Geomatics, Natural Hazards & Risk (Dec 2023)
Investigating kinematics and triggers of slow-moving reservoir landslide using an improved MT-InSAR method
Abstract
AbstractThe safety of the Three Gorges Reservoir, the world’s largest hydraulic project, has attracted significant attention. Among the landslides in the area, the Xinpu landslide complex stands out as one of the largest. However, accurately detecting and interpreting the timing and magnitude of its movement remains a challenge. This study introduces a framework using interferometric synthetic aperture radar (InSAR) for investigating the Xinpu landslide complex. In areas with dense vegetation like Xinpu, estimating landslide movement has traditionally been problematic. The proposed method, based on independent component analysis-assisted intermittent multi-temporal InSAR, overcomes this limitation by retrieving complete and reliable time-series deformation without prior deformation models. Through an analysis of Sentinel-1 datasets covering a period of 4 years, our approach significantly enhances deformation density and accuracy in comparison to other multi-temporal InSAR techniques. When contrasted with these methods, our proposed approach exhibits an average RMS improvement of up to 46.2% between GNSS measurements and the InSAR-derived results. The complete time-series deformation reveals diverse triggers at different sections of the Xinpu landslide complex. Precipitation, with a 45-day lag, primarily controls the landslide, while the toe of the landslide shows joint effects from precipitation and water level fluctuations, with a 14-day lag. Cross-correlation analysis estimates lag times for deformation responses to different triggers and enables discussion of the complex’s kinematic process. Overall, this approach enhances InSAR performance in vegetated areas and aids in uncovering landslide triggers for effective hazard mitigation.
Keywords
