Rahlves et al. (2025) Investigating the multi-millennial evolution and stability of the Greenland ice sheet using remapped surface mass balance forcing

Identification

• Journal: ˜The œcryosphere
• Year: 2025
• Date: 2025-12-02
• Authors: Charlotte Rahlves, Heiko Goelzer, Andreas Born, Petra M. Langebroek
• DOI: 10.5194/tc-19-6403-2025

Research Groups

• NORCE Research AS, Bjerknes Centre for Climate Research, Bergen, Norway
• Department of Earth Science, University of Bergen, Bjerknes Centre for Climate Research, Bergen, Norway

Short Summary

This study introduces and evaluates an SMB remapping procedure for stand-alone ice sheet models to efficiently simulate the multi-millennial evolution of the Greenland ice sheet. The remapping method effectively captures first-order climate-ice sheet feedbacks, preserving ablation zone structure and reducing biases compared to conventional parameterizations, leading to more realistic long-term mass loss projections.

Objective

• To introduce and evaluate an SMB remapping scheme for stand-alone ice sheet models to efficiently capture first-order climate-ice sheet feedbacks, comparing its impact on the multi-millennial evolution and stability of the Greenland ice sheet under various CMIP6 emission scenarios against conventional SMB-elevation feedback parameterizations.

Study Configuration

• Spatial Scale: Greenland ice sheet, simulated on a 4 km Cartesian grid, divided into 25 regional drainage basins, with SMB lookup tables constructed using 100 m elevation intervals.
• Temporal Scale: Multi-millennial simulations spanning 10,000 years, preceded by a 5000-year spin-up and a 1000-year relaxation, and a historical simulation from 1960 to 2014. Climate forcing beyond 2100 is extended by repeating the mean SMB from 2081–2100.

Methodology and Data

• Models used:
  • Ice Sheet Model: Community Ice Sheet Model (CISM) v2.1
  • Regional Climate Model (RCM): MAR v3.12 (Modèle Atmosphérique Régional)
  • Glacial-Isostatic Adjustment (GIA) Model: Elastic Lithosphere, Relaxing Asthenosphere (ELRA) model
• Data sources:
  • Climate forcing: Dynamically downscaled output from six CMIP6 Earth System Models (ESMs) and Shared Socio-economic Pathways (SSPs) (e.g., NorESM2-MM-SSP5-8.5, UKESM1-0-LL-SSP5-8.5, MPI-ESM1-2-HR-SSP5-8.5, NorESM2-MM-SSP2-4.5, CESM2-SSP1-2.6, MPI-ESM1-2-HR-SSP1-2.6).
  • Reference Surface Mass Balance (SMB) and Surface Temperature (ST): ERA5 reanalysis (1960–1989 mean).
  • Bedrock topography and present-day ice sheet geometry: BedMachine v3.
  • Observed outlet-glacier retreat: Prescribed using retreat masks based on observations.

Main Results

• On centennial timescales (up to 2100), differences in total ice mass loss between various SMB forcing methods are small (within 3.4 gigatonnes), indicating less criticality of SMB-elevation feedback treatment.
• On multi-millennial timescales (up to 10,000 years), differences become significant:
  • Simulations neglecting SMB-elevation feedback severely underestimate mass loss.
  • The parameterized SMB-elevation feedback method (based on runoff gradients) leads to premature ice sheet stabilization (around 1.6 x 10^6 gigatonnes after approximately 2000 years) as margins retreat to regions with weak gradients.
  • The SMB and SMB anomaly remapping approaches produce more extensive and persistent ablation zones, driving continued ice sheet retreat and preventing stabilization even after 10,000 years under high forcing.
  • Total SMB remapping and SMB anomaly remapping yield very similar large-scale mass loss outcomes, though total SMB remapping produces slightly steeper negative SMB gradients at retreating margins.
  • Including glacial-isostatic adjustment (GIA) significantly reduces ice loss, maintaining approximately 25% more ice mass after 10,000 years by uplifting bedrock and reducing melt rates.
• Under high-emission scenarios (e.g., UKESM1-0-LL-SSP5-8.5), the Greenland ice sheet undergoes complete disintegration shortly after 4000 years.
• Under low-emission scenarios (e.g., CESM2-SSP1-2.6), the ice sheet retains 85%–95% of its present-day mass and stabilizes within 2000 to 4000 years.
• Intermediate high-emission scenarios (NorESM2-MM-SSP5-8.5, MPI-ESM1.2-SSP5-8.5) show ongoing retreat beyond 10,000 years, with mass losses of 55% and 47% of initial mass, respectively, without reaching a new equilibrium.

Contributions

• Introduces and evaluates a novel SMB remapping scheme that dynamically adapts SMB forcing to evolving ice sheet geometry within a computationally efficient stand-alone ice sheet model.
• Demonstrates that this remapping method better preserves the spatial structure of the ablation zone and reduces non-physical biases compared to conventional SMB-elevation feedback parameterizations, particularly over multi-millennial timescales.
• Provides a valuable, computationally efficient alternative to fully coupled climate-ice sheet models for long-term and ensemble simulations of Greenland ice sheet evolution.
• Shows that a simplified extension of SMB forcing (using a 20-year average beyond 2100) does not significantly alter long-term simulation outcomes.

Funding

• The Research Council of Norway (project 324639, GREASE)
• European Union’s Horizon 2020 research and innovation programme (grant agreement 869304, PROTECT)
• Sigma2 – the National Infrastructure for High Performance Computing and Data Storage in Norway (projects NN8085K, NN8006K, NS5011K, NS8006K, NS8085K)

Citation

@article{Rahlves2025Investigating,
  author = {Rahlves, Charlotte and Goelzer, Heiko and Born, Andreas and Langebroek, Petra M.},
  title = {Investigating the multi-millennial evolution and stability of the Greenland ice sheet using remapped surface mass balance forcing},
  journal = {˜The œcryosphere},
  year = {2025},
  doi = {10.5194/tc-19-6403-2025},
  url = {https://doi.org/10.5194/tc-19-6403-2025}
}