A method for optimizing SBAS-InSAR interpretation results based on landslide susceptibility

ObjectiveThe interpretation results of small baseline subset interferometric synthetic aperture radar (SBAS-InSAR) exhibit multiple solutions, making it uncertain to directly use the interpreted deformation points for identifying potential landslide-prone areas. Therefore, taking the north bank of the Qingjiang River (Changyang Section) as the study area, this study proposed a method to comprehensively optimize SBAS-InSAR interpretation results by incorporating landslide susceptibility evaluation. MethodsFirstly, the deformation points interpreted by SBAS-InSAR were analyzed using clustering and outlier detection (Anselin Local Moran's I), and low-value cluster deformation points were retained. Subsequently, eight factors, including elevation, slope, slope aspect, engineering geological rock group, distance to fault, distance to water system, distance to road, topographic wetness index, were selected to evaluate and generate a landslide susceptibility zoning map using the information value method. The reliability of the landslide susceptibility evaluation was confirmed by a receiver operating characteristic (ROC) curve, with an AUC value of 0.844. ResultsThe optimized SBAS-InSAR results were obtained by filtering low-value cluster deformation points based on a threshold value (deformation rate (v) ≤ −10 mm/a) and incorporating the landslide susceptibility zoning map. Field verification in selected areas showed that the number of deformation points was reduced after optimization, and their distribution characteristics were more consistent with the historical landslide development in the study area. Additionally, taking the Yupingcun landslide group and the Pianshan landslide as typical cases, the surface deformation values monitored by SBAS-InSAR and GNSS at the same time were compared. In the case of the Yupingcun landslide, the difference between surface displacement values monitored by SBAS-InSAR and global navigation satellite system (GNSS) ranged from 0 to 7.87 mm, with an average difference of approximately 2.23 mm and an RMSE of 3.67. ConclusionThe proposed optimization method for SBAS-InSAR interpretation was demonstrated to be both practical and reliable, providing useful references for the application of InSAR technology in the field of geological disasters.