| Summary: | Ice shelves, the floating extensions of glaciers and ice sheets, create a safety band around Antarctica. They control the flow of ice that drains into the ocean by buttressing the upstream grounded ice. Loss of ice shelf stability and integrity results in reduced buttressing and leads to increased discharge, i.e. contribution to global sea level rise. Accurate predictions of rates of sea level rise therefore require monitoring of ice shelf dynamics. So far, the potential of SAR data for this task has hardly been exhausted. While comprehensive products exist for recent years, data of early SAR satellites has up to now only been used to a very limited extent. To fill this research gap, this thesis exploits the entire ERS and Envisat archive within West Antarctic Pine Island Bay, a region that requires particular attention due to drastic on-going changes. A deep learning based semantic segmentation approach is applied to derive a 20-year time series (1992-2011) of ice shelf front dynamics from nearly 2000 available scenes. The used HED-UNet framework proved to be transferable to this purpose. For selected test scenes, the detected ice shelf fronts deviate on average by 355 m from manual references, the segmentation accuracy is 96%. The resulting product of yearly, seasonal and monthly time series reveals individual dynamic patterns for all five investigated ice shelves. The most considerable fluctuations were found for Pine Island Ice Shelf in terms of frequency of calving events (multiple cycles of calving and re-advance) and Thwaites ice tongue in terms of size of break-up (80 km retreat in early 2002). Despite different change rates and magnitude, most ice shelves show similar signs of destabilisation. This not only manifests through retreating front positions, but also through the ice shelf geometry. Signs of weakening appear in the form of fracturing, rifting, disintegration events and the development of passive parts that no longer contribute to the buttressing effect.
|
|---|