Numerical Simulation of Landslide Dam Formation and Failure Process

Abstract

Landslide dams are typically formed by loose landslide deposits which can be vulnerable to overtopping failure. Subsequent flash flood can pose great threat to downstream area. Hong Kong is subject to subtropical weather and the terrain is hilly. Due to the changing climate, extreme rainfall is becoming more and more frequent which can cause a large number of landslides. Therefore, it is essential to evaluate possible landslide dam formation and failure process, especially on Hong Kong Island which is densely populated. In this paper, a physically-based distributed cell model, EDDA 2.0, is adopted which is capable to predict rainfall-induced landslides, debris flows, erosion and deposition process, and flash floods. Possible landslide dam formation and failure are predicted under an extreme rainfall condition (65% of the 24-h PMP, Probable Maximum Precipitation). Although the magnitude of the landslide and the landslide dam is not very large, the subsequent flash flood can pose much greater threat to people’s lives and properties especially in urban area such as Hong Kong Island where the population density is rather high.