Climate Change 2001:
Working Group I: The Scientific Basis
Other reports in this collection

11.2.3.4 Sensitivity to climatic change

The sensitivity of the ice sheet's surface mass balance has been studied with multiple regression analyses, simple meteorological models, and GCMs (Table 11.7). Most progress since the SAR has been made with several coupled AOGCMs, especially in the "time-slice" mode in which a high-resolution AGCM (Atmospheric General Circulation Model) is driven by output from a low-resolution transient AOGCM experiment for a limited duration of time. Model resolution of typically 100 km allows for a more realistic topography crucial to better resolving temperature gradients and orographic forcing of precipitation along the steep margins of the polar ice sheets. Even then, GCMs do not yet perform well in reproducing melting directly from the surface energy fluxes. The ablation zone around the Greenland ice sheet is mostly narrower than 100 km, and the important role played by topography therefore requires the use of downscaling techniques to transfer information to local and even finer grids (Glover, 1999). An additional complication is that not all melt water runs off to the ocean and can be partly retained on or in the ice sheet (Pfeffer et al., 1991; Janssens and Huybrechts, 2000).

Table 11.7: Mass balance sensitivity of the Greenland and Antarctic ice sheets to a 1°C local climatic warming.
Source dB/dT (mm/yr/°C) Method
Greenland ice sheet
Van de Wal (1996) +0.31a Energy balance model calculation on 20 km grid
Ohmura et al. (1996) +0.41c
[+0.04]cd
ECHAM3/T106 time slice [2xCO2 - 1xCO2 ]
Smith (1999) [-0.306]d CSIRO9/T63 GCM forced with SSTs 1950-1999
Janssens and Huybrechts (2000) +0.35a
[+0.26]b
Recalibrated degree-day model on 5 km grid with new precipitation and surface elevation maps
Wild and Ohmura (2000) +0.09c
[-0.13]cd
ECHAM4-OPYC3/T106 GCM time slice [2xCO2 -1xCO2 ]
Antarctic ice sheet
Huybrechts and Oerlemans (1990) -0.36 Change in accumulation proportional to saturation vapour pressure
Giovinetto and Zwally (1995b) -0.80e Multiple regression of accumulation to sea-ice extent and temperature
Ohmura et al. (1996) -0.41c ECHAM3/T106 time slice [2xCO2 - 1xCO2 ]
Smith et al. (1998) -0.40 CSIRO9/T63 GCM forced with SSTs 1950-1999
Wild and Ohmura (2000)

-0.48c

ECHAM4-OPYC3/T106 time slice [2xCO2 - 1xCO2]
dB/dT Mass balance sensitivity to local temperature change expressed as sea level equivalent. Note that this is not a sensitivity to global average
temperature change.
a Constant precipitation.
p Including 5% increase in precipitation.
c Estimated from published data and the original time slice results.
d Accumulation changes only.
e Assuming sea-ice edge retreat of 150 km per °C.

For Greenland, estimates of the sensitivity to a 1°C local warming over the ice sheet are close to 0.3 mm/yr (with a total range of +0.1 to +0.4 mm/yr) of global sea level equivalent. This range mainly reflects differences in the predicted precipitation changes and the yearly distribution of the temperature increase, which is predicted to be


Figure 11.4:
Estimates of global sea level change over the last 140,000 years (continuous line) and contributions to this change from the major ice sheets: (i) North America, including Laurentia, Cordilleran ice, and Greenland, (ii) Northern Europe (Fennoscandia), including the Barents region, (iii) Antarctica. (From Lambeck, 1999.)

larger in winter than in summer in the GCMs, but is assumed uniform in the studies of Van de Wal (1996) and Janssens and Huybrechts (2000). Another difference amongst the GCM results concerns the time window over which the sensitivities are assessed. The CSIRO9/T63 sensivities are estimated from high-resolution runs forced with observed SSTs for the recent past (Smith et al., 1998; Smith, 1999), whereas the ECHAM data are given as specific mass balance changes for doubled minus present atmospheric CO2. Thompson and Pollard (1997) report similar results to the ECHAM studies but the corresponding sensitivity value could not be calculated because the associated temperature information is not provided. Some palaeoclimatic data from central Greenland ice cores indicate that variations in precipitation during the Holocene are related to changes in atmospheric circulation rather than directly to local temperature (Kapsner et al., 1995; Cuffey and Clow, 1997), such that precipitation might not increase with temperature (in contrast with Clausen et al., 1988). For glacial-interglacial transitions, the ice cores do exhibit a strong positive correlation between temperature and precipitation (Dansgaard et al., 1993; Dahl-Jensen et al., 1993; Kapsner et al., 1995; Cuffey and Marshall, 2000), as simulated by AOGCMs for anthropogenic warming. Although other changes took place at the glacial-interglacial transition, this large climate shift could be argued to be a better analogue for anthropogenic climate change than the smaller fluctuations of the Holocene. To allow for changes in circulation patterns and associated temperature and precipitation patterns, we have used time-dependent AOGCM experiments to calculate the Greenland contribution (Section 11.5.1).

For Antarctica, mass-balance sensitivities for a 1°C local warming are close to -0.4 mm/yr (with one outlier of -0.8 mm/yr) of global sea level equivalent. A common feature of all methods is the insignificant role of melting, even for summer temperature increases of a few degrees, so that only accumulation changes need to be considered. The sensitivity for the case that the change in accumulation is set proportional to the relative change in saturation vapour pressure is at the lower end of the sensitivity range, suggesting that in a warmer climate changes in atmospheric circulation and increased moisture advection can become equally important, in particular close to the ice sheet margin (Bromwich, 1995; Steig, 1997). Both ECHAM3 and ECHAM4/OPYC3 give a similar specific balance change over the ice sheet for doubled versus present atmospheric CO2 to that found by Thompson and Pollard (1997).

In summary, the static sensitivity values suggest a larger role for Antarctica than for Greenland for an identical local temperature increase, meaning that the polar ice sheets combined would produce a sea level lowering, but the spread of the individual estimates includes the possibility that both ice sheets may also balance one another for doubled atmospheric CO2 conditions (Ohmura et al., 1996; Thompson and Pollard, 1997). For CO2 increasing according to the IS92a scenario (without aerosol), studies by Van de Wal and Oerlemans (1997) and Huybrechts and De Wolde (1999) calculated sea level contributions for 1990 to 2100 of +80 to +100 mm from the Greenland ice sheet and about -80 mm from the Antarctic ice sheet. On this hundred year time-scale, ice-dynamics on the Greenland ice sheet was found to counteract the mass-balance-only effect by between 10 and 20%. Changes in both the area-elevation distribution and iceberg discharge played a role, although the physics controlling the latter are poorly known and therefore not well represented in the models. Because of its longer response time-scales, the Antarctic ice sheet hardly exhibits any dynamic response on a century time-scale, except when melting rates below the ice shelves were prescribed to rise by in excess of 1 m/yr (O'Farrell et al., 1997; Warner and Budd, 1998; Huybrechts and De Wolde, 1999; see also Section 11.5.4.3).



Other reports in this collection