 |
 |
The air temperature is extrapolated from the climate model's topography to the true elevation by using a seasonally varying lapse rate (Ohmura (1987), Schwerdtfeger (1970)). This is particularly important for the MIT model which resolves land and ocean but has no topography; the correction applied to the ECHAM data is small and does not significantly affect the results. This simple adjustment allows an excellent reconstruction of the annually averaged temperature distribution over the ice sheet. Figure 3 compares the temperature predicted by the MIT model with that of the ECHAM model, the latter has been shown by Wild and Ohmura (1999) to be close to observations.
Because summer temperatures determine directly the ablation and runoff predicted by the positive degree-day and linear models and indirectly the melting calculated by the snowpack model through their control of the snow albedo, they play a much more important role than the annual average temperature on the Greenland ice sheet. Fig.4 shows the difference between the average summer temperature (June, July, August) predicted by the MIT and ECHAM. The ECHAM model is 4-6 oC warmer than the MIT model in the Northern half of the ice sheet, and a few degrees colder in the Southern coastal regions. The comparison to the monthly mean temperature maps derived from observations (Ohmura (1987)) show the 5oC isotherm at sea-level at about 74oN and the 0oC isotherm remaining fairly close to the coast along the northern shore, a situation in between what the two climate models are predicting.