Lggecnrs Bp96 38402 Saint Martin dHeres Cedex France

The Russian station of Vostok was established in 1957 during the International Geophysical Year at the south geomagnetic pole by the Soviet Antarctic Expeditions. The site (78°28'S, 106°48'E) is 1500 km from the ocean, at an altitude of 3488 m a.s.l. with a continental climate and a mean annual temperature of -55°C. The station overlies the southern end of a giant subglacial lake, Lake Vostok. The total ice thickness is 3750 m, and the snow accumulation rate is only 2.2 cm water equivalent per year. Since 1965, several ice cores have been retrieved by Russian drillers. A 2000-m-deep core was obtained in 1980 and provided the first climatic sequence covering the previous interglacial period (120,000yr ago) and the preceding glacial period at ca. 150,000 yr ago (Lorius et al., 1985). Since then, a borehole was extended down to 3623 m, and stopped 130 m above the water interface. The palaeocli-matic record deduced from analysis of the first 3310 m was the first long record covering the past 420,000yr (Petit et al., 1999). Below 3310 m the ice layering is disturbed by glacier dynamics, but the very basal ice from 3539 m to 3750 m is formed by freezing of the lake water (Jouzel et al., 1999).

Here, the main results from the palaeoclimate reconstruction from the first 3310 m of the core are summarized. The past temperature, greenhouse gas content (CO2, CH4) and concentrations of aerosols of marine and continental origin have been reconstructed (Fig. 79.1).

The variations of the temperature at the surface of the ice sheet are deduced from the stable isotope composition of the ice. Both simple and more complex isotopic models predict that deuterium (SD) and S18O values should vary linearly with temperature in mid- and high latitudes, and this is well documented from observations of modern Antarctic precipitation (Jouzel et al., 2003b). With respect to the present mean annual temperature (—55°C), the reconstructed Vostok temperature profile displays an overall amplitude of about 12°C, although slightly different estimates of climatic changes (up to 18°C) have been proposed (Salamatin et al., 1998). From the mean time separating each warm period (interglacial), the record suggests the presence of cycles with a periodicity of about 100,000 yr. Glacial periods dominate the records, representing almost 90% of cycle duration. The present warm period, referred to as the Holocene, started ca. 12,000 yr ago and appears to be a very stable period with respect to other interglacials. The overall climate record displays a 'saw tooth' structure with temperatures gradually decreasing from the interglacial period to reach the minimum (ca. -67 to -73°C) of the glacial period. This was followed by a more rapid deglaciation, taking place in just a few millennia (ca. 5 to 10kyr). The overall pattern of the Vostok temperature mimics the reversed global ice volume profile deduced from marine sediment studies, and the reconstructed atmospheric temperature clearly leads the ice volume by a few thousand years.

Spectral analysis of the record indicates the presence of three major periodicities of about 100,000, 40,000 and 20,000yr. These characterize orbital geometry variations and the Earth's move

Figure 79.1 The climatic record over the past 420,000yr deduced from the first 3310m of the Vostok ice core (adapted from Petit et al., 1999). (a) Global ice volume (in relative units, and inverted scale) as deduced from the marine sediment record (from Bassinot et al., 1994). (b) Temperature (difference °C with the present surface temperature) deduced from the stable isotope composition of the ice. (c & d) Records of greenhouse gases: CO2 (ppmv) and CH4 (ppbv) as deduced from entrapped air bubbles. Note the recent increase up to the present level for CO2 (360ppmv) and CH4 (1750ppbv) reflecting anthropogenic activity since the 1850s. (e) Profile of continental dust concentration (ppm) as plotted in log scale. (f) Profile of sodium concentration (ppb), representative of marine aerosols. (See www.blackwellpublishing.com/knight for colour version.)

Figure 79.1 The climatic record over the past 420,000yr deduced from the first 3310m of the Vostok ice core (adapted from Petit et al., 1999). (a) Global ice volume (in relative units, and inverted scale) as deduced from the marine sediment record (from Bassinot et al., 1994). (b) Temperature (difference °C with the present surface temperature) deduced from the stable isotope composition of the ice. (c & d) Records of greenhouse gases: CO2 (ppmv) and CH4 (ppbv) as deduced from entrapped air bubbles. Note the recent increase up to the present level for CO2 (360ppmv) and CH4 (1750ppbv) reflecting anthropogenic activity since the 1850s. (e) Profile of continental dust concentration (ppm) as plotted in log scale. (f) Profile of sodium concentration (ppb), representative of marine aerosols. (See www.blackwellpublishing.com/knight for colour version.)

ment around the sun, which trigger large climatic variations by modulating solar energy received by the Earth according to latitude and season (Berger, 1978). The Vostok record confirms the astronomical theory of palaeoclimate (i.e. the so-called 'Milankovic' theory) that was firstly supported by marine records (e.g. Bassinot et al., 1994).

