A satellite soil moisture– and radar rainfall–based methodology for slope-scale seepage–stability modelling of rainfall-induced landslides

Abstract

Satellite-derived surface soil moisture (SSM) data are rarely applied in landslide modelling due to their shallow sensing depth (< 5 cm) and large spatial footprint, limiting their perceived relevance for subsurface hydrological processes. This study presents an integrated approach combining SMAP-Sentinel L2 SSM with high-resolution gridded radar rainfall data for coupled subsurface flow–stability landslide modelling. The 2020 landslide in Khao Yai National Park, Thailand, was used as a case study, supported by two years (2022–2024) of slope monitoring, including soil moisture, rainfall, and tiltmeter-based deformation data. Monthly averaged SSM showed reasonable agreement with in-situ soil moisture (R² = 0.63) and was applied as an initial boundary condition in flow models to estimate soil moisture variation with depth and time. The SSM-based model produced comparable results to in-situ soil moisture-based model (R² = 0.96) but generally underpredicted the factor of safety (F_S) by 9% especially during the onset of rain events. Two-dimensional flow and stability analyses revealed that rainfall-induced subsurface flow parallel to the slope reduced F_S to near unity prior to 2020 landslide. A moderate correlation (R² = 0.62) between soil moisture change and tiltmeter deformation indicated slope contraction during dry periods and downslope movement during wet conditions. These findings support the use of SSM and radar rainfall for improving landslide prediction and early warning systems.