Improved Monitoring of Antarctic Ice Shelves Using the Sentinel-1 TOPSAR Acquisition Mode

The Antarctic Ice Sheet is the largest mass of ice on Earth. In contact with the ocean, the ice starts to float and forms an ice shelf. These ice shelves have an important role in the stability of the Antarctic Ice Sheet. They are acting as regulators by restraining the upwards ice flow. The thinning or damage of these ice shelves is directly translated into a decrease of their holding capacity and causes an acceleration of the ice flow and a retreat of the grounding line.

In parallel, the European Space Agency developed the Sentinel-1 Synthetic Aperture Radar – SAR satellite, part of the Copernicus program. Sentinel-1 aims at the systematic observation of the Earth in the C-Band radar frequency (5.54 GHz), with a revisit time of 6 to 12 days. This thesis aims to study the ability of Sentinel-1 to improve the monitoring of ice shelves, opening up new opportunities in terms of capturing small-scale events.

To begin with, we studied the ice flow velocity on ice shelves using speckle tracking or differential interferometry. SAR interferometry allows measuring local displacements between two acquisition dates. Classical interferometry allows deriving this velocity component only along the line-of-sight direction. This limitation may be overcome using the specificities of the TOPSAR acquisition mode, allowing a bidimensional estimation of the displacements.

The ice-motion estimation is far from trivial, and the specific case of ice shelves introduces additional problems in repeat-pass SAR studies. First, the rapid changes due to high displacements induce a rapid temporal decorrelation. We present a technique, coherence tracking, that aims to correct interferograms from these surface movements. Second, impacted by ocean tides and the inverse barometer effect, ice shelves are subject to vertical variations, inducing a bias to be removed. In the thesis, we present an empirical method using double interferograms and a climate model that corrects these false signals. These double interferograms can also be exploited to map the grounding line at a high temporal and spatial resolution over the Roi Baudouin Ice Shelf over multiple years.

Finally, focusing on amplitude information, we can analyze spatial patterns to monitor the ice shelves crevasses and calving front, and their evolution. Similarly, time-series analysis of SAR backscattering allows the determination of melt seasons and their spatial variability.

All these techniques participate in the derivation of new information about ice-shelves dynamics. From the results we obtained and the abundant literature, we conclude that the Sentinel-1 mission is a major step towards an improved monitoring of Antarctic ice-shelves dynamics.