Simulating the climate and surface mass balance of Greenland with a regional climate model

The surface mass balance (SMB) of the Greenland ice sheet represents one of the main uncertainties in past and future sea level changes. The sign of the contribution to sea level is unknown, but under future climate scenarios, the components of the SMB will respond differently to climate change. While precipitation may be expected to increase over the ice sheet, so too will ablation around the coastal margins. Recent studies suggest that the latter may dominate in the long-term.

In order to address the uncertainties in the current and future SMB of the Greenland ice sheet we are using the UK Hadley Centre atmospheric model version 3 (HadAM3) to simulate the climate of Greenland. The global model has a fairly coarse resolution of 3.75° by 2.5° that does not represent the orography of the island with sufficient detail to accurately simulate the strong dependence of the SMB components on the orography. To overcome this we are using a limited area configuration of the model with a 0.44° resolution. This regional model shows much more skill in reproducing observed snowfall and ablation patterns over the ice sheet.

The regional model is forced by lateral and sea-surface boundary conditions from the global model. The global model results have been extensively verified and agree well with observations and atmospheric analyses in the North-Atlantic region. However, the standard HadAM3 model has a number of deficiencies in terms of surface boundary conditions and snow and ice parameterisations.

The standard orography used is that the of the US Navy, which has been shown to have significant errors over large regions of Greenland. HadAM3 also assumes that the ice sheet covers Greenland entirely. Both these problems have been amended via the introduction of recent orography and ice sheet distributions. The sensitivity of the atmospheric circulation and the climate to these is considered, and both are shown to change the SMB components significantly. The other main source of error in the model is the parameterisation of the snow/ice surface albedo. The albedo has a weak dependence on temperature and therefore varies too little throughout the year, which introduces temperature errors that have a detrimental effect on the SMB. Improved snow/ice physics that take into account the aging of snow and the resultant changes in albedo are considered.