Glacier-surface velocities in alpine terrain from optical satellite imagery - Accuracy improvement and quality assessment

Abstract

The worldwide retreat of mountain glaciers has important consequences for the water, food, and power supply of large and densely populated areas in South and Central Asia. Successful mitigation of the hydrological impacts on societies as well as assessing glacier-related hazards require large-scale monitoring of glacier dynamics. However, detailed glaciological data from the Asian highlands are lacking, due to its size and difficult accessibility. We have applied a novel technique for precise orthorectification, co-registration, and sub-pixel correlation of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) satellite imagery to derive surface velocities of Himalayan glaciers. Our approach allows for the correction of offsets due to attitude effects and sensor distortions, as well as elevation errors if a digital elevation model (DEM) from the Shuttle Radar Topography Mission (SRTM) was used for orthorectification. After post-processing, the error on the displacements is on the order of 2–4 m per correlation. Translated into annual velocities, this error is reduced (increased) when the correlated images are more (less) than a year apart. Through application of a filtering procedure and several quality tests, the consistency of the results is validated to provide confidence in the remotely sensed velocity measurements, despite the lack of ground control. This novel approach allows fast, easy, and economically viable acquisition of detailed glaciological data in areas of difficult access and provides a means for large-scale monitoring of glaciers in high mountainous terrain.