中文摘要：

利用潇河流域51 a的径流量资料,选择每月之前的12个月径流量作为预报因子,采用逐步回归分析的方法,对枯水期径流量进行了分析研究。结果表明,可以利用月径流量之间的关系建立枯水期月径流量预测模型,所建立的预测模型包含的自变量不超过5个,每个预测模型中均包含前一个月的径流量,均没有常数项,可用于实际作业预报;模型的预测精度随月径流量的增加有降低的趋势,其中12月份到翌年4月份的预测模型精度可以达到甲等,10月份、11月份和5月份可以达到乙等。

英文摘要：

By use of the 51 years＇ runoff data of Xiao river basin,selecting the previous 12 months runoff of every runoff as the predictors,and by the stepwise regression analysis method,the runoff of the dry season is studied.The results show that the monthly runoff prediction model on the dry season can be established by means of the relationship between the monthly runoffs,and the independent variables of this prediction model are not more than 5.Each prediction model includes the previous monthly runoff,while there is no constant,and the model can be used in the practical operation forecast.The forecast accuracy of the model decreased with the monthly runoff increase,and reached class A from December to April in the following year,and reached class B in October,November and May.