Himalayan glaciers are key components of earth's cryosphere, acting as hydrological reservoirs vital to many human and natural systems. Most Himalayan glaciers are shrinking in response to changing climate, which will potentially impact water resources, natural hazards, sea level rise, and many other aspects. However, there is much uncertainty regarding the state of these glaciers, as direct field data are difficult to obtain. Accordingly, long-timespan remote sensing techniques are needed to measure changing glaciers, which have memory and often respond to climate on decadal timescales. This study uses declassified historical imagery from the Hexagon spy satellite database to fulfill this requirement. A new highly-automated, computer-vision based solution is used to extract historical terrain models from Hexagon imagery, which are used as a baseline to compute geomorphic change for glaciers in the Kingdom of Bhutan and Tibet Autonomous Region of the eastern Himalayas. In addition to glaciers, the new method is used to quantify changes resulting from the Thistle Creek Landslide (surface elevation changes resulting from the landslide show an average elevation decrease of 14.4 ± 4.3 meters in the source area, an increase of 17.6 ± 4.7 meters in the deposition area, and a decrease of 30.2 ± 5.1 meters resulting from a new roadcut) and Mount St. Helens eruption in western North America (results show an estimated 2.48 ± 0.03 km3 of material was excavated during the eruption-triggered debris slide). These additional results illustrate the applicability of Hexagon imagery to a variety of landscape processes. Regarding the primary application in the Himalayas, all studied glaciers show significant ice loss. Futhermore, the multi-decadal timespan reveals important aspects of glacier dynamics not detectable with temporally shorter datasets. Some glaciers exhibit inverted mass-balance gradients due to variations in debris-cover, while enhanced ice losses are prominent on glacier toes terminating in moraine-dammed proglacial lakes, resulting from calving caused by thermal undercutting. Remarkably, debris-covered glaciers show significant thinning despite insulating effects of the debris, likely due to poorly-understood ice cliff and melt pond mechanisms. The mean annual geodetic mass balance of 22 studied glaciers over a 32-year period is estimated to be -0.16 ± 0.03 m yr-1 water equivalent. Thus, these glaciers are not in equilibrium with current climate, and appear to be losing significant amounts of ice regardless of debris-cover.



College and Department

Physical and Mathematical Sciences; Geological Sciences



Date Submitted


Document Type





declassified imagery, computer vision, DEM, geomorphic change, climate, glaciers, Himalayas, Bhutan

Included in

Geology Commons