Accumulation

The snowfall, evaporation and accumulation (snowfall - evaporation) values predicted by the MIT and ECHAM models for Greenland and Antarctica are summarized in Table 1. Both models predict similar accumulation totals for Greenland in 1990 and both are sufficiently close to the value of 553 ^. 10¹² kg a^-1 derived from observations (Houghton et al. (1996)) for the estimates to be adequate for our modelling purposes. The current accumulation over Antarctica is however overestimated by both models, and by the MIT model in particular. It is also highly likely that these models will overestimate the changes in precipitation accompanying changes in the climatic forcing. Because snow accumulation is critical in determining the evolution of the mass balance of the ice sheets over the coming century, the precipitation amounts predicted by the MIT and ECHAM climate models for Antarctica which were used to calculate changes in sea-level were scaled by factors of 0.62 and 0.72 respectively in order to reproduce the best guess of the current total accumulation of 1810 ^. 10¹²kg a^-1 (Vaughan et al. (1999)). This scaling is justified by the following simple argument: The air temperatures predicted by the climate models are generally too warm because the model's topography underestimates the true elevation, this is the case in particular for the MIT model which has no topography. This bias in temperature leads to an excessive moisture holding capacity of the atmosphere and too much precipitation. The increase in saturation water vapor pressure in a changing climate, and to some degree the increase in precipitation, will therefore be too large because the baseline is wrong. The drawback of this approach is that it neglects the effect of changes in atmospheric circulation on the precipitation patterns, this effect can however be expected to be smaller for Antarctica than for Greenland. Alternatively, one could assume that, although the current precipitation is overestimated, this will not be the case for changes in precipitation; this was the approach taken by most past modelling efforts (Thompson and Pollard (1997), Ohmura et al. (1996), Wild and Ohmura (1999)).

 

Table: Snowfall, evaporation and accumlation over the Greenland and Antarctic ice sheets for 1 and 2 x CO₂ conditions. Units are 10¹²kg a^-1. Source for the observed is $\dag$: Houghton et al. (1996), $\ddag$: Vaughan et al. (1999).

	MIT			ECHAM			Observations
	Snow	Evap.	Accum.	Snow	Evap.	Accum.	Accum.
Greenland 1 x CO2	649	95	554	585	46	540	553$\dag$
Greenland 2 x CO2	727	118	609	739	67	672
Greenland - Change	78	23	55	153	21	132
Antarctica 1 x CO2	3121	246	2875	2732	241	2491	1810$\ddag$
Antarctica 2 x CO2	3553	313	3240	3087	288	2799
Antarctica - Change	432	67	365	355	47	308

The unscaled changes in accumulation in Greenland and Antarctica between the current climate and the time of CO₂ doubling as well as the increase between those two dates are summarized in Table 1.

The ECHAM model predicts a more rapid increase in the amount of snowfall over Greenland during the next century than the MIT model. This discrepancy cannot be linked to differences in the evolution of the annual mean air temperature (summarized in Table 2) since both models predict the same increase ( +3.8^oC) over that region. Dividing the increase in accumulation by the change in temperature gives a 2.6 % increase in accumulation per degree of temperature change for the MIT model and a 6.4 % increase for the ECHAM model. These numbers bracket the value which would have been expected, had the precipitation been controlled by local thermodynamics and the change in saturation vapor pressure associated with changes in temperature ( $\sim$ + 5.3${\frac{\%}{^\circ C}}$), an assumption often made by glaciologists (Huybrechts et al. (1989), Huybrechts (1990b)).

 

Table: Annual average temperature over Greenland and Antarctica for 1 and 2 x CO₂ conditions and the temperature difference between the current climate and the time of CO₂ doubling. Units are ^oC

	MIT			ECHAM
	1 x CO2	2 x CO2	Change	1 x CO2	2 x CO2	Change
Greenland	-23.0	-19.2	+3.8	-20.7	-17.3	+ 3.4
Antarctica	-33.2	-29.5	+ 3.8	-34.4	-32.6	+ 1.7

The ECHAM model also predicts a larger increase in accumulation in Antarctica than the MIT model, but a smaller change in both the annual mean and summer temperatures. The absence of a strong correlation between the patterns of change in the accumulation and temperature fields point to large scale modifications of the atmospheric circulation and of the poleward flux of moisture as the cause of the increase in snowfall. These dynamic effects and changes in the storm track location in the Atlantic have been pointed out by Ohmura et al. (1996) in simulations of the climate change over Greenland with the ECHAM 3 GCM.