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

11.2.4 Interaction of Ice Sheets, Sea Level and the Solid Earth

11.2.4.1 Eustasy, isostasy and glacial-interglacial cycles

On time-scales of 103 to 105 years, the most important processes affecting sea level are those associated with the growth and decay of the ice sheets through glacial-interglacial cycles. These contributions include the effect of changes in ocean volume and the response of the earth to these changes. The latter are the glacio-hydro-isostatic effects: the vertical land movements induced by varying surface loads of ice and water and by the concomitant redistribution of mass within the Earth and oceans. While major melting of the ice sheets ceased by about 6,000 years ago, the isostatic movements remain and will continue into the future because the Earth's viscous response to the load has a time-constant of thousands of years. Observational evidence indicates a complex spatial and temporal pattern of the resulting isostatic sea level change for the past 20,000 years. As the geological record is incomplete for most parts of the world, models (constrained by the reliable sea level observations) are required to describe and predict the vertical land movements and changes in ocean area and volume. Relative sea level changes caused by lithospheric processes, associated for example with tectonics and mantle convection, are discussed in Section 11.2.6.

Figure 11.4 illustrates global-average sea level change over the last 140,000 years. This is a composite record based on oxygen isotope data from Shackleton (1987) and Linsley (1996), constrained by the Huon terrace age-height relationships of Chappell et al. (1996a), the estimate of the LGM sea level by Yokoyama et al. (2000), the late-glacial eustatic sea level function of Fleming et al. (1998), and the timing of the Last Interglacial by Stirling et al. (1998). These fluctuations demonstrate the occurrence of sea level oscillations during a glacial-interglacial cycle that exceed 100 m in magnitude at average rates of up to 10 mm/year and more during periods of decay of the ice sheets and sometimes reaching rates as high as 40 mm/year (Bard et al., 1996) for periods of very rapid ice sheet decay. Current best estimates indicate that the total LGM land-based ice volume exceeded present ice volume by 50 to 53x106 km3 (Yokoyama et al., 2000).

Local sea level changes can depart significantly from this average signal because of the isostatic effects. Figure 11.5 illustrates typical observational results for sea level change since the LGM in regions with no significant land movements other than of a glacio-hydro-isostatic nature. Also shown are model predictions for these localities, illustrating the importance of the isostatic effects. Geophysical models of these isostatic effects are well developed (see reviews by Lambeck and Johnston, 1998; Peltier, 1998). Recent modelling advances have been the development of high-resolution models of the spatial variability of the change including the detailed description of ice loads and of the melt-water load distribution (Mitrovica and Peltier, 1991; Johnston, 1993) and the examination of different assumptions about the physics of the earth (Peltier and Jiang, 1996; Johnston et al., 1997; Kaufmann and Wolf, 1999; Tromp and Mitrovica, 1999).

Figure 11.5: Examples of observed relative sea level change (with error bars, right-hand side) and model predictions for four different locations. The model predictions (left-hand side) are for the glacio-hydro-eustatic contributions to the total change (solid line, right hand side). (a) Angermann River, Sweden, near the centre of the former ice sheet over Scandinavia. The principal contribution to the sea level change is the crustal rebound from the ice unloading (curve marked ice, left-hand side) and from the change in ocean volume due to the melting of all Late Pleistocene ice sheets (curve marked esl). The combined predicted effect, including a small water loading term (not shown), is shown by the solid line (right-hand side), together with the observed values. (b) A location near Stirling, Scotland. Here the ice and esl contributions are of comparable magnitude but opposite sign (left-hand side) such that the rate of change of the total contribution changes sign (right-hand side). This result is typical for locations near former ice margins or from near the centres of small ice sheets. (c) The south of England where the isostatic contributions from the water (curve marked water) and ice loads are of similar amplitude but opposite sign. The dominant contribution to sea level change is now the eustatic contribution. This behaviour is characteristic of localities that lie well beyond the ice margins where a peripheral bulge created by the ice load is subsiding as mantle material flows towards the region formerly beneath the ice. (d) A location in Australia where the melt-water load is the dominant cause of isostatic adjustment. Here sea level has been falling for the past 6,000 years. This result is characteristic of continental margin sites far from the former areas of glaciation. (From Lambeck, 1996.)

Information about the changes in ice sheets come from field observations, glaciological modelling, and from the sea level observations themselves. Much of the emphasis of recent work on glacial rebound has focused on improved calculations of ice sheet parameters from sea level data (Peltier, 1998; Johnston and Lambeck, 2000; see also Section 11.3.1) but discrepancies between glaciologically-based ice sheet models and models inferred from rebound data remain, particularly for the time of, and before, the LGM. The majority of ice at this time was contained in the ice sheets of Laurentia and Fennoscandia but their combined estimated volume inferred from the rebound data for these regions (e.g., Nakada and Lambeck, 1988; Tushingham and Peltier, 1991, 1992; Lambeck et al., 1998) is less than the total volume required to explain the sea level change of about 120 to 125 m recorded at low latitude sites (Fairbanks, 1989; Yokoyama et al., 2000). It is currently uncertain how the remainder of the ice was distributed. For instance, estimates of the contribution of Antarctic ice to sea level rise since the time of the LGM range from as much as 37 m (Nakada and Lambeck, 1988) to 6 to 13 m (Bentley, 1999; Huybrechts and Le Meur, 1999). Rebound evidence from the coast of Antarctica indicates that ice volumes have changed substantially since the LGM (Zwartz et al., 1997; Bentley, 1999) but these observations, mostly extending back only to 8,000 years ago, do not provide good constraints on the LGM volumes. New evidence from exposure age dating of moraines and rock surfaces is beginning to provide new constraints on ice thickness in Antarctica (e.g., Stone et al., 1998) but the evidence is not yet sufficient to constrain past volumes of the entire ice sheet.



Other reports in this collection