Deep Seating Slow-moving Landslides in Upper Mustang: Mapping, Kinematics and Triggering Factors

Upper Mustang, central Nepal, is a dry valley located between the Tibetan plateau and the High Himalayas. The Thakkhola fault system, which bounds the Thakkhola half-graben, gave its orientation to the valley, which is nearly perpendicular to the main Himalayan range. The Kali Gandaki River rises here and flows south through the high Himalayan peaks of Annapurna and Dhaulagiri, influencing the valley's landscape with its cycles of sediment aggradation and erosion. In the current phase of incision, the river has generated steep slopes that are further destabilized by the altered Tethyan shales below, creating a perfect setup for the emergence of large-scale slope deformations.

Although they have been recognized for a long time, these major deep-seated slope deformations have never been thoroughly investigated, and their activity has never been studied. Despite the area's low population, landslides have affected several settlements, including Muktinath, a significant Hindu pilgrimage destination, where deformations are destroying houses and roadways. At present, there are still questions concerning the relocation of some villages, including the monastery complex.

The landslides may be driven by spatial factors (aspect, elevation), climate factors (permafrost, snow melt, precipitation) and anthropogenic activity (irrigation). Using both remote sensing data and in-situ observations, this project aims to determine the rates and patterns of slope deformation in the Upper Mustang region and assess the possible temporal and spatial controls on the deformations.

In order to monitor landslides across a range of velocities, we use both correlation of optical satellite images from Sentinel-II (2016-2023), and InSAR time-series processing from Sentinel-I images (2015-2024). Initial mapping of the region indicated six significant deformation zones moving at varying rates, all located in the area where the Tethyan shale bedrock is found. We generate time series of displacement at finer resolution using correlation of Planet images (2016-2024), concentrating on specific landslides. On the Dhe landslide, a period of faster movement in early 2019 is found. Field observations have revealed numerous water sources in the landslides that could impact its kinematics. To supplement the kinematic analysis, seismic ambient noise from a single station seismometer is analysed to better characterize the subsurface properties of the landslides. We will present the first analyses and results from this multi-source dataset.