Retrieving 4D landslide displacement using Pléiades satellite stereo pairs on the La Valette landslide

Abstract

Slow-moving landslides pose a substantial threat to communities and infrastructure, with annual creeping distances ranging from a few mm to hundreds of meters. To protect local communities from the landslide motion, landslide displacement retrieving or monitoring is necessary. However, traditional field investigations are time- and labor-consuming, which may limit the understanding of the landslide evolution and thereby mitigation. We propose a new 4D landslide displacement framework, combining satellite-based structure-from-motion, cross-correlation feature tracking for horizontal ground-surface deformation measurements with COSI-Corr, and DEM differencing for vertical ground-surface deformation. We apply this method to very high resolution (0.5 m) optical stereo images acquired by the Pléiades satellite constellation. We use our method to monitor the annual movement of the ‘La Valette’ landslide in the French Alps, between 2012 and 2022. During this period, the landslide moved most actively during the years 2012 and 2013, with average 3D displacement rates of 1.22 and 0.89 cm / day, respectively. Furthermore, we found a decelerating trend in movement rates from 2012 to 2022, which we attribute to warmer weather, decreasing precipitation rates, drier air conditions, and the implementation of a drainage installation. Our study demonstrates the potential of very-high resolution satellite stereo imagery for near-real time accurate monitoring of 4D landslide displacement, which benefits the tracking and hazard management of slow-moving landslides.