基于LSTM神经网络方法的粤东近岸海温预报-海洋预报

基于LSTM神经网络方法的粤东近岸海温预报

单位：1. 国家海洋局汕尾海洋环境监测中心站, 广东汕尾 516600;
2. 国家海洋环境预报中心, 北京 100081;
3. 热带海洋环境国家重点实验室 (中国科学院南海海洋研究所), 广东广州 510301

分类号：P731.31

出版年·卷·期（页码）：2022·39·第五期（27-36）

摘要：

利用粤东沿岸各海洋站点的海温观测数据以及区域大气模式的气象预报数据，基于长短期记忆的神经网络方法，通过分析筛选训练数据，建立了高效的粤东近岸24 h海温预报方法。与观测资料进行对比，该方法在粤东近岸海域24 h海温预报的均方根误差和平均绝对误差分别为0.45℃和0.32℃，在珠江口沿岸的误差更小。进一步分析表明，气象要素与海温变化值的相关性比与高通滤波后的海温相关性更好。对海温变化值进行预报，而后叠加海温初始值，可以得到更加准确且稳定的预报结果。

The sea temperature observation data of marine stations along the coast of eastern Guangdong and meteorological forecast data of a regional atmospheric model is analyzed, selected and trained based on the Long Short-Term Memory neural network, and an efficient 24-hour nearshore sea temperature forecasting method along the coast of eastern Guangdong is established in this paper. Compared with observation, the root mean square error and the mean absolute error of 24-hour sea temperature forecast in the coastal waters along eastern Guangdong are 0.45℃ and 0.33℃, respectively, and the error is even smaller in the Pearl River Estuary. Further analysis shows that the correlation between meteorological factors and the variation of sea temperature is more significant than that with the sea temperature after high pass filtering. Therefore, more accurate and stable forecasting results could be achieved when the sea temperature variation is forecasted and superimposed with its initial value thereafter.

