Surface melt outweighs ice discharge over the next three centuries in fully coupled MAR-GISM simulations

When studying the future evolution and sea level contribution of the Greenland ice sheet, a realistic representation of ice sheet–atmosphere interactions in simulations is crucial. Therefore, to analyse the ice sheet evolution over the coming three centuries, we have performed several fully coupled ice sheet–regional climate model simulations. Our two-way coupled MAR–GISM simulations were driven by IPSL-CM6A-LR output under the SSP5-8.5 scenario, available up to 2300, and outlet glacier retreat was included through an empirical retreat parametrization.To disentangle the long-term importance of ice mass loss through surface mass balance (SMB) versus marine discharge, we compare simulations with only atmospheric or oceanic forcing to simulations with both forcings applied simultaneously. They indicate that both atmospheric and oceanic forcing individually still lead to a similar sea level contribution by 2100. But, by 2300 the SMB-driven ice mass loss is about five times larger than the discharge-driven ice mass loss in our simulations, as the ice sheet retreats on land and gradually loses contact with the ocean. Besides, an analysis of the SMB components and freshwater fluxes between simulations demonstrates that surface melting and ice discharge through the ice–ocean boundary are mutually competitive processes.Lastly, in terms of total ice mass loss, the importance of the chosen sensitivity parameter in the retreat parametrization increases over time. Whereas the difference in ice mass loss contribution from SMB versus discharge attenuates between simulations of differing sensitivity, because surface melting becomes increasingly dominant relative to marine discharge. Nevertheless, our simulations indicate that the applicability of the empirical retreat parametrization, which was developed for recent observations, becomes questionable over time.