构建准实时海面风场的一种智能算法

刘晓燕; 宋晓姜; 郭安博宇; 郝赛; 彭炜

doi:10.12284/hyxb2024051

构建准实时海面风场的一种智能算法

doi: 10.12284/hyxb2024051

刘晓燕^{1, 2,},
宋晓姜^{1, 2, ,},
郭安博宇^{1, 2},
郝赛^{1, 2},
彭炜^{1, 2}

1.
国家海洋环境预报中心，北京 100081
2.
自然资源部海洋灾害预报技术重点实验室，北京 100081

基金项目: 国家重点研发计划(2023YFC3107901)；自然资源部空间海洋遥感与应用研究重点实验室开放基金(202102004)。

详细信息

作者简介:
刘晓燕（1988—），女，山东省日照市人，从事海面风场资料同化与融合研究。E-mail：liuxiaoyan-de@163.com

通讯作者:
宋晓姜（1981—），女，北京市人，正高级工程师，从事海洋气象预报研究。E-mail：xjsong@nmefc.cn

中图分类号: P714⁺.2
计量
- 文章访问数: 87
- HTML全文浏览量: 39
- PDF下载量: 18
- 被引次数: 0
出版历程
- 收稿日期: 2024-01-24
- 修回日期: 2024-05-10
- 网络出版日期: 2024-07-12
- 刊出日期: 2024-06-01

An intelligent algorithm for constructing quasi-real-time sea surface wind field

Liu Xiaoyan^{1, 2
,},
Song Xiaojiang^{1, 2
, ,},
Guo Anboyu^{1, 2},
Hao Sai^{1, 2},
Peng Wei^{1, 2}

1.
National Marine Enviroment Forecasting Center, Beijing 100081 China
2.
Key Laboratory of Marine Hazards Forecasting, Ministry of Natural Resources, Beijing 100081, China

摘要

摘要: 本文基于深度学习U-Net网络构建了CMA-GFS数值模式风场订正模型，并以此订正模型订正后的风场为背景场（CMA-GFS_Unet），以HY-2B/2C/2D以及MetOp-B 4颗卫星的散射计海面风资料为观测资料，采用插补法快速完成准实时海面风场的构建。此智能算法可实现滞后3 h准实时生成空间分辨率为0.25°、时间分辨率为6 h的全球海面融合风场(Fusion_QRT)。分别使用CCMP融合风场数据和中国近海浮标10 m风矢量数据对CMA-GFS、CMA-GFS_Unet和Fusion_QRT 3组风场资料进行了评估，结果表明，CMA-GFS_Unet风场质量得到显著提升，Fusion_QRT风场风速质量得到进一步改善，但风向质量略有降低：相较于CCMP，3组风场的风速平均绝对误差（MAE）分别为1.13 m/s、0.89 m/s和0.84 m/s，CMA-GFS_Unet和Fusion_QRT相较于CMA-GFS分别提升了21.3%和25.7%；风向MAE分别为17.5°、15.5°和16°，分别提升了11.3%和8.6%；而相较于浮标，风速MAE分别为1.50 m/s、1.36 m/s和1.28 m/s，分别提升了9.3%和14.7%；风向MAE分别为23.3°、22.7°和24.0°，分别提升了3.0%和−3.9%。
- U-Net /
- CCMP /
- CMA-GFS /
- HY-2B/2C/2D /
- MetOp-B /
- 准实时 /
- 海面风场
Abstract: In this paper, the correction model of CMA-GFS numerical model wind field is constructed based on the deep learning U-Net network, and the construction of the quasi-real-time sea surface wind field is rapidly accomplished by interpolation method using the corrected wind field with the correction model as the background field (CMA-GFS_Unet), and using the scatterometer sea surface wind data from the four satellites, namely, HY-2B/2C/2D and MetOp-B as the observation data. This intelligent algorithm can realize the generation of global sea surface fusion wind field (Fusion_QRT) with a spatial resolution of 0.25° and a temporal resolution of 6 hours in quasi-real time with a lag of 3 hours. The CMA-GFS, CMA-GFS_Unet and Fusion_QRT wind fields are evaluated using the CCMP fusion wind field data and the 10 m wind vector data from the Chinese offshore buoys, respectively.The results show that the quality of the CMA-GFS_Unet wind field has been significantly improved, and the quality of the wind speed of the Fusion_QRT wind field has been further improved but the quality of the wind direction has been slightly reduced. The mean absolute errors (MAEs) of wind speed are 1.13 m/s, 0.89 m/s and 0.84 m/s for the three wind fields by using CCMP data as reference, and the CMA-GFS_Unet and Fusion_QRT wind fields have improved by 21.3% and 25.7% compared to the CMA-GFS, respectively; while the MAEs of wind direction are 17.5°, 15.5° and 16°, and have improved by11.3% and 8.6%, respectively.The MAEs of wind speed are 1.50 m/s, 1.36 m/s and 1.28 m/s for the three wind fields by using buoy data as reference, and have improved by 9.5% and 14.7% , respectively; while the MAEs of wind direction are 23.3°, 22.7° and 24.0°, and have improved by 3.0% and −3.9% , respectively.
- U-Net /
- CCMP /
- CMA-GFS /
- HY-2B/2C/2D /
- MetOp-B /
- quasi-real-time /
- sea surface wind field

