High-resolution modelling of the Antarctic surface mass balance, application for the twentieth, twenty first and twenty second centuries

About 75 % of the Antarctic surface mass gain occurs over areas below 2,000 m asl, which cover 40 % of the grounded ice-sheet. As the topography is complex in many of these regions, surface mass balance modelling is highly dependent on horizontal resolution, and studying the impact of Antarctica on the future rise in sea level requires physical approaches. We have developed a computationally efficient, physical downscaling model for high-resolution (15 km) long-term surface mass balance (SMB) projections. Here, we present results of this model, called SMHiL (surface mass balance high-resolution downscaling), which was forced with the LMDZ4 atmospheric general circulation model to assess Antarctic SMB variability in the twenty first and the twenty second centuries under two different scenarios. The higher resolution of SMHiL better reproduces the geographical patterns of SMB and increase significantly the averaged SMB over the grounded ice-sheet for the end of the twentieth century. A comparison with more than 3200 quality-controlled field data shows that LMDZ4 and SMHiL reproduce the observed values equally well. Nevertheless, field data below 2,000 m asl are too scarce to efficiently show the added value of SMHiL and measuring the SMB in these undocumented areas should be a future scientific priority. Our results suggest that running LMDZ4 at a finer resolution (15 km) may give a future increase in SMB in Antarctica that is about 30 % higher than by using its standard resolution (60 km) due to the higher increase in precipitation in coastal areas at 15 km. However, a part (∼15 %) of these discrepancies could be an artefact from SMHiL since it neglects the foehn effect and likely overestimates the precipitation increase. Future changes in the Antarctic SMB at low elevations will result from the competition between higher snow accumulation and runoff. For this reason, developing downscaling models is crucial to represent processes in sufficient detail and correctly model the SMB in coastal areas.