A simplified model of the ocean-atmosphere system: role of the atmosphere and atmospheric energy transport in climate evolution.

Summary A simplified surface ocean-atmosphere coupled model already existing at the lmd has been adapted here to make a tool strictly comparable to the ipsl global coupled model. A comparative analysis of climate simulations conducted with these two parallel ocean-atmosphere coupled models, using the same atmosphere model but ocean models of different complexity, has allowed us to highlight common systematic biases, which are also present in other climate models (kiehl, 1998). These are a cooling of the western tropical Pacific basin and a drying of the central Pacific accompanied by the penetration of a tongue of cold water from the east of the basin. This bias is also marked by a cooling of the tropical upper troposphere above 30°C. In order to reveal the role of the atmosphere in the tropical thermal drift of the simulated climate, we have performed sensitivity experiments which we believe significantly affect the energetics of the upper troposphere. These sensitivity experiments concern: the high ice clouds, the representation of convection and the ozone content of the upper atmospheric layers. The results of these simulations allow us to suggest that the tropical thermal drift at the ocean surface is controlled by an energetic imbalance at the top of the atmosphere. The transmission of these biases to the ocean surface is then made by the subsidence branches of the hadley-walker cells. This analysis of atmospheric feedbacks is completed by a study of atmospheric energy transport (tae). We validate the tae simulated by the lmd mcga against the transports derived from the ecmwf reanalyses and then focus on the variation of these transports in a changing climate. For the different energy types, the response of atmospheric energy transport to a prescribed change in ocean surface temperatures corresponding to a 2co 2 is directly related to the equator-tropic thermal gradient. However, the variations in the different energy terms partially offset each other in terms of the wet static energy transport. The analysis in terms of divergence of the atmospheric energy transport allows to quantify the importance of the dynamic feedbacks compared to the radiative feedbacks. The wet static energy transport varies by 10-20 wm 2 . These variations are of the same order of magnitude as the variations in radiative fluxes induced at the tropopause in a 2co 2 experiment. On the other hand, the variations in the different types of energy are of an order of magnitude larger. Finally, we show that the hydrological component of the tae is affected by the parameterization of the precipitation process.