Multi-satellite InSAR combination to support multi-scale analyses of hillslope processes 

Abstract

<p>Landslide inventory maps represent a preliminary step toward landslide susceptibility, hazard, and risk assessment. The increasing enhancement of A-DInSAR (Advanced Differential Synthetic Aperture Radar Interferometry) techniques facilitates the detection of Earth’s surface displacements over large or remote areas. Moreover, applying post-processing tools to the measurements retrieved by PS-InSAR analyses (i.e., one of the most common multitemporal A-DinSAR techniques) permits the representation of gravity-driven processes evolution in both spatial and temporal terms. Nevertheless, geometric distortions linked to the orbit and acquisition parameters of the SAR sensors, along with insufficient site coverage and spatial density of the PS-InSAR analyses, may lead to a lack of information, especially in mountainous areas. To address this problem, we processed the data using different InSAR tool packages and exploited the combination of orbital geometries for different satellites at the regional and local scales. These analyses were applied over an area encompassing four regions in the Central Apennines (Italy), within the framework of a broader national project which aims at mapping and updating landslide-prone slopes interacting with urban centers. For each processed dataset, we compared the spatial coverage and the accuracy of the displacements, providing statistical correlation tests to establish the relationship between the different InSAR tool packages. Therefore, we were able to verify the possible underestimation of the velocity and coherency measurements, and then select the best dataset (or the best combination) for further analyses. Based on the comparison between the dataset and through a semi-automatic approach, we then selected several areas that exceeded specific velocity thresholds and were densely covered by PS. In these areas, classified with a high priority level, detailed analyses were performed through a set of post-processing plugins designed for the software QGIS. Spatial and temporal deformation trends of the PSI results, along with subtle surface patterns within the landslide area, were highlighted by the post-processing analyses. Thus, we derived a detailed geomorphological characterization for the high priority phenomena interacting with cities and infrastructures. While at the regional scale findings from our work help the validation and integration of multi-satellite datasets, at the local scale the proposed workflow can also support the prioritization of site-specific monitoring and intervention planning.</p>