HTML全文

图 1 浮标站点位置分布

Fig. 1 Distribution of buoys location

下载: 全尺寸图片幻灯片

图 2 准实时海面风场构建流程

Fig. 2 Flow chart of quasi-real-time surface wind field construction

下载: 全尺寸图片幻灯片

图 3 U-Net网络结构示意

Fig. 3 Diagram of U-Net network structure

下载: 全尺寸图片幻灯片

图 4 CMA-GFS风场订正效果

Fig. 4 CMA-GFS wind field correction effect diagram

下载: 全尺寸图片幻灯片

图 5 CMA-GFS风场订正效果（西北太平洋区域）

Fig. 5 CMA-GFS wind field correction effect diagram (Northwest Pacific region)

下载: 全尺寸图片幻灯片

图 6 准实时海面风场构建过程示意

Fig. 6 Process of quasi-real-time surface wind field construction

下载: 全尺寸图片幻灯片

图 7 海面融合风场构建过程示意

Fig. 7 Process of surface fuisonwind field construction

下载: 全尺寸图片幻灯片

图 8 4组数据风速/风向误差时间序列

Fig. 8 Time series of wind speed/direction error for four sets of data

下载: 全尺寸图片幻灯片

图 9 5组数据与浮标观测的风速/风向散点图

Fig. 9 Scatter plots of wind speed/direction of five sets of data vs buoys

下载: 全尺寸图片幻灯片

图 10 5组数据与浮标观测风速在不同海域的泰勒图

Fig. 10 Tylar diagram of wind speed of five sets of data in different sea areas

下载: 全尺寸图片幻灯片

图 11 5组数据在不同浮标站点的风速/风向平均绝对误差情况

Fig. 11 MAE of wind speed/direction at different buoy sites for five sets of data

下载: 全尺寸图片幻灯片

表 1 卫星散射计数据平均覆盖率

Tab. 1 Average coverage of satellite scatterometer data

时间	00 UTC	06 UTC	12 UTC	18 UTC
覆盖率	33.05%	42.88%	28.38%	16.51%

下载: 导出CSV

表 2 4组风场数据风速风向检验结果

Tab. 2 Results of wind speed and direction for four sets of wind field data

检验数据	风速				风向
检验数据	ME/(m·s⁻¹)	MAE/(m·s⁻¹)	RMSE/(m·s⁻¹)	CC	ME/(°)	MAE/(°)	RMSE/(°)
CMA-GFS	−0.06	1.13	1.53	0.925	0.5	17.5	31.2
CMA-GFS_Unet	−0.06	0.89	1.21	0.955	0.2	15.5	28.1
Fusion_QRT	−0.06	0.84	1.16	0.958	0.3	16.0	28.8
Fusion	−0.05	0.78	1.09	0.963	0.3	16.4	29.5

