Progress and challenges of deep learning techniques in intelligent grid weather forecasting
-
摘要: 中国智能网格天气预报已初步建立0—30 d涵盖基本气象要素的无缝隙气象预报业务体系。近年深度学习技术兴起,给不同领域带来前所未有的变革。同样,深度学习的非线性映射能力、海量信息提取能力、时空建模能力等优势为进一步提升智能网格预报的准确性和精细化水平提供了新的思路和方法。越来越多的研究将深度学习技术应用于智能网格预报的各个方面,包括数值预报订正和解释应用、集合天气预报、相似集合、统计降尺度、纯数据驱动的预报模型和极端天气预报等,并展示出良好的应用潜力。然而,目前深度学习技术在天气预报领域的应用仍处于起步阶段,将其引入智能网格预报业务体系还面临诸多挑战,主要包括算法的选择、算法的数据基础、多源数据融合以及模型的可解释性、可信度、可用性和工程化等。通过回顾近年来深度学习技术在智能网格预报中的应用进展和前景,同时对面临的挑战与应对进行探讨,将有利于促进深度学习技术在天气客观预报领域更好、更稳定的发展。Abstract: 0—30 d seamless fine gridded weather forecasts have been initially established to cover fundamental forecast elements in China. In recent years, the advances and applications of deep learning have brought unprecedented changes to different fields. The capabilities of nonlinear mapping, massive information extraction, spatial-temporal modeling and other advantages of deep learning provide new concepts and methods for further improvement of forecast accuracy and refinement. The growing studies on deep learning techniques have been applied widely to weather forecasting, including statistical postprocessing, ensemble forecasting, analog ensemble, statistical downscaling, data-driven forecasting models and extreme weather forecasting. The deep learning techniques have demonstrated a great application potential. However, the application of deep learning in gridded weather forecasting is still at the initial stage. The challenges include algorithm selection, benchmark dataset, multi-source data blending, interpretability, reliability, availability and operational implementation, etc., when introducing it into current Intelligent Grid Forecast System. Review of the progress and challenges of the deep learning at fine gridded weather forecasting in recent years will be helpful for us to better understand deep learning techniques and their application in weather forecasting.
-
表 1 目前中国无缝隙智能网格预报技术体系
Table 1. Current technologies of seamless intelligent grid weather forecasting in China
预报时效 技术方法 临近预报
(0—4 h)多尺度变幅光流临近预报技术
循环网络和时空记忆的深度学习
基于迎风格式的平流外推
残差衰减滚动订正短时预报
(4—24 h)基于GRAPES-3 km 实时频率匹配订正技术
快速分析和预报系统(RAFS,Rapid Analysis and Forecast System)短时动态频率拟合
动态双因子滚动预报模型
格点化输出统计快速更新系统(GMOSRR,Gridded Model Output Statistics Rapid Refresh)逐时滚动系统
全球、中尺度匹配融合中短期预报
(1—10 d)基于最优背景场生成技术的定量降水预报
自适应全球及中尺度集成技术
集合最优百分位预报技术
频率匹配订正预报技术
基于逻辑回归(LR,Logistic Regression)配料法降水预报技术
基于最优路径的台风暴雨预报技术
降水相态客观预报技术
基于背景分析的全球模式输出统计(GMOS,Global Model Output Statistics)技术
基于区域建模的非连续变量技术
超前空间实况信息融入技术
海洋风、能见度客观预报技术延伸期预报
(10—30 d)气温最优分位组合方法
降水分级位相组合方法
基于集合的逻辑回归方法
基于集合的卡尔曼滤波方法表 2 深度学习技术在智能网格预报应用的主要算法和模型列表
Table 2. List of main algorithms and models of deep learning in intelligent grid weather forecasting
应用模块 算法/模型 应用效果、优势 主要参考文献 模式解释应用 人工神经网络(ANN) 构建数据间非线性关系,允许加入更多预报因子等数据 McGovern,et al,2017
Salazar,et al,2021卷积神经网络(CNN) 基于卷积等模块计算,相较于ANN,有更好的空间特征提取能力 门晓磊等,2019
陈锦鹏等,2021基于CNN算法的U-Net结构 U-Net作为语义分割模型应用于预报误差订正,优势是保留原始场的空间信息 Dupuy,et al,2021
Han,et al,2021
Grönquist,et al,2021长短期记忆(LSTM) 考虑了时间序列特征在预报解释应用上的影响 Zhang C J,et al,2020 集合预报 参数化统计后处理 人工神经网络(ANN) 利用神经网络自动学习预报变量和分布参数之间的非线性关系 Rasp,et al,2018
Ghazvinian,et al,2021非参数化统计后处理 基于神经网络的分位数回归方法 利用神经网络得到分位数函数的近似值,从而拟合预报变量的概率分布 Cannon,2018
Bremnes,2020基于神经网络的直方图法 基于神经网络预测预报要素在指定分仓值(bins)的概率来估计概率密度分布 Veldkamp,et al,2021
Scheuerer,et al,2020Transformer Transformer算法能够处理全部集合成员,从集合成员之间提取额外信息来订正集合预报 Finn,2021 量化不确
定性卷积神经网络(CNN) 只需少量集合成员就可以提供比原始集合成员更有效的预报不确定性信息 Grönquist,et al,2021 生成集合
预报卷积神经网络(CNN) 替代数值模式,利用深度学习直接生成集合预报 Scher,et al,2018 相似集合(AnEn) 人工神经网络(ANN) 利用ANN生成确定性预报,作为附加预报因子组成相似集合 Cervone,et al,2017 条件变分自编码器(CVAE) 利用CVAE生成预报要素的概率分布,对概率预报进行订正 Fanfarillo,et al,2021 长短期记忆(LSTM) 优化了AnEn计算流程,基于深度学习算法避免了原始AnEn方法在特征选择和权重优化上的缺陷 Hu,et al,2021 统计降尺度 长短期记忆递归神经网络
(RNN-LSTM)能够反映局地气象要素的时空相关性 Misra,et al,2018 基于CNN的超分辨率模型 SRCNN和DeepSD已成为深度学习在统计降尺度中的基准模型 Chao,et al,2014
Vandal,et al,2017
Kumar,et al,2021基于生成式对抗网络(GAN)的超分辨率模型 基于GAN产生更精确的生成样本,代表模型有ESRGAN Wang,et al,2018
Singh,et al,2019
Harris,et al,2022数据驱动的预报模型 卷积长短时记忆模型 (ConvLSTM) 基于时序特征的临近降水预报模型,是该领域里程碑式的工作 Shi,et al,2015 基于注意力机制的MetNet 实现了1 km空间分辨率和2 min时间分辨率美国范围内提前7—8 h的降水预报 Kaae Sønderby,et al,2020 预测循环神经网络(PredRNN)、
