基于机器学习的海表温度对中国降水的预测研究

作者：李良伟1 2  邹斌1 3 4  石立坚1  刘鹏1 2

单位：1. 国家卫星海洋应用中心, 北京 100081; 2. 国家海洋环境预报中心, 北京 100081; 3. 自然资源部空间海洋遥感与应用重点实验室, 北京 100081; 4. 南方海洋科学与工程广东省实验室, 广东 广州 511458

关键词：K-means 聚焦时延神经网络 主成分分析 降水 海表温度

分类号：P731.11;P426.6

出版年·卷·期（页码）：2021·38·第三期（29-37）

摘要：

利用K-means聚类算法将36 a的中国周降水数据聚类为7个区域，其中非季风区的中国西北和青藏高原的大部分地区聚类为两个区域，其余区域聚类为5个区域。将各区域的降水量数据经过指数平滑之后，输入聚焦时延神经网络（FTDNN），求解其与经过主成分分析降维处理的SST之间的关系。结果表明：季风区的降水量最佳延迟时间比远离海洋的非季风区域的最佳延迟时间大，同时也得到了预测各区域降水的最佳延迟时间。在对各区域的降水量进行短期和中长期的预测中得出，FTDNN神经网络在利用降维后的SST预测降水量上显示出很好的预测效果，尤其是中长期趋势的预测，与传统的小波变换重建原序列以及均生函数方法的预测结果相比，相关系数和均方根误差都有了很大的提高，新疆地区测试结果的相关系数比小波变换结果提高了0.05。

In this paper, 36 years of weekly precipitation data in China are clustered into 7 regions by using K-means clustering algorithm. Most areas of the northwest China and Qinghai-Tibet Plateau, which belong to non-monsoon areas, are clustered into two regions, and the rest areas of China are clustered into five regions. The precipitation data of each region are input into Focused Time-Delay Neural Network (FTDNN) after exponential smoothing, and the relationship between FTDNN and SST reduced by Principal Component Analysis (PCA) is solved. At the same time, the optimal delay time for predicting precipitation in each region is also obtained. The results show that the FTDNN neural network has a good effect on the prediction of precipitation, especially for its medium and long term trend, compared with the traditional wavelet transform to reconstruct the original series and the mean generating function method. The correlation coefficient and root mean square error of the prediction results in this paper is greatly improved. The correlation coefficient of the test results in Xinjiang increases by 0.05 compared with the results of wavelet transform.

参考文献：

[1] Zahraie B, Roozbahani A. SST clustering for winter precipitation prediction in southeast of Iran:Comparison between modified K-means and genetic algorithm-based clustering methods[J]. Expert Systems with Applications, 2011, 38(5):5919-5929. [2] Cordery I, McCall M. A model for forecasting drought from teleconnections[J]. Water Resources Research, 2000, 36(3):763-768. [3] Higgins R W, Kim H K, Unger D. Long-lead seasonal temperature and precipitation prediction using Tropical Pacific SST consolidation forecasts[J]. Journal of Climate, 2004, 17(17):3398-3414. [4] Krishnamurthy V, Goswami B N. Indian Monsoon-ENSO relationship on interdecadal timescale[J]. Journal of Climate, 2000, 13(3):579-595. [5] Meehl G A, Hu A X. Megadroughts in the Indian monsoon region and Southwest North America and a mechanism for associated multidecadal Pacific Sea surface temperature anomalies[J]. Journal of Climate, 2006, 19(9):1605-1623. [6] 徐建军, 朱乾根. ENSO及其年代际异常对全球及亚洲季风降水影响的数值研究[J]. 气象学报, 1999, 57(3):301-315. [7] 陶诗言, 张庆云. 亚洲冬夏季风对ENSO事件的响应[J]. 大气科学, 1998, 22(4):399-407. [8] 王绍武, 赵宗慈. 我国旱涝36年周期及其产生的机制[J]. 气象学报, 1979, 37(1):64-73. [9] 于淑秋, 林学椿. 北太平洋海温的气候跃度及其对中国汛期降水的影响[J]. 热带气象学报, 1997, 13(3):265-275. [10] 朱益民, 杨修群. 太平洋年代际振荡与中国气候变率的联系[J]. 气象学报, 2003, 61(6):641-654. [11] 孙柏民, 孙淑清. 海温在东亚冬季风影响江淮流域旱涝中的作用[C]//"八五"灾害性气候的过程及诊断. 北京:气象出版社, 1996:46-53. [12] 杨一鸣, 潘嵘, 潘嘉林, 等. 时间序列分类问题的算法比较[J]. 计算机学报, 2007, 30(8):1259-1266. [13] 周爱武, 于亚飞. K-Means聚类算法的研究[J]. 计算机技术与发展, 2011, 21(2):62-65. [14] Rousseeuw P J. Silhouettes:a graphical aid to the interpretation and validation of cluster analysis[J]. Journal of Computational and Applied Mathematics, 1987, 20:53-65. [15] 保丽红. 主成分分析与线性判别分析降维比较[J]. 统计学与应用, 2020, 9(1):47-52. [16] Messié M, Chavez F. Global modes of sea surface temperature variability in relation to regional climate indices[J]. Journal of Climate, 2011, 24(16):4314-4331. [17] Waibel A, Hanazawa T, Hinton G, et al. Phoneme recognition using time-delay neural networks[J]. IEEE Transactions on Acoustics, Speech, and Signal Processing, 1989, 37(3):328-339. [18] 石拓. 基于数据挖掘技术的短期风电预测[J]. 吉林省教育学院学报, 2015, 31(5):79-80. [19] Weigend, A. S., and Gershenfeld N A. Time Series Prediction:Forecasting The Future And Understanding The Past[M]. Reading, MA:Addison-Wesley, 1994. [20] 陈咸吉. 中国气候区划新探[J]. 气象学报, 1982, 40(1):35-48. [21] Eilers P H C. A perfect smoother[J]. Analytical Chemistry, 2003, 75(14):3631-3636. [22] 张存杰, 李栋梁, 王小平. 东北亚近100年降水变化及未来10~15年预测研究[J]. 高原气象, 2004, 23(6):919-929.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】