LI Xiaoyong,CHEN Keyi,LI Xichen.Application of Convolutional Neural Network in ENSO Prediction[J].Journal of Chengdu University of Information Technology,2022,37(01):81-87.[doi:10.16836/j.cnki.jcuit.2022.01.014]
卷积神经网络在ENSO预报中的应用
- Title:
- Application of Convolutional Neural Network in ENSO Prediction
- 文章编号:
- 2096-1618(2022)01-0081-07
- 分类号:
- P457.8
- 文献标志码:
- A
- 摘要:
- 为提高对ENSO的预报能力，同时针对用机器学习方法做气候预报时观测资料不足的问题，基于深度卷积神经网络(convolutional neural networks，CNN)架构，以CMIP6模式资料和GODAS观测资料为数据集，训练出一个应用于ENSO预报的神经网络。结果表明，在训练神经网络时引入CMIP6模式资料能提高数据量，解决了机器学习中观测资料不足的问题。在时效为1～9个月的后报实验中，神经网络的表现优于传统的动力模式和统计模式。对照实验显示模式数据的加入以及采用集合预报的方法有利于改善预报效果，热含量数据的加入则表现出负面效果。对后报实验的结果分析显示，神经网络的预报准确度存在年内和年际变化，其中年内变化与ENSO预报中普遍存在的春季预报障碍有关。实验结果显示卷积神经网络在ENSO预报中的有效性。
- Abstract:
- In order to improve the forecasting of ENSO and to address the problem of insufficient observational data when using machine learning methods for climate forecasting, a neural network based on a deep convolutional neural network(CNN)architecture was trained for ENSO forecasting using CMIP6 model data and GODAS observations as the dataset. The results show that the introduction of CMIP6 model data in training the neural network can improve the amount of data and solve the problem of insufficient observation information in machine learning. In the hindcast experiment with a lead time of 1to 9 months, the performance of the network is better than the traditional dynamic models and statistical models. Control experiments show that the introduction of model data and the use of ensemble forecast are conducive to improving the prediction effect, while the addition of heat content data shows a negative effect. The analysis of the results of the hindcast experiment shows that there are annual and interannual variation in the accuracy, and the interannual variation is related to the spring predictability barrier which is generally present in ENSO forecast. The results of the experiments show the effectiveness of convolutional neural networks in ENSO forecasting.
参考文献/References:
[1] Bjerknes J A.Atmospheric Teleconnections from the Equatorial Pacific[J].Monthly Weather Review,1969,97(3):163-172.
[2] 李崇银.气候动力学引论[M].2版.北京:气象出版社,2000:245-262.
[3] Johnson S J,Stockdale T N,Ferranti L,et al.SEAS5: The new ECMWF seasonal forecast system[J].Geoscientific Model Development Discussions,2019,12(3):1087-1117.
[4] 吴洪宝,吴蕾.气候变率诊断和预测方法[M].北京:气象出版社,2010:99-263.
[5] Xue Y,Leetmaa A,Ji M. ENSO Prediction with Markov Models:The Impact of Sea Level[J].Journal of Climate,2000,13(4):849-871.
[6] Kondrashov D,Kravtsov Sergey,Robertson Andrew,et al.A hierarchy of data-based ENSO models[J].Journal of Climate,2005,18(21):4425-4444.
[7] Hecht Nielsen R.Theory of the backpropagation neural network[J].International 1989 Joint Conference on Neural Networks,1988,1(S1):445-448.
[8] 许柏宁,姜金荣,郝卉群,等.一种基于区域海表面温度异常预测的ENSO预报深度学习模型[J].科研信息化技术与应用,2017,8(6):65-76.
[9] Dandan He,Pengfei Lin, Hailong Liu,et al.DLENSO:A Deep Learning ENSO Forecasting Model[R].PRICAI 2019:Trends in Artificial Intelligence,2019:12-23.
[10] Broni Bedaiko C,Katsriku F A,Unemi T,et al.El ni o-southern oscillation forecasting using complex networks analysis of LSTM neural networks[R].Artificial Life and Robotics,2019,24(4):445-451.
[11] Ham Y,Kim J,Luo J,et al.Deep learning for multi-year ENSO forecasts[J].Nature,2019,573(7775):568-572.
[12] Behringer D W,Xue Y.Evaluation of the global ocean data assimilation system at NCEP:The Pacific Ocean[C].Proceedings of Eighth Symposium on Integrated Observing and Assimilation Systems for Atmosphere,Oceans,and Land Surface,2004:11-15.
