Investigation of slope instabilities in NW Bhutan as derived from systematic DInSAR analyses

Abstract

In this work we present a methodology based on remote sensing data to map and classify unstable slopes in high alpine areas and apply this methodology to the Himalaya of northwestern Bhutan. We leverage on radar images acquired by the ESA Envisat and the JAXA ALOS-1 satellites between 2007 and 2011. Synthetic aperture radar differential interferometry (DInSAR) is used to identify and quantify recent ground surface displacements associated with potentially unstable slopes. Optical satellite images and a high-resolution digital surface model is used for the geomorphological classification of individual landforms and deposits associated with displacements. The analysis of the geomorphological characteristics reveals that displacements are associated with: rock slides, mountain slope deformation and rock slope deformation (77.5%), soil creep (4%), soil slides (1.9%) and rock glaciers (12.3%); 4.3% of the detected displacements remain of unknown origin. The method applied here for the assessment of activity likelihood uses a very large number of interferograms obtained with images acquired by different satellites. The method, which is new and replicable, takes advantage of the combination of different acquisition geometries and then combines indices of vegetation cover, layover, velocity estimate and temporal sampling associated with each detected active area to assign a weight. The analysis shows that a cluster of unstable slopes exists in the northwest, potentially related to the presence of an important regional lineament and the distribution of permafrost. Comparatively little ongoing activity is detected in the south, but field validation documents substantial landslide activity in the recent geological past. The dataset here generated, containing 700 potentially unstable areas with an associated activity likelihood, is a novel dataset that makes an essential basis of future analysis to investigate rock slope instability distribution and predisposing factors. Moreover, this dataset represents a new important step towards future landslide hazard assessment in the Himalaya of Bhutan.