Comparison of two depth-averaged numerical models for debris flow runout estimation

This paper analyses an important aspect of the continuum numerical modelling of rapid landslides as debris flows: “By using the same rheological parameter values, are the results obtained with codes that implement the same constitutive equations, but a different numerical solver, equal?” To answer this question, the two numerical codes RASH3D and GeoFlow_SPH are used here to back-analyse the debris flow event that occurred in the Nora stream (northwestern Italian Alps) in October 2000. Comparison of results evidenced that the RASH3D best-fit rheological values for the Nora event back-analysis overestimated both the final depositional heights and the simulated flow velocities if used in GeoFlow_SPH. To obtain thickness values comparable with those measured in situ, it was necessary to re-calibrate GeoFlow_SPH rheological parameter values. This way, with the exception of a larger lateral spreading of the sliding mass given by RASH3D, both thickness and velocity values were similar for the two numerical codes.