[13] Mckenna S,Santoso A,Gupta A S,et al.Indian Ocean Dipole in CMIP5 and CMIP6: characteristics,biases,and links to ENSO[J].Scientific Reports,2020,10(1).
[14] Santoso A,Mcphaden M J,Cai W. The defining characteristics of ENSO extremes and the strong 2015/16 El Niño:ENSO extremes[J].Reviews of Geophysics,2017,55(4):1079-1129.
[15] Gergis J L,Fowler A M. Classification of synchronous oceanic and atmospheric El Nio-Southern Oscillation(ENSO)events for palaeoclimate reconstruction[J]. International Journal of Climatology,2010,25(12):1541-1565.
[16] Simonyan K,Zisserman A.Very Deep Convolutional Networks for Large-Scale Image Recognition[J].Computer Science-Computer Vision and Pattern Recognition,2014:1409-1556.
[17] Pappenberger F,Beven K J,Hunter N M,et al.Cascading model uncertainty from medium range weather forecasts(10 days)through a rainfall-runoff model to flood inundation predictions within the European Flood Forecasting System(EFFS)[J].Hydrology and Earth System Sciences,2005,9(4):381-393.
[18] Thomas Tomkins Warner.Numerical Weather and Climate Prediction[M].Cambridge:Cambridge University Press,2011:252-280.
[19] Joseph Lee Rodgers,W Alan Nicewander.Thirteen ways to look at the correlation coefficient[J].The American Statistician,1988,42(1):59-66.
[20] McPhaden, Michael J.Tropical Pacific Ocean heat content variations and ENSO persistence barriers[J].Geophysical Research Letters,2003,30(9):319-338.
[21] Weyn J A,Durran D R,Caruana R.Can Machines Learn to Predict Weather? Using Deep Learning to Predict Gridded 500 hPa Geopotential Height From Historical Weather Data[J].Journal of Advances in Modeling Earth Systems,2019,11(8):2680-2693.
[22] Anthony G Barnston,Michael K Tippett, Michelle L L’Heureux, et al. Skill of Real-Time Seasonal ENSO Model Predictions during 2002-11: Is Our Capability Increasing? [J].Bulletin of the American Meteorological Society,2012,93(5):631-651.
[23] Mu M,Duan W,Wang B.Season-dependent dynamics of nonlinear optimal error growth and El Nino-Southern Oscillation predictability in a theoretical model[J].Journal of Geophysical Research Atmospheres,2007,112(D10):D10113(1-10).
[24] Mu M,Xu H,Duan W,et al.A kind of initial errors related to “spring predictability barrier” for El Niño events in Zebiak-Cane model[J]. Geophysical Research Letters,2007,34(3):1-6.
[25] Zheng F,Fang X H,Zhu J,et al.Modulation of Bjerknes feedback on the decadal variations in ENSO predictability[J].Geophysical Research Letters,2016,43(24):12560-12568.
[26] 张雅乐,俞永强,段晚锁.四个耦合模式ENSO后报试验的“春季预报障碍”[J].气象学报,2012,70(3):506-519.
