留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

数值天气预报多要素深度学习融合订正方法

张延彪 陈明轩 韩雷 宋林烨 杨璐

张延彪,陈明轩,韩雷,宋林烨,杨璐. 2022. 数值天气预报多要素深度学习融合订正方法. 气象学报,80(1):1-15 doi: 10.11676/qxxb2021.066
引用本文: 张延彪,陈明轩,韩雷,宋林烨,杨璐. 2022. 数值天气预报多要素深度学习融合订正方法. 气象学报,80(1):1-15 doi: 10.11676/qxxb2021.066
Zhang Yanbiao, Chen Mingxuan, Han Lei, Song Linye, Yang Lu. 2022. Multi-element deep learning fusion correction method for numerical weather prediction. Acta Meteorologica Sinica, 80(1):1-15 doi: 10.11676/qxxb2021.066
Citation: Zhang Yanbiao, Chen Mingxuan, Han Lei, Song Linye, Yang Lu. 2022. Multi-element deep learning fusion correction method for numerical weather prediction. Acta Meteorologica Sinica, 80(1):1-15 doi: 10.11676/qxxb2021.066

数值天气预报多要素深度学习融合订正方法

doi: 10.11676/qxxb2021.066
基金项目: 国家重点研发计划项目(2018YFF0300102)、北京市自然科学基金项目(8212025)
详细信息
    作者简介:

    张延彪,主要从事机器学习与气象应用研究,E-mail:17854229337@163.com

    通讯作者:

    陈明轩,主要从事短期临近预报技术研究。E-mail: mxchen@ium.cn

  • 中图分类号: P435

Multi-element deep learning fusion correction method for numerical weather prediction

  • 摘要: 数值天气预报作为现代天气预报的主流技术方法,近年来不断朝着精细化方向发展,但预报误差至今仍无法避免。文中在CU-Net模型中引入稠密卷积模块形成数值预报要素偏差订正模型Dense-CUnet,在此基础上进一步融合多种气象要素和地形特征构建了Fuse-CUnet模型,开展不同模型的偏差订正试验和对比分析。以均方根误差(RMSE)和平均绝对误差(MAE)作为评分标准,通过与ECMWF原始预报结果、ANO方法订正结果以及CU-Net方法订正结果进行对比,证明Dense-CUnet模型可有效改进数值预报订正效果,融合多个要素的Fuse-CUnet模型能使订正效果有更大提升。

     

  • 图  1  训练集 (a)、验证集和测试集 (b) 划分 (色阶为研究区域地形高度)

    Figure  1.  Partition of training set (a),verification set and testing set (b) (the color shadings show the terrain height in the research area)

    图  2  CU-Net网络结构

    Figure  2.  The network structure of CU-Net

    图  3  稠密卷积模块结构

    Figure  3.  The network structure of Denseblock

    图  4  Dense-CUnet网络结构

    Figure  4.  The network structure of Dense-CUnet

    图  5  Fuse-CUnet基本网络结构 (以温度订正为例)

    Figure  5.  The network structure diagram of Fuse-CUnet (with temperature correction as an example)

    图  7  2 m气温订正可信度曲线 (a. −20—0℃、b. 0—30℃)

    Figure  7.  Reliability curve for 2 m temperature correction (a. −20—0℃;b. 0—30℃)

    图  8  2018年1月11日12时温度订正结果 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

    Figure  8.  Corrected results of temperature at 12:00 UTC 11 January 2018 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

    图  10  2 m相对湿度订正可信度曲线

    Figure  10.  Reliability curve for 2 m relative humidity correction

    图  11  2018年10月1日12时相对湿度订正结果 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

    Figure  11.  Corrected results of relative humidity at 12:00 UTC 1 October 2018 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

    图  12  2018年全年U10风速均方根误差空间分布 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet)

    Figure  12.  Spatial distribution of U10 wind speed RMSE in 2018 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet)

    图  14  10 m风速订正可信度曲线 (a. U10,b. V10

    Figure  14.  Reliability curves for 10 m wind speed correction (a. U10,b. V10

    图  15  2018年3月15日12时全风速订正结果 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

