A near-global eddy-resolving OGCM for climate studies

Abstract. Eddy-resolving global ocean models are highly desired for spatially-improved climate studies, but this is challenging because they require careful configuration and substantial computational resources. Model drift, partially related to insufficient model spin-up, imperfect model physics or bias in surface forcing, can be problematic, leading to contamination of climate change signals. In this study, we adapt a near-global eddy-resolving ocean general circulation model, originally developed for short-range ocean forecasting, for climate studies. The Ocean Forecasting Australia Model version 3 (OFAM3) is spun up for 20 years, with repeated year 1979 forcing and adaptive relaxation (Newtonian nudging) of temperature and salinity in the deep ocean to an observation-based climatology. In addition, surface heat fluxes from the JRA-55 atmospheric reanalysis are adjusted during the spin-up experiment to minimise excessive net heat uptake in the ocean. In the historical experiment, spanning 1979–2014, a non-adaptive relaxation is applied by repeating the same relaxation rates derived from the last five years of the spin-up experiment, and the surface heat flux adjustment diagnosed during the spinup experiment is also maintained. We demonstrate that the historical experiment driven by the JRA-55 reanalysis does not have significant drifts (e.g., as shown by simulated global ocean heat content), and also provides an eddy-resolving simulation of the global ocean circulation over the period 1979–2014. Decadal changes, such as the strengthening of the subtropical gyre circulation, are also reasonably simulated. A biogeochemical model is coupled with OFAM3 to produce patterns of primary productivity and carbon fluxes that are consistent with observations. Experiences gained from our numerical experiments will be helpful to other modelling groups who are interested in running global eddy-resolving OGCMs for climate studies.