基于多要素 3D 特征提取的短期定量降水预报技术研究
作者单位:

1.南京信息工程大学;2.国家气象中心

基金项目:

新疆维吾尔自治区重点研发任务专项(编号 2022B03027)、国家重点研究发展计划灾害性天气公里级无缝隙 快速滚动预报技术研究项目(2021YFC3000903)、国家自然科学基金项目(重点项目)(U20B2061)、中国气象局创新发展 项目(CXFZ2023J001)


Research on Short-Term Quantitative Precipitation Forecasting Technology Based on Multi-Factor 3D Feature Extraction
Author:
Affiliation:

1.Nanjing University of Information Science and Technology;2.National Meteorological Centre

Fund Project:

Xinjiang Uygur Autonomous Region Key Research and Development Task Special (No. 2022B03027), National Key Research and Development Plan Disaster Weather Kilometer Level seamless Rapid Rolling Forecasting Technology Research Program (2021YFC3000903), National Natural Science Foundation of China (Key Program) (U20B2061), and Innovative Development of China Meteorological Administration Program (CXFZ2023J001)

  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献
  • | |
  • 引证文献
  • | |
  • 文章评论
    摘要:

    传统业务数值天气模式 (Numerical Weather Prediction,NWP)在定量降水预报任务中,由于空间分辨率有限、物理参数化方案不够完善、泛化性较弱等原因,使得 NWP 预报中存在固有偏差,而深度学习神经网络有着强大的非线性拟合能力、能够自主性学习到任务相关的关键特征、泛化性较高等优势,可以有望改善现状。为了解决上述问题,本文提出基于多要素 3D 特征提取的短期定量降水预报技术研究,基于欧洲中期天气预报中心(ECMWF)提供的高分辨率 ECMWF-HRES(EC-Hres)模式预报数据,构建3D-QPF(3D-Quantitative Precipitation Forecast)语义分割模型,通过先分类后回归的耦合框架,捕捉多种降水相关要素数据的 3D 空间特征,并得到与降水实况数据间的相非线性关系,最后在损失函数中增加 PR(Pre and Rec)损失函数进一步提升模型对偏态数据的预报效果。实验结果表明,3D-QPF 的逐日累积降水预报不仅在晴雨量级(0.1mm/24h)准确率评分稳定增长,在暴雨量级(50mm/24h)的准确率评分也有明显提升:暴雨量级较 EC-Hres 的 TS(Threat Score)评分最高提升了 15.8%,均方根误差(RMSE,Root Mean Square Error)优化达到 18.71%。经过长期检验,3D-QPF 模型与 EC-Hres、中国气象局全球模式(CMA-GFS)预报以及 2D-Unet 和 3D-Unet 等经典网络模型相比做出了有效的预报订正效果。此外,随着预报时效延长至 3 天,模型的优化效果仍能够保持相较稳定。

    Abstract:

    Traditional numerical weather prediction (NWP) models have inherent biases in quantitative precipitation forecasting tasks due to limited spatial resolution, incomplete physical parameterization schemes, and weak generalization, Deep learning neural networks, with their strong non-linear fitting capabilities, autonomous learning of task-relevant features, and high generalization, hold the potential to address these issues and improve the current state of NWP forecasting. To address the aforementioned issues, this paper proposes a research on short-term quantitative precipitation forecasting technology based on multi factor 3D feature extraction. Based on high-resolution ECMWF HRES (EC Hs) model forecasting data provided by the European Centre for Medium Range Weather Forecasts (ECMWF), a 3D-QPF (3D-Quantitative Precision Forecast) semantic segmentation model is constructed. Through a coupled framework of classification and regression, the 3D spatial features of various precipitation related element data are captured, And obtain a non-linear relationship with the actual precipitation data, and finally add the PR (Pre and Rec) loss function to the loss function to further improve the predictive performance of the model on skewed data.The experimental results show that the accuracy score of 3D-QPF daily cumulative precipitation forecast not only increases steadily on the sunny and rainy scale (0.1mm/24h), but also improves significantly on the rainstorm scale (50mm/24h): the rainstorm scale is 15.8% higher than the EC-Hres 'TS(Threat Score) score, and the root mean square error (RMSE,Root Mean Square Error) optimization reaches 18.71%. After long-term testing, the 3D- QPF model has achieved effective prediction correction compared to classic network models such as ECHres, China Meteorological Administration Global Model (CMA-GFS) forecasting, 2D Unet, and 3D-Unet. In addition, as the forecast time is extended to 3 days, the optimization effect of the model can still remain relatively stable.