运动循环神经网络(MotionRNN)显著提升了捕捉和预测雷达回波运动的能力 Wang,et al,2017
Wu,et al,2021条件生成对抗网络(CGAN) 基于GAN从雷达数据场的条件分布中生成临近降水概率预报 Ravuri,et al,2021 极端天气预报 堆叠的CNN模型 以堆叠的方式构建深度模型,输出降水等级的概率分布,预报临近极端降水发生的概率 Franch,et al,2020 胶囊神经网络 (CapsNets) 能够提取数据特征的属性关系,提高极端气温的预报准确率 Chattopadhyay,et al,2020 -
曹勇,刘凑华,宗志平等. 2016. 国家级格点化定量降水预报系统. 气象,42(12):1476-1482 doi: 10.7519/j.issn.1000-0526.2016.12.005Cao Y,Liu C H,Zong Z P,et al. 2016. State-level gridded quantitative precipitation forecasting system. Meteor Mon,42(12):1476-1482 (in Chinese) doi: 10.7519/j.issn.1000-0526.2016.12.005 曹勇,包红军,张恒德等. 2021. 基于快速滚动更新的无缝隙定量降水预报模型. 河海大学学报(自然科学版),49(4):303-308Cao Y,Bao H J,Zhang H D,et al. 2021. Seamless quantitative precipitation forecasting model based on rapid rolling update technique. J Hohai Univ Nat Sci,49(4):303-308 (in Chinese) 陈锦鹏,冯业荣,蒙伟光等. 2021. 基于卷积神经网络的逐时降水预报订正方法研究. 气象,47(1):60-70 doi: 10.7519/j.issn.1000-0526.2021.01.006Chen J P,Feng Y R,Meng W G,et al. 2021. A correction method of hourly precipitation forecast based on convolutional neural network. Meteor Mon,47(1):60-70 (in Chinese) doi: 10.7519/j.issn.1000-0526.2021.01.006 代刊,朱跃建,毕宝贵. 2018. 集合模式定量降水预报的统计后处理技术研究综述. 气象学报,76(4):493-510 doi: 10.11676/qxxb2018.015Dai K,Zhu Y J,Bi B G. 2018. The review of statistical post-process technologies for quantitative pre-cipitation forecast of ensemble prediction system. Acta Meteor Sinica,76(4):493-510 (in Chinese) doi: 10.11676/qxxb2018.015 金荣花,代刊,赵瑞霞等. 2019. 我国无缝隙精细化网格天气预报技术进展与挑战. 气象,45(4):445-457 doi: 10.7519/j.issn.1000-0526.2019.04.001Jin R H,Dai K,Zhao R X,et al. 2019. Progress and challenge of seamless fine gridded weather forecasting technology in China. Meteor Mon,45(4):445-457 (in Chinese) doi: 10.7519/j.issn.1000-0526.2019.04.001 李扬,刘玉宝,许小峰. 2021. 基于深度学习改进数值天气预报模式和预报的研究及挑战. 气象科技进展,11(3):103-112 doi: 10.3969/j.issn.2095-1973.2021.03.012Li Y,Liu Y B,Xu X F. 2021. Advances and challenges for improving numerical weather prediction models and forecasting using deep learning. Adv Meteor Sci Technol,11(3):103-112 (in Chinese) doi: 10.3969/j.issn.2095-1973.2021.03.012 刘娜,熊安元,张强等. 2021. 强对流天气人工智能应用训练基础数据集构建. 应用气象学报,32(5):530-541 doi: 10.11898/1001-7313.20210502Liu N,Xiong A Y,Zhang Q,et al. 2021. Development of basic dataset of severe convective weather for artificial intelligence training. J Appl Meteor Sci,32(5):530-541 (in Chinese) doi: 10.11898/1001-7313.20210502 马雷鸣. 2020. 天气预报中的人工智能技术进展. 地球科学进展,35(6):551-560Ma L M. 2020. Development of artificial intelligence technology in weather forecast. Adv Earth Sci,35(6):551-560 (in Chinese) 门晓磊,焦瑞莉,王鼎等. 2019. 基于机器学习的华北气温多模式集合预报的订正方法. 气候与环境研究,24(1):116-124 doi: 10.3878/j.issn.1006-9585.2018.18049Men X L,Jiao R L,Wang D,et al. 2019. A temperature correction method for multi-model ensemble forecast in north china based on machine learning. Climatic Environ Res,24(1):116-124 (in Chinese) doi: 10.3878/j.issn.1006-9585.2018.18049 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: 10.11898/1001-7313.20210101Sun 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 孙全德,焦瑞莉,夏江江等. 2019. 基于机器学习的数值天气预报风速订正研究. 气象,45(3):426-436 doi: 10.7519/j.issn.1000-0526.2019.03.012Sun Q D,Jiao R L,Xia J J,et al. 2019. Adjusting wind speed prediction of numerical weather forecast model based on machine learning methods. Meteor Mon,45(3):426-436 (in Chinese) doi: 10.7519/j.issn.1000-0526.2019.03.012 王在文,陈敏,Delle Monache L等. 2019. 相似集合预报方法在北京区域地面气温和风速预报中的应用. 气象学报,77(5):869-884 doi: 10.11676/qxxb2019.044Wang Z W,Chen M,Delle Monache L,et al. 2019. Application of analog ensemble method to surface temperature and wind speed prediction in Beijing area. Acta Meteor Sinica,77(5):869-884 (in Chinese) doi: 10.11676/qxxb2019.044 许小峰. 2018. 从物理模型到智能分析—降低天气预报不确定性的新探索. 气象,44(3):341-350 doi: 10.7519/j.issn.1000-0526.2018.03.001Xu X F. 2018. From physical model to intelligent analysis:A new exploration to reduce the uncertainty of weather forecast. Meteor Mon,44(3):341-350 (in Chinese) doi: 10.7519/j.issn.1000-0526.2018.03.001 俞小鼎,郑永光. 2020. 中国当代强对流天气研究与业务进展. 