Characterization and Classification of River Network Types

In nature, rivers are always connected in various forms to constitute a specific type of river network. The identification and classification of river network types in watersheds is the premise of hydrological research. In this study, the Yellow River Basin, Huaihe River Basin, Haihe River Basin and Yangtze River Basin are divided into 71 sub-basins. According to the definition of river network types, the sub-basin river networks are qualitatively divided into 7 types. By comparing and analysing three river network characteristic parameters, which are river network density, river flow direction and river sinuosity, this study found that the types of river networks can be preliminarily determined according to the statistical data distribution of river sinuosity. The Cauchy distribution is used to fit the distribution characteristics of river sinuosity to further accurately determine the types of river networks. Except for the average R2 of the rectangular river network, which is 0.66, the R2 values of the fitting curves of the other river network types all range from 0.86 to 0.97. This method is applied to the four major watersheds, and the results are consistent with the hierarchical clustering analysis, with an accuracy of 82.86%. The method proposed in this study has application potential and can be applied to the automatic classification of river network types with high accuracy and efficiency.