    参考文献
    [1] 杜行远. 数值天气预报简介 [J]. 气象, 1979, (03): 5-8.
    [2] 杨绚, 代刊, 朱跃建. 深度学习技术在智能网格天气预报中的应用进展与挑战 [J]. 气象学报, 2022, 80(05): 649-67.
    [3] LECUN Y, BENGIO Y, HINTON G. Deep learning [J]. nature, 2015, 521(7553): 436-44.
    [4] CHANTRY M, CHRISTENSEN H, DUEBEN P, et al. Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI [J]. Philosophical Transactions of the Royal Society A, 2021, 379(2194): 20200083.
    [5] LI H, CHEN L, ZHONG X, et al. FuXi: A cascade machine learning forecasting system for 15-day global weather forecast [J]. npj Clim Atmos Sci, 2023, (6): 190.
    [6] BI K, XIE L, ZHANG H, et al. Accurate medium-range global weather forecasting with 3D neural networks [J]. Nature, 2023, 619(7970): 533-8.
    [7] S?NDERBY C K, ESPEHOLT L, HEEK J, et al. Metnet: A neural weather model for precipitation forecasting [J]. arXiv preprint arXiv:200312140, 2020.
    [8] TOMPSON J J, JAIN A, LECUN Y, et al. Joint training of a convolutional network and a graphical model for human pose estimation [J]. Advances in neural information processing systems, 2014, 27.
    [9] YAN Y, LEE F, WU X, et al. Face recognition algorithm using extended vector quantization histogram features [J]. Plos one, 2018, 13(1): e0190378.
    [10] BEN-BOUALLEGUE Z, CLARE M C, MAGNUSSON L, et al. The rise of data-driven weather forecasting [J]. arXiv preprint arXiv:230710128, 2023.
    [11] WANG Y, LONG M, WANG J, et al. Predrnn: Recurrent neural networks for predictive learning using spatiotemporal lstms [J]. Advances in neural information processing systems, 2017, 30.
    [12] WANG Y, GAO Z, LONG M, et al. Predrnn++: Towards a resolution of the deep-in-time dilemma in spatiotemporal predictive learning; proceedings of the International Conference on Machine Learning, F, 2018 [C]. PMLR.
    [13] 王绍武. 短期气候预测的可预报性与不确定性 [J]. 地球科学进展, 1998, (01): 9-15.
    [14] 智协飞, 王田, 季焱. 基于深度学习的中国地面气温的多模式集成预报研究 [J]. 大气科学学报, 2020, 43(03): 435-46.
    [15] 左洪超, 吕世华, 胡隐樵. 中国近 50 年气温及降水量的变化趋势分析 [J]. 高原气象, 2004, (02): 238-44.
    [16] 施能, 陈绿文, 林振敏. 全球降水与区域性季风降水相关分析 [J]. 气象科技, 2002, (02): 73-9.
    [17] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11): 2278-324.
    [18] PRICE I, RASP S. Increasing the accuracy and resolution of precipitation forecasts using deep generative models; proceedings of the International conference on artificial intelligence and statistics, F, 2022 [C]. PMLR.
    [19] CIVITARESE D S, SZWARCMAN D, ZADROZNY B, et al. Extreme precipitation seasonal forecast using a transformer neural network [J]. arXiv preprint arXiv:210706846, 2021.
    [20] CHEN L, DU F, HU Y, et al. SwinRDM: integrate SwinRNN with diffusion model towards high-resolution and high-quality weather forecasting; proceedings of the Proceedings of the AAAI Conference on Artificial Intelligence, F, 2023 [C].
    [21] 史加荣, 缑璠. 基于 EEMD-CNN-GRU 的短期风向预测 [J]. 南京信息工程大学学报(自然科学版), 2023, 15(05): 568-73.
    [22] SCHER S, MESSORI G. Weather and climate forecasting with neural networks: using general circulation models (GCMs) with different complexity as a study ground [J]. Geoscientific Model Development, 2019, 12(7): 2797-809.
    [23] SONG K, YANG G, WANG Q, et al. Deep learning prediction of incoming rainfalls: An operational service for the city of Beijing China; proceedings of the 2019 International Conference on Data Mining Workshops (ICDMW), F, 2019 [C]. IEEE.
