High-resolution sea ice in long-term global ocean GCM integrations

Assessment of radiation forcing data sets for large-scale sea ice models in the Southern Ocean

“…This in turn deteriorates the simulated interannual variations. Some of these features were found in other Antarctic sea ice simulations (Fichefet et al, 2003;Timmermann et al, 2005;Stössel et al, 2007;Zhang, 2007;Mathiot, 2009;Timmermann et al, 2009). Another dilemma is that sea ice simulations performed with coupled climate models used in the last IPCC climate assessment show the same tendency of lower performance for the Antarctic than for the Arctic (Arzel et al, 2006).…”

Interactive momentum flux forcing over sea ice in a global ocean GCM

“…The instantaneous patterns of all variables shown thus far move swiftly eastward following the synoptic wind field. Since only the sea‐ice component has high resolution (see section 2), the coarse ocean grid pattern emerges where the upper‐ocean temperature has a large impact on sea ice, such as along the ice edge [see Stössel et al , 2007]. The variables of Figure 2 shown as winter (JAS) means (Figure 3) do not show much of a distinct pattern when plotted for the same isoline intervals as in Figure 2.…”

“…The way the momentum flux is modified by the simulated sea ice, in particular its subgrid‐scale surface heterogeneity, follows the approach of SC93. Following Stössel et al [2007], the Southern Ocean sea‐ice component of the ocean model has a horizontal resolution of about 22 km, while the ocean model is coarse resolution (on the order of 200 km). This setup allows for detailed investigations of air–ice–ocean interactions, in particular along the Antarctic coast‐ or ice‐shelf line, which is critical for a realistic representation of coastal polynyas, thermodynamic sea‐ice growth, and ultimately AABW formation.…”

“…The question of whether a heterogeneous ASL parameterization causes long‐term changes in the deep ocean thus seems to narrow down to the question which of the form‐drag assumptions, those of MFH or MFHS4 are more realistic. Depending on which satellite passive microwave algorithm to believe [e.g., Markus and Cavalieri , 2000], our simulations lie in the highest range [ Stössel et al , 2007], implying that a somewhat lower (3–6%) ice concentration than simulated may be more realistic. While a minimum ice freeboard of 0.5 m is too high for typical Southern Ocean sea ice, it could be justified due to the fact that we did not account for the form drag due to pressure ridges and sastrugi.…”

“…The discretization and grid resolution follow Drijfhout et al [1996] and Maier‐Reimer et al [1993], using an Arakawa E grid with a coarse horizontal resolution of 3.5° × 3.5°, 11 vertical layers with actual topography‐following bottom‐layer depth, and a prognostic free sea surface. The resolution of the southern hemisphere sea‐ice component of this model is ninefold higher in both principle horizontal directions, yielding a grid spacing of about 22km [ Stössel et al , 2007]. Furthermore, the model includes a plume‐convection parameterization following Paluszkiewicz and Romea [1997], as introduced by Kim and Stössel [2001].…”