Impurities in the Antarctic ice are found at very low concentration (10-9gg-1), mostly comprising the small fraction of primary aerosols emitted by the ocean (sea salt) and the continents (dust). Sulphuric acid is present in ice, resulting from the biogenic emission of dimethyl sulphide (DMS) and gas to particle conversion (Legrand & Mayewski, 1997). Volcanic inputs include discrete events of ash and H2SO4 peaks from the SO2 oxidation, whereas anthropogenic activity is revealed through heavy metals (e.g. Pb) for the most recent period (Boutron et al., 1994).

Sodium (Na) characterizes sea spray whereas dust, for which the mass is mostly represented by particles with sizes between 0.8 and 3 |lm, originated mostly from the continents and arid zones. During glacial periods, the colder atmosphere was also drier than during interglacials, reducing the precipitation rate and the hydrological cycle, and therefore the capability of the atmosphere to be naturally washed out by precipitation. This is of importance because it significantly increases the residence time of aerosols and their ability to be disseminated worldwide. For glacial periods, the measured sodium concentration is three to four times higher than throughout the Holocene period, an effect partly due to the reduced precipitation and longer residence time in the atmosphere, but also to a probable more cyclonic activity around Antarctica. These factors, taken together, more than account for the greater sea-ice extent.

It has been hypothesized recently that the Antarctic sea-salt record is mostly influenced by salt blown off fresh sea ice, and could be used as an indicator for ice coverage (Wolff et al., 2003). Indeed, this may be applicable for coastal sites but questionable for inland sites. In fact the sodium records from Vostok and from the new EPICA Dome C appear almost synchronous and closely anticorrelated with isotopic temperature (r2 ~ 0.70 over the past 420,000 yr), supporting the idea that sodium concentrations are firstly linked to atmospheric processes: for example, cyclonic activity around Antarctica, residence time of aerosols, etc. On the other hand, as the sea ice extent is probably also influenced by the heat provided by the deep ocean surrounding Antarctica, some significant leads or lags between a sea-ice indicator and atmospheric temperature are expected, but are not observed for the sodium profile.

For glacial periods, the dust record shows concentrations of up to 750 ppb, 50 times greater than for interglacial periods (15 ppb), a magnitude now documented by several other East Antarctic ice cores (Delmonte et al., 2002, 2004a). Again a drier atmosphere and a probably more efficient meridional transport toward Antarctica are suggested as main factors, to which a source effect is added by the significant extension of continental aridity. Indeed, transport effect is still unclear: general circulation model (GCM) simulations suggest the transport change is less significant by comparison with the source extension effect (Andersen et al., 1998). In addition, recent studies provide evidence for opposite changes in the size distribution of dust between sites during the deglaciation, suggesting that the dust advection to the East Antarctic plateau has a regional character that varies with time (Delmonte et al., 2004b). The source effect and the residence time of aerosols appear, therefore, the most important factors.

The geochemical composition of the dust (isotopes from Sr and Nd) from several East Antarctic ice cores made it possible to determine their geographical origin, mostly in southern South America and in particular Patagonia (Delmonte et al., 2004a). This area is very sensitive to climate change because it is under the influence of sea-ice extension in the Drake passage between South America and the Antarctic Peninsula. Moreover, the Southern Andes, on which ice caps can develop, act as a meteorological barrier to precipitation. Together, this leads to intense periglacial erosion in the mountains, efficient transport of sediment to the plains by seasonal runoff and intense aeolian surface aggradation of the outwash plain during the dry seasons.

Air bubbles entrapped within the ice provide atmospheric air samples that are naturally and continuously collected during snow and firn densification. Vostok ice contains ca. 8% of air by volume. Between the different glacial and interglacial periods, greenhouse gas contents varied similarly as did the temperature.

For CO2 the concentrations varied from 180 to 300ppmv (parts per million by volume) and for CH4 from 360 to 700ppbv. For the recent period, since 1850, CO2 and CH4 have both steadily increased and have reached the highest concentrations ever observed for the past 420,000yr: 360ppmv and 1750ppbv respectively. Such changes are the consequence of anthropogenic activity with the use of fossil fuels for energy as well as the intensification of agricultural activity.

Over the 420,000 yr preceding this recent period, the record shows clear correlations between temperature and greenhouse gases, indicating that theses gases participated and contributed to the amplitude of the temperature variations. During the last deglaciation, an initial forcing (insolation?) leading to a small increase in greenhouse gas content (CO2, but also CH4 and other gases such as N2O, etc.) gave first a very small radiative forcing change, which was amplified by positive reactions from the climate system: for example, decrease of sea ice, increase of the atmospheric water content and its own associated greenhouse effect, increase of surface albedo, warming the ocean temperature which releases more CO2, and warming of soils which increases the biological CH4 emissions. Finally up to 50% of the global temperature change, which was about 4°C from the glacial to interglacial period, may be due to the resulting greenhouse gas effect (Lorius et al., 1990). Such results and amplification effects are now also supported by climate simulations using general circulation models (GCMs).

Information derived from the ice-core records makes it possible to describe the natural changes of the climate, and is useful to test climatic models with scenarios from the past (e.g. glacial period). The estimate of climate sensitivity to greenhouse gases as deduced from the past and from GCMs is of importance for predicting future global warming and the environmental changes due to the almost inescapable doubling of atmospheric CO2 by the end of this century.

Was this article helpful?

0 0

Post a comment