    [24] 曹伟华, 南刚强, 陈明轩, et al. 基于深度学习的京津冀地区精细尺度降水临近预报研究 [J]. 气象学报, 2022, 80(04): 546-64.
    [25] ZHOU K, SUN J, ZHENG Y, et al. Quantitative Precipitation Forecast Experiment Based on Basic NWP Variables Using Deep Learning [J]. Advances in Atmospheric Sciences, 2022, 39(9): 1472-86.
    [26] BAUER P, THORPE A, BRUNET G. The quiet revolution of numerical weather prediction [J]. Nature, 2015, 525(7567): 47-55.
    [27] HERSBACH H, BELL B, BERRISFORD P, et al. The ERA5 global reanalysis [J]. Quarterly Journal of the Royal Meteorological Society, 2020, 146(730): 1999-2049.
    [28] 宫宇, 代刊, 徐珺, et al. GRAPES-GFS 模式暴雨预报天气学检验特征 [J]. 气象, 2018, 44(09): 1148-59.
    [29] 邱辉, 郭云谦, 李帅, et al. ECMWF 集合降水预报在长江流域应用性能评估 [J]. 人民长江,2020, 51(02): 71-6.
    [30] 陆振宇, 孔小翠, 张恒德, et al. 机器学习对延伸期降水集合预报的订正分析 [J]. 现代电子技术, 2022, 45(07): 86-91.
    [31] 韩帅, 师春香, 姜立鹏, et al. CLDAS 土壤湿度模拟结果及评估 [J]. 应用气象学报, 2017, 28(03): 369-78.
    [32] 师春香, 潘旸, 谷军霞, et al. 多源气象数据融合格点实况产品研制进展 [J]. 气象学报, 2019, 77(04): 774-83.
    [33] HAN S, LIU B, SHI C, et al. Evaluation of CLDAS and GLDAS datasets for Near-surface Air Temperature over major land areas of China [J]. Sustainability, 2020, 12(10): 4311.
    [34] 郭阳, 师春香, 徐宾, et al. CLDAS 陆面融合实况数据对天津雾和霾判识的准确性分析 [J]. 干旱气象, 2023, 41(04): 657-65.
    [35] SUN J, TAO Z. Some essential issues connected with severe convective weather analysis and forecast [J]. Meteor Mon, 2012, 38(2): 164-73.
    [36] ?I?EK ?, ABDULKADIR A, LIENKAMP S S, et al. 3D U-Net: learning dense volumetric segmentation from sparse annotation; proceedings of the Medical Image Computing and Computer-Assisted Intervention–MICCAI 2016: 19th International Conference, Athens, Greece, October 17-21, 2016, Proceedings, Part II 19, F, 2016 [C]. Springer.
    [37] WOO S, PARK J, LEE J-Y, et al. Cbam: Convolutional block attention module; proceedings of the Proceedings of the European conference on computer vision (ECCV), F, 2018 [C].
    [38] LIN T-Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection; proceedings of the Proceedings of the IEEE international conference on computer vision, F, 2017 [C].
    [39] 陈桂英, 赵振国. 短期气候预测评估方法和业务初估 [J]. 应用气象学报, 1998, (02): 51-8.
    [40] LI X, SUN X, MENG Y, et al. Dice loss for data-imbalanced NLP tasks [J]. arXiv preprint arXiv:191102855, 2019.
    [41] LARRAONDO P R, RENZULLO L J, VAN DIJK A I, et al. Optimization of deep learning precipitation models using categorical binary metrics [J]. Journal of Advances in Modeling Earth Systems, 2020, 12(5): e2019MS001909.
    [42] TREBING K, STA?CZYK T, MEHRKANOON S. SmaAt-UNet: Precipitation nowcasting using a small attention-UNet architecture [J]. Pattern Recognition Letters, 2021, 145: 178-86.
    [43] KINGMA D P, BA J. Adam: A method for stochastic optimization [J]. arXiv preprint arXiv:14126980, 2014.
    相似文献
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

熊文睿,张恒德,陆振宇,郭云谦.基于多要素 3D 特征提取的短期定量降水预报技术研究[J].南京信息工程大学学报,,():

复制
分享
文章指标
  • 点击次数:167
  • 下载次数: 0
  • HTML阅读次数: 0
  • 引用次数: 0
历史
  • 收稿日期:2023-12-22
  • 最后修改日期:2024-01-23
  • 录用日期:2024-01-23

地址:江苏省南京市宁六路219号    邮编:210044

联系电话:025-58731025    E-mail:nxdxb@nuist.edu.cn

南京信息工程大学学报 ® 2025 版权所有  技术支持:北京勤云科技发展有限公司