Development of an empirical model for predicting peak breach flow of landslide dams considering material composition

The existing empirical models do not consider the influence of material composition of landslide deposits on the peak breach flow due to the uncertainty in the material composition and the randomness of its distribution. In this study, based on the statistical analyses and case comparison, the factors influencing the peak breach flow were comprehensively investigated. The highlight is the material composition-based classification of landslide deposits of 86 landslide cases with detailed grain-size distribution information. In order to consider the geometric morphology of landslide dams and the potential energy of dammed lakes, as well as the material composition of landslide deposits in an empirical model, a multiple regression method was applied on a database, which comprises of 44 documented landslide dam breach cases. A new empirical model for predicting the peak breach flow of landslide dams was developed. Furthermore, for the same 44 documented landslide dam failures, the predicted peak breach flow obtained by using the existing empirical models for embankment and landslide dams and that obtained by using the newly developed model were compared. The comparison of the root mean square error ( E rms ) and the multiple coefficient of determination ( R 2 ) for each empirical model verifies the accuracy and rationality of the new empirical model. Furthermore, for fair validation, several landslide dam breach cases that occurred in recent years in China and have reliable measured data were also used in another comparison. The results show that the new empirical model can reasonably predict the peak breach flow, and exhibits the best performance among all the existing empirical models for embankment and landslide dam breaching.