基于LSTM网络的海水温度剖面预报研究-海洋预报

基于LSTM网络的海水温度剖面预报研究

单位：1. 海军潜艇学院, 山东青岛 266199;
2. 青岛海洋科学与技术试点国家实验室, 山东青岛 266000

分类号：P731.31

出版年·卷·期（页码）：2024·41·第三期（33-43）

摘要：

基于海水温度历史观测和海洋模式数值预报数据，利用长短期记忆（LSTM）神经网络开展了海水温度剖面短时预报方法研究。以南海17°46.91'N，112°03.24'E处海水温剖面预报为研究对象，利用模式预报数据和观测数据，构建了观测、预报、观测和预报混合 3个样本数据集。基于LSTM神经网络模型，建立了由编码器-解码器组成的多对多海水温度剖面序列预报模型，并开展了模型的训练和性能验证分析工作。结果表明：该模型具有较高的预报精度，对小样本问题的处理具有良好的稳定性；采用深度分层预报，可有效提高模型预报精度，改善模型泛化能力；与只使用观测数据集的预报相比，混合样本集预报误差下降明显，为基于小样本观测数据的海洋环境要素预报研究提供了一种思路。

Based on ocean temperature historical observations and ocean model data, the short-term forecasting method for ocean temperature profile is studied using Long Short -Term Memory (LSTM) neural network. For the ocean temperature profile forecasts at (17°46.91'N, 112°03.24'E) in the South China Sea, three sample sets of observation, forecasts, mixed observation-forecasts are constructed using the observational and forecasting data. Based on the LSTM neural network, a many-to-many ocean temperature profile forecasting model composed of encoder and decoder is established, and model training and verification are carried out. The results show that the model has a high forecasting accuracy and a good stability for processing small sample problems. Utilizing deep layered forecasting method can effectively improve the forecasting accuracy and generalization ability. The forecasting error of mixed sample set decreases significantly in comparison with that of observation sample set, which provides an idea for marine environment forecasts of small sample problems.

