The Greenland-Ice-Sheet evolution over the last 24 000 years: insights from model simulations evaluated against ice-extent markers

Continental ice sheets retain a long-term memory stored in their geometry and thermal properties. In Greenland, this creates a disequilibrium with the present climate, as the ice sheet is still adjusting to past changes that occurred over millennial timescales. Data-consistent modelling of the paleo Greenland-Ice-Sheet evolution is thus important for improving model initialisation in future projection experiments. Open questions also remain regarding the ice sheet's former volume, extent, flux, internal flow dynamics, thermal conditions, and how such properties varied in space since the last glaciation. Here, we conduct a modelling experiment that aims to produce simulations in agreement with empirical data on Greenland's ice-margin extent and timing over the last 24 000 years. Given uncertainties in model parameterisations, we apply an ensemble of 100 ice-sheet-wide simulations at 5×5 km resolution using the Parallel Ice Sheet Model, forced by simulations from the isotope-enabled Community Earth System Model. Using a new Greenland-wide reconstruction of former ice margin retreat (PaleoGrIS 1.0), we score each simulation's fit from 24 000 years ago to 1850 AD. The results provide insights into the dynamics, drivers, and spatial heterogeneities of the local Last Glacial Maximum, Late-glacial, and Holocene evolution of the Greenland Ice Sheet. For instance, we find that between 16 and 14 thousand years ago, the ice sheet lost most ice grounded on the continental shelf. This marine-sector retreat, associated with mass loss rates up to seven times greater than today's, was likely mainly driven by ocean warming. Our model–data comparison results also show regional heterogeneities in fit and allows estimating agreement-score sensitivity to parameter configurations, which should prove useful for future paleo-ice-sheet modelling studies. Finally, we report remaining model-data misfits in ice extent, here found to be largest in northern, northeastern, and central-eastern Greenland, and discuss possible causes for these.