Viscoplastic modelling of rainfall-driven slow-moving landslides: application to California Coast Ranges

Slow-moving landslides are widely observed in mountainous areas worldwide. While most of these landslides move slowly downslope over long periods of time, some ultimately accelerate rapidly and fail catastrophically. Simulating the landslide creep movement triggered by environmental factors such as precipitation, is therefore necessary to anticipate potential damaging effects on proximal infrastructure, habitat, and life. Here, we present a physically-based model that links pore-water pressure changes in the landslide mass with a new viscoplastic constitutive law designed to capture different temporal trends in slow-moving landslides. The model accounts for landslide velocity changes caused by rainfall infiltration through the Terzaghi's effective stress principle, thus directly resolving the deformation of the active shear zone. Calibration and validation of the computations benefited from both ground-based and remote sensing data for three active landslides in the California Coast Ranges, USA. We find that our model can accurately describe both slow quasi-continuous and episodic movement commonly displayed by active landslides. Although inherent limitations of the viscoplasticity framework did not enable us to describe catastrophic landslide acceleration, our model provides versatile tools that can be used to analyze and describe distinct types of slow-moving landslide dynamics.

Automated summary

Slow-moving landslides are common in mountainous areas worldwide. Understanding the factors that trigger their movement is important for assessing potential risks to infrastructure and life. In this study, a physically-based model was developed to simulate the deformation of slow-moving landslides caused by changes in pore-water pressure. The model accurately described the slow movement of active landslides, but was unable to capture catastrophic acceleration.