Downscaled surface mass balance in Antarctica: impacts of\nsubsurface processes and large-scale atmospheric circulation

Abstract

Abstract. Antarctic surface mass balance (SMB) is largely determined by precipitation over the continent and subject to regional climate variability related to the Southern Annular Mode (SAM) and other climatic drivers at the large scale. Locally however, firn and snow pack processes are important in determining SMB and the total mass balance of Antarctica and global sea level. Here, we examine factors that influence Antarctic SMB and attempt to reconcile the outcome with estimates for total mass balance determined from the GRACE satellites. This is done by having the regional climate model HIRHAM5 forcing two versions of an offline subsurface model, to estimate Antarctic ice sheet (AIS) SMB from 1980 to 2017. The Lagrangian subsurface model estimates AIS SMB of 2473.5\u2009±\u2009114.4\u2009Gt per year, while the Eulerian subsurface model variant results in slightly higher modelled SMB of 2564.8\u2009±\u2009113.7\u2009Gt per year. The majority of this difference in modelled SMB is due to melt and refreezing over ice shelves and demonstrates the importance of firn modelling in areas with substantial melt. Both the Eulerian and the Lagrangian SMB estimates are within uncertainty ranges of each other and within the range of other SMB studies. However, the Lagrangian version has better statistics when modelling the densities. There is a mean bias in modelled density of −24.0\u2009±\u200918.4\u2009kg\u2009m−3 and −8.2\u2009±\u200915.3\u2009kg\u2009m−3 for layers less than 550\u2009kg\u2009m−3 for the Eulerian and Lagrangian framework, respectively. For layers with a density above 550\u2009kg\u2009m−3 the bias is −31.7\u2009±\u200923.4\u2009kg\u2009m−3 and −35.0\u2009±\u200923.7\u2009kg\u2009m−3 for the Eulerian and Lagrangian framework, respectively. The mean firn 10 m temperature bias is 0.42–0.52\u2009°C. Further, analysis of the relationship between SMB in individual drainage basins and the SAM, is carried out using a bootstrapping approach. This shows a robust relationship between SAM and SMB in half of the basins (13 out of 27). In general, when SAM is positive there is a lower SMB over the Plateau and a higher SMB on the westerly side of the Antarctic Peninsula, and vice versa when the SAM is negative. Finally, we compare the modelled SMB to GRACE data by subtracting the solid ice discharge, and find that there is a good agreement in East Antarctica, but large disagreements over the Antarctic Peninsula.There is a large difference between published estimates of discharge that make it challenging to use mass reconciliation in evaluating SMB models on the basin scale.\n