InSAR displacement time series post-processing to back-analyze a slope failure

Abstract

In this study, the failure of a slope adjacent to a motorway was back-analyzed based on InSAR data. The location of the slided area and the exact date of the event were not known in advance. A post-processing strategy was applied on the displacement time series in order to aid the detection of instability signs and to enable the identification of the location and the narrowing down of the time interval of the slide. InSAR displacement time series were obtained following the small baseline subset approach implemented on an automatic processing platform. Distributed scatterers were clustered based on the similarity of their displacement time series, in order to form clusters of scatterers with similar behavior. This procedure allowed the computation of displacement time series representative of each cluster, aiding the detection of instability signs on the slope. One of the clusters showed a sudden movement away from the SAR sensor. It was later confirmed that the slide had occurred at the location of the scatterers belonging to that cluster and during the time interval between the two observation epochs corresponding to the break in the time series. In conclusion, the proposed method was effective in the back-analysis of the slope failure, hopefully contributing to the uptake of InSAR technology by structural safety experts.