Use of PSInSAR for Long-Term Surface Displacement Monitoring as a Complement to the Official Landslide Susceptibility Map in the Saguenay–Lac-Saint-Jean Region, Quebec

Abstract

Monitoring ground surface movements is crucial for slope failure hazard assessment. A total of 168 Sentinel-1A SAR ascending orbit datasets were analyzed (January 2018 to May 2024) to identify slope failure-susceptible locations in the Saguenay–Lac-Saint-Jean (SLSJ) region in Quebec, Canada, using persistent scatterer interferometric synthetic aperture radar (PSInSAR). Validation against four regional Global Navigation Satellite System (GNSS) stations shows strong agreement, with correlation coefficients (r > 0.9) across stations and root-mean-square error (RMSE) values of 10.8 and 12.7 mm between the PSInSAR and GNSS cumulative displacement time series over the six-year period. Annual Line-of-Sight (LOS) surface displacement rates generally ranged from 0 to 20 mm/yr, and the median absolute deviation (MAD) method was subsequently applied to the PSInSAR time series to identify anomalous persistent scatterer (PS) points within PS-detectable areas as candidate displacement signals. Integrating them with historical slope failure records (HSFRs) highlights partial spatial correspondence with the official landslide susceptibility map (LSM). Results show that the official LSM covers about 45% of HSFRs and 50% of anomalous PS points, with the remainder located outside its mapped boundaries. Conversely, our approach achieved over 80% overlap between anomalous PS points and HSFRs within areas where PS observations are available, identifying candidate slope failure-susceptible areas beyond those currently mapped by the official LSM, thereby warranting further investigations. This is exemplified by a 2022 landslide detected beyond the LSM boundaries. The proposed approach could serve as a complementary framework for mapping slope failure-susceptible locations within the study region.