A depth-averaged SPH-FV landslide dynamic model for evaluating hazard zones

Abstract

Thousands of landslides worldwide lead to significant casualties and property damage. A numerical model is crucial for simulating the runout of possible future landslides and generating reliable hazard zone maps. Two main methods can be adopted to simulate motion: one based on the Eulerian description and the other based on the Lagrangian description. Each description offers varying adaptability to simulating landslides. In this paper, we proposed a method of coupling the depth-averaged smoothed particle hydrodynamics (SPH) method and a finite volume scheme with the van Leer splitting (FV-VLS). For the solid phase, we utilized the depth-averaged SPH method to simulate the movement process. The SPH method can track the movement of key parts. In contrast, we employed the FV-VLS based on the Eulerian description to reduce the calculation amount. In 1-D dam break simulation, both the depth-averaged SPH and FV-VLS methods yielded similar results. The 1-D two-phase dam break simulation shows that buoyancy reduces friction and lateral solid stress, increases hydraulic stress, and generates a complex motion process. The Yigong landslide simulation indicates that solid fractions will affect the motion trajectory in the 2-D simulation. A smaller solid fraction will cause the landslide to affect a larger area.