下载: 导出CSV

表 3 5组数据风速风向检验结果

Tab. 3 Results of wind speed and direction for five sets of wind field data

海区	检验数据	风速				风向
海区	检验数据	ME/(m·s⁻¹)	MAE/(m·s⁻¹)	RMSE/(m·s⁻¹)	CC	ME/(°)	MAE/(°)	RMSE/(°)
渤海	CCMP	−0.51	1.19	1.55	0.92	−14.7	24.6	32.9
	CMA-GFS	0.71	1.53	1.99	0.85	−2.7	26	37.1
	CMA-GFS_Unet	−0.42	1.37	1.87	0.87	−4.3	25.6	36.3
	Fusion_QRT	−0.76	1.44	1.85	0.89	−9.6	28	39.1
	Fusion	−0.76	1.41	1.81	0.90	−12.2	28.7	39.1
黄海	CCMP	0.65	1.43	1.91	0.83	−2.1	19.5	25.9
	CMA-GFS	1.06	1.72	2.2	0.8	6.6	21.3	30.7
	CMA-GFS_Unet	0.64	1.55	2.05	0.79	3.2	21.1	29.5
	Fusion_QRT	0.5	1.4	1.86	0.81	−1.2	23.1	31.2
	Fusion	0.48	1.36	1.82	0.82	−2.4	23.5	31.6
东海	CCMP	0.04	1.01	1.31	0.89	4.3	23.6	37.7
	CMA-GFS	0.16	1.33	1.74	0.82	9.8	25.4	38.7
	CMA-GFS_Unet	−0.12	1.24	1.61	0.85	6.5	23.7	36.7
	Fusion_QRT	−0.24	1.13	1.5	0.87	2.6	24.3	37.3
	Fusion	−0.21	1.09	1.45	0.88	3.4	24.5	37.6
南海	CCMP	0.08	0.97	1.44	0.86	−7.1	19.7	28
	CMA-GFS	0.03	1.36	1.84	0.75	−0.5	22.8	34.3
	CMA-GFS_Unet	−0.14	1.2	1.6	0.82	−5.2	22.1	32.4
	Fusion_QRT	−0.26	1.15	1.54	0.85	−5.2	22.8	34.1
	Fusion	−0.20	1.12	1.50	0.86	−6.7	22.6	34.0

下载: 导出CSV

参考文献(33)

