Rainfall infiltration frequently triggers slope failures by elevating pore water pressure and compromising the shear strength of unsaturated soil layers. Modeling these severe geological events remains computationally challenging. Standard grid-based techniques, such as the Finite Element Method (FEM), typically fail due to severe mesh distortion under large deformations, whereas the Discrete Element Method (DEM) demands excessive computational resources. Addressing this gap, we develop an advanced theoretical framework utilizing the Material Point Method (MPM) integrated with a liquid-solid-gas three-phase mechanics model. We first verify the algorithmic accuracy using a 1D unsaturated soil column test. Subsequently, the framework is deployed to capture the dynamic displacement and mechanical responses of a 2D rainfall-induced landslide. Benchmarking against FEM data confirms that our multiphase MPM accurately models the infiltration process and subsequent structural collapse. Ultimately, this approach offers a highly robust computational strategy for analyzing large-scale landslide deformations and improving predictive assessments.
Rainfall-induced slope instability poses a significant risk to dam infrastructure in tropical regions, where intense and prolonged precipitation frequently alters subsurface stress conditions. This study aims to evaluate the effects of rainfall infiltration on slope stability in the Rajui Dam area, Aceh, Indonesia, by quantifying changes in safety factor and deformation behavior under existing and high-rainfall conditions. A finite element–based numerical approach was employed using two-dimensional slope models representing critical slopes near the dam access road and spillway. Soil mechanical and hydraulic properties were derived from field and laboratory data, while rainfall was modeled as surface infiltration corresponding to a high-intensity scenario of 150 millimeters over ten days. Slope stability was evaluated using the strength reduction method, with safety factor and total displacement adopted as primary indicators of stability. The results indicate that rainfall infiltration consistently reduces the safety factor and increases slope displacement at both locations. One slope exhibited lower safety factor values, indicating higher susceptibility to instability, while the other showed greater deformation despite maintaining a marginally higher safety factor. Displacement patterns concentrated near slope toes and propagated along potential shear zones, suggesting progressive instability driven by increased pore water pressure and reduced effective stress. These findings demonstrate that identical rainfall conditions can produce different stability responses depending on slope geometry and material characteristics, highlighting the importance of site-specific evaluation.
Waste dump slopes may appear stable after an initial landslide, yet their renewed instability potential under subsequent hydraulic disturbance can remain significant. This study investigates a landslide-affected waste dump slope in Ma’anshan City, Anhui Province, China, by integrating pre-landslide and post-landslide UAV-based three-dimensional reconstruction, CAD-derived representative profiles, field investigation, and measured pre-failure rainfall data. Two-dimensional coupled seepage-stability models were established in Midas GTS NX, and the shear strength reduction method was used to evaluate slope responses under pre-landslide natural and rainfall-infiltration conditions, post-landslide natural and surface-reaching phreatic-line conditions, and reinforced conditions. The results show that, under the adopted upper-bound effective infiltration ratio, the factor of safety of the pre-landslide slope decreased from 1.17 under natural conditions to 0.98 after the recorded rainfall process, indicating that the slope could evolve from a low-reserve stable state to a near-limit or unstable state under high effective infiltration. After the initial landslide, the reconfigured slope reached temporary equilibrium with a factor of safety of 1.55 under natural conditions, but decreased to 0.92 under the prescribed surface-reaching phreatic-line scenario, revealing its conditional renewed-instability potential under an upper-bound hydraulic state. The numerically interpreted instability-controlled section showed good profile-scale agreement with the UAV-derived landslide morphology, with a rear-scarp offset of 0.3 m and a main affected-zone length error of 7.7%. For the reconfigured residual slope, a stabilization scheme dominated by prestressed anchor cables and assisted by local shallow bolts increased the factor of safety to 2.28 under natural conditions and 1.28 under the surface-reaching phreatic-line scenario. The proposed framework provides a practical basis for post-landslide reassessment and residual slope stabilization of rainfall-sensitive waste dump slopes.
The traditional tank model-based landslip early warning system (LEWS) calculates the soil water index (SWI) as a single time series driven by basin-averaged rainfall, which cannot capture spatial heterogeneity along linear highway infrastructures. To overcome this limitation, this study proposes an integrated model that couples the tank model with an ordinary differential equation (ODE) form stormwater runoff simulation model: namely, the distributed runoff model (DRM). The DRM-computed distributed surface water depth replaces the first-layer water height of the tank model to generate spatially varying SWI values. The proposed framework is validated against the 2016 Typhoon No. 10 event that triggered five landslides (L1–L5) along Highway 274 in Hokkaido, Japan. Quantitative results show the following: (1) at all five landslide locations, the peak SWI values exceed 225 mm, while at a non-landslide reference point (L0) the peak SWI is only 158 mm, demonstrating clear spatial differentiation; (2) the predicted landslide initiation times from the integrated model deviate by less than 1.5 h from the actual occurrence times, whereas the shallow water equations (SWEs) and tank-coupled model advances predictions by over 7 h (L3, L4 and L5); (3) after revising the critical line (CL) based on the event data, the proposed model demonstrates a 100% identification rate for the five landslide sites with zero false alarms at L0 in this case study, indicating its potential for practical application. Compared with the tank + SWEs, the proposed tank + DRM approach maintains comparable spatial resolution but significantly improves temporal accuracy and computational efficiency, making it practical for real-time early warning along elongated highway projects. This study provides a spatially differentiated and temporally reliable decision-support tool for rainfall-induced landslide risk assessment along transportation corridors.