Widespread near-coastal bedrock erosion surfaces in East Antarctica and their implications for long-term ice-sheet behaviour

The sub-ice topography of East Antarctica provides a crucial record of the long-term geological, geomorphological, and glaciological evolution of the continent. In particular, the morphology of the East Antarctic Ice Sheet (EAIS) bed is a valuable and hitherto underexploited archive of past ice-sheet behaviour. Analysis of the subglacial landscape can therefore help improve our understanding of the response of the ice sheet to episodes of warming in the geological past that serve as analogues for current and projected future climate change. Here, we conduct a systematic search of the extensive repositories of airborne ice-penetrating radar data acquired in the past two decades to map the distribution of low-relief subglacial bed surfaces close to the East Antarctic ice margin between Princess Elizabeth Land and George V Land (70°E to 160°E). Individual surfaces are characterised by consistent elevations over distances of 10s to 100s of kilometres and relatively low-amplitude, high-frequency roughness (i.e., valleys and inselbergs). We map 31 separate low-relief bed surfaces, which range from 500 to 50,000 km2 in area and comprise ~40% of the perimeter of this sector of the East Antarctic margin. The surfaces are typically overlain by cold-based, slow-moving ice and bounded by deep subglacial troughs that host fast-flowing ice streams and outlet glaciers. Underneath the modern-day EAIS, these low-relief bed surfaces are situated at a broad range of elevations. However, when the elevations are isostatically adjusted for the removal of the EAIS, the distribution narrows substantially and, alongside cluster analysis of the morphology of the surfaces, indicates that they constitute a single, statistically consistent population around the entirety of this sector of the EAIS margin. Under ice-free conditions, the coastal surfaces would be situated above sea level and gently dipping in a seaward direction, and we suggest that they are remnants of a widespread fluvial planation surface formed following Gondwana break-up and preserved with only minor geomorphological modification since EAIS inception. The presence of these ancient surfaces has important implications for the past, present, and future behaviour of this sector of the EAIS.