Antarctic contribution to future sea level from ice shelf basal melt as constrained by ice discharge observations

Abstract

Antarctic mass loss is the largest contributor to uncertainties in sea level projections on centennial time scales. In this study we aim to constrain future projections of the contribution of Antarctic dynamics by using ice discharge observations. The contribution of Antarctica's ice discharge is computed with ocean thermal forcing from 14 Earth system models (ESMs) and linear response functions (RFs) from 16 ice sheet models for 3 shared socioeconomic pathway (SSP) scenarios. New compared to previous studies, basal melt sensitivities to ocean temperature changes were calibrated on four decades of observed ice discharge changes rather than using observation-based basal melt sensitivities. Calibration improved historical performance but did not reduce the uncertainty in the projections. The results show that even with calibration the acceleration during the observational period is underestimated for the Amundsen Region, indicating that ice and/or ocean processes are not well represented. Also the relative contribution of the Amundsen Region is underestimated. The Amundsen Region contribution and sea level acceleration are improved by choosing an Amundsen Region-specific calibration (rather than Antarctic-wide), quadratic basal melt parameterisation (rather than linear) and thermal forcing near the ice shelf base (rather than the deepest layer above the continental shelf). With these methodological choices we arrive at a median dynamic sea level contribution of 0.12 m for SSP1-2.6, 0.14 m for SSP2-4.5 and 0.17 m for SSP5-8.5 in 2100 relative to 1995–2014, sitting in between projections of previous multimodel studies (ISMIP6 emulator and LARMIP-2). Our results show that constraining the basal melt parameterisation on Amundsen Region ice discharge rather than applying the median basal melt sensitivities used in LARMIP-2 and the mean Antarctic distribution of ISMIP6 leads to higher sea level contributions. However, differences in basal melt sensitivities alone cannot explain the differences in our projections compared to the emulated ISMIP6 and LARMIP-2. We conclude that uncertainties associated with ESMs and ice sheet models affect the projected sea level contribution more than our methodological choices in the calibration and basal melt computation.