Skip to Main content Skip to Navigation

Simulations climatiques régionales couplées atmosphère - océan - glace de mer en Antarctique.

Abstract : In th econtext of global warming, predicting sea level remains chalanging. The surface mass balance of Antarctica may be the only negative contribution. On the other hand, the ice sheet dynamics could exhibit a non-linear behaviour and it could therefore accelerate and thin some glaciers (Meehl et al. 2007). For these 2 reasons, climate of Antarctica needs a lot of onvestigations. Global climate models badly reproduce some important features of antarctic climate : precipitations are over-estimated because of the too smooth topography ; the surface energy balance is badly simulated because physical parametrizations are too poor. That is why the regional modelling is used in this work. It allows a better resolution and more detailed parametrizations.

Sea ice has a key role in the antarctic climate for exemple because its extension modfies available humidity for the atmosphere. The whole ocean also play an important role since the coastal deep water formation increases exchanges netween surface and deep ocean. That is why the aim of the present work is to create and to evaluate a new ocean - sea ice - atmosphere coupled model.

Regarding the atmospheric part, the model MAR (Modèle Atmosphérique Régional, Gallée et al. 2005) is used here. This model has been especially developped for polar regions. It differs from other polar regional models by its ellaborated snow representation, and an interactive representation of blowing snow. Ocean and sea-ice are simulated with NEMO (Nucleus for European Modeling of the Ocean), made of OPA-9 (Océan PArallélisé, Madec 2007) and of LIM-2 (Louvain Ice Model, Fichefet 1997). The ocean model uses an ellaborated representation of turbulent diffusion along isopycnes and of vertical diffusion. The sea ice model uses a 3 layers thermodynamical model, and a viscous-plastic rheology. MAR and NEMO are coupled via OASIS-3 (Valcke et al. 2003). The resulting model is called TANGO, it stands for Triade Atmosphere-Neige, Glace de mer, Océan.

Before analysing TANGO simulations, one has to know the behaviour of each model separately. In a first step, MAR's sensitivity to orographic roughness representation. By simulating a case already analysed in some papers, it is shown that MAR is able to simulate mesocyclones. It is also shown that the role of katabatic winds is small compared to synoptic forcing, as opposed to the conclusions of precedent papers. As coastal katabatic winds strongly depend on the orographic roughness of the Transarctic Mountains, polynias in TANGO could be influenced by this parametrization ; that is why it is tunned with authomatic weather stations. It is eventually shown that the sea ice fraction does not strongly influence the atmospheric circulation.

Estimating the role of coupling appears complicated since part of TANGO behaviour comes from phyyscial feedbacks whereas another part comes from a change in "forcing". MAR actually sees sea ice fraction from SSM/I and NEMO sees atmospheric fields from ERA-40. In TANGO, MAR receives the defaulf from NEMO and reversely. To evaluate the ability of TANGO to simulate feedbacks, a set of experiment is used. MAR is forced by fields from NEMO and reversely. Comparisons of these experiments to TANGO experiments show that physical feedbacks are simulated.
Document type :
Complete list of metadatas
Contributor : Nicolas Jourdain <>
Submitted on : Tuesday, March 25, 2008 - 5:55:06 AM
Last modification on : Thursday, November 19, 2020 - 3:54:26 PM
Long-term archiving on: : Thursday, May 20, 2010 - 10:26:38 PM


  • HAL Id : tel-00266564, version 1



Nicolas Jourdain. Simulations climatiques régionales couplées atmosphère - océan - glace de mer en Antarctique.. Climatologie. Université Joseph-Fourier - Grenoble I, 2007. Français. ⟨tel-00266564⟩



Record views


Files downloads