Stability Assessment of a Large-Scale Landslide in the Minjiang River Basin Based on UAV Oblique Photogrammetry and Numerical Simulation

Abstract

The West Route of the South-to-North Water Diversion Project traverses geologically fragile alpine terrain where large pre-existing landslides pose long-term risks to infrastructure safety and sustainable corridor development. Taking the Xianiba landslide in the upper Minjiang River Basin (plan area approximately 8.6 × 104 m2) as a case study, this study integrates unmanned aerial vehicle (UAV) oblique photogrammetry, three-dimensional geological modeling, and FLAC3D numerical analysis to investigate the relationship between short-term rainfall infiltration and deep-seated hydraulic reactivation. Under natural conditions, the three-dimensional model reaches mechanical equilibrium, and the strength-reduction method (SRM) gives a factor of safety (Fs) of 1.656, while the rear and upper portions of the sliding interface show stronger localized shear response. A separate one-dimensional infiltration analysis was used to constrain the depth of direct short-term matrix wetting, showing that the 24 h rainfall-affected depth increases from approximately 0.43 m at 20 mm/d to 1.00 m at 150 mm/d. Prescribed hydraulic-head scenarios were then used to quantify the stability evolution of the deep sliding interface. Across the re-equilibrated prescribed-head cases (h = 1.0–2.25 m), Fs remains close to 1.64, whereas interface shear displacement and current shear-yield response increase markedly with hydraulic head. The model can re-establish equilibrium under a hydraulic head of 2.25 m but develops sustained sliding at 2.5 m and above, defining a model-conditioned transition range of approximately 2.25–2.50 m. These results indicate that short-term rainfall mainly controls shallow wetting, whereas potential reactivation of the deep-seated landslide is more strongly associated with hydraulic weakening along the pre-existing sliding interface. The 24 h rainfall cases are therefore interpreted as short-term matrix-infiltration scenarios rather than complete field-scale recharge pathways. The proposed framework supports landslide risk assessment, targeted monitoring, and more sustainable planning and operation of major infrastructure in fragile alpine regions.