参考文献：

[1] 王辉,万莉颖,秦英豪,等.中国全球业务化海洋学预报系统的发展和应用[J].地球科学进展, 2016, 31(10):1090-1104. WANG H, WAN L Y, QIN Y H, et al. Development and application of the Chinese global operational oceanography forecasting system[J]. Advances in Earth Science, 2016, 31(10):1090-1104.
[2] 刘娜,王辉,凌铁军,等.全球业务化海洋预报进展与展望[J].地球科学进展, 2018, 33(2):131-140. LIU N, WANG H, LING T J, et al. Review and prospect of global operational ocean forecasting[J]. Advances in Earth Science, 2018, 33(2):131-140.
[3] 张志远,孙立尹,何锡玉,等.海洋环境数值预报业务综合运维平台设计与实现[J].海洋预报, 2019, 36(3):63-70. ZHANG Z Y, SUN L Y, HE X Y, et al. Design and implementation of marine environment numerical forecast integrated operation and maintenance platform[J]. Marine Forecasts, 2019, 36(3):63-70.
[4] 马继瑞,韩桂军,李威,等.海洋三维温盐流数值模拟研究的有关进展和问题[J].海洋学报, 2014, 36(1):1-6. MA J R, HAN G J, LI W, et al. The relevant progress and issues in numerical simulations of the oceanic 3-D temperature, salinity and current[J]. Acta Oceanologica Sinica, 2014, 36(1):1-6.
[5] 宋振亚,刘卫国,刘鑫,等.海量数据驱动下的高分辨率海洋数值模式发展与展望[J].海洋科学进展, 2019, 37(2):161-170. SONG Z Y, LIU W G, LIU X, et al. Research progress and perspective of the key technologies for ocean numerical model driven by the mass data[J]. Advances in Marine Science, 2019, 37(2):161-170.
[6] 刘文如.零基础入门Python深度学习[M].北京:机械工业出版社, 2020. LIU W R. Zero basic introduction to Python deep learning[M]. Beijing:China Machine Press, 2020.
[7] GÉRON A.机器学习实战-基于Scikit-Learn、Keras和Tensorflow[M].宋能辉,李娴,译.北京:机械工业出版社, 2020. GÉRON A. Hands-on machine learning with Scikit-Learn, Keras&TensorFlow[M]. SONG N H, LI X, trans. Beijing:China Machine Press, 2020.
[8] 倪铮,梁萍.基于LSTM深度神经网络的精细化气温预报初探[J].计算机应用与软件, 2018, 35(11):233-236. NI Z, LIANG P. Fine temperature forecast based on LSTM deep neural network[J]. Computer Applications and Software, 2018, 35(11):233-236.
[9] 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(3):435-446. ZHI X F, WANG T, JI Y. Multimodel ensemble forecasts of surface air temperature over China based on deep learning approach[J]. Transactions of Atmospheric Sciences, 2020, 43(3):435-446.
[10] 马景奕,刘维成,闫文君.基于深度学习的气象要素预测方法[J].热带气象学报, 2021, 37(2):186-193. MA J Y, LIU W C, YAN W J. Meteorological elements forecasting method based on deep learning[J]. Journal of Tropical Meteorology, 2021, 37(2):186-193.
[11] 唐旺,马尚昌,李程.基于滑动窗口的LSTM地温预测方法[J].成都理工大学学报(自然科学版), 2021, 48(3):377-384. TANG W, MA S C, LI C. LSTM ground temperature prediction method based on sliding window[J]. Journal of Chengdu University of Technology (Science&Technology Edition), 2021, 48(3):377-384.
[12] 沈皓俊,罗勇,赵宗慈,等.基于LSTM网络的中国夏季降水预测研究[J].气候变化研究进展, 2020, 16(3):263-275. SHEN H J, LUO Y, ZHAO Z C, et al. Prediction of summer precipitation in China based on LSTM network[J]. Climate Change Research, 2020, 16(3):263-275.
[13] 王国松,王喜冬,侯敏,等.基于观测和再分析数据的LSTM深度神经网络沿海风速预报应用研究[J].海洋学报, 2020, 42(1):67-77. WANG G S, WANG X D, HOU M, et al. Research on application of LSTM deep neural network on historical observation data and reanalysis data for sea surface wind speed forecasting[J]. Acta Oceanologica Sinica, 2020, 42(1):67-77.
[14] 徐凌宇,张高唯,江湾湾,等.深度学习神经网络及其在海洋环境信息挖掘预测中的应用[J].海洋信息, 2018, 33(1):17-23. XU L Y, ZHANG G W, JIANG W W, et al. Deep learning neural network and its application in marine environmental information mining prediction[J]. Marine Information, 2018, 33(1):17-23.
[15] 查铖,贺琪,宋巍,等.结合注意力机制的区域型海表面温度预报算法[J].海洋通报, 2020, 39(2):191-199. ZHA C, HE Q, SONG W, et al. Regional sea surface temperature prediction algorithm combined with attention mechanism[J]. Marine Science Bulletin, 2020, 39(2):191-199.
[16] 焦艳,黄菲,高松,等.基于长短时记忆神经网络的辽东湾海冰延伸期预报方法研究[J].中国海洋大学学报, 2020, 50(6):1-11. JIAO Y, HUANG F, GAO S, et al. Research on extended-range forecast model of sea ice in the Liaodong bay based on long short term memory network[J]. Periodical of Ocean University of China, 2020, 50(6):1-11.
[17] 林琪凡,耿旭朴,谢婷,等.基于长短期记忆神经网络的西太平洋暖池变化预测[J].厦门大学学报(自然科学版), 2021, 60(5):927-936. LIN Q F, GENG X P, XIE T, et al. Trend prediction of Western Pacific warm pool based on long short-term memory neural networks[J]. Journal of Xiamen University (Natural Science), 2021, 60(5):927-936.
[18] 笪良龙.海洋水声环境效应建模与应用[M].北京:科学出版社, 2012. DA L L. Modeling and application of underwater acoustic environmental effect[M]. Beijing:Science Press, 2012.
[19] LIU H L, LIN P F, ZHENG W P, et al. A global eddy-resolving ocean forecast system in China-LICOM Forecast System (LFS)[J]. Journal of Operational Oceanography, 2023, 16(1):15-27.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】