Abstract
The Interferometric Synthetic Aperture Radar (InSAR) technique is a well-developed remote sensing tool which has been widely used in the investigation of landslides. Average deformation rates are calculated by weighted averaging (stacking) of the interferograms to detect small-scale loess landslides. Heifangtai loess terrace, Gansu province China, is taken as a test area. Aiming to generate multi-temporal landslide inventory maps and to analyze the landslide evolution features from December 2006 to November 2017, a large number of Synthetic Aperture Radar (SAR) datasets acquired by L-band ascending ALOS/PALSAR, L-band ascending and descending ALOS/PALSAR-2, X-band ascending and descending TerraSAR-X and C-band descending Sentinel-1A/B images covering different evolution stages of Heifangtai terrace are fully exploited. Firstly, the surface deformation of Heifangtai terrace is calculated for independent SAR data using the InSAR technique. Subsequently, InSAR-derived deformation maps, SAR intensity images and a DEM gradient map are jointly used to detect potential loess landslides by setting the appropriate thresholds. More than 40 active loess landslides are identified and mapped. The accuracy of the landslide identification results is verified by comparison with published literatures, the results of geological field surveys and remote sensing images. Furthermore, the spatiotemporal evolution characteristics of the landslides during the last 11 years are revealed for the first time. Finally, strengths and limitations of different wavelength SAR data, and the effects of track direction, geometric distortions of SAR images and the differences in local incidence angle between two adjacent satellite tracks in terms of small-scale loess landslides identification, are analyzed and summarized, and some suggestions are given to guide the future identification of small-scale loess landslides with the InSAR technique.
Highlights
Loess occupies an area of about 630,000 km2, equivalent to approximately 6.6% of the total area of China [1]
According to the height-to-phase formula [63], the topographic phase is proportional to the perpendicular baseline B⊥ between two Synthetic Aperture Radar (SAR) images, and the digital elevation model (DEM) error ∆Z can be estimated with following equation: δφ = [A, B]
DEM errors occur in differ1e0notifa2l8 interferograms generated from advanced land observing satellites (ALOS)/PALSAR-2, TerraSAR-X and Sentinel-1A/B datasets
Summary
Loess occupies an area of about 630,000 km, equivalent to approximately 6.6% of the total area of China [1]. A series of advanced multi-temporal InSAR (MT-InSAR) algorithms, including Persistent Scatterers Interferometry (PSI) [15,16,17,18,19,20,21,22] and the Small Baseline Subset (SBAS) [23,24,25,26], have been developed to overcome the major limitations of the traditional InSAR method and improve the applicability of the InSAR technique These advanced MT-InSAR techniques have been widely used in the identification of landslides [27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43]. Multi-sensor InSAR datasets including L-band ALOS/PALSAR and ALOS/PALSAR-2 data, X-band TerraSAR-X data and C-band Sentinel-1A/B data with various spatial-resolutions, revisiting periods and wavelengths, make it possible to detect small-scale loess landslides and generate multi-temporal inventory maps
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
