Landslide risk assessment with multi pass DInSAR analysis and error suppressing approach

Landslide is one of the most dreadful natural hazards and the prime risk source causing lethal damages in many countries. In spite of various attempts to measure the landslide susceptibility by the remote sensed method including Differential Interferometric SAR (DInSAR) analysis, the construction of reliable forecasting systems still remains unsolved. Thus, we tackled the problem of DInSAR analysis for monitoring landslide risk over the mountainous areas where InSAR observations are usually contaminated by the orographic effects and other error elements. In order to measure the correct surface deformation which might be a prelude of landslide, time series analysis and atmospheric correction of DInSAR interferograms were conducted and crossly validated. The target area of this experiment is the eastern part of Korean peninsula centered in Uljin. In there, the landslide originated by the geomorphic factors such as high sloped topography and localized torrential down pour is critical issue. The landslide cases frequently occurred in the cutting side of mountainous area by the anthropogenic construction activities. Although high precision DInSAR measurements for monitoring the landslide risks are essential in such circumstances, it is difficult to attain sufficient enough accuracy because of the external factors inducing the error component in electromagnetic wave propagation. For instance, the local climate characteristics such as orographic effect and the proximity to seashore can produce the significant anomalies in the water vapor distribution and consequently result in the error components of InSAR phase angle measurements. Moreover the high altitude parts of target area cause the stratified tropospheric delay error in DInSAR measurement. After all, the improved DInSAR approaches to cope all above obstacles are highly necessary. Thus we employed two approaches i.e. StaMPS/MTI (Stanford Method for Persistent Scatterers/Multi-Temporal InSAR, Hopper et al., 2007) which was newly developed for extracting the reliable deformation values even with the presence of error terms and two pass DInSAR with the error term compensation based on the external weather information using L band ALOS PALSAR and C band ENVISAT ASAR. Since MERIS Reduced Resolution (RR) coverage over target areas includes a few cloud free scenes, the water vapor map constructions were feasible with 1.2km spatial resolutions in two pass DInSAR pairs and enable to assess the deformation measurement of StaMPS/MTI by the inter comparison. Although the correlation between deformation patterns from two pass DInSAR and StaMPS was not very clearly identified in this study, the deformation values and the landslide triggering factors showed some agreements. Thus the quantitative landslide monitoring scheme is supposedly feasible on the condition that the high accuracy atmospheric error map and the methodology effectively compensating it from DInSAR interferograms are available. The scheme in this study will be further upgraded for the application of future C, X and L band SAR by incorporating the spaceborne radiometer and/or weather forecasting model to establish electromagnetic wave delay map.