Timing and pathways of East Antarctic Ice Sheet retreat

A lack of reliable chronology often leaves Antarctic ice sheet reconstructions incomplete. Despite being the most heavily investigated region of the Antarctic continental shelf, the post-Last Glacial Maximum (LGM) record of marine-based ice sheet dynamics in the Ross Sea has remained largely elusive and at odds with fringing terrestrial records of ice sheet decay. Issues with radiocarbon dating stem primarily from poorly preserved carbonate, contamination by glacially-recycled carbon, and insufficient consideration for glacial geomorphic context and sedimentary facies. We assess nearly 70 newly acquired and more than 200 published radiocarbon ages in the context of a consistent sedimentary facies and geomorphic framework to interpret the timing of the local glacial maximum and subsequent retreat of the East Antarctic Ice Sheet (EAIS) in the Ross Sea and potential causes of retreat. Compound-specific radiocarbon ages help to provide good constraints on open marine onset, and we demonstrate that, with proper corrections and considerations, acid insoluble organic (AIO) radiocarbon ages can be reliable in open marine sediments and carbonate ages can be reliable in grounding-zone proximal sediments. We find that the ice flowing through Pennell and JOIDES troughs remained at maximum extent for approximately 10,000 years before retreating at 15.1 and 13 cal ka BP, respectively. An ice shelf covered the outer continental shelf until ∼9.5 cal ka BP, after which it collapsed in JOIDES Trough, creating deep iceberg furrows. Retreat through Pennell Trough was more episodic, marked by grounding-zone wedges that backstep upstream and onto lateral banks creating deglacial emergent ice rises. As ice retreated from ∼9 to ∼4 cal ka BP, there was considerable reorganization of drainage and at least one episode of readvance through southern JOIDES Trough that began drawdown of terrestrial ice near Ross Island at ∼7.8 cal ka BP. The ice shelf appears to have retreated to a near-present configuration by 2 cal ka BP. However, due to inconsistencies in calendar ages, grounding line position must be inferred from relative chronology based on glacial landforms, which suggests grounded ice retreated nearly to Drygalski Ice Tongue and lingered atop Crary Bank where it was maintained by ice from David Glacier. Our reconstruction further suggests that the EAIS in the western Ross Sea was not a discernible contributor to Meltwater Pulse 1A, but likely contributed to eustatic sea level rise well through the middle—perhaps even late—Holocene.