Identifying reservoir landslide deformation evolution stages with time series InSAR: application to the Xinpu landslide in China’s three Gorges region

• An InSAR-based method is proposed to identify landslide evolution stages. • Xinpu landslide was active before the reservoir impoundment. • The landslide showed alternating initial and uniform deformation stages. • Water levels and precipitation are key triggers for the landslide’s movement. Reservoir landslides generally exhibit complex spatiotemporal deformation characteristics. Accurately identifying their evolution stages is essential for risk assessment and early warning systems. Interferometric Synthetic Aperture Radar (InSAR) technology have shown significant potential in reservoir landslide detection and monitoring. However, the technology’s low observation frequency and high deformation fluctuations introduce notable challenges in the identification of evolution stage. In this context, the study proposes an enhanced method for identifying reservoir landslide evolution stages. Time series deformation along the slope direction was derived from multi-sensor Synthetic Aperture Radar (SAR) systems. Following this, the probability distribution of the deformation velocity, as inferred from InSAR, was estimated. After that, the landslide’s deformation evolution stage was delineated by employing the Comprehensive Standardized Deformation Index (CSDI). The method was applied to the Xinpu Landslide in China’s Three Gorges Region, where nearly three decades of time-series deformation from 1993 to 2024 were analyzed using 259 SAR images. The analysis revealed that the Xinpu landslide was active before the Three Gorges Reservoir’s impoundment in 2003, with an average deformation rate of approximately 70 mm/year in the line-of-sight (LOS) direction and a cumulative deformation of 8 m over 30 years in the slope direction. The analysis of the landslide velocity’s probability distribution revealed a fluctuating deformation pattern of Xinpu landslide, marked by alternating initial and uniform deformation stages. The proposed methodology surpasses the tangent angle and rate ratio methods, which are susceptible to overestimating the accelerated stage. The study provides scientific support for early warning and prevention of landslide risks in reservoir areas.