Developing a Radar based Landslide Monitoring System (LMS) Along the Eastern Lake Michigan Bluffs

Abstract:

The goal of the proposed research is to develop a cost-effective and reliable landslide monitoring system (LMS) to detect land deformation that precedes landslide development along the Eastern Lake Michigan bluffs using temporal Sentinel-1 radar data (revisit time 12 days).

The system will download Sentinel-1 scenes and associated orbital correction files as they become available on the European Space Agency (ESA) website and apply radar interferometric techniques using SARscape software and interactive data language (IDL) modules. This includes conducting pre-processing steps, generating interferograms between the recently downloaded scenes with each of the previously acquired scenes, and visual inspection of the generated interferograms and omission of problematic interferograms. These steps are followed by: (1) application of small baseline subset (SBAS) and Persistent Scatter Interferometry (PSI) time series analyses methods to the selected stack of interferograms to extract deformation rates throughout the investigated period, and (2) uploading the updated deformation rate maps and time series data into an existing web-based GIS service, (3) ) identify the targets that demonstrate changes in deformation pattern when new sets of images are added.

For validation purposes of the proposed LMS, sixty Sentinel-1 SLC images acquired during a temporal span of 3 years were used and processed. PSI and SBAB InSAR techniques were utilized to extract line-of-sight (LOS) deformation velocity (mm/year) of a landslide that occurred in Miami Park on August 8, 2019. Spatial correlation of the extracted InSAR deformation rates with observations extracted from geological/geomorphological data in a multilayer GIS environment has been analyzed. An increase in LOS velocities (up to 8 mm/year) was observed a month prior to the onset of the August 8, 2019 landslide.