In culminating ice caps near the centre of the Antarctic ice sheet, the horizontal flow of ice is negligible and annual snow layers accumulate without melting, growing thinner and thinner with no discontinuities. As ice deformation occurs by vertical compression rather than shearing, the age of the ice is a function of depth and the snow accumulation rate. The physical transformation of snow into ice does not change the chemical composition of the snow or aerosols, or of volcanic and cosmic dust deposited on the surface. Owing to very low accumulation rates and vertical compression, the individual layers are not apparent in most Antarctic ice cores. Dating should therefore be done through indirect methods such as the analysis of isotopes and other components of ice and air bubbles, ash and acidic aerosols from known volcanic eruptions, and cosmogenic radionuclides (Paterson 1994; Legrand and Mayewski 1997; Petit et al. 1999; Archer et al. 2000; Wagner et al. 2000).
The balance of a number of isotopic constituents within the atmosphere, cryosphere and oceans is peculiar to particular climate conditions and varies according to climate change. During the evaporation of seawater, the water molecules composed of light isotopes (e.g. 160,17O or 1H2) vaporise more easily than those containing 180 or deuterium (D). The evaporation rate is also related to water temperature (cold water allows relatively more 160 to evaporate than 180). In equilibrium conditions, atmospheric water vapour contains 1 % less 180 and 10% less D than average ocean water. During vapour condensation, molecules containing heavy isotopes precipitate more readily, and the remaining vapour is therefore depleted in heavy isotopes. The progressive cooling of water vapour, such as during its passage from the Southern 0cean to the colder Antarctic landmass, will result in precipitation with increasing concentrations of lighter isotopes. Thus, although many factors can affect the isotopic composition of precipitation, the effect of temperature is remarkably predictable. A mass spectrometer can be used to determine variations in 180/160 and/or 1H/D ratios in ice melted from different layers of a core; the difference between the measured ratios and those of Standard Mean 0cean Water (SM0W) are expressed as S180 and SD. Low values of S180 and SD indicate low palaeotemperatures, because snow is enriched in 160 and the ocean in 180 during periods of glaciation.
The first ice core from the Russian Vostok station (2,083 m) was obtained during a series of drillings in the early 1970s and 1980s. Lorius et al. (1985) performed the first isotopic analysis of the ice core, and two years later Jouzel et al. (1987) analysed an ice record spanning a full glacial-interglacial cycle. Drilling continued at Vostok until January 1998, reaching a depth of 3,623 m. By measuring the continuous deuterium profile along the ice core, the Vostok temperature record was extended to the past 420,000 years and four glacial-interglacial cycles (Petit et al. 1999; Fig. 13). Although the third and fourth climate cycles in the ice core show a shorter duration than the first two
cycles, all cycles are characterised by a similar sequence of a warm interglacial followed by a cold glacial period, which ends with a rapid return to an interglacial period. The overall glacial-interglacial temperature change in surface temperatures is about 12 °C. Climate cycles deduced from the Vostok ice core appear to be more uniform than those in deep-sea core records (Petit et al. 1999). However, more recent measurements (Petit et al. 2000) show that the Vostok climate record may be disturbed below 3,311-m depth.
At Dome C (75° 06'S and 123° 24'E, about 3,233 m above sea level), the EPICA research team (a consortium of European countries) drilled the ice to a depth of more than 3,100 m during the 2002-2003 summer season. The ice is believed to be older than 500,000 years, and a longer historical climate record will probably be obtained through further drilling down to a depth of about 3,300 m.
Cosmic rays and solar irradiation impinging on the upper atmosphere produce 10Be and 36Cl. After their formation, these isotopes become quickly attached to aerosols and are removed from the atmosphere by precipitation. Their concentrations in ice cores can thus be used both as stratigraphic markers to compare different ice cores and as markers of long-term changes in the amount of snow deposition (Paterson 1994).
