Iceberg Census and Tracking in the Weddell Sea Utilizing Satellite Data
Final Report Abstract
Within the project, an iceberg detection method was developed and applied to a region with difficult sea ice conditions. The development of the iceberg detection method required more time than expected. The first results, presented in the progress report after one year, were improved and a new commercial software package was applied for pre-processing. Problems with this software package required that all single-channel SAR images (ENVISAT and ERS-2) had to be pre-processed twice. After some performance tests the development of a detection method was finalized using the theoretical K-distribution for modeling the backscatter intensities of icebergs. The derivation of threshold values for separating icebergs, sea ice, and open water was based on ENVISAT images at different resolutions, which were recorded over the southern Weddell Sea region, and on ERS-2 images recorded over the tip of the Antarctic Peninsula. A first estimation of iceberg masses within a selection of SAR scenes was provided in the article and revealed very large uncertainties in our areas of investigation. On average, the iceberg areas were overestimated in the automated procedure. Another problem was that the sea ice in the area of investigation was strongly deformed, which induced higher backscattering coefficients similar to the ones of the icebergs and hence caused wrong classification results. To check whether this problem can be reduced, the potential of polarimetric SAR data for iceberg detection was tested. Five fully-polarized RADARSAT-2 images in two areas of investigation (Bellingshausen Sea and southern Weddell Sea, see Table 1) were available for this study. The polarimetric SAR improves the performance of iceberg detection to a certain degree, but there are still uncertainties in the clear distinction between icebergs and their surroundings. Parallel to the development of the iceberg detection method, the iceberg drift model was implemented into a Finite Elemente Sea ice and Ocean Model (FESOM). Within the framework of this work, sensitivity studies on the influence of the different forces (e.g. wind drag, ocean drag) on the iceberg drift have been carried out. The buoy data presented in Schodlok et al. [2006] were used as a reference dataset. Additionally, the drift of the huge iceberg D18 was modeled and compared with SAR observations. To improve the results provided by both studies, the model is currently linked to the latest FESOM version, which has a higher spatial resolution. Results are yet not available. The determination of the calving rates could not be carried out due to lack of time. This is now one of the goals of our current work. This new DFG project is regarded a follow-up of the work described here.
Publications
-
(2009), Detection of icebergs using ERS SAR images, Cryospheric Science Workshop 4.11.-6.11.2009 Sheffield, UK
Wesche, C. and Dierking, W.
-
(2010), Separation of icebergs and sea ice in SAR images, Proceedings of ESA Living Planet Symposium, 28.6.-2.7.2010 Bergen, Norway. European Space Agency, H. Lacoste-Francis, (ed.) CD-ROM, Nordwijk (ESA-SP 686)
Wesche, C. and Dierking, W.
-
(2012). Iceberg signatures and detection in SAR images in two test regions of the Weddell Sea, Antarctica, Journal of Glaciology, 58(208), 325-339
Wesche, C. and W. Dierking
