Evaluating Simulated Antarctic Ice Sheet Mass Balance in the NASA GISS ModelE GCM

The Antarctic ice sheet (AIS) represents a large source of uncertainty for future sea level rise. AIS surface mass balance (SMB) is a key component of overall ice sheet mass balance and modulates dynamic ice flow. In this context it is important that earth system models (ESMs) used to project future ice sheet mass change, feedbacks between ice and climate, and future sea level rise, are able to capture ice sheet SMB, its spatial variability and components. Here we examine the simulated SMB, SMB components and drivers in simulations with the NASA GISS ModelE 2.1 ESM (GISS-E2.1) in comparison with high-resolution regional climate model (RCM) simulations from the Modèle Atmosphérique Régionale (MAR), which are taken as a best approximation to the true values. While snowfall and sublimation are the largest components of SMB, surface melt is also important along coastal ice shelves, and future enhanced melt may contribute ice shelf collapse. Therefore, we pay particular attention to simulated melt, including evaluation of melt patterns against satellite-derived estimates. We also investigate the climatic drivers of SMB and the ability of GISS-E to capture patterns of atmospheric circulation in free-running simulations. Preliminary results indicate that over the grounded and floating AIS, ModelE agrees with MAR regarding overall AIS SMB (at ~2,500 Gt yr-1). ModelE precipitation is higher than MAR precipitation by ~900 Gt yr-1, but this is compensated for by sublimation that is larger by an order of magnitude (~900 vs. 90 Gt yr-1). Simulated melt is comparable between the two models, but ModelE does not include meltwater refreezing, which is of the same order of magnitude as melt in MAR. The results are consistent with a recent assessment over the Greenland Ice Sheet, where the large sublimation difference was found to be influenced by simulated surface roughness length, and highlight both strengths and areas of improvement in the model simulations.