Slope activity and processes in the Himalaya of Northern Bhutan

Bhutan is a small Himalayan country, located between India and Tibet. Its territory is characterised by active tectonics, erosion and extreme topographic gradients as well as highly varied climatic zones. Such geographical setting makes for a landscape prone to geological mass movements. In this project we focus on 1) fundamental contributions to scientific knowledge regarding processes conditioning large rock slope movements in the High Himalaya of Bhutan and 2) landslide hazard assessment by exploiting the state-of-the art in Differential Interferometry of Synthetic Aperture Radar (DInSAR) image processing and post-processing. We analysed high resolution optical images (GoogleEarth) and high resolution DTM (ALOS World 3D) in order to detect unstable slopes with clear geomorphological features. Moreover, we processed satellite-based radar images (2 Envisat ASAR tracks and 4 ALOS Palsar-1 frames) with DInSAR techniques, covering an observation period between August 2005 and March 2011. Validation of specific cases was carried out with more recent data (ALOS Palsar-2 and Sentinel-1, between September 2014 and July 2017). Specifically, we generated more than 500 differential wrapped interferograms and analysed them individually for actively deforming areas. Successively, we created a weight based on acquisition parameters and number of observations for each individual object, in order to generate a likelihood of activity. We also processed 5 of the available tracks with the SBAS (Small Baseline Subset) approach and obtained velocity and displacement maps which were also analysed for potentially active areas and displacement time series. We thus obtained a large inventory of rockslides and rock glaciers (more than 1300 objects) based on optical images and two inventories based on DInSAR information containing more than 1000 objects classified by landslide type including: 1) rockslides, 2) rock glaciers, 3) moraines, 4) talus or slope debris. The underlying driving processes were systematically investigated through the analysis of the time series trends. For this purpose, we also analysed the seasonal cycles within the time series and identified two cycles that were interpreted as uncorrected atmospheric delays. Outside of such atmospheric cycles we were able to detect spatially well constrained areas characterised by seasonal cycles likely caused by reversible ground deformation. This, combined with the analysis of the overall trends seen in the time series, allowed for the classification of the processes driving slope instability: 1) gravitational, 2) reversible related to permafrost freeze and thaw and 3) reversible related to groundwater recharge and depletion. The extent of the investigated area (roughly 8000 km2) and the amount of systematically analysed data are unprecedented. We show how to generate a classification of instabilities types and processes and to maximise the extraction of information on slope instability for an extensive and largely inaccessible region.