Detection and displacement characterization of landslides using multi-temporal satellite SAR interferometry: A case study of Danba County in the Dadu River Basin

Abstract

Detection of potential slope instabilities across wide areas using SAR interferometry (InSAR) techniques can aid prevention and mitigation of landslide disasters. Nevertheless, the low spatial density of coherent radar targets detected by conventional time series InSAR methods in mountainous areas with complex topography and vegetation cover introduces considerable uncertainties into the deformation measurements, and thus results in a high probability of omissions for landslide detection. In this study, a new time series InSAR method named Coherent Scatterers InSAR (CSI) is proposed to solve this problem through the joint exploitation of Persistent Scatterers (PS) and Distributed Scatterers (DS) to increase the quantity of Measurement Points (MPs) in rural environments. We applied this approach to detect potentially unstable slopes at a catchment scale, delineate sliding boundaries, and measure the deformation of major landslides. The archived ALOS PALSAR and ENVISAT ASAR data stacks covering the Danba County in the upper reach of the Dadu River Basin in southwest China, were processed. The PALSAR-measured deformation rate map revealed 17 suspected landslides and their sliding boundaries. By contrast, only six out of these 17 landslides were detected from the ASAR results, as the other 11 landslides were undetected due to the much lower density of coherent points. These detection results were indirectly verified by two different approaches. Two differential interferograms for recently acquired L-band ALOS-2 PALSAR-2 data pairs were visually inspected to check whether these landslides are still in active deformation. The reliability of the InSAR results was further evaluated by comparisons against field survey and in-situ GPS measurements. The spatial and temporal patterns of surface displacements at several slow-moving landslides were analyzed in detail, and general guidelines on the application of the CSI method for landslide investigations were discussed and summarized.