The recent observed warming on the Antarctic Peninsula and Bellingshausen Sea has been a critical climatic change, the fastest regional warming in the Southern Hemisphere (SH), with an order of magnitude greater than global mean warming. The warming trend around the Antarctic Peninsula correlates with the major reduction in sea-ice extension in the Bellingshausen Sea, at least during the satellite era (i.e. since 1979). The mechanisms that caused this regional warming are still unknown and therefore it is difficult to anticipate whether the trend will remain in the next 100 years. A possible mechanism is related with changes in the large-scale atmospheric circulation over the Southern Ocean and particularly around the Drake Passage. But the changes in the sea-ice cover and of the oceanic circulation and their interaction with the atmosphere are other possible major contributors to the observed trends.

The main issue of this study is to examine the ability of a realistic global climate model to reproduce this observed regional change and to assess these different processes. We use a state-of-art ocean-atmosphere general circulation model (IPSL/CM2) which includes coupled-to-land and -ocean carbon models to simulate the evolution of climate from 1860 to 2100 with no flux corrections. Two 241-y simulations were carried out: a control simulation in which no anthropogenic CO2 sources is considered and a climate-change scenario simulation in which CO2 emissions are prescribed following the IPCC SRES98 A2 scenario.

The control run displays no significant drift of the global mean surface temperature while the scenario run present an increase of the global temperature consistent with the observations. Furthermore the most important features of the Southern Hemisphere climate are also well reproduced by the model in the control simulation as well as in the scenario simulation. Moreover, the two simulations present close realistic representations of the leading patterns of SH circulation low-frequency variability.

The response of the SH circulation to transient anthropogenic greenhouse warming mainly consists in an increase of the meridional pressure gradient between the high and mid latitudes, modulated by a wave-3 like pattern. That structure, which characterizes the linear trend of the SH circulation resembles the leading mode of interannual variability, known as the High Latitude Mode.

The model is able to reproduce some of the observed climate change in the vicinity of the Antarctic Peninsula. The regional change in the near surface temperature simulated around the Antarctic Peninsula is related to the modification of the mean atmospheric circulation but also to change in its variability. In particular the scenario simulation shows a decrease of the amplitude of the SAO and the High Latitudinal Mode presents more frequent events of positive phase. These two changes might have an influence on the Antarctic Peninsula temperature positive change. The mechanisms acting in the coupled atmosphere-ocean-ice system that explain those changes will be further discussed at the Conference.