Depth-Integrated Two-Phase Modeling of Two Real Cases: A Comparison between r.avaflow and GeoFlow-SPH Codes

Abstract

Due to the growing populations in areas at high risk of natural disasters, hazard and risk assessments of landslides have attracted significant attention from researchers worldwide. In order to assess potential risks and design possible countermeasures, it is necessary to have a better understanding of this phenomenon and its mechanism. As a result, the prediction of landslide evolution using continuum dynamic modeling implemented in advanced simulation tools is becoming more important. We analyzed a depth-integrated, two-phase model implemented in two different sets of code to stimulate rapid landslides, such as debris flows and rock avalanches. The first set of code, r.avaflow, represents a GIS-based computational framework and employs the NOC-TVD numerical scheme. The second set of code, GeoFlow-SPH, is based on the mesh-free numerical method of smoothed particle hydrodynamics (SPH) with the capability of describing pore pressure’s evolution along the vertical distribution of flowing mass. Two real cases of an Acheron rock avalanche and Sham Tseng San Tsuen debris flow were used with the best fit values of geotechnical parameters obtained in the prior modeling to investigate the capabilities of the sets of code. Comparison of the results evidenced that both sets of code were capable of properly reproducing the run-out distance, deposition thickness, and deposition shape in the benchmark exercises. However, the values of maximum propagation velocities and thickness were considerably different, suggesting that using more than one set of simulation code allows us to predict more accurately the possible scenarios and design more effective countermeasures.