Crevasses on Svalbard Glaciers: Distribution and Dynamic Controls
Not peer reviewed
MetadataShow full item record
Crevasses play a key role in understanding a wide range of glaciological processes, but the current understanding is mainly based on a few key studies. Recent years, studies on crevasse formation have been motivated by the understanding of processes of calving and the breakup of ice shelves. Dynamics of glaciers can help us to reconstruct long-term environmental changes and describe a glaciers response to climate changes. In addition to the glaciological perspective, crevasses can be major hazard to travelers. Glaciers are important travel routes on Svalbard, but currently there are no systematic data on their distribution or their relationship to glaciological variables. From ground truth measurements of crevasse positions and crevasse widths on Borebreen and Tunabreen, and manual mapping of crevasses in ArcMap, TerraSAR-X images are confirmed to be a suitable tool for defining crevasse areas on glaciers. However, there are some restrictions to the method based on the variations of backscatter for different glacier surface conditions. This can lead to an underestimation of the crevasse area extent. Delineation of the crevasse area limit line on 32 images in a 22- day cycle in the period from February 9th 2013 to April 1st 2015, shows that the terminal crevasse field of Tunabreen is migrating up- glacier during spring and summer, and stabilizes during winter. This pattern coincides with the pattern of retreat and still stand or advance of the glacier front during the same time period. By using velocity maps and strain rate maps based on feature tracking from two TerraSAR-X images in a 11-day cycle, it was aimed to find threshold velocity and strain rate for crevasse opening. The velocity maps used only contain velocity information in the lower part of the terminal crevasse field on Tunabreen, and no velocity was detected within the upper part of the crevasse field. Values of the 1st principal strain rate in the upper crevasse area and along the crevasse area limit line could indicate a threshold strain rate for crevasse opening between 0.1 and 0.2 day^-1 x 10^-3. The resolution and quality of the data used is considered to be too low to define and confirm this as a strain rate threshold for crevasse opening. The velocity mapping produced by feature tracking of TerraSAR-X images only detects velocity in about half the size of the mapped terminal crevasse field on Tunabreen. When applying the method to the fast-flowing glacier Austre Torellbreen, it is seen that the area of detected velocity is almost equal in extent to the mapped crevasse area. In a safety perspective, crevasse maps based on TerraSAR-X data give a more accurate indication of potential hazardous areas compared to velocity maps.
PublisherThe University of Bergen
Copyright the Author. All rights reserved