Simulating the variations of carbon dioxide in the global atmosphere on the hexagonal grid of DYNAMICO coupled with the LMDZ6 model

Abstract. Efforts to monitor the emissions and absorptions of atmospheric carbon dioxide (CO₂) over the globe and to understand their varying regional patterns with greater accuracy have intensified in recent years. This study evaluates the performance of a new model coupling, ICO, built around the Laboratoire de Météorologie Dynamique atmospheric general circulation model (LMDZ) for simulating CO₂ transport. ICO utilizes the new icosahedral hydrostatic dynamical core called Dynamico running on an unstructured grid, which enables potential improvements in spatial resolution at the Equator while removing artificial distortions and numerical filters at the poles. Comparisons with a reference configuration using a structured latitude-longitude grid reveal that ICO well captures seasonal variations in CO₂ concentrations at surface stations. While not significantly enhancing the capture of complex seasonal patterns, ICO maintains comparable accuracy. Both configurations exhibit similar vertical CO₂ concentration profiles and display a consistent bias in the lower stratosphere relative to observational data. ICO demonstrates advantages in computational efficiency and storage, thanks to its reduced cell count per level and a homogeneous grid structure. It holds promise for future developments, including with the LMDZ offline model and associated inversion system, which contribute to the Copernicus Atmosphere Monitoring Service. Overall, the ICO configuration showcases the efficacy of utilizing an unstructured grid for the physics, and the capability of Dynamico in accurately simulating CO₂ transport. This study emphasizes the importance of advanced modeling approaches and high-resolution innovative grids in enhancing our understanding of the global carbon cycle and refining climate models.