径流的形成是受多种因素影响的,其中有些因素是已知的,另外也有不少因素是未知的.通常采用的长期预报方法都是基于已知的信息,通过以下两类统计方法建立预报模型:一类是分析径流的前期影响因素,运用多元回归、聚类分析、主分量分析等方法达到客观分型,浓缩信息的目的建立预报模型;另一类是直接根据预报对象实测信息,通过时间序列分析,把水文序列看作是一个离散化的过程,建立自回归模型进行外推,以达到预报的目的.无论怎样,上述两类方 The formation of runoff is affected by a number of factors, some of which are known, and many of which are unknown. The long-term forecasting methods commonly used are based on known information and are based on the following two types of statistical methods Model: The first category is the analysis of runoff early factors, the use of multiple regression, cluster analysis, principal component analysis and other methods to achieve objective classification, the purpose of enrichment information to establish a forecasting model; the other is based directly on the forecast object of the measured information through Time Series Analysis, the hydrological sequence as a discrete process, the establishment of the autoregressive model extrapolation, in order to achieve the purpose of forecasting. In any case, the two types of parties