ICE VELOCITY ESTIMATION OF SOUTHWEST GREENLAND USING MULTI-TEMPORAL SYNTHETIC APERTURE RADAR OBSERVATIONS

Abstract:

Glaciers have been focused on climate change and sea-level rise directly related to its loss from global warming. Spaceborne synthetic aperture radar (SAR) observations enable us to detect a long term of dramatic changes over a very wide swath of the glacier. Although differential radar interferometry (DInSAR) has been used for a very powerful method to estimate glacier movement with high accuracy of from cm to mm scale, it is often limited by degraded coherence due to significant temporal and volume decorrelation from fast ice velocity. On the contrary, because offset tracking does not rely on interferometric phase coherence, it could be applied for fast-moving areas such as glacier surface. Hence, we adopted intensity offset tracking, which is capable to apply even if the displacement is large, using estimation of the displacement determined by cross-correlation between two SAR amplitude images. The accuracy and computing efficiency of the offset tracking depends on moving window patch size to compare the same features between the two images. In this study, ice velocity of Narsap sermia and Akullersuup Sermia located in southwest Greenland has been observed by using TerraSAR-X/TanDEM-X and Sentinel-1A/B SAR observations. These two glaciers are tidal outlet glaciers that are characterized by flowing directly and terminating to the sea and its terminus calves to form iceberg. We acquired 253 images acquired from July 22, 2013 to August 10, 2019 at ascending direction with VV and HH linear polarization. The multi-temporal SAR observations of the ice velocity changes over about six years present. A time-series of ice velocity calculated with a 6 to 11-day temporal baseline shows seasonal and spatial velocity variations. We extracted profiles by stacking the results from the multi-temporal observations for statistical analysis. The Nuuk’s temperature data, sea and land surface temperature product from Moderate Resolution Imaging Spectroradiometer (MODIS) sensor was utilized to compare with our result.