The aim of this task is to model Antarctic SMB and to quantify the processes affecting its evolution in a changing climate, with a focus on melt at low elevation coastal areas and low lying areas (coastal regions and ice shelves) where melt may increase in the future.

Back-trajectory calculations will be performed using meteorological reanalyses to interpret ice core content delivered by task 3. Then, surface data from task 2, 3 and 4 will be used to offer “field-based” SMB maps and information on the spatial distribution of isotopes. This information will be used to evaluate the water isotope distribution and surface mass balance given by the stretched grid model LMDz-iso with an improved snow module (Punge et al., 2012). Over specific regions, the regional atmospheric circulation model (MAR) and a combined statistical-dynamical-physical downscaling method will be used at higher resolution. CMIP5 and, if available, CORDEX simulations under the RCP4.5 and RCP8.5 scenarios will used to drive regional high-resolution SMB projections over the Antarctic.

Proposed method :

We will combine the isotopic version of LMDZ with the new detailed ice sheet snow module originally developed for MAR and recently included in a non-isotopic version of LMDZ (Punge et al., 2012). This will enable us to produce present-day simulations that can be used to evaluate conjunctly the isotopic content of Antarctic surface snow and the Antarctic SMB which is represented in a more detailed manner with the improved snow module that explicitly simulates processes such as water percolation, refreezing, etc.

Following the approach described by Krinner et al. (2008, 2014), we will use sea-surface condition changes (sea-surface temperature and sea-ice concentration) as simulated by CMIP5 coupled models under the RCP4.5 and RCP8.5 scenarios as forcings for LMDZ in order to produce simulations at regional high resolution (about 40 to 50 km) over the Antarctic. The anomaly method developed by Krinner et al. (2014) implies that the choice of the forcing global model is less critical than when a regional climate model is used. We take advantage of this to produce future simulations using the multi-model mean CMIP5 sea-surface condition changes as forcing data, along with outputs from one or two specific coupled models to be determined, in order to assess uncertainties linked to the choice of the forcing models.

For Adélie Land, the Antarctic Peninsula (including the Larsen A and B ice shelves), and the Filchner-Ronne ice shelf, specific analyses will be carried out using the regional climate model MAR. Moreover, specific downscaling methods based on Gerbaux (2005) and Agosta et al. (2013), applied to a range of CMIP5 outputs and to outputs from our higher-resolution-simulations will allow a more detailed analysis of melt rates

Updated on 15 novembre 2016