    Figure  15.  Corrected results of wind speed at 12:00 UTC 15 March 2018 (a. ECMWF,b. ANO,c. CU-Net,d. Dense-CUnet,e. Fuse-CUnet,f. ERA5)

  • [1] 郭瀚阳,陈明轩,韩雷等. 2019. 基于深度学习的强对流高分辨率临近预报试验. 气象学报,77(4):715-727 doi: 10.11676/qxxb2019.036

    Guo H Y,Chen M X,Han L,et al. 2019. High resolution nowcasting experiment of severe convections based on deep learning. Acta Meteor Sinica,77(4):715-727 (in Chinese) doi: 10.11676/qxxb2019.036
    [2] 钱维宏. 2012. 如何提高天气预报和气候预测的技巧?. 地球物理学报,55(5):1532-1540 doi: 10.6038/j.issn.0001-5733.2012.05.010

    Qian W H. 2012. How to improve the skills of weather and climate predictions?. Chinese J Geophys,55(5):1532-1540 (in Chinese) doi: 10.6038/j.issn.0001-5733.2012.05.010
    [3] 孙健,曹卓,李恒等. 2021. 人工智能技术在数值天气预报中的应用. 应用气象学报,32(1):1-11 doi: 10.11898/1001-7313.20210101

    Sun J,Cao Z,Li H,et al. 2021. Application of artificial intelligence technology to numerical weather prediction. J Appl Meteor Sci,32(1):1-11 (in Chinese) doi: 10.11898/1001-7313.20210101
    [4] 曾晓青,薛峰,赵瑞霞等. 2019. 几种格点化温度滚动订正预报方案对比研究. 气象,45(7):1009-1018 doi: 10.7519/j.issn.1000-0526.2019.07.011

