Pre-collapse motion of the February 2021 Chamoli rock-ice avalanche, Indian Himalaya

Abstract

Abstract. On the 7th of February 2021, a large rock-ice avalanche triggered a debris flow in Chamoli district, Uttarakhand, India, leaving over 200 dead or missing. The rock-ice avalanche originated from a steep, glacierized north-facing slope. In this work, we assess the precursory signs exhibited by this slope prior to the catastrophic collapse. We evaluate monthly slope motion from 2015 to 2021 through feature tracking of high-resolution optical satellite imagery. We then combine these data with a time series of pre- and post-event DEMs, which we use to evaluate elevation change over the same area. Both datasets show that the 26.9 Mm3 collapse block moved over 10 m horizontally and vertically in the five years preceding collapse, with particularly rapid motion occurring in the summers of 2017 and 2018. We propose that the collapse results from a combination of snow-loading in a deep headwall crack and permafrost degradation in the heavily jointed bedrock. Our observation of a clear precursory signal highlights the potential of satellite imagery for monitoring the stability of high-risk slopes. We find that the timing of the Chamoli rock-ice avalanche could likely not have been forecast from satellite data alone.

Detection of slow‐moving landslides through automated monitoring of surface deformation using Sentinel‐2 satellite imagery

“…The median total displacement of this landslide over the past 7 years Slow-moving landslides are sensitive to changes in climate, both directly through changes in their hydrology and water availability (Ardizzone et al, 2023;Gariano & Guzzetti, 2016;Handwerger et al, 2019Handwerger et al, , 2022 and indirectly through changes in their boundary conditions (Dai et al, 2020;Mititelu-Ionus¸et al, 2011;Van Wyk de Vries, Bhushan, et al, 2022;Yang et al, 2021). The close match between the average displacement of the Oriental lateral moraine and lateral retreat rate of the glacier is an example of this second category.…”

“…Rapid climate warming is a driving factor for the current high rates of ice loss (Abdel Jaber et al, 2019;Willis et al, 2012), which in turn impact the stability of surrounding slopes. Other changes in local conditions such as vegetation cover, snow accumulation, rivers and landslide dam or glacial outburst floods can also affect the long-term motion of slow-moving landslides (Cook et al, 2018;Mititelu-Ionus ȩt al., 2011;Van Wyk de Vries, Bhushan, et al, 2022;Yang et al, 2021). In many instances, slow-moving landslide motion is driven by the availability of water which increases internal pore-water pressure and decreases the frictional strength of slopes (Bogaard & Greco, 2016;Lacroix et al, 2020).…”

“…Because slow-moving landslides often cannot be detected through changes in surface properties on a scale observable by satellites, alternative methods are necessary to identify them (Figure 1). Three main techniques have been used: topographic change detection, interferometric synthetic aperture radar (InSAR) and feature tracking (Dai et al, 2020;Dille et al, 2021;Lacroix et al, 2022;Shugar et al, 2021;Stumpf et al, 2014;Stumpf et al, 2017;Van Wyk de Vries, Bhushan, et al, 2022). Topographic change detection can be used to identify vertical change associated with slope deformation F I G U R E 1 Range of common spatial scales and deformation velocities for slow-moving landslides.…”

“…Feature tracking measures the horizontal displacement of coherent patterns between a pair of images, which can be derived from optical or radar sensors (Bickel et al, 2018). These methods have been widely used for historical and ongoing monitoring of known slow-moving landslides (Casagli et al, 2023;Delacourt et al, 2007;Ferrigno et al, 2017;Wasowski & Bovenga, 2014), but low signal-to-noise ratios complicate their use for automated detection of previously unknown landslides, particularly in steep mountainous terrain (Bekaert et al, 2020;Lacroix, 2016;Lacroix et al, 2022;Rosi et al, 2018;Van Wyk de Vries, Bhushan, et al, 2022).…”