Modeling Pack Ice as a Cavitating Fluid

Polar ocean circulation is influenced by fluxes of salt and freshwater at the surface as ice freeze in one location, is transported by the winds and currents, and melts again elsewhere. The motion of sea ice, moreover, is strongly affected by internal stresses that arise from the mechanical strength of the ice cover. A simple sea-ice dynamics model, allowing these effects to be included in large-scale climate studies, is presented. In this model a cavitating fluid behaviour is assumed whereby the ice pack does not resist divergence or shear, but does resist convergence. While less realistic than other rheologies that include shear strength, this assumption has certain advantages for long-term climate studies. First, it allows a simple and efficient numerical scheme, in both rectangular and spherical coordinates, which as developed here along with a generation to include shear strength via the Mohr-Coulomb failure criteria. Second, realistic ice transport is maintained, even when the model is driven by smoothed wind forcing–a feature that may be useful in coupled ice-ocean climate models using mean monthly or mean annual winds. Finally, the lack of shear strength allows smooth flow past an obstacle, making the scheme attractive for coupling to a global ocean circulation model using an artificial island to avoid the mathematical singularity at the North Pole. Noteworthy. however, is the fact that the numerical scheme developed here does not require an island at the pole, making the model equally suited for coupling to a global atmospheric circulation model. Three-year dynamic-themodynamic simulations using observed forcing from 1981 to 1983 are performed using the cavitating fluid model and a more complete viscous-plastic model for comparison. The thickness buildup patterns, net ice growth, atmospheric heat flux, and total ice volume calculated by the cavitating fluid model are very similar to the viscous-plastic model results; however, the cavitating fluid model substantially overestimates local ice drift when compared to observed buoy drift. A 3-year simulation using the spherical grid version of the model, both with and without an artificial island at the pole, shows that the island has little impact on the thickness buildup and ice transport. Overall, the cavitating fluid approximation is shown to be a useful simplification, allowing essential feedback between ocean circulation and ice transport to be efficiently included in large-scale climate studies.