气温和降水的均生外延主分量预测模型

PRINCIPAL COMPONENT FORECAST MODEL OF TEMPERATURE AND PRECIPITATION BASED ON MEAN GENERATING FUNCTION IN PERIOD EXTRAPOLATION

本文采用了对窗区滑动的气温和降水的时间序列,构造均生函数外延矩阵。对外延矩阵进行主分量分析,建立了以主分量作自变量的气温或降水的长期预测模型。实例分析表明,该模型应用于气温长期预测精度较高,用于降水趋势预测亦是可行的。

In this study, for ne time series of temperature and precipitation, a new procedure to build a period extrapolation based on the ihean generating function is suggested. A long-range forecast model of temperature or precipitation based on the principal component analysis to the period extrapolation matrix is presented. The forecast experiments show that the result of temperature forecast is satisfactory and that of precipitation forecast trend is suitable.