The isotope analysis of ice cores from Vostok station (Antarctica), to the depth of 950 m N. I. Barkov, E. S. Korotkevich, F. G. Gordienko and V. M. Kotlyakov Abstract. Interpretation of the results of oxygen isotope analysis of ice cores obtained from the borehole at Vostok station is carried out. The borehole reached the depth of 950 m. To interpret the age of the samples the model of ice deformation with non-uniform strain rates by Nye was used. The series of samples represented 46 500 years. It is shown that the isotope profile of the Vostok borehole is mainly formed under the influence of a climatic agent and is not complicated by the proper effect of glacial cover. At the beginning of the Holocene, climatic warming occurred in Central Antarctica between 15 000 and 11 000 years ago; isotope shift made up 5%0, which is in keeping with a temperature increase of about 5°C. Four cold periods are distinguished during the period between 46 500 and 15 00 years ago. They are separated by three warmer periods. Mean 8(I80) makes up -58.6°/oo in cold periods and -56.8%o in 'warm' periods. During the last 11 000 years mean S(lsO) equals -53.8%o- The main result of the study is that for the last 50 000 years the basic temperature changes occurred synchronously in polar areas of both hemispheres. Analyse isotopique des carottes de glace obtenues à la station Vostok (Antarctique) jusqu'à une profondeur de 950 m Résumé. On présente les résultats de l'analyse de la composition isotopique (oxygène) des carottes de glace obtenues à la station Vostok jusqu'à une profondeur de 950 m. Pour obtenir l'âge des échantillons, on utilise le modèle de déformation de la glace de Nye (taux de déformation non uniforme). La série d'échantillons représente 46 500 ans. On montre que le profil isotopique obtenu est principalement dû au facteur climatique et n'est pas perturbé par l'effet propre de la couverture de glace. Au début de l'Holocène le réchauffement climatique dans l'Antarctique central s'est produit entre 15 000 et 10 000 ans avant la période actuelle; la variation isotopique est de 5°/oo, ce qui suggère un réchauffement de l'ordre de 5°C. De — 46 500 à - 15 000 ans on distingue quatre périodes froides. Elles sont séparées par trois 18 périodes plus chaudes. Les S( 0) moyens sont de l'ordre de -58.6%0 dans les périodes froides et de -56.8°/oo dans les périodes 'chaudes'. Durant les 11 000 dernières années, la moyenne est de — 53.8%o- La principale conclusion de cette étude est que pour les derniers 50 000 ans les principaux changements de température se sont produits simultanément dans les régions polaires des deux hémisphères. This paper contains the results of isotope analyses of ice samples from the borehole at Vostok station, drilled in 1972-1973. The drilling reached the depth of 950 m. For the entire borehole a good core, 12 cm in diameter, was obtained. The measurements were carried out in 1972 and 1975 with laboratory equipment of the Moscow University. The samples, delivered to the laboratory in solid state, were analysed for lsO-variations with mass spectrometers MI-1305 (in 1972) and MI-1309 (in 1975). The samples were melted at room temperature immediately before the analyses. For the isotope analyses the methods by Cohn and Urey (1938) were used with some changes. All 180-values are expressed in form of deviations S(lsO) in °/oo relative to 'standard mean ocean water' (SMOW). The measured accuracy for not fewer than three measurements on each sample reached ± 3%0. The selection of samples from the Vostok borehole, their transport to Leningrad and delivery to the laboratory was fulfilled by the Soviet Antarctic Expedition under the guidance of N. I. Barkov and E. S. Korotkevich. Laboratory studies and interpretation of the results were carried out under the leadership of V. M. Kotlyakov and F. G. Gordienko. 382 The isotope analysis of ice cores from Vostok 383 The results of the isotope analysis of the upper 507 m of the Vostok borehole core have already been published (Barkov et ah, 1974, 1975). In the first series of experiments small samples measuring 5-50 cm vertically were used. They correspond to 2-20 years of ice accumulation, but very often the samples contained sediments of 3-4 years. In the second series of experiments each sample was 0.5-5 m long, i.e. it covered 20-200 years, which produced reliable mean data and excluded accidental results. The samples of the second series were obtained from another borehole drilled at Vostok station near the first one ; they contained the core from the depth of 500 m as well. A number of analyses from the upper 500 m were repeated because the quality of samples from the