Evaluation of the CMIP5 models in the aim of regional modelling of the Antarctic surface mass balance

The surface mass balance (SMB) of the Antarctic Ice Sheet cannot be reliably deduced from global climate models (GCMs), both because their spatial resolution is insufficient and because their physics are not adapted for cold and snow-covered regions. By contrast, regional climate models (RCMs) adapted for polar regions can physically and dynamically downscale SMB components over the ice sheet using large-scale forcing at their boundaries. Polar-oriented RCMs require appropriate GCM fields for forcing because the response of the cryosphere to a warming climate is dependent on its initial state and is not linear with respect to temperature increase. In this context, we evaluate the current climate in 41 climate models from the Coupled Model Intercomparison Project Phase 5 (CMIP5) data set over Antarctica by focusing on forcing fields which may have the greatest impact on SMB components simulated by RCMs. Our inter-comparison includes six reanalyses, among which ERA-Interim reanalysis is chosen as a reference over 1979–2014. Model efficiency is assessed taking into account the multi-decadal variability of the fields over the 1850–1980 period. We show that fewer than 10 CMIP5 models show reasonable biases compared to ERA-Interim, among which ACCESS1-3 is the most pertinent choice for forcing RCMs over Antarctica, followed by ACCESS1-0, CESM1-BGC, CESM1-CAM5, NorESM1-M, CCSM4 and EC-EARTH. Finally, climate change over the Southern Ocean in CMIP5 is less sensitive to the global warming signal than it is to the present-day simulated sea-ice extent and to the feedback between sea-ice decrease and air temperature increase around Antarctica