基于主成分分析法的嵌套流域水文相似性研究

Research on Hydrological Similarity among Nested Basins Based on Principal Component Analysis

以全国不同地区的4个典型嵌套流域为研究对象,利用主成分分析法将流域地形地貌、土壤植被等32个特征指标降维后进行水文相似性分析,发现下会流域子流域的水文特征基本相似,其他3个流域相似度都较低。为验证基于主成分的水文相似性分析结果,建立各子流域的降雨径流相关图,并分别将每个嵌套流域下子流域的相关图综合到一张图上,计算综合前后径流深合格率,仅有下会流域的合格率变幅在可接受范围内。结果表明,基于主成分分析嵌套流域的水文相似性是合理的,嵌套流域内水文相似度较高的子流域可用同一条降雨径流关系线来表示。

The Principal Component Analysis is used to reduce the dimensions of 32 characteristic indexes related to topography and geomorphology as well as soil and vegetation in order to analyze the hydrological similarity of four typical nested basins nationwide. The basic similarity is shown among sub-basins of Xiahui nested basin while low similarity appeared in other three nested basins. To verify the results of Principal Component Analysis, the rainfall-runoff correlation diagram of sub-basins are developed and then comprehensively assessed for each nested basin. Comparing the qualified rate of runoff depth of each sub- basin with the comprehensive assessment one, only the variation in Xiahui basin is acceptable. It is concluded that the analysis of hydrological similarity among nested basins based on principal component is reasonable and a common rainfall-runoff relationship can describe the sub-basins with high hydrological similarity in one nested basin.