Cluster analysis for factor classification and riverbank collapse along the desert wide valley reach of the upper Yellow River

An index system was developed using both natural factors and human factors to study riverbank collapse along the desert wide valley reach (Ning-Meng Section) of the upper Yellow River. Remote sensing and field investigation data sets are combined with hydrological, topographic and meteorological data from 2006 to 2010 to identify the factors that influence river bank collapse along the channel with a regression analysis to identify the key factors. The bank collapses along the monitored reaches are classified using a clustering analysis module. The results indicate that the hydrodynamic conditions and slope of river bank are the key factors affecting bank collapse. The clustering results for bank collapse are consistent with the results of the field investigation, so they provide a reliable evaluation method for predicting riverbank collapse along the Ning-Meng Section of the Yellow River. ©, 2014, Press of Tsinghua University. All right reserved.