Investigation of ice features with unknown glaciological origin in the ablation zone of southwest Greenland

Kurzfassung

In recent years, remote sensing methods have become an essential tool for monitoring polar regions. Microwaves are able to penetrate dry snow and ice, rendering Synthetic Aperture Radar (SAR) measurements sensitive to the subsurface structure of glaciers and ice sheets, particularly at lower frequencies. In the ablation zone of southwest Greenland, SAR data have revealed ice features of unknown glaciological origin. These ice features, here referred to as radar-dark and radar-bright features, initially distinguished by different backscatter characteristics, were investigated using multi-modal SAR data collected during an airborne campaign with DLR’s F-SAR system. A multi-modal approach was essential to combine complementary insights, allowing the integration of advanced SAR techniques such as polarimetry, interferometry, tomography, temporal analysis, and modeling to characterize surface and subsurface processes of these features and explore their origins. Radar-dark features exhibit low backscatter, dominated by scattering at the surface. They typically appear as round or oval shapes, ranging from 100 to 500 m and frequently form interconnected spatial patterns. Conversely, radar-bright features demonstrate high backscatter, predominantly originating from the subsurface, which results in pronounced spatial contrasts between both features. The use of SAR polarimetry has revealed key differences in scattering mechanisms. Radar-dark features were observed to exhibit low entropy and mean alpha angles, which is consistent with the dominance of surface scattering. In contrast, radar-bright features demonstrated higher entropy and alpha angles, indicating the presence of more complex scattering processes with significant volume scattering. The analysis of interferometric SAR data provided insights into the phase center heights, confirming the shallow radar penetration observed in radar-dark features and the deeper penetration evident in radar-bright features. Tomographic reconstructions provided a three-dimensional perspective, revealing that very weak subsurface structures were present in radar-dark features, while pronounced subsurface structures were present in radar-bright features. Further analysis demonstrated that radar-bright features consistently contain a subsurface scattering layer situated between 15 and 50 m below the surface. Scattering models, validated against observed coherence values, supported these findings. For radar-bright features, a two-component model corresponding to the surface and subsurface scattering components was suitable, while radar-dark features required a three-component model to account for surface dominance and weak subsurface contributions. The temporal analysis of ALOS-2 data revealed that radar-dark features are relatively stable over time, despite significant surface melt rates, moving with glacier flow. The absence of correlation between these features and surface topography or roughness indicates a subsurface origin. It is hypothesized that these features result from the formation of a weathering crust caused by the infiltration of meltwater and subsequent refreezing, which creates a porous layer at the near-surface level. The presence of residual liquid water within this crust, when it is in a frozen state, is likely to attenuate radar signals, thereby reducing both penetration and subsurface scattering. The spatial distribution of radar-dark features may also be influenced by the presence of former supraglacial lakes, which may affect the location of weathering crust formation. In contrast, radar-bright features are characterized by frozen ice without liquid water contents. These regions are presumed to be areas of bare ice, where the absence of a weathering crust serves to minimize signal attenuation and enhance penetration into the subsurface. Overall, this study emphasizes the importance of multi-modal SAR data in the characterization of ice features within the ablation zone of southwest Greenland. By integrating multiple SAR techniques, we were able to identify fundamental differences in scattering mechanisms, surface and subsurface processes, and temporal changes. Nevertheless, to confirm the proposed hypothesis regarding the formation of radar-dark and radar-bright features, additional ground measurements and field observations are required. This further validation would serve to refine these findings, thereby enhancing their applicability to glacier modeling.