气象学报,78(3):391-418 doi: 10.11676/qxxb2020.035Yu X D,Zheng Y G. 2020. Advances in severe convective weather research and operational service in China. Acta Meteor Sinica,78(3):391-418 (in Chinese) doi: 10.11676/qxxb2020.035 张小玲,杨波,盛杰,et al. 2018. 中国强对流天气预报业务发展. 气象科技进展,8(3):8-18Zhang X L,Yang B,Sheng J,et al. 2018. Development of operations on forecasting severe convective weather in China. Adv Meteor Sci Technol,8(3):8-18 (in Chinese) 张延彪,陈明轩,韩雷等. 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):153-167 doi: 10.11676/qxxb2021.066Zhang Y B,Chen M X,Han L,et al. 2022. Multi-element deep learning fusion correction method for numerical weather prediction. Acta Meteor Sinica,80(1):153-167 (in Chinese) doi: 10.11676/qxxb2021.066 周康辉,郑永光,韩雷等. 2021a. 机器学习在强对流监测预报中的应用进展. 气象,47(3):274-289Zhou K H,Zheng Y G,Han L,et al. 2021a. Advances in application of machine learning to severe convective weather monitoring and forecasting. Meteor Mon,47(3):274-289 (in Chinese) 周康辉,郑永光,王婷波. 2021b. 利用深度学习融合NWP和多源观测数据的闪电落区短时预报方法. 气象学报,79(1):1-14Zhou K H,Zheng Y G,Wang T B. 2021b. Very short-range lightning forecasting with NWP and observation data:A deep learning approach. Acta Meteor Sinica,79(1):1-14 (in Chinese) Agrawal S, Barrington L, Bromberg C, et al. 2019. Machine learning for precipitation nowcasting from radar images. arXiv: 1912.12132 Alessandrini S,Sperati S,Delle Monache L. 2019. Improving the analog ensemble wind speed forecasts for rare events. Mon Wea Rev,147(7):2677-2692 doi: 10.1175/MWR-D-19-0006.1 Ayzel G,Scheffer T,Heistermann M. 2020. RainNet v1.0:A convolutional neural network for radar-based precipitation nowcasting. Geosci Model Dev,13(6):2631-2644 doi: 10.5194/gmd-13-2631-2020 Baño-Medina J,Manzanas R,Gutiérrez J M. 2020. Configuration and intercomparison of deep learning neural models for statistical down-scaling. Geosci Model Dev,13(4):2109-2124 doi: 10.5194/gmd-13-2109-2020 Bau D,Zhou B L,Khosla A,et al. 2017. Network dissection:Quantifying interpretability of deep visual representations∥Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu:IEEE,3319-3327 Bauer P,Dueben P D,Hoefler T,et al. 2021. The digital revolution of Earth-system science. Nat Comput Sci,1(2):104-113 doi: 10.1038/s43588-021-00023-0 Boeing G. 2016. Visual analysis of nonlinear dynamical systems:Chaos,fractals,self-similarity and the limits of prediction. Systems,4(4):37 doi: 10.3390/systems4040037 Boukabara S A,Krasnopolsky V,Penny S G,et al. 2021. Outlook for exploiting artificial intelligence in the earth and environmental sciences. Bull Amer Meteor Soc,102(5):E1016-E1032 doi: 10.1175/BAMS-D-20-0031.1 Brehmer J R,Strokorb K. 2019. Why scoring functions cannot assess tail properties. Electron J Statist,13(2):4015-4034 Bremnes J B. 2004. Probabilistic forecasts of precipitation in terms of quantiles using NWP model output. Mon Wea Rev,132(1):338-347 doi: 10.1175/1520-0493(2004)132<0338:PFOPIT>2.0.CO;2 Bremnes J B. 2020. Ensemble postprocessing using quantile function regression based on neural networks and bernstein polynomials. Mon Wea Rev,148(1):403-414 doi: 10.1175/MWR-D-19-0227.1 Buehner M,Jacques D. 2020. Non-Gaussian deterministic assimilation of radar-derived precipitation accumulations. Mon Wea Rev,148(2):783-808 doi: 10.1175/MWR-D-19-0199.1 Cannon A J. 2018. Non-crossing nonlinear regression quantiles by monotone composite quantile regression neural network,with application to rainfall extremes. Stoch Environ Res Risk Assess,32(11):3207-3225 doi: 10.1007/s00477-018-1573-6 Cervone G,Clemente-Harding L,Alessandrini S,et al. 2017. Short-term photovoltaic power forecasting using artificial neural networks and an analog ensemble. Renew Energy,108:274-286 doi: 10.1016/j.renene.2017.02.052 Chantry M,Christensen H,Dueben P,et al. 2021. Opportunities and challenges for machine learning in weather and climate modelling:Hard,medium and soft AI. