Description of the resolution hierarchy of the global coupled HadGEM3-GC3.1 model as used in CMIP6 HighResMIP experiments

Roberts, M. J., Baker, A. ORCID: https://orcid.org/0000-0003-2697-1350, Blockley, E. W., Calvert, D., Coward, A., Hewitt, H. T., Jackson, L. C., Kuhlbrodt, T. ORCID: https://orcid.org/0000-0003-2328-6729, Mathiot, P., Roberts, C. D., Schiemann, R. ORCID: https://orcid.org/0000-0003-3095-9856, Seddon, J., Vanniere, B. ORCID: https://orcid.org/0000-0001-8600-400X and Vidale, P. L. ORCID: https://orcid.org/0000-0002-1800-8460 (2019) Description of the resolution hierarchy of the global coupled HadGEM3-GC3.1 model as used in CMIP6 HighResMIP experiments. Geoscientific Model Development, 12 (12). pp. 4999-5028. ISSN 1991-9603 doi: 10.5194/gmd-12-4999-2019

Abstract/Summary

CMIP6 HighResMIP is a new experimental design for global climate model simulations that aims to assess the impact of model horizontal resolution on climate simulation fidelity. We describe a hierarchy of global coupled model resolutions based on the HadGEM3-GC3.1 model that range from an atmosphere-ocean resolution of 130 km-1° to 25 km-1/12°, all using the same forcings and initial conditions. In order to make such high resolution simulations possible, the experiments have a short 30 year spinup, followed by at least century-long simulations with both constant forcing (to assess drift and the focus of this work), and historic forcing. We assess the change in model biases as a function of both atmosphere and ocean resolution, together with the effectiveness and robustness of this new experimental design. We find reductions in the biases in top of atmosphere radiation components and cloud forcing. There are significant reductions in some common surface climate model biases as resolution is increased, particularly in the Atlantic for sea surface temperature and precipitation, primarily driven by increased ocean resolution. There is also a reduction in drift from the initial conditions both at the surface and in the deeper ocean at higher resolution. Using an eddy-present and eddy-rich ocean resolution enhances the strength of the North Atlantic ocean circulation (boundary currents, overturning circulation and heat transports), while an eddy-present ocean resolution has a considerably reduced Antarctic Circumpolar Current strength. All models have a reasonable representation of El Nino – Southern Oscillation. In general the biases present after 30 years of simulations do not change character markedly over longer timescales, justifying the experimental design.

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/86816
Item Type	Article
Refereed	Yes
Divisions	Science > School of Mathematical, Physical and Computational Sciences > NCAS Science > School of Mathematical, Physical and Computational Sciences > Department of Meteorology
Publisher	European Geosciences Union
Download/View statistics	View download statistics for this item