An alternative representation of Synthetic Aperture Radar images as an aid to the interpretation of englacial observations

Abstract. Ground-penetrating radar reveals subsurface geometry and ice stratigraphy that contains information about past and present dynamics of the cryosphere. Synthetic Aperture Radar (SAR) is a processing technique based on averaging the radar echoes received at multiple locations as the radar moves relative to the target. Due to this averaging, the directional properties of the back-scattering from the target received at these multiple locations are lost. We introduce an alternative representation of SAR images that preserves directional information encoded in its Doppler spectrum: the Doppler frequency accounts for the time-delay variation from the radar to the target. With this technique, called Red–Green–Blue Doppler Decomposition (RGB-DD), the Doppler spectrum of a SAR image is split into three equalised bands, each band representing a primary direction of arrival of the radar echoes. A primary colour is assigned to each band to allow joint representation in a single RGB image. We apply our representation framework to several datasets acquired with the British Antarctic Survey (BAS) airborne ice-sounding radar over three Antarctic ice streams. Compared to the standard SAR method that is based solely on the averaged intensity level, this method facilitates the enhanced interpretation of englacial features such as ice stratigraphy, crevasses, tephra layers, and along-flow transitions in strain rate. The technique may be extended to other sensors and applications.