摘要
海温是认知和研究大气、海洋物理性质和演变规律的重要海洋环境动力要素之一,也是海洋预报组成中不可或缺的组成部分,对海温的分布及演变规律进行研究对海洋经济、环保和国防安全具有重要的意义。本文提出了一种基于相似系数的海温统计预报方法,利用数理统计分析方法,构建基于相似系数的统计预报模型来实现海温的单点时间序列预报。在南海海域海温预报中长期实验中,该方法比气候态预报和ARIMA预报方法的结果更优,证明该方法的有效性并为海温的中长期预报提供了新思路。
Sea Surface Temperature(SST)is one of the important marine environmental dynamics elements in the cognition and study of the physical properties and evolution of the atmosphere and ocean,and is also an indispensable component in the composition of marine forecasting.The study of the distribution and evolution of SST is of great significance to the marine economy,environmental protection and national defense security.In this study,a statistical forecasting method based on the similarity coefficient was proposed,and a statistical forecasting model based on the similarity coefficient was constructed to realize the single-point time series forecasting of SST.In the medium and long-term experiments of SST forecasting in the South China Sea,this method has better results than those from Optimal Climate Normals forecasting and ARIMA forecasting methods,which proves the effectiveness of this method and provides a new idea for the medium and long-term forecasting of SST.
作者
李科
苑福利
刘厂
LI Ke;YUAN Fuli;LIU Chang(Naval Research Institute of PLA,Tianjin 300061,China;College of Intelligent SystemsScience and Engineering,Harbin Engineering University,Harbin 150001,China)
出处
《海洋学研究》
CSCD
北大核心
2021年第1期67-78,共12页
Journal of Marine Sciences
关键词
时间序列预报
相似系数分析
海温中长期预报
预报误差订正
time series forecasting
similarity coefficient analysis
medium and long-term SST forecasts
forecast error correction
作者简介
李科(1980-),男,湖南省长沙市人,副研究员,主要从事海洋测绘和海洋环境保障论证研究。E-mail:Like235@sina.com。