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200083 Chao D,Loy C C,He K M,et al. 2014. Learning a deep convolutional network for image super-resolution∥Proceedings of the 13th European Conference on Computer Vision. Zurich:Springer,184-199 Chattopadhyay A,Nabizadeh E,Hassanzadeh P. 2020. Analog forecasting of extreme-causing weather patterns using deep learning. J Adv Model Earth Syst,12(2):e2019MS001958 Chattopadhyay A,Mustafa M,Hassanzadeh P,et al. 2022. Towards physics-inspired data-driven weather forecasting:Integrating data assimilation with a deep spatial-transformer-based U-NET in a case study with ERA5. Geosci Model Dev,15(5):2221-2237 doi: 10.5194/gmd-15-2221-2022 de Ruiter B. 2021. Post-processing multi-model medium-term precipitation forecasts using convolutional neural networks. arXiv: 2105.07043 Delle Monache L,Nipen T,Liu Y B,et al. 2011. Kalman filter and analog schemes to postprocess numerical weather predictions. Mon Wea Rev,139(11):3554-3570 doi: 10.1175/2011MWR3653.1 Delle Monache L,Eckel F A,Rife D L,et al. 2013. Probabilistic weather prediction with an analog ensemble. Mon Wea Rev,141(10):3498-3516 doi: 10.1175/MWR-D-12-00281.1 Demaeyer J,Vannitsem S. 2020. Correcting for model changes in statistical postprocessing:An approach based on response theory. Nonlin Processes Geophys,27(2):307-327 doi: 10.5194/npg-27-307-2020 Deng J,Dong W,Socher R,et al. 2009. ImageNet:A large-scale hierarchical image database∥Proceedings of 2009 IEEE Conference on Computer Vision and Pattern Recognition. Miami:IEEE,248-255 Dosovitskiy A, Beyer L, Kolesnikov A, et al. 2021. An image is worth 16×16 words: Transformers for image recognition at scale. arXiv: 2010.11929 Dupuy F,Duine G J,Durand P,et al. 2019. Local-scale valley wind retrieval using an artificial neural network applied to routine weather observations. J Appl Meteor Climatol,58(5):1007-1022 doi: 10.1175/JAMC-D-18-0175.1 Dupuy F,Mestre O,Serrurier M,et al. 2021. ARPEGE cloud cover forecast postprocessing with convolutional neural network. Wea Forecasting,36(2):567-586 doi: 10.1175/WAF-D-20-0093.1 Ebert-Uphoff I,Hilburn K. 2020. Evaluation,tuning,and interpretation of neural networks for working with images in meteorological applications. Bull Amer Meteor Soc,101(12):E2149-E2170 doi: 10.1175/BAMS-D-20-0097.1 Fanfarillo A,Roozitalab B,Hu W M,et al. 2021. Probabilistic forecasting using deep generative models. GeoInformatica,25(1):127-147 doi: 10.1007/s10707-020-00425-8 Finn T S. 2021. Self-attentive ensemble transformer: Representing ensemble interactions in neural networks for earth system models. arXiv: 2106.13924 Foresti L,Sideris I V,Nerini D,et al. 2019. Using a 10-year radar archive for nowcasting precipitation growth and decay:A probabilistic machine learning approach. Wea Forecasting,34(5):1547-1569 doi: 10.1175/WAF-D-18-0206.1 Franch G,Nerini D,Pendesini M,et al. 2020. Precipitation nowcasting with orographic enhanced stacked generalization:Improving deep learning predictions on extreme events. Atmosphere,11(3):267 doi: 10.3390/atmos11030267 Frei C,Isotta F A. 2019. Ensemble spatial precipitation analysis from rain gauge data:Methodology and application in the european alps. J Geophys Res,124(11):5757-5778 doi: 10.1029/2018JD030004 French M N,Krajewski W F,Cuykendall R R. 1992. Rainfall forecasting in space and time using a neural network. J Hydrol,137(1-4):1-31 doi: 10.1016/0022-1694(92)90046-X Gardner M W,Dorling S R. 1998. Artificial neural networks (the multilayer perceptron):A review of applications in the atmospheric sciences. Atmos Environ,32(14-15):2627-2636 doi: 10.1016/S1352-2310(97)00447-0 Geer A J. 2021. Learning earth system models from observations:Machine learning or data assimilation?. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200089 Germann U,Zawadzki I. 2002. Scale-dependence of the predictability of precipitation from continental radar images. Part Ⅰ:Description of the methodology. Mon Wea Rev,130(12):2859-2873 doi: 10.1175/1520-0493(2002)130<2859:SDOTPO>2.0.CO;2 Germann U,Zawadzki I. 2004. Scale dependence of the predictability of precipitation from continental radar images. Part Ⅱ:Probability forecasts. J Appl Meteor Climatol,43(1):74-89 doi: 10.1175/1520-0450(2004)043<0074:SDOTPO>2.0.CO;2 Ghazvinian M,Zhang Y,Seo D J,et al. 2021. A novel hybrid artificial neural network:Parametric