Sentinel-1 Data Analysis for Landslide Detection and Mapping: First Experiences in Italy and Spain

Abstract

Abstract:One of the key inputs to landslide susceptibility and hazard analyses is provided by a precise inventory map, including the information of landslide activity. In the last decade the satellite SAR Interferometry has been demonstrated to be a powerful tool for landslide mapping and monitoring. However, until now, the systematic use of the technique has been strongly limited by different aspects like the image availability, the revisit time and the loss of coherence. In this context, the Sentinel-1 constellation provides interesting characteristics for landslide mapping and monitoring: the wavelength (55.5 mm) and the short temporal baseline (6 days when Sentiel-1B data will be available). The latter one is expected to be a key feature for increasing coherence and for defining monitoring and updating plans. However, the exploitation of these high coherent data for landslide purposes is not straightforward and demands of new tools and methods to properly discriminate the real displacements. The aim of this work is to show which are, according to the authors experience, the main pros and cons of landslides mapping and monitoring with Sentinel-1 data. These conclusions are based on the analysis over three different test sites with different characteristics from the SAR view point: the Canarias Islands (Spain), the Molise region (Italy) and the Toscana region (Italy). The Canarias Islands test site shows a good example of how Sentinel-1 can systematically provide deformation maps over wide areas and with a very high sampling density. The Molise region application shows how the short temporal revisit time of Sentinel-1 allows the detection and monitoring of landslides also over agricultural areas. In particular, for this test site, are shown some examples of detection and update of active landslide phenomena. Finally, in the third example, the Toscana area, we show another example of difficult scenario, where the high temporal sampling has allowed the detection of some landslide phenomena. However, this test site is also a good example to show one of the main problems to be solved when applying short temporal baseline interferometry in rural environments: false deformation trends related to soil changes. AbstractThe differential interferometric SAR (DInSAR) technique is a powerful tool to detect and monitor ground deformation. In this paper we address an important DInSAR application, which is the detection and mapping of landslides. The potential of DInSAR to detect and monitor landslides has been extensively documented in the literature, mainly using the Cband data from the European Remote Sensing (ERS-1 and -2), Envisat and Radarsat missions. A significant improvement in landslide monitoring is expected by the SAR data of the two satellites Sentinel-1A and -1B of the European Space Agency. This paper describes the authors' first experience using Sentinel-1 for landslide monitoring. The paper describes the data processing and analysis strategy, and then illustrates some deformation measurement results obtaine...

Quantitative Assessment of Digital Image Correlation Methods to Detect and Monitor Surface Displacements of Large Slope Instabilities

“…The application of remote sensing techniques enables also the extraction of information about the evolution of the slope instability. Within the last decade, air-and spaceborne remote sensing methods using active sensors, such as differential synthetic aperture radar (SAR) interferometry (DInSAR) [5,10,11], or methods using input from passive sensors, such as digital image correlation (DIC) [12][13][14][15][16], have been established to detect and monitor surface displacements at regional scales.…”

New Perspectives in Landslide Displacement Detection Using Sentinel-1 Datasets

“…These techniques, commonly referred to as Advanced DInSAR (A-DInSAR) or Time Series Radar Interferometry (TS-InSAR) are similar implementations of two main approaches: the Persistent Scatterer Interferometry (PSI) [4], based on the identification of natural point-like targets in stacks of interferograms, and the Small Baseline Subsets (SBAS) [5], centered on the privileged spatial and temporal relations of the acquisitions. So far, A-DInSAR analyses have proven to be effective in the detection and measurement of landslide classified as very slow (few centimeters per year) [11][12][13][14][15][16]. This study demonstrates that processing data collected by the S1 mission, with properly designed interferometric approaches, can extend the field of application to the upper class of slow landslides (few metres per year), thus providing new perspectives in landslides detection.…”

The Maoxian landslide as seen from space: detecting precursors of failure with Sentinel-1 data

“…They mount SAR sensors suitable for large-scale mapping (Kalia et al 2017;Novellino et al 2017;Vecchiotti et al 2017), with short revisiting times (6 days, when both sensors are active over the same area) and quite good spatial resolution (Geudtner et al 2014;Barra et al 2017;Casagli et al 2017). The enhanced temporal resolution of Sentinel-1 compared to all previous ESA missions and the regularity of the acquisitions using the same acquisition mode are now paving the way to the use of InSAR data for early detection of movements in places where ground-based monitoring would be unfeasible, like the remote landslide-prone areas in Sichuan.…”