Isolated ice bubbles trapped in ice contain "fossil air"; by placing a thin slice of ice core in a vacuum chamber and cracking the ice, the concentrations of escaping gases, such as CO2 and other greenhouse gases, can be determined. Since the finding by Barnola et al. (1987), many ice cores from Greenland and Antarctica have shown remarkable similarities between the greenhouse gas curve and that of S18O and SD. Peak CO2 concentrations occur during warm periods and low concentrations mark glaciations. This trend probably reflects feedback mechanisms between glacial, oceanic, atmospheric and biological systems.
According to Siegert (2001), whether CO2 records are synchronous or occur before or after S18O variations remains unresolved. This knowledge is necessary to understand the effective role of greenhouse gases in forcing glacial activity and/or their development as a result of glaciation. Once the role of CO2 greenhouse gases in the behaviour of ice sheets is established, it will be possible to evaluate the relative importance of feedback mechanisms involving greenhouse gases in recent climate change. A study of CO2 glacial/interglacial cycles recorded in ice cores (Archer et al. 2000) indicates that this gas potentially forced climate change in the last two glaciations (i.e. its concentrations increased prior to the decay of ice, as shown in the S18O signal). The most likely driver of CO2 change over a glacial cycle is the ocean, but it is still not clear what processes (e.g. increased solubility of CO2 in cold water, biological productivity, pH variations) are responsible for CO2 uptake and delivery across glacial cycles. Moreover, numerical modelling of the role of CO2 variations in the last glacial-interglacial cycle (from the Vostok ice core) in forcing the Earth's climate show that CO2 variations alone cannot reproduce the ice-age cycle (Loutre and Berger 2000). This result seems to indicate that CO2 variations may not themselves be forcers of climate change, but that they may be influential as part of a feedback mechanism.
Ice impurities can be analysed to obtain information about deep ice sheets and their former environment. Glacial periods are characterised by an increase in windborne concentrations of fine sand, silt and clay particles. In contrast to atmospheric circulation in Greenland, which was probably affected by rapid change during the last glacial cycle (Svensson et al. 2000) - which in Antarctica was rather stable - dust concentrations in the
Vostok ice core show a strong periodicity of 100,000 and 41,000 years (Petit et al. 1999).
Large volcanic events can be recorded in ice cores as ash layers or as increases in acidity (tephra horizons) produced by the transformation in the atmosphere of sulphur dioxide into sulphuric acid aerosols. Ash and acidity peaks in ice core layers constitute useful stratigraphic markers because they match with historical eruptions (e.g. Hammer et al. 1980; Francis 1993). Three ash layers 3.3 km below the ice-sheet surface have been detected in the Vostok core (Petit et al. 1999) and, together with acidic layers detected in deep ice sheets by airborne radar sounding (Millar 1981), these records reveal no obvious relationship between major volcanic events and ice-age cycles.
The chemical composition of the atmosphere has been dramatically altered by human activity. Lead isotope measurements in Greenland ice cores indicate that early large-scale atmospheric pollution of the Northern Hemisphere by mining of this metal in Spain began between 150 B.c. and 50 a.d. (i.e. during the Carthaginian and Roman civilisations; Rosman et al. 1997). In the last 200 years, the world population has increased by more than 500 % and ice cores worldwide contain higher concentrations of CO2,CH4,N2O and persistent pollutants from atmospheric nuclear bombs, and industrial and agricultural activity. As will be discussed in Chapter 4, some of these pollutants have a global distribution, and the chemical composition of Antarctic snow and ice cores reflects the impact of heavy metals, radionuclides and persistent organic pollutants (POPs) from remote anthropogenic sources and/or human activity in Antarctica. Radionuclides and persistent pollutants also provide useful stratigraphic marker horizons which can be used to date snow and ice cores and to reconstruct the mass balance of glaciers (e.g. Lefauconnier et al. 1994). Moreover, ice cores can be used to study the possible relationship between changes in atmospheric composition and past global changes. A strong correlation between concentrations of Na+, Ca2+, SO42- or the value of the ratio Cl/Na+ and the S18O signal has been found in Antarctic ice cores (Legrand and Mayewski 1997). As discussed in the next chapter, ocean-atmosphere interactions and the sulphur cycle are among the processes linking atmospheric chemistry to global climate change. Measurements of soluble and insoluble constituents in snow and ice over Antarctica are valuable not only as indicators of changes in their source strength but also as tracers of atmospheric circulation over the continent. In the simplest case, for instance, marine versus continental air masses can be differentiated on the basis of sea salt (e.g. NaCl) versus continental dust (e.g. Al, CaSO4).
