Development of a large-scale coupled sea-ice model for interannual simulations of ice cover in the Arctic

Abstract

A coupled ice-ocean numerical model is developed which improves the simulation of the annual cycle and interannual variations in ice cover in the Arctic. The model is a further development of the work by Semtner (1987). Although the accuracy of the simulated ice concentration is increased, the annual cycle of ice coverage is still exaggerated. Several experiments are conducted to determine the importance of incorporating a fully interactive ocean, to select an optimum strength parameter for use in the ice rheology, to investigate the model's sensitivity to changes in the albedo of the frozen surface and to determine the relative importance of the various dynamic and thermodynamic forcing mechanisms. The regional dependence of these mechanisms and an assessment of two statistical analysis techniques used to measure model improvement are also examined. Inclusion of a fully prognostic ocean component vice a ten-year mean ocean cycle in the model improves the correlation of simulated ice concentration fields with observed data. This is the case for all regions in the Arctic; for both the annual cycle and interannual variations of the ice cover. A reduced strength parameter value, p*=hxl04 , is found to improve the simulation of the ice thickness distribution with increased overall thickness and better compression north of the Canadian Archipelago and Greenland. In contrast to results using ice models without a fully prognostic ocean component, this model is quite insensitive to changes in the frozen surface albedo. Exceptions are evident where the ocean heat flux into the mixed layer is small and the ice is thin. At the spatial (110 km) and temporal (monthly) scales used here, the heat provided by the ocean appears to be the dominant mechanism controlling the position of the ice edge and the extent of the ice pack. Within the pack, it is the dynamic forcing and, in particular, the wind forcing which controls the ice thickness and thickness distribution. The ocean circulation below the mixed layer appears to position the heat underneath the MIZ. The MIZ is also the region where the ice thickness tends to decrease through divergence. The linkage between the subsurface heat and the thinned ice cover is apparently controlled by conditions at the surface and the resulting response of the mixed layer.