Kara-Bogaz-Gol huge landslide in Turkmenistan: Activity monitoring using Sentinel-1 TOPS time series interferometry
As a major natural hazard, landslides modify the morphology of earth surface and cause fatalities and property damage worldwide every year. Detection and monitoring of slope movements and linking them with the potential inducing factors are of great importance for better understanding the landslide kinematics and preventing landslide disasters. In this study, we report on a previously unknown a giant coastal landslide along the lagoon of the Caspian Sea which is the largest landslide in the world. Monitoring the activity of such an immense slope instability is hence critical. The landslide activity is in the eastern coast of the Kara-Bogaz-Gol (KBG) shallow inundated depression in the northwestern corner of Turkmenistan has been studied using the Persistent Scatterer Interferometry (PSI) technique. In the present study, the Stanford Method for PS package (StaMPS) is employed to process the series of Sentinel 1-A and 1-B Synthetic Aperture Radar (SAR) images in TOPS mode acquired between 2014 and 2020 from two descending and one ascending orbits. The investigations of the surface displacement obtained from three different data sets are presented in a spatio-temporal extend. We constructed the mean velocity field and the time-series of the deformation obtained from all data sets and decomposed them into east-west and vertical component. We utilised InSAR-derived velocity field using elastic dislocation modeling to find a possible dislocation plane resembling the landslide basal décollement and calculating the volume of the landslide which allowed us better defining a model capable to describe the dynamics of the slow moving coastal landslide. Our results show that ~20-km-long section of the eastern bank of the lagoon is sliding towards the lake at a rate of 25 mm/yr with a clear dominance of horizontal motion with a series of embedded circular slip surfaces. We also investigated the temporal response of the slide to the seasonal and multi-annual hydrological parameters obtained from different satellite missions. The comparative temporal analysis reveals week-long accelerating creep events at similar amplitudes occurring seasonally which is consistent from three independent InSAR time-series. This suggest a good correlation between landslide movement episodes and its sensitivity to hydrological forcing. The time-series results show that some sectors of the landslide are still active. Our study also reveals the importance of remotely characterizing often inaccessible slopes in a large scale for better understanding the landslide dynamics and reduce the associated geohazard.