Land-based glaciers or ice streams flowing into the Southern Ocean may originate floating glacier tongues or ice shelves which accumulate snow on their surfaces. The two largest embayments of Antarctica are occupied by the Ross Ice Shelf in the Ross Sea and by the Filchner-Ronne Ice Shelf in the Weddell Sea. These two shelves and the Amery Ice Shelf, which is fed by the Lambert Glacier, drain a combined area of more than 60 % of the Antarctic continent, with a flow of 0.8-2.4 km/year. Together with many other smaller shelves occurring along the coast and comprising about 47 % of the coastline,Antarc-tic shelves have a surface area of about 1.72x106 km2. Their seaward edges are marked by cliffs rising 30-50 m above sea level, with an overall thickness of about 200 m. As shelves are grounded and constrained by promontories and islands, their thickness increases landwards and in places where ice streams enter the floating ice mass. The grounding line of the Ross Ice Shelf, for instance, is about 1,000 m thick. Based on basal melting and freezing patterns, this shelf can be divided into three zones (Souchez and Lorrain 1991): an area of enhanced bottom freezing near the grounding line, one with slow bottom freezing, and an outer zone affected by stronger circulation, greater heat exchange and net basal melting.
The stability of ice shelves depends on ice discharge from feeding glaciers, the morphology of the coast, location of bedrock "pinning points", net snow accumulation at the surface, and freezing at the base. Ablation is due to basal melting of the lower surface (about 3 m year-1 in the ice front of the Ross Ice Shelf; Jacobs et al. 1986), net surface ablation and, above all, to the calving of icebergs. Jacobs et al. (1992) estimated that calving accounts for 77% of ice lost from the Antarctic ice sheet. Calving from Antarctic shelves is generally episodic, releasing large tubular icebergs which can be tracked over vast distances and for several years.
The Antarctic ice shelves are regularly monitored through satellite imagery, and the calving of giant icebergs is increasingly attracting the attention of mass media, contributing to a growing general concern over the possible effects of global warming. Too much importance is probably attributed to calving events in Antarctica. The breaking-off of icebergs as large as small countries from the Ronne-Filchner and Ross ice shelves may be part of their normal lifecycle. The edge of a shelf may retreat on one side and advance on the other. The B-15 iceberg (about 290 km long and 37 km wide), calved from the Ross Ice Shelf in March 2000, was the largest iceberg recorded by the US National Ice Centre, and the images made the rounds of the world. However, the ice shelf was in a northerly advanced state, and even greater bergs were probably calved from its edge in the past. For instance, in 1956 during an expedition to Antarctica, the US icebreaker Glacier reported a berg about 330 km long and 95 km wide in the Ross Sea.
As discussed in the previous chapter, the Antarctic Peninsula is experiencing an enhanced warming trend, and the progressive retreat of its shelves (overall about 13,000 km2 since 1974) is probably linked to changing climate conditions. In contrast to the calving of large icebergs from the Ronne-Filch-ner and Ross ice shelves, in 1995 the northern Larsen Ice Shelf (A; about
2,000 km2) disintegrated into thousands of small icebergs. Further south, the Larsen (B) Ice Shelf and other shelves began to break up, receding past their historical minimum extent. Satellite imagery shows that the retreat is continuing, and that the northern section of the Larsen (B) Ice Shelf has completely shattered and separated from the continent; in February 2002 about 3,250 km2 disintegrated in a plume of thousands of icebergs adrift in the Weddell Sea.
Was this article helpful?