相似文献/References:
[1]廖洪涛,肖天贵,魏 微,等.东亚梅雨季低频波波包传播特征[J].成都信息工程大学学报,2019,(02):143.[doi:10.16836/j.cnki.jcuit.2019.02.008]
LIAO Hongtao,XIAO Tiangui,WEI Wei,et al.Low Frequency Wave Packet Propagation
Characteristics in East Asian Meiyu Season[J].Journal of Chengdu University of Information Technology,2019,(01):143.[doi:10.16836/j.cnki.jcuit.2019.02.008]
[2]王雨歌,郑佳锋,朱克云,等.一次西南涡过程的云-降水毫米波云雷达回波特征分析[J].成都信息工程大学学报,2019,(02):172.[doi:10.16836/j.cnki.jcuit.2019.02.011]
WANG Yuge,ZHENG Jiafeng,ZHU Keyun,et al.Analysis of Cloud-Precipitation Echo Characteristics of a Southwest Vortex[J].Journal of Chengdu University of Information Technology,2019,(01):172.[doi:10.16836/j.cnki.jcuit.2019.02.011]
[3]青 泉,罗 辉,陈刚毅.基于L波段秒级探空数据V-3θ图形的四川盆地暴雨预报模型研究[J].成都信息工程大学学报,2019,(02):186.[doi:10.16836/j.cnki.jcuit.2019.02.013]
QING Quan,LUO Hui,CHEN Gangyi.Forecasting Model of Torrential Rain in Sichuan Basin based on V-3θ
Structural Graphs of L-Band Second Level Sounding Data[J].Journal of Chengdu University of Information Technology,2019,(01):186.[doi:10.16836/j.cnki.jcuit.2019.02.013]
[4]吴秋月,华 维,申 辉,等.基于湿位涡与螺旋度的一次西南低涡强降水分析[J].成都信息工程大学学报,2019,(01):63.[doi:10.16836/j.cnki.jcuit.2019.01.013]
WU Qiuyue,HUA Wei,SHEN Hui,et al.Diagnostic Analysis of a Southwest Vortex Rainstormbased on Moist Potential Vorticity and Helicity[J].Journal of Chengdu University of Information Technology,2019,(01):63.[doi:10.16836/j.cnki.jcuit.2019.01.013]
[5]李潇濛,赵琳娜,肖天贵,等.2000-2015年青藏高原切变线统计特征分析[J].成都信息工程大学学报,2018,(01):91.[doi:10.16836/j.cnki.jcuit.2018.01.016]
LI Xiao-meng,ZHAO Lin-na,XIAO Tian-gui,et al.Statistical Characteristics Analysis of the Shear Linein the Qinghai-Tibet Plateau from 2000 to 2015[J].Journal of Chengdu University of Information Technology,2018,(01):91.[doi:10.16836/j.cnki.jcuit.2018.01.016]
[6]喻乙耽,马振峰,范广洲.华西秋雨气候特征分析[J].成都信息工程大学学报,2018,(02):164.[doi:10.16836/j.cnki.jcuit.2018.02.011]
YU Yi-dan,MA Zhen-feng,FAN Guang-zhou.The Analysis of Climatic Feature of Autumn Rainfall in West China[J].Journal of Chengdu University of Information Technology,2018,(01):164.[doi:10.16836/j.cnki.jcuit.2018.02.011]
[7]孙康慧,巩远发.20世纪70年代末云南省雨季降水的突变及原因分析[J].成都信息工程大学学报,2018,(02):177.[doi:10.16836/j.cnki.jcuit.2018.02.012]
SUN Kang-hui,GONG Yuan-fa.Abrupt Change of Precipitation in Rainy Season in YunnanProvince in Late 1970s and its Cause Analysis[J].Journal of Chengdu University of Information Technology,2018,(01):177.[doi:10.16836/j.cnki.jcuit.2018.02.012]
[8]吴树炎,顾建峰,刘海文,等.高原冬季雪深与重庆夏季降水的年际关系研究[J].成都信息工程大学学报,2018,(02):184.[doi:10.16836/j.cnki.jcuit.2018.02.013]
WU Shu-yan,GU Jian-feng,LIU Hai-wen,et al.Interannual Relationship between Winter Snow Depth over TibetanPlateau and Summer Precipitation over Chongqing[J].Journal of Chengdu University of Information Technology,2018,(01):184.[doi:10.16836/j.cnki.jcuit.2018.02.013]
[9]魏 凡,李 超.利用气象雷达信息划设雷暴飞行限制区的方法研究[J].成都信息工程大学学报,2018,(02):205.[doi:10.16836/j.cnki.jcuit.2018.02.016]
WEI Fan,LI Chao.Study on the Method of Setting Up Limited Flying area ofThunderstorm by Using Weather Radar Information[J].Journal of Chengdu University of Information Technology,2018,(01):205.[doi:10.16836/j.cnki.jcuit.2018.02.016]
[10]朱 莉,张腾飞,李华宏,等.云南一次短时强降水过程的中尺度特征及成因分析[J].成都信息工程大学学报,2018,(03):335.[doi:10.16836/j.cnki.jcuit.2018.03.017]
ZHU Li,ZHANG Teng-fei,LI Hua-hong,et al.Analysis on Meso-scale Features and Forming Reasons of a Short TimeIntensive Precipitation Case in Yunnan Province[J].Journal of Chengdu University of Information Technology,2018,(01):335.[doi:10.16836/j.cnki.jcuit.2018.03.017]
备注/Memo
收稿日期:2021-01-10
基金项目:国家自然科学基金资助项目(41875039)