Transient Climate Change in the CSIRO Coupled Model with Dynamic Sea Ice

The CSIRO coupled model has been used in a "transient" greenhouse experiment. This model contains atmospheric, oceanic, comprehensive sea-ice (dynamic/thermodynamic plus leads), and biospheric submodels. The model control run (over 100 years long) employed flux corrections and displayed only a small amount of cooling, mainly at high latitudes. The transient experiment (1% increase in CO2 compounding per annum) gave a 2°C warming at time of CO2 doubling. The model displayed a "cold start" effect with a (maximum) value estimated at 0.3°C. The warming in the transient run had an asymmetrical response as seen in other coupled models, with the Northern Hemisphere (NH) warming more than the Southern Hemisphere (SH). However, the land surface response in this model is different from some other transient experiments in that there is not a pronounced drying of the midlatitudes in the NH in summer. In the control run the ice model gave realistic ice distributions at both poles, with the NH ice in particular displaying considerable interdecadal variability. In the transient run the ice amount decreased more in the NH than the SH (corresponding with a greater NH warming). The NH ice extent and volume in summer was considerably reduced in depth and extent compared to the control. However, the model ice dynamics and thermodynamics allowed for a successful regrowth of the ice during the winter season to give a coverage comparable to that of the control run, although thinner. During the transient run there is a freshening of the surface salinity in the oceans at high latitudes. In the SH this is caused mainly by increases in precipitation over evaporation. The same is true for the NH, but it is found that there is a similar magnitude contribution to the polar freshening from ice melt and land runoff changes. The freshening in the North Atlantic reduces the strength of the meridional overturning by 35% at time of doubling of CO2. Other changes in the global climate at the end of the transient run relative to the control run are also investigated.