The extreme atmospheric boundary layer over the Antarctic Plateau and its representation in climate models

Observation of the Atmospheric Boundary Layers (ABL) above the Antarctic Plateau has revealed the strongest near-surface temperature stratifications on the Earth. A correct parametrization of the very stratified Antarctic ABLs in General Circulation Models (GCM) is critical since they exert a strongcontrol on the continental scale temperature inversion, on the coastal katabatic winds and subsequently on the Southern Hemisphere circulation. The previous Gewex Atmospheric Boundary Layer Studies (GABLS) highlighted that the parametrization of the very stratified, or very stable, ABLs isone of the most critical challenge in the atmospheric modelers community. Indeed, the nature of the mixing processes are not completely understood and the commonly used similarity laws, on which the model’s parametrization are usually based, are no longer valid. The aim of this PhD work is to evaluate and improve the modelling of the ABL over the Antarctic Plateau by the Laboratoire de Meteorologie Dynamique-Zoom (LMDZ) GCM, the atmospheric component of the IPSL Earth System Model in preparation for the sixth Coupled Models Intercomparison Project. Before the model evaluation itself, an in-depth study of the dynamics of the atmospheric surface layer and of the stable ABL over the Antarctic Plateau was carried out from in situ measurements at Dome C. The analysis enabled the first estimations of the roughness length and of the surface fluxes during the polar night at this location as well as the characterization of very frequent occurences of near-surface moisture supersaturations with respect to ice. Investigation of meteorological measure-ments along a 45 m tower also revealed two distinct dynamical regimes of the stable ABL at this location. In particular, the relation between the near surface inversion amplitude and the wind speed takes a typical ’reversed S-shape’, suggesting a system obeing with an hysteresis. A further analysisshowed that this is a clear illustration of a general and robust feature of the stable ABL systems, corresponding to a ‘critical transition’ between a steady turbulent and a steady ‘radiative’ regime. LMDZ was then run on 1D simulations during a typical clear-sky summertime diurnal cycle in the framework of the fourth GABLS case. Sensitivity tests to surface parameters, vertical grid and turbulent mixing parametrizations were performed leading to significant improvements of the model and to a new configuration better adapted for Antarctic conditions. 3D simulations were then carried outwith the ’zooming capability’ of the horizontal grid and with nudging. These simulations enabled a further evaluation of the model over a full year and extending the analysis beyond Dome C. In particular, this study raised the importance of the radiative scheme and of the surface layer scheme forthe modelling of the ABL during the polar night over the Plateau. Finally, the PhD work extented toward the modelling of the stable ABL over the other continents, assessing how the frequently underestimated subgrid mixing of momentum and heat can be compensated by a transfer of large scalekinetic energy toward turbulent kinetic energy when the flow is slowed down by orographic gravity wave drag.