Speaker
Description
In an experimental airborne SAR dataset, a complex and heterogeneous scattering pattern was found at the Russel glacier, also known as K-Transect, in western Greenland. In a first analysis, two potential ice types could be classified based on their distinct differences in scattering processes. A first glaciological theory for these two ice types is, on the one hand, temperate ice with low liquid water content and, on the other hand, cold and homogeneous glacial ice without liquid water content. A better understanding of the distribution and development of these two ice types could be a crucial information for surface-mass balance, glacier dynamics and hydro-glaciological processes [1].
Polarimetric and interferometric analyses of SAR data are a unique tool to investigate these ice types based on their sensitivity to different dielectric properties and scattering structures. In L-band, with a deep signal penetration into the ice sheet, the difference between the ice types is particularly pronounced which indicates a significant difference in subsurface structure of the two ice types [2]. In a Pauli decomposition of the research area, one ice type demonstrates a weak surface scattering, which is referred to as "black patches" due to its color in the RGB Pauli image. The other ice type shows as bright green which indicates strong volume scattering. The interferometric analysis confirms the first impression of the Pauli decomposition. A shallow penetration can be seen in the “black patches” with phase center depths between 0 and -2 m in HH and VV (0 to -5 m in HV), which aligns with high coherences, low volume decorrelation and surface scattering as the main scattering mechanism. The second ice type shows, in accordance with the high amount of volume scattering, a phase center depth up to -30m in HV (0 to -15m in HH and VV). The coherence decreases significantly over increasing baselines due a higher volume decorrelation.
Conventional tomographic analysis of the “black patches” presents mainly surface scattering in all polarizations, which triggered the question whether this phenomenon is only related to the surface or if there is scattering below. Therefore, additional tomographic analysis with further steps for surface cancelation [3] and noise filtering was conducted and multiple ice layers in the subsurface with a low intensity could be determined. However, possible artefacts due to the filtering, ambiguities and sidelobes have to be further investigated to separate them from the actual subsurface signal. To validate the results of the tomographic investigation, (Pol-)InSAR modeling of the coherences and phase center depth over all baselines was executed. Hereby, the vertical backscattering distribution was modelled for the two ice types. The ice type with high volume scattering could be well explained through a “box” as a surface component which accounts for the rough surface and a uniform volume in the subsurface. The modeling of the “black patches” still remains a challenge, especially the weak but non-negligible subsurface and will be further investigated with a combination of SAR methods.
For the glaciological context, an analysis of ALOS data indicates that this pattern originates higher up the ice sheet, starting possibly in the super-imposed ice zone. Further analysis of additional spaceborne and in-situ data will help to understand the glaciological origin and its formation processes.
[1] Paterson, W. S. B., “The physics of glaciers”, 3rd edition, Oxford (1994), doi: 10.1016/c2009-0-14802-x
[2] Pardini, Matteo et al., “A Multi-Frequency SAR Tomographic Characterization of Sub-Surface Ice Volumes”, European Conference on Synthetic Aperture Radar: Hamburg, Germany (2016)
[3] Joerg, Hannah et al., “On the Separation of Ground and Volume Scattering Using Multibaseline SAR Data”, IEEE Geoscience and Remote Sensing Letters, Vol. 14, No. 9 p. 1570-1574 (2017)
