Impact of Grids and Dynamical Cores in CESM2.2 on the Surface Mass Balance of the Greenland Ice Sheet

• Published in: Journal of Advances in Modeling Earth Systems
• Main Authors: Adam R. Herrington, Peter H. Lauritzen, Marcus Lofverstrom, William H. Lipscomb, Andrew Gettelman, Mark A. Taylor
• Format: Article in Journal/Newspaper
• Language: English
• Published: American Geophysical Union (AGU) 2022
• Subjects: global climate models; ice sheets; modeling; Physical geography; GB3-5030; Oceanography; GC1-1581; Greenland; Ice Sheet
• Online Access: https://doi.org/10.1029/2022MS003192; https://doaj.org/article/344ed1292bc447bd8ef334e0bf75a821

Abstract Six different configurations, a mixture of grids and atmospheric dynamical cores available in the Community Earth System Model, version 2.2 (CESM2.2), are evaluated for their skill in representing the climate of the Arctic and the surface mass balance of the Greenland Ice Sheet (GrIS). The finite‐volume dynamical core uses structured, latitude‐longitude grids, whereas the spectral‐element dynamical core is built on unstructured meshes, permitting grid flexibility such as quasi‐uniform grid spacing globally. The 1°–2° latitude‐longitude and quasi‐uniform unstructured grids systematically overestimate both accumulation and ablation over the GrIS. Of these 1°–2° grids, the latitude‐longitude grids outperform the quasi‐uniform unstructured grids because they have more degrees of freedom to represent the GrIS. Two Arctic‐refined meshes, with 1/4° and 1/8° refinement over Greenland, were developed for the spectral‐element dynamical core and are documented here as newly supported configurations in CESM2.2. The Arctic meshes substantially improve the simulated clouds and precipitation rates in the Arctic. Over Greenland, these meshes skillfully represent accumulation and ablation processes, leading to a more realistic GrIS surface mass balance. As CESM is in the process of transitioning away from conventional latitude‐longitude grids, these new Arctic‐refined meshes improve the representation of polar processes in CESM by recovering resolution lost in the transition to quasi‐uniform grids, albeit at increased computational cost.