This work has used an eddy resolving coupled simulation with a 1/12 degree NEMO ocean model, coupled to 25km atmosphere model, to study air-sea interactions over a 20 year timeframe. It is compared to a similarly configured model with an eddy-permitting ¼ degree resolution, in order to examine the role of eddies and model resolution in the climate system.
We find important improvements in the large-scale circulation due to enhanced ocean resolution, several of which can be traced to air-sea interactions. Hourly coupling in these simulations, compared to reference models using three hourly coupling, suggest an impact on the upper levels of the Atlantic Meridional Overturning Circulation, driving an increase of several Sverdrups.
The eddy resolving coupled simulation has enhanced heat loss adjacent to the North Atlantic Current that acts to reduce a large model bias. This increased heat loss is due primarily to an increased latent heat flux. Following the work of Chelton and Xie (2010) and Bryan et al (2010), we then examined the relationship between SST anomalies and those in surface wind and latent heat flux in order to understand the mechanism of the enhanced heat loss. Using a box-car filter to remove the mean component of the spatial field, there are indications that the more turbulent nature of the eddy resolving ocean allows stronger SST perturbations and hence enhanced wind variability and latent heat flux. Further tests were made by using different filters and different atmospheric resolutions in the coupled model to understand the sensitivity of the results.
Changes to the air-sea fluxes are then implicated in changes to the large-scale circulation. The enhancement of the heat loss over the North Atlantic Current increases the heat transport required by the ocean to balance the atmospheric heat transport implied by the surface fluxes, and indeed the eddy-resolving simulations have a greatly enhanced northward ocean heat transport in much better agreement with observations.
These results have important implications for climate simulations in terms of heat and energy flows within the system, and also for ocean spin-up and equilibration we find that with an eddy-resolving ocean, the ocean is able to maintain the large-scale structure from the initial conditions much more effectively, and in essence it spins down much less than the eddy-permitting simulation.