Antarctica

Little data are available that directly records Antarctic Ice Sheet variations during the last few million years. Most climate and ice sheet modelling of Pleistocene Antarctica has been directed at individual times: LGM, modern and future (e.g. Oerlemans, 1982a; Warner and Budd, 1998; Huybrechts et al., 2004). Throughout the Pliocene and Pleistocene the EAIS is usually thought to have been stable with few deeply submerged marine margins, narrow continental shelf areas limiting expansion (Anderson et al., 2002) and marginal summer temperatures that are too cold to allow significant surface melt even with favourable orbital configurations. One regional exception is the Prydz Bay-Lambert Graben drainage system, where geologic and seismic studies have found significant regional variations in glacial extent through the later Cenozoic (McKelvey et al., 2001; Whitehead et al., 2004; O'Brien et al., 2007). On continental scales, though, the EAIS has been insensitive to the ~120m sea-level variations caused by Northern Hemispheric ice cycles, and by orbitally and CO2-induced local air temperature changes. Contrary views have been proposed by Raymo et al. (2006) involving smaller and more variable EAISs prior to the MPT (~ 1 Ma, see above), and by several researchers (e.g. Webb and Harwood, 1991) who suggest that Sirius Group deposits in the Transantarctic mountains imply drastic EAIS retreat in the Wilkes Bay sector during the Pliocene (see Miller and Mabin, 1998).

In contrast, the WAIS is grounded well below sea level with extensive deep grounding lines and wide continental shelves in the Weddell and Ross embayments. Consequently it is much more sensitive than the EAIS to variations in sea level and ocean temperatures via ice shelf basal melt. The WAIS is thought to have repeatedly expanded to the continental shelf limits and back during the Pleistocene, driven primarily by sea-level variations of ~120m caused by NH ice growth and decay. A few three-dimensional Antarctic ice models have been run in this way through the last 400 kyr (Ritz et al., 2001; Huybrechts, 2002), forced by prescribed temperature and sea-level variations from ice core and deep sea core records. These three-dimensional ice models have the capacity to some degree to simulate floating ice and advancing and retreating marine grounding lines. They capture what is known of the WAIS and EAIS behaviour since the LGM reasonably well, including the retreat of WAIS grounding lines over the Ross and Weddell embayments (Conway et al., 1999; Anderson et al., 2002), as illustrated in Fig. 11.9. However, the models' treatments of basal hydrology and grounding line migration are simplified and open to question, especially in the light of recent theoretical developments in grounding line treatment that suggests coarse-grid models are not capable of realistic migration (Schoof, 2007), and fuller-stress models are needed to capture drawdown in ice stream regions (Payne et al., 2004).

Most of the variation in total grounded ice volume in the Huybrechts and Ritz glacial-cycle models is due to expansion and contraction of grounding lines across continental shelves, mostly in the WAIS Ross and Weddell sectors, but also to the west of the Antarctic Peninsula and other areas. The modelled G-I (~LGM vs. modern) change in Antarctic ice volume is equivalent to ~ 15-20 m of sea level. The EAIS interior responds in the opposite sense, contracting slightly at Northern Hemispheric glacial maxima due to lower model snowfall rates (which are reduced as temperatures fall). At these times, modelled surface elevations in the central EAIS plateau decrease by ~ 100-150 m compared to the present, with minor volume changes equivalent to just a few metres of sea level.

Just as for the NH, more sophisticated regional climate, ocean and ice models have been applied to Antarctica but only for particular 'snapshot' times, mainly the present and LGM. These include RCMs (Bailey and Lynch, 2000; Bromwich et al., 2004), ice-sheet models including basal sediment and hydrology (Vogel et al., 2003), longitudinal stresses (Payne et al., 2004), subglacial lakes (Siegert, 2005; Pattyn and Siegert, 2007) and regional ocean models capable of resolving sub-ice shelf circulation (Jenkins and Holland, 2002; Holland et al., 2003). The combination of additional physical processes and higher model resolution could well be important for

Figure 11.9: Snapshots of Antarctic Ice Sheet Evolution during the last G-I cycle, simulated by a three-dimensional ice-sheet model driven by parameterised forcing, from Huybrechts (2002; his Fig. 3). Shown is surface ice elevation relative to present sea level. Contour interval is 250 m; thick lines are for every 1,000 m; the lowest contour is for 250 m and generally close to the grounding line (s.l.e., sea-level equivalent). The x- and y-axis-coordinates are arbitrary spatial values used by the model.

Figure 11.9: Snapshots of Antarctic Ice Sheet Evolution during the last G-I cycle, simulated by a three-dimensional ice-sheet model driven by parameterised forcing, from Huybrechts (2002; his Fig. 3). Shown is surface ice elevation relative to present sea level. Contour interval is 250 m; thick lines are for every 1,000 m; the lowest contour is for 250 m and generally close to the grounding line (s.l.e., sea-level equivalent). The x- and y-axis-coordinates are arbitrary spatial values used by the model.

the long-term evolution of Antarctic ice, but their use in time-evolving simulations awaits future work.

0 0

Post a comment