Impacts of Model Resolution on the Simulation of Sea-level Variability by a Global Ocean-Sea Ice Model

This is the authors' abstract. We don't add key points for this paper.

Abstract Previous studies have explored the impacts of model resolution on sea-level simulations, but no quantitative conclusion has been derived. To this end, in this research, we analyze the effects of model resolution on the simulation of sea-level variability based on the second-generation climate system ocean model from the State Key Laboratory of Numerical Modeling for Atmospheric Science and Geophysical Fluid Dynamics, Institute of Atmosphere Physics (LICOM2) with resolutions of 1° (LICOM2-L) and 0.1° (LICOM2-H). The results show that the spatial distribution and variation trends of the time-averaged dynamic sea level (DSL) simulated by the LICOM2-L and LICOM2-H are consistent with the observations. The magnitude of the zonal-averaged DSL from the LICOM2-H is closer to the observations, especially in high latitudes. Furthermore, the LICOM2-H simulations can also depict mesoscale characteristics in the western boundary currents and Antarctic Circumpolar Current (ACC) regions. The interannual variability, decadal variability, and variation trends of the DSL are estimated using a multivariate linear regression model based on the LICOM2-L and LICOM2-H datasets during 1958–2007. The distributions of the interannual variability, decadal variability, and variation trends of the LICOM2-L and LICOM2-H simulations are identical in the tropics and mid-latitudes, and the differences are also concentrated in the regions of the western boundary currents and ACC, which are primarily caused by the thermosteric sea level (TSSL) and the halosteric sea level. In contrast, the DSL variability differences in the Southern Ocean are mainly due to the TSSL. Analyses of ocean heat content (OHC) budgets suggest that the variations in the OHC drive variations in the TSSL. The differences between the LICOM2-L and LICOM2-H simulations are mainly in decadal variability and variation trends. Both large-scale mean advection and eddy-induced transport are responsible for the differences.