Regional grid refinement in an Earth system model: impacts on the simulated Greenland surface mass balance

• Published in: The Cryosphere
• Main Authors: van Kampenhout, Leonardus, Rhoades, Alan M., Herrington, Adam R., Zarzycki, Colin M., Lenaerts, Jan T. M., Sacks, William J., van den Broeke, Michiel R.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2019
• Subjects: article; Verlagsveröffentlichung; Greenland; Ice Sheet; The Cryosphere
• Online Access: https://doi.org/10.5194/tc-13-1547-2019; https://noa.gwlb.de/receive/cop_mods_00002043; https://noa.gwlb.de/servlets/MCRFileNodeServlet/cop_derivate_00002001/tc-13-1547-2019.pdf; https://tc.copernicus.org/articles/13/1547/2019/tc-13-1547-2019.pdf

In this study, the resolution dependence of the simulated Greenland ice sheet surface mass balance (GrIS SMB) in the variable-resolution Community Earth System Model (VR-CESM) is investigated. Coupled atmosphere–land simulations are performed on two regionally refined grids over Greenland at 0.5∘ (∼55 km) and 0.25∘ (∼28 km), maintaining a quasi-uniform resolution of 1∘ (∼111 km) over the rest of the globe. On the refined grids, the SMB in the accumulation zone is significantly improved compared to airborne radar and in situ observations, with a general wetting (more snowfall) at the margins and a drying (less snowfall) in the interior GrIS. Total GrIS precipitation decreases with resolution, which is in line with best-available regional climate model results. In the ablation zone, CESM starts developing a positive SMB bias with increased resolution in some basins, notably in the east and the north. The mismatch in ablation is linked to changes in cloud cover in VR-CESM, and a reduced effectiveness of the elevation classes subgrid parametrization in CESM. Overall, our pilot study introduces VR-CESM as a new tool in the cryospheric sciences, which could be used to dynamically downscale SMB in scenario simulations and to force dynamical ice sheet models through the CESM coupling framework.