[1]	张巍, 杜超凡, 郭安博宇, 等. 一种机器学习海面风场快速融合的方法[J]. 海洋学报, 2022, 44(11): 144−158. Zhang Wei, Du Chaofan, Guo Anboyu, et al. Sea surface wind field smart fusion base on machine learning method[J]. HaiyangXuebao, 2022, 44(11): 144−158.
[2]	柳婧, 宋晓姜, 王彰贵. 海面风场融合技术综述[J]. 海洋预报, 2018, 35(3): 81−87. Liu Jing, Song Xiaojiang, Wang Zhanggui. Review of fusion technology of sea surface wind[J]. Marine Forecasts, 2018, 35(3): 81−87.
[3]	旷芳芳, 张友权, 张俊鹏, 等. 3种海面风场资料在台湾海峡的比较和评估[J]. 海洋学报, 2015, 37(5): 44−53. Kuang Fangfang, Zhang Youquan, Zhang Junpeng, et al. Comparison and evaluation of three sea surface wind products in Taiwan Strait[J]. HaiyangXuebao, 2015, 37(5): 44−53.
[4]	詹思玙, 齐琳琳, 卢伟. 基于CCMP资料和现场观测资料的西北太平洋海面风场特征分析[J]. 海洋预报, 2017, 34(2): 10−20. Zhan Siyu, Qi Linlin, Lu Wei. Analysis of sea surface wind in northwest Pacific based on CCMP satellite data[J]. Marine Forecasts, 2017, 34(2): 10−20.
[5]	金铸钰, 刘凯, 郭安博宇, 等. 基于CCMP的西北太平洋海面风场特征分析[J]. 海洋预报, 2022, 39(2): 20−33. Jin Zhuyu, Liu Kai, Guo Anboyu, et al. Characteristics analysis of the sea surface wind in the Northwest Pacific basedon CCMP satellite data[J]. MarineForecasts, 2022, 39(2): 20−33.
[6]	林晓娟, 李响, 刘晓燕, 等. EC预报数据在中国近海的适用性研究[J]. 海洋预报, 2023, 40(6): 51−66. Lin Xiaojuan, Li Xiang, Liu Xiaoyan, et al. Applicability of EC forecast data in China’s offshore waters[J]. Marine Forecasts, 2023, 40(6): 51−66.
[7]	郑崇伟. 基于CCMP风场的近22年中国海海表风场特征分析[J]. 气象与减灾研究, 2011, 34(3): 41−46. Zheng Chongwei. Sea surface wind field analysis in the China seaduring the last 22 years with CCMP wind field[J]. Meteorology and Disaster Reduction Research, 2011, 34(3): 41−46.
[8]	柳婧. 基于最优插值方法的中国近海海面风场资料融合研究[D]. 北京: 国家海洋环境预报中心, 2018. Liu Jing. Research on data fusion of sea surface wind in China’s offshore based on optimal interpolation method[D]. Beijing: NationalMarine Environmental Forecasting Center, 2018.
[9]	唐焕丽, 姚琴, 吕晓莹, 等. 多源卫星融合的广东海域海面风场特征[J]. 遥感信息, 2020, 35(1): 117−122. Tang Huanli, Yao Qin, Lü Xiaoying, et al. Characteristics of sea surface wind field in Guangdong sea area with multi-source satellite fusion[J]. Remote Sensing Information, 2020, 35(1): 117−122.
[10]	吕思睿. 多源卫星海面风场融合研究[D]. 南京: 南京信息工程大学, 2023. Lü Sirui. Multi-source satellite sea surface wind field fusion study[D]. Nanjing: Nanjing University of Information Science & Technology, 2023.
[11]	齐亚琳, 林明森. 数据融合技术在海洋二号卫星数据中的应用[J]. 航天器工程, 2012, 21(3): 117−123. Qi Yalin, Lin Mingsen. Application of the data fusion technique in HY-2 satellite data[J]. Spacecraft Engineering, 2012, 21(3): 117−123.
[12]	Zou Juhong, Lin Minsen, Zou Bin, et al. Fusion of sea surface wind vector data acquired by multi-source active and passive sensors in China sea[J]. International Journal of Remote Sensing, 2017, 38(23): 6477−6491. doi: 10.1080/01431161.2017.1356486
[13]	刘宇昕. 星载微波散射计数据和数值模式产品海面风场混合分析研究[D]. 北京: 国家海洋环境预报研究中心, 2013. Liu Yuxin. A study on the mixing analysis of sea surface wind field with satellite microwave scatterometer data and numerical model products[D]. Beijing: NationalMarine Environmental Forecasting Center, 2013.
[14]	李莎, 舒红, 徐正全. 利用时空Kriging进行气温插值研究[J]. 武汉大学学报•信息科学版, 2012, 37(2): 237−241. Li Sha, Shu Hong, Xu Zhengquan. Interpolation of temperature based on spatial-temporal kriging[J]. Geomatics and Information Science of Wuhan University, 2012, 37(2): 237−241.
[15]	许遐祯, 高健, 张康宇, 等. 基于多源数据的中国近海风场融合方法研究[J]. 杭州师范大学学报(自然科学版), 2016, 15(3): 325−330. Xu Xiazhen, Gao Jian, Zhang Kangyu. Fusion method of China’s offshore wind field based on multi-source data[J]. Journal of Hangzhou Normal University (Natural Science Edition), 2016, 15(3): 325−330.
[16]	项杰, 王慧鹏, 王春明, 等. 南海海面风场变分融合的初步研究[J]. 热带气象学报, 2015, 31(2): 153−160. Xiang Jie, Wang Huipeng, Wang Chunming, et al. Preliminary study on variational blending of surface vector winds in South China Sea[J]. Journal of Tropical Meteorology, 2015, 31(2): 153−160.
[17]	邹婷. 2DVAR算法在局地海表流场和风场融合中的应用[D]. 南京: 南京信息工程大学, 2020. Zou Ting. Application of 2DVAR algorithms to the fusion of localized sea surface current and wind fields[D]. Nanjing: Nanjing University of Information Science and Technology, 2020.
