Three-dimensional deformation inversion of pre-sliding Baige landslide by multiple adaptive SAR offset tracking method

Abstract The Baige landslide, located on the west bank of the Jinsha River, in the Tibet Autonomous Region, China occurred on October 11 and November 3, 2018, which resulted in a blockage of the river, posing a significant threat to the safety of life and property of the people living downstream. Therefore, it is of importance to invert the high-precision, high spatiotemporal resolution three-dimensional pre-sliding deformation to analyze its developing process and reveal the precursory signal for the early warning the potential risks. Synthetic aperture radar (SAR) offset tracking was employed with adaptive window sizes to calculate the deformations prior to the landslide event from different tracks, including ascending and descending Sentinel- 1 A tracks, and ascending ALOS/PALSAR- 2 track. The three-dimensional deformation time series was inverted from two dimensional ones from three tracks. The results show that the Baige landslide moved continuously eastward, experiencing an initial acceleration, followed by deceleration; and it moved southward with continuously acceleration, even more pronounced when it approached the sliding time. Meanwhile, it moved vertically downward, with continuously increasing deformation rates. Then, the shear and dilation strains are inverted based on three-dimensional deformation time series. The results show that they were reversed on both edges of the landslide and increased over time, with dilation strains being larger than shear strains. Lastly, the thickness and volume of the landslide estimated through three-dimensional deformation are consistent at the order of magnitude with the digital elevation model (DEM) difference, which indicates that SAR deformation technique can provide key parameters for landslide analysis and assessment with high efficiency.