Landslide Monitoring Using Multi-Temporal SAR Interferometry with Advanced Persistent Scatterers Identification Methods and Super High-Spatial Resolution TerraSAR-X Images

Feng Zhao,Jordi Mallorqui,Josep Gili,Jordi Corominas,Rubén Iglesias

doi:10.3390/rs10060921

Abstract

Landslides are one of the most common and dangerous threats in the world that generate considerable damage and economic losses. An efficient landslide monitoring tool is the Differential Synthetic Aperture Radar Interferometry (DInSAR) or Persistent Scatter Interferometry (PSI). However, landslides are usually located in mountainous areas and the area of interest can be partially or even heavily vegetated. The inherent temporal decorrelation that dramatically reduces the number of Persistent Scatters (PSs) of the scene limits in practice the application of this technique. Thus, it is crucial to be able to detect as much PSs as possible that can be usually embedded in decorrelated areas. High resolution imagery combined with efficient pixel selection methods can make possible the application of DInSAR techniques in landslide monitoring. In this paper, different strategies to identify PS Candidates (PSCs) have been employed together with 32 super high-spatial resolution (SHR) TerraSAR-X (TSX) images, staring-spotlight mode, to monitor the Canillo landslide (Andorra). The results show that advanced PSI strategies (i.e., the temporal sub-look coherence (TSC) and temporal phase coherence (TPC) methods) are able to obtain much more valid PSs than the classical amplitude dispersion (DA) method. In addition, the TPC method presents the best performance among all three full-resolution strategies employed. The SHR TSX data allows for obtaining much higher densities of PSs compared with a lower-spatial resolution SAR data set (Sentinel-1A in this study). Thanks to the huge amount of valid PSs obtained by the TPC method with SHR TSX images, the complexity of the structure of the Canillo landslide has been highlighted and three different slide units have been identified. The results of this study indicate that the TPC approach together with SHR SAR images can be a powerful tool to characterize displacement rates and extension of complex landslides in challenging areas.

Highlights

Every year, with the onset of rains and snow melting, landslides represent one of the major natural threats to human life and infrastructures in natural and urbanized environments
The line of sight (LOS) displacement rate maps derived by the three methods are shown in Figure 8a–c, respectively
To make a fair comparison, the pixel selection thresholds for all the three methods were established based on a phase standard deviation of 15◦

Summary

Introduction

With the onset of rains and snow melting, landslides represent one of the major natural threats to human life and infrastructures in natural and urbanized environments In this context, different surveying techniques, such as inclinometers, extensometers, piezometers, jointmeters, photogrammetry, LiDAR or Global Positioning Satellite System, are typically employed to address landslide monitoring problems [1,2,3,4,5,6,7,8]. They are labor intensive, expensive and usually require skillful users for data interpretation They typically provide poor spatial sampling and coverage, which hinder the characterization of complex landslides. The new generation of X-band SAR sensors, like the German TerraSAR-X and TanDEM-X satellites or the Italian constellation Cosmo-Skymed, have led to a scientific breakthrough presenting a lower revisiting time (up to few days) and an improved spatial resolution (even below the meter), compared with their predecessors ERS-1/2, ENVISAT-ASAR and RADARSAT-1 or the recently Sentinel-1, which worked at the C-band

Methods

Results

Conclusion