    Zeng X Q,Xue F,Zhao R X,et al. 2019. Comparison study on several grid temperature rolling correction forecasting schemes. Meteor Mon,45(7):1009-1018 (in Chinese) doi: 10.7519/j.issn.1000-0526.2019.07.011
    [5] Bauer P,Thorpe A,Brunet G. 2015. The quiet revolution of numerical weather prediction. Nature,525(7567):47-55 doi: 10.1038/nature14956
    [6] Bjerknes V. 1904. Das problem der Wettervorhersage,betrachtet vom Standpunkte der Mechanik und der Physik. Meteor Zeits,21:1-7
    [7] Cho K, van Merriënboer B, Gulcehre C, et al. 2014. Learning phrase representations using RNN encoder-decoder for statistical machine translation∥Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Doha, Qatar: Association for Computational Linguistics, 1724-1734
    [8] Glahn H R,Lowry D A. 1972. The use of model output statistics (MOS) in objective weather forecasting. J Appl Meteor,11(8):1203-1211 doi: 10.1175/1520-0450(1972)011<1203:TUOMOS>2.0.CO;2
    [9] Goodfellow I J, Pouget-Abadie J, Mirza M, et al. 2014. Generative adversarial nets∥Proceedings of the 27th International Conference on Neural Information Processing Systems. Montreal, Canada: MIT Press, 2672-2680
    [10] Graves A. 2012. Supervised Sequence Labelling with Recurrent Neural Networks. New York: Springer
    [11] Han L,Sun J Z,Zhang W,et al. 2017. A machine learning nowcasting method based on real-time reanalysis data. J Geophys Res:Atmos,122(7):4038-4051 doi: 10.1002/2016JD025783
    [12] Han L,Chen M X,Chen K K,et al. 2021. A deep learning method for bias correction of ECMWF 24-240 h forecasts. Adv Atmos Sci,38(9):1444-1459 doi: 10.1007/s00376-021-0215-y
    [13] Hersbach H,Bell B,Berrisford P,et al. 2020. The ERA5 global reanalysis. Quart J Roy Meteor Soc,146(730):1999-2049 doi: 10.1002/qj.3803
    [14] Hochreiter S,Schmidhuber J. 1997. Long short-term memory. Neural Comput,9(8):1735-1780 doi: 10.1162/neco.1997.9.8.1735
    [15] Homleid M. 1995. Diurnal corrections of short-term surface temperature forecasts using the Kalman filter. Wea Forecasting,10(4):689-707 doi: 10.1175/1520-0434(1995)010<0689:DCOSTS>2.0.CO;2
    [16] Huang G,Liu Z,Van Der Maaten L,et al. 2017. Densely connected convolutional networks∥2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu,HI,USA:IEEE,2261-2269
    [17] Ioffe S,Szegedy C. 2015. Batch normalization:Accelerating deep network training by reducing internal covariate shift∥Proceedings of the 32nd International Conference on International Conference on Machine Learning. Lille,France:JMLR. org,448-456
    [18] Kingma D P, Ba J L. 2015. Adam: A method for stochastic optimization∥The 3rd International Conference for Learning Representations, San Diego. http: //arxiv.org/abs/1412.6980
    [19] Klein W H,Lewis B M,Enger I. 1959. Objective prediction of five-day mean temperatures during winter. J Meteor,16(6):672-682 doi: 10.1175/1520-0469(1959)016<0672:OPOFDM>2.0.CO;2
    [20] Kusiak A,Wei X P,Verma A P,et al. 2013. Modeling and prediction of rainfall using radar reflectivity data:A data-mining approach. IEEE Trans Geosci Remote Sens,51(4):2337-2342 doi: 10.1109/TGRS.2012.2210429
    [21] Long J,Shelhamer E,Darrell T. 2015. Fully convolutional networks for semantic segmentation∥2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston MA:IEEE,3431-3440
    [22] Leinonen J,Guillaume A,Yuan T L. 2019. Reconstruction of cloud vertical structure with a generative adversarial network. Geophys Res Lett,46(12):7035-7044 doi: 10.1029/2019GL082532
    [23] Marzban C. 2003. Neural networks for postprocessing model output:ARPS. Mon Wea Rev,131(6):1103-1111 doi: 10.1175/1520-0493(2003)131<1103:NNFPMO>2.0.CO;2
    [24] Peng X D,Che Y Z,Chang J. 2013. A novel approach to improve numerical weather prediction skills by using anomaly integration and historical data. J Geophys Res:Atmos,118(16):8814-8826 doi: 10.1002/jgrd.50682
    [25] Rodrigues E R,Oliveira I,Cunha R L F,et al. 2018. DeepDownscale:A deep learning strategy for high-resolution weather forecast∥2018 IEEE 14th International Conference on e-Science (e-Science). Amsterdam,Netherlands:IEEE,415-422
    [26] Shi W Z,Caballero J,Huszár F,et al. 2016. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network∥2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas,NV,USA:IEEE,1874-1883
    [27] Shi X J, Chen Z R, Wang H, et al. 2015. Convolutional LSTM network: A machine learning approach for precipitation nowcasting∥Proceedings of the 28th International Conference on Neural Information Processing Systems. Montreal, Canada: MIT Press, 802-810
    [28] Shi X J, Gao Z H, Lausen L, et al. 2017. Deep learning for precipitation nowcasting: A benchmark and a new model∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: Curran Associates Inc. , 5622-5632
    [29] Shorten C,Khoshgoftaar T M. 2019. A survey on image data augmentation for deep learning. J Big Data,6(1):60 doi: 10.1186/s40537-019-0197-0
    [30] Wang Y B, Long M S, Wang J M, et al. 2017. PredRNN: Recurrent neural networks for predictive learning using spatiotemporal LSTMs∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: Curran Associates Inc. , 879-888
    [31] Wang Y B, Gao Z F, Long M S, et al. 2018. PredRNN++: Towards a resolution of the deep-in-time dilemma in spatiotemporal predictive learning∥Proceedings of the 35th International Conference on Machine Learning, PMLR 80: 5123-5132
    [32] Wang Y B,Zhang J J,Zhu H Y,et al. 2019. Memory in Memory:A predictive neural network for learning higher-order non-stationarity from spatiotemporal dynamics∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach,CA,USA:IEEE,9146-9154
  • 加载中
图(16)
计量
  • 文章访问数:  136
  • HTML全文浏览量:  10
  • PDF下载量:  36
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-05-27
  • 修回日期:  2021-09-02
  • 网络出版日期:  2021-10-25

目录

    /

    返回文章
    返回