scheme for postprocessing medium-range pre-cipitation forecasts. Adv Water Resour,151:103907 doi: 10.1016/j.advwatres.2021.103907 Glahn H R,Lowry D A. 1972. The use of model output statistics (MOS) in objective weather forecasting. J Appl Meteor Climatol,11(8):1203-1211 doi: 10.1175/1520-0450(1972)011<1203:TUOMOS>2.0.CO;2 Gneiting T,Raftery A E. 2005. Weather forecasting with ensemble methods. Science,310(5746):248-249 doi: 10.1126/science.1115255 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:MIT Press,2672-2680 Grönquist P,Yao C Y,Ben-Nun T,et al. 2021. Deep learning for post-processing ensemble weather forecasts. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200092 Guastavino S, Piana M, Tizzi M, et al. 2021. Prediction of severe thunderstorm events with ensemble deep learning and radar data. arXiv: 2109.09791 Hamill T M,Bates G T,Whitaker J S,et al. 2013. NOAA's second-generation global medium-range ensemble reforecast dataset. Bull Amer Meteor Soc,94(10):1553-1565 doi: 10.1175/BAMS-D-12-00014.1 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 Harris L, McRae A T T, Chantry M, et al. 2022. A Generative Deep Learning Approach to Stochastic Downscaling of Precipitation Forecasts. arXiv: 2204.02028 Haupt S E,Chapman W,Adams S V,et al. 2021. Towards implementing artificial intelligence post-processing in weather and climate:Proposed actions from the Oxford 2019 workshop. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200091 He K M,Zhang X Y,Ren S Q,et al. 2016. Deep residual learning for image recognition∥Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE,770-778 Hinton G E,Osindero S,Teh Y W. 2006. A fast learning algorithm for deep belief nets. Neural Comput,18(7):1527-1554 doi: 10.1162/neco.2006.18.7.1527 Hochreiter S,Schmidhuber J. 1997. Long short-term memory. Neural Comput,9(8):1735-1780 doi: 10.1162/neco.1997.9.8.1735 Höhlein K,Kern M,Hewson T,et al. 2020. A comparative study of convolutional neural network models for wind field downscaling. Meteor Appl,27(6):e1961 Horn B K P,Schunck B G. 1981. Determining optical flow. Artif Intell,17(1-3):185-203 doi: 10.1016/0004-3702(81)90024-2 Hu W M, Cervone G, Young G, et al. 2021. Weather analogs with a machine learning similarity metric for renewable resource forecasting. arXiv: 2103.04530 Huang G,Liu Z,van der Maaten L,et al. 2017. Densely connected convolutional networks∥Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu:IEEE,2261-2269 Irrgang C,Boers N,Sonnewald M,et al. 2021. Towards neural earth system modelling by integrating artificial intelligence in earth system science. Nat Mach Intell,3(8):667-674 doi: 10.1038/s42256-021-00374-3 Jacques-Dumas V, Ragone F, Borgnat P, et al. 2021. Deep learning-based extreme heatwave forecast. arXiv: 2103.09743 Junk C,Delle Monache L,Alessandrini S. 2015. Analog-based ensemble model output statistics. Mon Wea Rev,143(7):2909-2917 doi: 10.1175/MWR-D-15-0095.1 Kaae Sønderby C, Espeholt L, Heek J, et al. 2020. MetNet: A neural weather model for precipitation forecasting. arXiv: 2003.12140 Kashinath K,Mustafa M,Albert A,et al. 2021. Physics-informed machine learning:Case studies for weather and climate modelling. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200093 Kasim M F, Watson-Parris D, Deaconu L, et al. 2020. Building high accuracy emulators for scientific simulations with deep neural architecture search. arXiv: 2001.08055 Keisler R. 2022. Forecasting Global Weather with Graph Neural Networks. arXiv: 2202.07575 Keller J D,Delle Monache L,Alessandrini S. 2017. Statistical downscaling of a high-resolution precipitation reanalysis using the analog ensemble method. J Appl Meteor Climatol,56(7):2081-2095 doi: 10.1175/JAMC-D-16-0380.1 Kendall A, Gal Y. 2017. What uncertainties do we need in bayesian deep learning for computer vision?∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 5580-5590 Krizhevsky A,Sutskever I,Hinton G E. 2017. ImageNet classification with deep convolutional neural networks. Commun ACM,60(6):84-90 doi: 10.1145/3065386 Kudo A. 2021. Statistical post-processing for gridded temperature prediction using encoder-decoder-based deep convolutional neural networks. arXiv: 2103.01479 Kumar B,Chattopadhyay R,Singh M,et al. 2021. Deep learning–based downscaling of summer monsoon rainfall data over Indian region. Theor Appl Climatol,143(3-4):1145-1156 doi: 10.1007/s00704-020-03489-6 Lagerquist R,McGovern A,Gagne II D J. 2019. Deep learning for spatially explicit prediction of synoptic-scale fronts. Wea Forecasting,34(4):1137-1160 doi: 10.1175/WAF-D-18-0183.1 Lakshminarayanan B, Pritzel A, Blundell C. 2017. Simple and scalable predictive uncertainty estimation using deep ensembles∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 6405-6416 Lang M N,Lerch S,Mayr G J,et al. 2020. Remember the past:A comparison of time-adaptive training schemes for non-homogeneous regression. Nonlin Processes Geophys,27(1):23-34 doi: 10.5194/npg-27-23-2020 Larraondo P R,Renzullo L J,van Dijk A I J M,et al. 2020. Optimization of deep learning precipitation models using categorical binary metrics. J Adv Model Earth Syst,12(5):e2019MS001909 LeCun Y,Bengio Y,Hinton G. 2015. Deep learning. Nature,521(7553):436-444 doi: 10.1038/nature14539 Leinonen J,Nerini D,Berne A. 2021. Stochastic super-resolution for downscaling time-evolving atmospheric fields with a generative adversarial network. IEEE Trans Geosci Remote Sens,59(9):7211-7223 doi: 10.1109/TGRS.2020.3032790 Liu Y J, Racah E, Prabhat J C, et al. 2016. Application of deep convolutional neural networks for detecting extreme weather in climate datasets. arXiv: 1605.01156 Long J,Shelhamer E,Darrell T. 2015. Fully convolutional networks for semantic segmentation∥Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE,3431-3440 Lundberg S M, Lee S I. 2017. A unified approach to interpreting model predictions∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 4768-4777 Marzban C,Stumpf G J. 1998. A neural network for damaging wind prediction. Wea Forecasting,13(1):151-163 doi: 10.1175/1520-0434(1998)013<0151:ANNFDW>2.0.CO;2 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 McGovern A,Elmore K L,Gagne Ⅱ D J,et al. 2017. Using artificial intelligence to improve real-time decision-making for high-impact weather. Bull Amer Meteor Soc,98(10):2073-2090 doi: 10.1175/BAMS-D-16-0123.1 McGovern A,Lagerquist R,Gagne Ⅱ D J,et al. 2019. Making the black box more transparent:Understanding the physical implications of machine learning. Bull Amer Meteor Soc,100(11):2175-2199 doi: 10.1175/BAMS-D-18-0195.1 Meinshausen N. 2006. Quantile regression forests. J Mach Learn Res,7:983-999 Misra S,Sarkar S,Mitra P. 2018. Statistical downscaling of precipitation using long short-term memory recurrent neural networks. Theor Appl Climatol,134(3/4):1179-1196 Montavon G,Lapuschkin S,Binder A,et al. 2017. Explaining nonlinear classification decisions with deep Taylor decomposition. Pattern Recogn,65:211-222 doi: 10.1016/j.patcog.2016.11.008 Pan B X,Hsu K,AghaKouchak A,et al. 2019. Improving precipitation estimation using convolutional neural network. Water Resour Res,55(3):2301-2321 doi: 10.1029/2018WR024090 Pathak J, Subramanian S, Harrington P, et al. 2022. FourCastNet: A global data-driven high-resolution weather model using adaptive fourier neural operators. arXiv: 2202.11214 Pfaff T, Fortunato M, Sanchez-Gonzalez A, et al. 2020. Learning mesh-based simulation with graph networks. arXiv: 2010.03409 Price I, Rasp S. 2022. Increasing the accuracy and resolution of precipitation forecasts using deep generative models. arXiv: 2203.12297v1 Prudden R, Adams S, Kangin D, et al. 2020. A review of radar-based nowcasting of precipitation and applicable machine learning techniques. arXiv: 2005.04988 Raghu M, Schmidt E. 2020. A survey of deep learning for scientific discovery. arXiv: 2003.11755 Rasp S,Lerch S. 2018. Neural networks for postprocessing ensemble weather forecasts. Mon Wea Rev,146(11):3885-3900 doi: 10.1175/MWR-D-18-0187.1 Rasp S,Dueben P D,Scher S,et al. 2020. WeatherBench:A benchmark data set for data-driven weather forecasting. J Adv Model Earth Syst,12(11):e2020MS002203 Rasp