Influence of Spatial Resolution and Treatment of Orography on GCM Estimates of the Surface Mass Balance of the Greenland Ice Sheet

The major ice sheets are of considerable importance in the evolution of climate, and in order to predict their future state, or understand their past history, it is necessary to understand the forcing imposed upon their evolution by the atmosphere. Ice sheet models have already been coupled to atmospheric models with some degree of success, but before such tools are used to predict the past and future evolution of the ice sheets, it is important that the atmospheric model is capable of reproducing the current mass balance distribution of the ice sheets reasonably. At first inspection, the surface mass balance of the Greenland ice sheet predicted by the United Kingdom Universities Global Atmospheric Modelling Project general circulation model (UGCM) is typical of that produced by the majority of GCMs, with a moderate overestimation of accumulation and a large overestimation of ablation. Further investigation reveals that there are a number of systematic errors involving the treatment of orography and radiation within the UGCM, some of which are likely to affect most other GCMs. A postprocessing method is developed that allows these errors to be corrected without performing further GCM integrations. This procedure is based on the downscaling of GCM output to a finer grid and correction for altitude changes. The degree by which the resolution of the fine grid exceeds that of the GCM grid has a strong effect on the predicted mass balance although these changes become smaller as very high resolutions are reached. A resolution of about 0.7° in each direction appears to be sufficient for such a reanalysis and provides results for the components of the surface mass balance of the Greenland ice sheet that are significantly closer to observations than the initial estimates. A sensitivity study suggests that uncertainties related to the the surface forcing, or the choice of atmospheric lapse rate, are of secondary importance.