参考文献：

[1] 张建华.海温预报知识讲座:第一讲海水温度预报概况[J].海洋预报, 2003, 20(4):81-85. ZHANG J H. Lecture on sea surface temperature prediction:lecture 1 overview of sea surface temperature prediction[J]. Marine Forecasts, 2003, 20(4):81-85.
[2] 王兆毅, 李云, 王旭.中国近岸海域基础预报单元海温预报指导产品研制[J].海洋预报, 2020, 37(4):59-65. WANG Z Y, LI Y, WANG X. Development of forecast guidance product for sea temperature of basic forecast units in the Chinese coastal waters[J]. Marine Forecasts, 2020, 37(4):59-65.
[3] 李云, 刘钦政, 张建华, 等.最优插值方法在西北太平洋海温同化中的应用研究[J].海洋预报, 2008, 25(2):25-32. LI Y, LIU Q Z, ZHANG J H, et al. Optimal interpolation and its application to assimilation of sea temperature in Northwest Pacific Region[J]. Marine Forecasts, 2008, 25(2):25-32.
[4] 张权, 王燕, 王关锁, 等.高分辨率浪-潮-流耦合数值预报系统的业务化运行与检验分析[J].海洋科学进展, 2011, 29(2):129-136. ZHANG Q, WANG Y, WANG G S, et al. Operation and verification for a wave-tide-current coupling numerical-forecast system with fine resolution[J]. Advances in Marine Science, 2011, 29(2):129-136.
[5] 吴玲娟, 高松, 徐江玲.黄渤海近岸精细化三维温盐流业务化数值预报系统研发[J].防灾科技学院学报, 2015, 17(4):83-91. WU L J, GAO S, XU J L. Development of Huanghai and Bohai coastal sea high resolution and operational numerical temperaturesalinity-current forecasting system[J]. Journal of Institute of Disaster Prevention, 2015, 17(4):83-91.
[6] 张培军, 周水华, 梁昌霞.基于卫星遥感海温数据的南海SST预报误差订正[J].热带海洋学报, 2020, 39(6):57-65. ZHANG P J, ZHOU S H, LIANG C X. Study on the correction of SST prediction in South China Sea using remotely sensed SST[J]. Journal of Tropical Oceanography, 2020, 39(6):57-65.
[7] 李燕, 张建华, 刘钦政, 等.单站海温短期预报自动化[J].海洋预报, 2007, 24(4):33-41. LI Y, ZHANG J H, LIU Q Z, et al. The automation of single sea station's surface sea temperature short term forecasting[J]. Marine Forecasts, 2007, 24(4):33-41.
[8] 张志刚, 蔡东明, 骆永军.周期外推方法在海水水温预报中的应用[J].海洋技术, 2007, 26(4):69-71. ZHANG Z G, CAI D M, LUO Y J. Application of distinct periodicity extrapolation to sea water temperature forecast[J]. Ocean Technology, 2007, 26(4):69-71.
[9] 修树孟, 逄爱梅, 申辉, 等.黄海水温垂直剖面的遥感反演与数值预报模型[J].海洋科学进展, 2009, 27(4):444-451. XIU S M, PANG A M, SHEN H, et al. Remote sensing inversion and numerical prediction model for the temperature vertical profile in the Yellow Sea[J]. Advances in Marine Science, 2009, 27(4):444-451.
[10] 匡晓迪, 王兆毅, 张苗茵, 等.基于BP神经网络方法的近岸数值海温预报释用技术[J].海洋与湖沼, 2016, 47(6):1107-1115. KUANG X D, WANG Z Y, ZHANG M Y, et al. An interpretation scheme of numerical near-shore sea-water temperature forecast based on BPNN[J]. Oceanologia et Limnologia Sinica, 2016, 47(6):1107-1115.
[11] 韩玉康, 余丹丹, 申晓莹, 等. HYCOM模式SST的预报误差订正[J].海洋预报, 2018, 35(3):76-80. HAN Y K, YU D D, SHEN X Y, et al. Study on the correction of SST prediction of HYCOM[J]. Marine Forecasts, 2018, 35(3):76-80.
[12] 齐义泉, 张志旭, 李志伟, 等.人工神经网络在海浪数值预报中的应用[J].水科学进展, 2005, 16(1):32-35. QI Y Q, ZHANG Z X, LI Z W, et al. Application of artificial neural network to numerical wave prediction[J]. Advances in Water Science, 2005, 16(1):32-35.
[13] 王建华, 于红兵, 宋运法.人工神经网络在潮汐数值预报中的应用[J].海洋预报, 2007, 24(2):47-51. WANG J H, YU H B, SONG Y F. Application of artificial neural network to numerical tidal prediction[J]. Marine Forecasts, 2007, 24(2):47-51.
[14] 廖德春, 王琪洁, 周永宏, 等.应用神经网络技术预报海表温度变化异常[J].中国科学院上海天文台年刊, 2008:11-20. LIAO D C, WANG Q J, ZHOU Y H, et al. Prediction of the sea surface temperature anomaly by neural networks[J]. Annals of Shanghai Astronomical Observatory Chinese Academy Sciences, 2008:11-20.
[15] HAQUE A U, MANDAL P, MENG J L, et al. Wind speed forecast model for wind farm based on a hybrid machine learning algorithm[J]. International Journal of Sustainable Energy, 2015, 34(1):38-51.
[16] KUMAR N K, SAVITHA R, AL MAMUN A. Regional ocean wave height prediction using sequential learning neural networks[J]. Ocean Engineering, 2017, 129:605-612.
[17] ZHENG G, LI X F, ZHANG R H, et al. Purely satellite data- driven deep learning forecast of complicated tropical instability waves[J]. Science Advances, 2020, 6(29):eaba1482.
[18] 周水华, 洪晓, 梁昌霞, 等.基于人工神经网络的台风浪高快速计算方法[J].热带海洋学报, 2020, 39(4):25-33. ZHOU S H, HONG X, LIANG C X, et al. A method of tropical cyclone wave height calculation based on Artificial Neural Network[J]. Journal of Tropical Oceanography, 2020, 39(4):25-33.
[19] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780.
[20] 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]. Haiyang Xuebao, 2020, 42(1):67-77.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】