S,Thuerey N. 2021. Data-driven medium-range weather prediction with a resnet pretrained on climate simulations:A new model for weather bench. J Adv Model Earth Sy,13(2):e2020MS002405 Ravuri S,Lenc K,Willson M,et al. 2021. Skilful precipitation nowcasting using deep generative models of radar. Nature,597(7878):672-677 doi: 10.1038/s41586-021-03854-z Reichstein M,Camps-Valls G,Stevens B,et al. 2019. Deep learning and process understanding for data-driven Earth system science. Nature,566(7743):195-204 doi: 10.1038/s41586-019-0912-1 Ronneberger O,Fischer P,Brox T. 2015. U-Net:Convolutional networks for biomedical image segmentation∥Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention. Munich:Springer,234-241 Roscher R,Bohn B,Duarte M F,et al. 2020. Explainable machine learning for scientific insights and discoveries. IEEE Access,8:42200-42216 doi: 10.1109/ACCESS.2020.2976199 Sabour S, Frosst N, Hinton G E. 2017. Dynamic routing between capsules∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 3859-3869 Sachindra D A,Ahmed K,Rashid M,et al. 2018. Statistical downscaling of precipitation using machine learning techniques. Atmos Res,212:240-258 doi: 10.1016/j.atmosres.2018.05.022 Salazar A A, Che Y Z, Zheng J F, et al. 2021. Multivariable neural network to postprocess short-term, hub-height wind forecasts. Energy Sci Eng, doi: 10.1002/ese3.928 Schaumann P,Hess R,Rempel M,et al. 2021. A calibrated and consistent combination of probabilistic forecasts for the exceedance of several precipitation thresholds using neural networks. Wea Forecasting,36(3):1079-1096 Scher S,Messori G. 2018. Predicting weather forecast uncertainty with machine learning. Quart J Roy Meteor Soc,144(717):2830-2841 doi: 10.1002/qj.3410 Scheuerer M,Switanek M B,Worsnop R P,et al. 2020. Using artificial neural networks for generating probabilistic subseasonal precipitation forecasts over California. Mon Wea Rev,148(8):3489-3506 doi: 10.1175/MWR-D-20-0096.1 Schultz M G,Betancourt C,Gong B,et al. 2021. Can deep learning beat numerical weather prediction?. Philos Trans Roy Soc A Math Phys Eng Sci,379(2194):20200097 Selvaraju R R,Cogswell M,Das A,et al. 2020. Grad-CAM:Visual explanations from deep networks via gradient-based localization. Int J Comput Vision,128(2):336-359 doi: 10.1007/s11263-019-01228-7 Sha Y K,Gagne Ⅱ D J,West G,et al. 2020a. Deep-learning-based gridded downscaling of surface meteorological variables in complex terrain. Part Ⅰ:Daily maximum and minimum 2-m temperature. J Appl Meteor Climatol,59(12):2057-2073 doi: 10.1175/JAMC-D-20-0057.1 Sha Y K,Gagne Ⅱ D J,West G,et al. 2020b. Deep-learning-based gridded downscaling of surface meteorological variables in complex terrain. Part Ⅱ:Daily precipitation. J Appl Meteor Climatol,59(12):2075-2092 doi: 10.1175/JAMC-D-20-0058.1 Shapley L S. 1953. A value for n-person games∥Kuhn H, Tucker A. Contributions to the Theory of Games Ⅱ. Princeton: Princeton University Press 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:MIT Press,802-810 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: Curran Associates Inc. , 5622-5632 Shi X J, Yeung D Y. 2018. Machine learning for spatiotemporal sequence forecasting: A survey. arXiv: 1808.06865 Simonyan K, Vedaldi A, Zisserman A. 2014. Deep inside convolutional networks: Visualising image classification models and saliency maps. arXiv: 1312.6034 Singh A, Albert A, White B. 2019. Downscaling numerical weather models with GANs∥Proceedings of the 9th International Workshop on Climate Informatics. Paris: Ecole Normale Superieure Sohn K,Yan X C,Lee H. 2015. Learning structured output representation using deep conditional generative models∥Proceedings of the 28th International Conference on Neural Information Processing Systems. Montreal:MIT Press,3483-3491 Sperati S,Alessandrini S,Delle Monache L. 2017. Gridded probabilistic weather forecasts with an analog ensemble. Quart J Roy Meteor Soc,143(708):2874-2885 doi: 10.1002/qj.3137 Sun J Z. 2005. Convective-scale assimilation of radar data:Progress and challenges. Quart J Roy Meteor Soc,131(613):3439-3463 doi: 