second borehole was higher. The results differed insignificantly from the earlier ones presented in the above-mentioned papers. The accumulation rate, necessary to calculate the age of the ice, was obtained from data on stake observations for the last 10 years, and also from results of isotope analyses of ice samples, 2-5 mm thick. The latter were obtained in entire sequences from five horizons. During preparation of the samples utmost care was taken to prevent isotopic exchange between the samples and ambient water vapour. Some scatter of the data might be caused by inadequate methods of sample selection from the core. The layers, only a few millimetres thick, were scraped along the whole core, then they were melted and mixed. It could be that this scraping was not ideal and the amount of ice from different areas varied slightly. According to the series of samples taken from the horizons at depths of 48, 82 and 160 m, accumulation made up 2.9 cm of ice per year for the last 5000 years. Surveys by one of the authors at Vostok station showed, that in this area snow accumulation is missing once in 5-6 years. If we take into account such omissions, the mean thickness of an annual layer equals 2.4 cm of ice or 2.2 g/cm2 per year, which is similar to the present-day accumulation rate of atmospheric precipitations. In order to interpret the age of the' data, we used the model of ice deformation with non-uniform strain rates proposed by Nye (Dansgaard and Johnsen, 1969). Two variants of conditions on the glacier bed were considered : freezing of the ice to the bedrock or exceeding the temperature of the pressure melting point. They have already been analysed by Dansgaard et al. (1971) while interpreting the data on the Camp Century borehole. Nye's model presupposes constancy in time of parameters influencing ice dynamics. Of course this assumption is rather rough and will require more refined models for the determination of the age of the ice. We used the above-mentioned model mainly because it had been applied to an interpretation of isotopic data from the Camp Century borehole. The use of the same methods to determine the age of the ice permits the data to be compared in a more consistent way. With the model used, the age of the core from Vostok station (T) can be expressed in a first approximation by T _ (2H - h)r lp2ff - h 2ajj 2y — h where aK is the thickness of the primary layer of annual accumulation; H is the thickness of the ice cover (3450 m, the upper 100 m make a firn layer); h = 400 m is the height above the bed above which the horizontal component of the strain rate becomes constant; y is the height of the level of the core sample above the bed ; and 384 N. i. Barkov eta/. T is the time unit, here equal to one year. The error of this age determination is estimated to be +5 per cent. It is well known that the shape of isotope profiles determined by data from deep boreholes in the ice depends on climatic changes (primarily on temperature conditions) and on variability of the absolute height of the surface of glacier covers, which is not constant in time. The absolute height of the Greenland and West Antarctica ice sheets was, evidently, greater during Wisconsin glaciation than it is now. In the area of Camp Century the glacial cover was 1200-1300 m higher than the present one. This is also confirmed by results of the analysis of ice samples from this borehole for gas content (Raynaud and Lorius, 1973). In the area of Byrd station the ice cap surface was 400-600 m higher during the Wisconsin stage than it is now. The bigger height of the surface caused lower temperature of precipitation deposited on the ice cap, which, in its turn, brought about the formation of glacier ice of lighter isotopic composition. This reason may be called the 'glaciologicaP agent in formation of isotope profiles of glacial cover in contrast to proper climatic factors. What is the contribution of the glaciological agent to the formation of the isotope profile of the Vostok borehole? Reconstructions of the Antarctic ice sheet for the stages of the Wisconsin and the maximum glaciations (see Voronov, 1960) show considerable fluctuations of its height in West Antarctica and peripheral areas of East Antarctica up to 1500-2000 m. Proceeding from the general laws of ice deformation and possible variability of the shape and thickness of the Antarctic ice sheet accompanying its larger dimensions, we may conclude, that the height of the surface in the area of Vostok station, which is not far from the ice divide of the East Antarctic dome, grew up by 100 m in the Wisconsin and changed by not more than 300 m in the period of the Quaternary maximum.