The role of resolution in modeling the Arctic Ocean circulation and dynamics

The Arctic Ocean has been traditionally a challenging region for modeling, in part due to the presence of the seasonal and multi-year ice pack and the grid singularity at the North Pole. The bathymetry of this region with the vast shallow shelves, submarine ridges, and troughs and its geometry with narrow straits controlling main exchanges with the North Atlantic (e.g. Fram Strait and Denmark Strait) and the North Pacific (i.e. Bering Strait) has constituted additional complexity for modelers. Finally, the horizontal resolution between order of 0(1km) and 0(10km) is required to resolve not only eddies but also large scale, narrow boundary currents in the Arctic Ocean.

Regional models of the Arctic Ocean have been developed in part to resolve some of the above mentioned issues. In contribution to the international Arctic Ocean Model Intercomparison Project (AOMIP) we compare results of our two coupled ice-ocean models of the Arctic Ocean, configured at 18-km and 30-level and at 9-km and 45-level grids. The two models are forced with the same atmospheric fields and were initialized with similar climatological and bathymetry data sets. Results for selected regions of the Arctic region are compared with emphasis on representation of the main circulation features and the mean and eddy kinetic energy.

Our findings indicate that the change of model grid resolution not only changes an amount of details represented in a model but more importantly it produces significantly different large scale circulation patterns. Doubling of the horizontal resolution from 18 km to 9 km increases the mean eddy kinetic energy (EKE) in a region by an order of magnitude or more. The knowledge of absolute magnitudes of EKE in the Arctic region is rather limited and further increases of EKE levels in models at even higher resolutions are expected. Significance of such findings is discussed in conclusions.