动态回归方程在衡水汛期降水预测中的应用

以衡水市三个站点（桃城区、饶阳、故城）6—8月份平均降水为因变量，以三站点的气候因子的平均值为自变量，利用动态回归方程对2001—2006年6—8月份三站平均降水量进行预报检验，其预报效果好于桃城区单站所得的预报结果。如果将因变量改为三站汛期总降水的平方根，自变量不变，则预报效果又好于三站汛期平均降水。结果表明：利用统计方法对区域性降水的预报可行，三站降水状况较单站降水更能代表该区域的气候状况。

Application of Dynamic Regression Function to Rainy Precipitation Forecast in HengShui

Average precipitation of three stations （ TaoChengQu, RaoYang, GuCheng） in June to August being as independent and average value of three stations climate factors being as dependent , using dynamic regression functions to test precipitations of three stations in rainy season .in 2001 to 2006 year,the forecast effect is better than TaoChengQu precipitations. If independent is square root of three stations total precipitations in rainy season and dependent is the same, its forecast effect is better than average precipitation of three stations. It show that regional precipitation forecast effect is affirmable, at the same time , it means that three stations precipitation much more respects this area climate than single station .