一、引言 表层海水温度预报的途径目前在我国仍以统计预报为主,数值预报和数值动力统计预报尚在探索中。表层海温统计预报一般又分为单要素时间序列预报和多要素相关预报两类。这些方法各有长处和短处,因此在预报实施中,应根据各种方法的特点,互相补充,综合发布预报。我们在应用经验正交函数分析预报表层海水温度场的基础上,也曾对表层海温场的多因子相关预报方法进行了多种试验。结果表明,在单点多因子回归预报中 I. INTRODUCTION The means of forecasting the surface sea temperature is still mainly based on statistical forecast in our country. Numerical forecast and numerical dynamic forecast are still under exploration. Surface temperature and temperature statistical forecasting is generally divided into single factor time series forecasting and multi-factor related forecasting two categories. Each of these methods has its strengths and weaknesses. Therefore, in the implementation of forecasting, the methods should be based on the characteristics of each method to complement each other and make comprehensive forecasts. Based on the empirical orthogonal function analysis and prediction of surface seawater temperature field, we conducted various experiments on the multi-factor correlation forecasting method of surface SST field. The results show that in the single point multi-factor regression prediction