Extending the range and reach of physically-based Greenland ice sheet sea-level projections

Abstract. We present an ensemble of physically-based ice sheet model projections for the Greenland ice sheet (GrIS) that was produced as part of the European project PROTECT. Our ice sheet model (ISM) simulations are forced by high-resolution regional climate model (RCM) output and other climate model forcing, including a parameterisation for the retreat of marine-terminating outlet glaciers. The experimental design builds on the Ice Sheet Model Intercomparison Project for CMIP6 (ISMIP6) protocol and extends it to more fully account for uncertainties in sea-level projections. We include a wider range of CMIP6 climate model output, more climate change scenarios, several climate downscaling approaches, a wider range of sensitivity to ocean forcing and we extend projections beyond the year 2100 up to year 2300, including idealised overshoot scenarios. GrIS sea-level rise contributions range from 16–76 mm (SSP1-2.6/RCP2.6), 22–163 mm (SSP2-4.5) and 27–354 mm (SSP5-8.5/RCP8.5) in the year 2100 (relative to 2014). The projections are strongly dependent on the climate scenario, moderately sensitive to the choice of RCM, and relatively insensitive to the ice sheet model choice. In year 2300, contributions reach 49 to 3127 mm, indicative of large uncertainties and a potentially very large long-term response. Idealised overshoot experiments to 2300 produce sea-level contributions in a range from 49 to 201 mm, with the ice sheet seemingly stabilised in a third of the experiments. Repeating end of the 21st century forcing until 2300 results in contributions of 58–163 mm (repeated SSP1-2.6), 98–218 mm (repeated SSP2-4.5) and 282–1230 mm (repeated SSP5-8.5). The largest contributions of more than 3000 mm by year 2300 are found for extreme scenarios of extended SSP5-8.5 with unabated warming throughout the 22nd and 23rd century. We also extend the ISMIP6 forcing approach backwards over the historical period and successfully produce consistent simulations in both past and future for three of the four ISMs. The ensemble design of ISM experiments is geared towards the subsequent use of emulators to facilitate statistical interpretation of the results and produce probabilistic projections of the GrIS contribution to future sea-level rise.