhang Kefei, Wu Suqin, et al. Real-time GNSS-derived PWV for typhoon characterizations: a case study for super typhoon Mangkhut in Hong Kong[J]. Remote Sensing, 2019, 12(1): 104.
[15]	Kang I, Park J. Use of GNSS-derived PWV for predicting the path of typhoon: case studies of Soulik and Kongrey in 2018[J]. Journal of Surveying Engineering, 2021, 147(4): 04021018.
[16]	He Qimin, Shen Zhen, Wan Moufeng, et al. Precipitable water vapor converted from GNSS-ZTD and ERA5 datasets for the monitoring of tropical cyclones[J]. IEEE Access, 2020, 8: 87275-87290.
[17]	Shi Jiaqi, Li Min, Steiner A K, et al. Evaluation of reanalysis precipitable water vapor under typhoon conditions using multi-source observations[J]. Atmospheric Chemistry and Physics, 2026, 26(7): 4633-4650.
[18]	Mei Xinyu, Xu Keke, Wang Xue. Monitoring water vapor variations during Super Typhoon Saola's impact on Hong Kong using GNSS observations[J]. Measurement Science and Technology, 2025, 36(8): 085801.
[19]	Chabibi F F, Putri N S E, Wijaya D D. Tropical cyclone intensity prediction using BP-RNN from GPS-derived precipitable water vapor and surface meteorological data[J]. IOP Conference Series: Earth and Environmental Science, 2024, 1418(1): 012020.
[20]	Dach R, Lutz S, Walser P, et al. Bernese GNSS software version 5.2[M]. Bern: Bern Open Publishing, 2015.
[21]	Herring TA, Floyd MA, Perry M, et al. Introduction to GAMIT/GLOBK[R]. Cambridge: Massachusetts Institute of Technology, 2010.
[22]	Saastamoinen J. Contributions to the theory of atmospheric refraction[J]. Bulletin Géodésique, 1972, 105(1): 279-298.
[23]	Landskron D, Böhm J. VMF3/GPT3: refined discrete and empirical troposphere mapping functions[J]. Journal of Geodesy, 2018, 92(4): 349-360.
[24]	Ding Junsheng, Chen Junping, Tang Wenjie, et al. Spatial-temporal variability of global GNSS-derived precipitable water vapor (1994-2020) and climate implications[J]. Remote Sensing, 2022, 14(14): 3493.
[25]	马华锋, 李黎, 张志杰, 等. 基于GNSS、ERA5和机器学习的区域天顶对流层延迟校正和大气可降水量预测模型[J]. 地球物理学报, 2026, 69(1): 55-68.
	Ma Huafeng, Li Li, Zhang Zhijie, et al. The Regional zenith tropospheric delay correction and precipitable water vapor prediction models based on the GNSS, ERA5 and machine learning[J]. Chinese Journal of Geophysics, 2026, 69(1): 55-68.
[26]	Jiang Chunhua, Chen Shaoni, Gao Xiang, et al. A direct conversion model from ZTD to PWV based on the random forest[J]. Measurement Science and Technology, 2025, 36(2): 025802.
[27]	Li Junyu, Li Haojie, Liu Lilong, et al. An improved typhoon monitoring model based on precipitable water vapor and pressure[J]. Geodesy and Geodynamics, 2024, 15(3): 276-290.
[28]	刘伟. 基于惯性/多普勒组合导航的风速风向测量研究[D]. 南京: 南京航空航天大学, 2011.
	Liu Wei. Research on wind speed and direction measurement based on INS/Doppler integrated navigation[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2011.
[29]	赵庆志. 地基GNSS水汽反演关键技术研究及其应用[D]. 武汉: 武汉大学, 2017.
	Zhao Qingzhi. Studies on the key technologies in water vapor inversion using ground-based GNSS and its applications[D]. Wuhan: Wuhan University, 2017.
[30]	Huler S. Defining the wind: the beaufort scale and how a 19th-century admiral turned science into poetry[M]. New York: Crown, 2005.
[31]	逯智科, 徐万海, 王玮, 等. 台风极值风速预测研究综述[J]. 力学与实践, 2025, 47(5): 906-919.
	Lu Zhike, Xu Wanhai, Wang Wei, et al. A review of research on prediction of typhoon extreme wind speeds[J]. Mechanics in Engineering, 2025, 47(5): 906-919.
[32]	Zhang Yuxin, Cao Shuyang, Cao Jinxin. A framework for efficient simulation of urban strong wind field during typhoon process using coupled WRF-UCM and embedded LES model[J]. Journal of Wind Engineering and Industrial Aerodynamics, 2024, 250: 105757.
[33]	Zhao Lin, Cui Wei, Fang Genshen, et al. State-of-the-art review on typhoon wind environments and their effects on long-span bridges[J]. Advances in Wind Engineering, 2024, 1(1): 100007.

风级	风速区间/（m/s）	风级	风速区间/（m/s）
0	[0，0.3）	7	[13.9，17.2）
1	[0.3，1.6）	8	[17.2，20.8）
2	[1.6，3.4）	9	[20.8，24.5）
3	[3.4，5.5）	10	[24.5，28.6）
4	[5.5，8.0）	11	[28.6，32.7）
5	[8.0，10.8）	12	[32.7，36.9）
6	[10.8，13.9）	>12	≥36.9

模型	台风类型	地面风速/（m/s）			风级
模型	台风类型	MAE	RMSE	STD	MAE	RMSE	STD
LPWV-WS	热带风暴	1.82	2.41	2.13	0.88	1.19	1.08
	强热带风暴	1.56	1.92	1.70	0.79	1.09	0.94
	台风	2.10	2.89	2.83	1.01	1.34	1.30
	强台风	1.95	2.58	2.41	0.92	1.24	1.16
	超强台风	2.50	3.21	3.04	1.15	1.48	1.37
PPWV-WS	热带风暴	1.82	2.39	2.12	0.85	1.14	1.06
	强热带风暴	1.54	1.91	1.69	0.76	0.99	0.93
	台风	1.98	2.75	2.71	0.93	1.25	1.23
	强台风	1.91	2.56	2.43	0.93	1.24	1.16
	超强台风	2.37	3.04	2.77	1.11	1.41	1.25
PWV-WS3	热带风暴	0.82	1.12	1.12	0.41	0.67	0.67
	强热带风暴	0.75	1.00	0.99	0.38	0.63	0.63
	台风	0.91	1.29	1.29	0.45	0.72	0.72
	强台风	0.86	1.17	1.17	0.42	0.68	0.68
	超强台风	0.90	1.27	1.26	0.44	0.71	0.71

模型	地面风速/（m/s）			风级
模型	MAE	RMSE	STD	MAE	RMSE	STD
LPWV-WS	2.27	2.99	2.83	1.06	1.39	1.30
AR_LPWV-WS	1.34	1.37	1.36	0.66	0.83	0.82
PPWV-WS	2.16	2.83	2.65	1.02	1.32	1.31
AR_PPWV-WS	1.21	1.28	1.27	0.56	0.78	0.77
AR_WS	1.57	1.63	1.62	0.73	0.89	0.89
PWV-WS3	0.89	1.25	1.24	0.44	0.71	0.70

变量	系数估计值	t值	p值
WS（t－1）	0.908	721.064	<0.001
PWV（t）	0.059	31.317	<0.001
PWV（t－1）	－0.046	－24.725	<0.001

方案	MAE	RMSE	STD
无加密	3.42	3.95	3.78
PWV加密	2.03	2.74	2.71