10.1256/qj.05.149 Taillardat M,Mestre O,Zamo M,et al. 2016. Calibrated ensemble forecasts using quantile regression forests and ensemble model output statistics. Mon Wea Rev,144(6):2375-2393 doi: 10.1175/MWR-D-15-0260.1 Toms B A,Barnes E A,Ebert-Uphoff I. 2020. Physically interpretable neural networks for the geosciences:Applications to earth system variability. J Adv Model Earth Syst,12(9):e2019MS002002 van Schaeybroeck B,Vannitsem S. 2015. Ensemble post-processing using member-by-member approaches:Theoretical aspects. Quart J Roy Meteor Soc,141(688):807-818 doi: 10.1002/qj.2397 Vandal T,Kodra E,Ganguly S,et al. 2017. DeepSD:Generating high resolution climate change projections through single image super-resolution∥Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Halifax:ACM,1663-1672 Vandal T,Kodra E,Ganguly A R. 2019. Intercomparison of machine learning methods for statistical downscaling:The case of daily and extreme precipitation. Theor Appl Climatol,137(1-2):557-570 doi: 10.1007/s00704-018-2613-3 Vannitsem S, Wilks D S, Messner J W. 2018. Statistical Postprocessing of Ensemble Forecasts. Amsterdam: Elsevier Vannitsem S,Bremnes J B,Demaeyer J,et al. 2021. Statistical postprocessing for weather forecasts:Review,challenges,and avenues in a big data world. Bull Amer Meteor Soc,102(3):E681-E699 doi: 10.1175/BAMS-D-19-0308.1 Vaswani A, Shazeer N, Parmar N, et al. 2017. Attention is all you need∥Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 6000-6010 Veldkamp S,Whan K,Dirksen S,et al. 2021. Statistical postprocessing of wind speed forecasts using convolutional neural networks. Mon Wea Rev,149(4):1141-1152 doi: 10.1175/MWR-D-20-0219.1 Wang X T,Yu K,Wu S X,et al. 2018. ESRGAN:Enhanced super-resolution generative adversarial networks∥Proceedings of the 15th European Conference on Computer Vision. Munich:Springer,63-79 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: Curran Associates Inc. , 879-888 Wang Z H,Chen J,Hoi S C H. 2021. Deep learning for image super-resolution:A survey. IEEE Trans Pattern Anal Mach Intell,43(10):3365-3387 doi: 10.1109/TPAMI.2020.2982166 Weyn J A,Durran D R,Caruana R,et al. 2021. Sub-seasonal forecasting with a large ensemble of deep-learning weather prediction models. J Adv Model Earth Syst,13(7):e2021MS002502 Witt C S d, Tong C, Zantedeschi V, et al. 2020. RainBench: Towards global precipitation forecasting from satellite imagery. arXiv: 2012.09670 Wu H X,Yao Z Y,Wang J M,et al. 2021. MotionRNN:A flexible model for video prediction with spacetime-varying motions∥Proceedings of 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Nashville:IEEE,15430-15439 Xu Q F,Deng K,Jiang C X,et al. 2017. Composite quantile regression neural network with applications. Expert Syst Appl,76:129-139 doi: 10.1016/j.eswa.2017.01.054 Yang C L,Wang N L,Wang S J,et al. 2018. Performance comparison of three predictor selection methods for statistical downscaling of daily precipitation. Theor Appl Climatol,131(1-2):43-54 doi: 10.1007/s00704-016-1956-x Yuan H L,Gao X G,Mullen S L,et al. 2007. Calibration of probabilistic quantitative precipitation forecasts with an artificial neural network. Wea Forecasting,22(6):1287-1303 doi: 10.1175/2007WAF2006114.1 Zhang C J,Zeng J,Wang H Y,et al. 2020. Correction model for rainfall forecasts using the LSTM with multiple meteorological factors. Meteor Appl,27(1):e1852 Zhang X L,Sun J H,Zheng Y G,et al. 2020. Progress in severe convective weather forecasting in China since the 1950s. J Meteor Res,34(4):699-719 doi: 10.1007/s13351-020-9146-2 Zhang Y,Tiňo P,Leonardis A,et al. 2021. A survey on neural network interpretability. IEEE Trans Emerging Top Comput Intell,5(5):726-742 doi: 10.1109/TETCI.2021.3100641 Zhou K H,Zheng Y G,Li B,et al. 2019. Forecasting different types of convective weather:A deep learning approach. J Meteorol Res,33(5):797-809 doi: 10.1007/s13351-019-8162-6 Zhou K H,Zheng Y G,Dong W S,et al. 2020. A deep learning network for cloud-to-ground lightning nowcasting with multisource data. J Atmos Ocean Tech,37(5):927-942 doi: 10.1175/JTECH-D-19-0146.1 -