非线性逐步回归在宜昌站年径流量预报中的应用  

Application of nonlinear stepwise regression in prediction of annual discharge at Yichang station

作　　者：胡己坤[1,2] 曹丽青[1,2] 葛朝霞[1,2] 郑苏娟[3] 黄美花[1,2] 

机构地区：[1]河海大学水文水资源与水利工程科学国家重点实验室,江苏南京210098 [2]河海大学水文水资源学院,江苏南京210098 [3]河海大学理学院,江苏南京210098

出　　处：《河海大学学报（自然科学版）》2011年第1期1-4,共4页Journal of Hohai University(Natural Sciences)

基　　金：河海大学自然科学基金(2009426411)

摘　　要：为了分析非线性关系的引入对中长期水文预报精度的影响,以对数变换后得到的非线性逐步回归模型为基础,利用长江宜昌站年径流量进行中长期水文预报,并与传统的逐步回归模型预报值进行比较.结果表明:宜昌站年径流量与该站上一年汛期7月、8月、10月流量存在明显非线性关系,并对太平洋副高活动以及厄尔尼诺效应反应敏感;非线性逐步回归模型拟合和试预报准确度均高于线性模型;引入非线性关系能在一定程度上提高模型的预报精度,具有实用价值.In order to analyze the influences of introducing nonlinear relationship on the precision of middle-and long-term hydrological prediction,the mean annual discharge at Yichang station on Changjiang River was employed for the hydrological prediction based on the nonlinear stepwise regression model through log-transformation.The predicted values were compared with those by the traditional stepwise regression models.The results show that the annual discharge at Yichang station has a remarkable nonlinear relationship with its discharge in July,August and October in pervious year,and it is sensitive to the activity of Pacific Subtropical High and El Nino Effect.The precision of the fitting values of the nonlinear stepwise regression model and the predicted values is higher than that of the linear models.Introducing the nonlinear relationship may improve the model precision to a certain extent and is applicable.

关 键 词：非线性 逐步回归 中长期水文预报 宜昌站 

分 类 号：P338.2[天文地球—水文科学]