[18]	吴新荣, 王喜冬, 李威, 等. 海洋数据同化与数据融合技术应用综述[J]. 海洋技术学报, 2015, 34(3): 97−103. Wu Xinrong, Wang Xidong, Li Wei, et al. Review of the application of ocean data assimilation and data fusion techniques[J]. Journal of Ocean Technology, 2015, 34(3): 97−103.
[19]	张东翔. 多源卫星海面风场产品检验及融合研究[D]. 长沙: 国防科技大学, 2018. Zhang Dongxiang. Research of multi-source satellite sea surface wind validation and data fusion[D]. Changsha: National University of Defense Technology, 2018.
[20]	李明悝, 侯一筠. 利用QuikSCAT/NCEP混合风场及WAVEWATCH模拟东中国海风浪场[J]. 海洋科学, 2005, 29(6): 9−12. Li Mingkui, HouYijun. Simulating wind-wave field of the East China Seas with QuikSCAT/NCEP blended wind and WAVEWATCH[J]. Marine Sciences, 2005, 29(6): 9−12.
[21]	Atlas R, Hoffman R N, Ardizzone J, et al. Across-calibrated, multiplatform ocean surface wind velocity product for meteorological and oceanographic applications[J]. Bulletin of the American Meteorological Society, 2011, 92(2): 157−174.
[22]	Mears C A, Scott J, Wentz F J, et al. A near-real-time version of the Cross-Calibrated Multiplatform (CCMP) ocean surface wind velocity data set[J]. Journal of Geophysical Research: Oceans, 2019, 124(10): 6997−7010. doi: 10.1029/2019JC015367
[23]	Mears C, Lee T, Ricciardulli L, et al. Improving the accuracy of the Cross-Calibrated Multi-Platform (CCMP) ocean vector winds[J]. Remote Sensing, 2022, 14(17): 4230. doi: 10.3390/rs14174230
[24]	王慧, 隋伟辉. 基于CCMP风场的中国近海18个海区海面大风季节变化特征分析[J]. 气象科技, 2013, 41(4): 720−725. Wang Hui, Sui Weihui. Seasonal variation analysis of sea surface winds in China Sea areas with CCMP wind field data[J]. Meteorological Science and Technology, 2013, 41(4): 720−725.
[25]	刘建强, 叶小敏, 兰友国. 我国海洋卫星遥感大数据及其应用服务[J]. 大数据, 2022, 8(2): 75−88. Liu Jianqiang, Ye Xiaomin, Lan Youguo. Remote sensing big data from Chinese oceansatellites and its application service[J]. Big Data Research, 2022, 8(2): 75−88.
[26]	林明森, 何贤强, 贾永君, 等. 中国海洋卫星遥感技术进展[J]. 海洋学报, 2019, 41(10): 99−112. Lin Mingsen, He Xianqiang, JiaYongjun, et al. Advances in marine satellite remote sensing technology in China[J]. HaiyangXuebao, 2019, 41(10): 99−112.
[27]	Park S H, Yoo J, Son D, et al. Improved calibration of wind estimates from advanced scatterometer MetOp-B in Korean Seas using deep neural network[J]. Remote Sensing, 2021, 13(20): 4164. doi: 10.3390/rs13204164
[28]	Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation[C]//Proceedings of the18th International Conferenceon Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Munich: Springer, 2015.
[29]	朱延和. 基于U-net神经网络的新疆地面气温和风的预报订正研究[D]. 南京: 南京信息工程大学, 2023. Zhu Yanhe. Prediction correction of surface temperature and wind in Xinjiang based on U-net neural network[D]. Nanjing: Nanjing University of Information Science &Technology, 2023.
[30]	张延彪, 陈明轩, 韩雷, 等. 数值天气预报多要素深度学习融合订正方法[J]. 气象学报, 2022, 80(1): 153−167. doi: 10.11676/qxxb2021.066 Zhang Yanbiao, Chen Mingxuan, Han Lei, et al. Multi-element deep learning fusion correction method for numerical weather prediction[J]. ActaMeteorologicaSinica, 2022, 80(1): 153−167 doi: 10.11676/qxxb2021.066
[31]	Han Lei, Chen Mingxuan, Chen Kangkai, et al. A deep learning method for bias correction of ECMWF 24-240 h forecasts[J]. Advances in Atmospheric Sciences, 2021, 38(9): 1444−1459. doi: 10.1007/s00376-021-0215-y
[32]	张延彪, 宋林烨, 陈明轩, 等. 基于卷积神经网络的京津冀地区高分辨率格点预报偏差订正试验[J]. 大气科学学报, 2022, 45(6): 850−862. Zhang Yanbiao, Song Linye, Chen Mingxuan, et al. A study of error correction for high-resolution gridded forecast based on a convolutional neural network in the Beijing-Tianjin-Hebei Region[J]. Transactions of Atmospheric Sciences, 2022, 45(6): 850−862.
[33]	Lagerquist R, Stewart J Q, Ebert-Uphoff I, et al. Using deep learning to nowcast the spatial coverage of convection from Himawari-8 satellite data[J]. Monthly Weather Review, 2021, 149(12): 3897−3921. doi: 10.1175/MWR-D-21-0096.1