ADVANCES IN INSAR LEADING TO CONTINUOUSLY UPDATED GROUND DEFORMATION MEASUREMENTS

InSAR has been widely used to study landslides over the last 20 years. It uses satellite radar imagery to measure ground movements with millimeter precision and has the advantage of observing large areas without the need for fieldwork or the installation of instrumentation. It is used to monitor large slow-moving landslides in terms of deformation rate, spatial extent, directions of movement and to identify periods of acceleration. In the case of sudden catastrophic collapses, InSAR has been used as a forensic tool to investigate precursor ground movement, but only after the failure has already taken place.

New developments in data processing algorithms, radar satellite systems and distributed computing are rapidly advancing InSAR to a point where it can now be used for the early detection and warning of impending ground movement. Algorithms have been developed for the automated ingestion of new radar images within minutes of their acquisition by the satellite. Optimized codes allow processes to run on distributed, parallel systems to take advantage of the possibilities offered by cloud computing. Finally, recent satellite systems offer worldwide coverage with a site revisit frequency of 12 days and with a data policy that allows free use of the imagery. However, these advances generate an enormous amount of data, making it necessary to develop novel approaches to automatically and quickly sift through the data to detect changes in ground movement behaviour, both spatially and in time. These innovations make it possible to use InSAR as an early detection tool for identifying changes in landslide behaviour and providing timely warnings to the authorities.

Here we present at least 2 case studies. One is in British Columbia, Canada, in an area affected by numerous geohazards that include landslides and volcanoes, with research funding from the Canadian Space Agency. It has the objective of developing an algorithm for analyzing time series of displacement and automatically identifying changes in ground movement every time a new satellite image is acquired. A second case study regards the May 2017 Mud Creek landslide in California that closed Highway 1 indefinitely. InSAR shows that an acceleration of the landslide was visible over 1 year earlier and identifies several other areas of potential instability along the highway.