GSA Connects 2022 meeting in Denver, Colorado

Paper No. 7-3
Presentation Time: 8:40 AM

MEASURING 3D SURFACE DISPLACEMENT OF LANDSLIDES USING COMBINED OPTICAL AND SAR IMAGERY


MUHAMMAD, Mahmud1, WILLIAMS-JONES, Glyn1, STEAD, Doug1 and DONATI, Davide2, (1)Earth Sciences, Simon Fraser University, Earth Sciences, 8888 Universit, Burnaby, BC V5A1S6, Canada, (2)Department of Civil, Chemical, Environmental, and Materials Engineering, University of Bologna, Via Zamboni, 33, Bologna, 40126, Italy

The geomorphic evolution of landslides depends on many factors ranging from the geological and structural configuration of the slope, which controls the style of deformation, and direction and magnitude of slope displacement. Monitoring the deformation and displacement of landslides is critical to identify the failure mechanism, assess and forecasts their potential evolution. In this respect, satellite-based remote sensing techniques developed in the last two decades have proven a very effective tool in monitoring unstable slopes and are today routinely employed for the analysis of landslides. However, some significant shortcomings still affect current remote-sensing datasets and methodologies. For instance, satellite radar remote sensing only measures the slope deformation along the satellite Line-of-Sight, thus failing to provide the magnitude of the 3D slope deformation vector. Satellite optical image datasets can be used to measure horizontal deformation via optical flow and image pixel correlation techniques but cannot provide the vertical component of the slope deformation due to the unfavorable point of view.

In this study, we present a method that combines radar and satellite optical image datasets, allowing the 3D slope surface deformation vector to be computed. We apply this method to investigate two different case studies, 1) the slowly deforming Mount Currie (British Columbia, Canada), and the rapid, flow-like Mud Creek landslide (California, USA).

We first use InSAR techniques, such as SqueeSAR and Mintpy, to processes the phase component for stacks of SAR images and derive the East-West and vertical component of the slope deformation. Then, we process Planet Labs® satellite images using Akh-Defo software, a python-based, in-house software that exploits optical flow algorithms. The software is capable of measuring the horizontal component of the slope deformation in a geographical coordinate space (NS, EW). Finally, we use standard python GIS scripts to derive parameters such as trend and plunge of slope movement.

The presented workflow can be used as a tool for expanding the capabilities of satellite-based radar monitoring, by allowing an enhanced understanding of the style of deformation of slopes, as well as to validate and constrain geomechanical numerical modelling of landslides.