Emergence of diverse failure patterns in weathering-induced landslides: Insights from high-fidelity particle finite element simulations

Landslides, as a ubiquitous type of mass wasting phenomena, occur under various geological and environmental conditions and exhibit diverse failure patterns. Among various factors, weathering has been widely recognised as one of the primary drivers on landslide evolution over geological timescales. However, how weathering induces slope instabilities, including how pre-failure rock mass degradation contributes to the landslide failure development and post-failure deposition of mobilised geomaterials, has not been fully understood. In this study, we develop a novel, physics-based unified computational framework to capture weathering-induced landslide evolution over multiple time scales, from the long-term pre-failure rock mass deformation to the short-term slope rupture and post-failure runout dynamics. Weathering laws and failure criteria are coupled to capture the combined effects of time-dependent strength degradation and strain-driven damage processes, while a frictional velocity-weakening law is adopted to characterise the rapid movement of yielded masses. The non-linear governing equations of landslide dynamics are solved using an implicit particle finite element framework that can model all the landslide stages from the long-term material degradation to short-term failure and runout. We further investigate the effects of weathering conditions (type and rate), geological properties (fracture sets and rock matrix) and slope geometry (angle and shape) on the failure patterns. Our high-fidelity numerical simulations capture the emergence of diverse landslide failure patterns resulting from the complex interplay among rock lithology, fracture distribution, weathering process, and gravitational forcing. Our numerical results show that matrix-dominated weathering tends to produce shallow landslides, while fracture-dominated weathering promotes the occurrence of deep landslides. For fracture-dominated weathering, the orientation of pre-existing fractures and the slope ratio significantly control the failure mode (e.g. falling, toppling, sliding, etc.), which further affects post-failure runout behaviour. Our computational framework opens the door to investigating and understanding weathering-induced rock slope failure evolution across spatial and temporal scales.