Time Dependent Simulations

Both Huybrechts and Ritz run their models over 400,000 years encompassing the last four glacial-interglacial cycles. The forcing in each case consists of time-series of air temperature change and eustatic sea-level variation. Both studies use temperature records derived from the 8D (deuterium) signal of the Vostok ice core. In each case accumulation is based on the present day pattern with the magnitude of variation calculated as a function of air temperature by the method described in Lorius et al. (1985). Huybrechts uses the SPECMAP reconstruction of sea-level change while Ritz et al. prefer that of Bassinot et al. (1994); the two records have some differences but sensitivity studies by Ritz et al. suggest that this does not affect their model results.

Overall the two reconstructions agree on several important aspects of Quaternary Antarctic evolution:

• The present day ice sheet is slightly larger than the minimum reached during the last interglacial. Ritz suggests that this is because the present ice sheet is still in decline, while Huybrechts determines that the ice-sheet retreat from the Weddell Sea was more advanced during the last interglacial.

• General responses that dominate during a glacial period are grounding line advance and interior ice growth in West Antarctica with surface lowering and little margin advance in East Antarctica.

• Eustatic sea-level variation is the forcing factor driving change in the position of the ice sheet grounding line while changes in temperature (and consequently rate of accumulation) have the largest effect on surface altitude in the interior of East Antarctica.

• Grounding line advance occurs more readily in the Weddell Sea embayment than in the Ross Sea. Both models cite the differences in bed topography - the depth of the Ross Sea increases more rapidly with distance from the present coastline - as the cause of this asymmetry.

• The patterns of ice flow and positions of the ice domes in East Antarctica are not significantly changed from their present configuration at the LGM. Ice flow in West Antarctica would have been very different due to the advanced margins and thicker inland ice.

There are several points at which results from the two models diverge. Huybrechts (2002) puts the timing of the LGM in Antarctica at 15 ka BP when the ice sheets would have contained 19.2m sea level equivalent greater mass than at present, equivalent to an increase of 5 m sea level equivalent on the sea level contribution predicted with the same model in an earlier study (Huybrechts, 1991). Ritz et al. (2001) place the LGM at around 18 ka BP and find a difference in mass of just 5.9 m sea level equivalent from today. The Ritz model generally has a lower surface slope around the margins due to its inclusion of a grounded 'dragging shelf' transition zone between inland and ice shelf flow intended to represent ice streaming. This factor is probably responsible for the majority of the differences between the two models, with the Ritz model placing thinner ice over much of West Antarctica at the

LGM, especially in the Ross Sea and also notably in the present day area of Prydz Bay in East Antarctica (c.f. Fig. 12.8C and D).

Interestingly, the reconstruction of Ritz et al. (2001) fails to predict any significant expansion of ice covering the Antarctic Peninsula except in the very southern part where it joins the West Antarctic Ice Sheet (Fig. 12.8D). This is in contrast to the results consistently obtained with Huybrechts model, which in agreement with geological evidence outlined earlier predicts an expansion of the Antarctic Peninsula Ice Sheet to the edge of the continental shelf under LGM conditions (Fig. 12.8C).

12.4.4. Ice-Sheet Sensitivity

Both Huybrechts and Ritz analysed the sensitivity of their respective models to variations in climatic forcing and to uncertainty of the parameter values by repeating the time dependent simulations described above and comparing snapshots of the ice-sheet at fixed time-slices. These experiments dealt with aspects of the ice-sheet dynamics including bedrock adjustment, thermo-dynamic coupling, basal sliding and the response of ice shelves to ocean warming as well as to variations in air temperature and accumulation forcing. The elements of sensitivity common to both models include:

• The ice-sheet volume is relatively sensitive to changes in LGM snow accumulation. Ritz et al. found that reducing glacial accumulation values to 40% of present day (c.f. 50% for the standard simulation) decreased the area of the LGM ice-sheet by 18% and the volume by 43%.

• Grounding line advance during glaciations is sensitive to the way ice shelves respond to temperature change. Not allowing the ice temperature (which controls viscosity) to cool and hence stiffen the ice or preventing the basal melt rate from decreasing significantly reduces the advancement of the grounding line during the LGM especially in West Antarctica. Sea level lowering appears to be necessary for grounding line advance but so does a decrease in surface temperature.

• The rate of crustal response appears to have an intricate relationship with the amount of grounding line retreat occurring at the glacial-interglacial transition. A more rapid isostatic response leads to earlier re-grounding of a retreating ice shelf in either the Ross or Weddell Sea embayments thereby reducing the total retreat.

• Finally, a lower value for the geothermal heat flux reduces the temperature of basal ice and limits the area of the base reaching the pressure melting point. The colder ice flows more slowly and the result is to increase ice-sheet thicknesses at the LGM.

0 0

Post a comment