¹⁰Be is a radioisotope with a half-life of 1.5 x 10⁶ a. It is produced continuously in the atmosphere by cosmic-ray induced spallation reactions on nitrogen and oxygen. Within a short time it becomes attached to aerosols and after a residence time of 1 to 2 a it is removed from the atmosphere by precipitation and stored in different archives on the Earth’s surface. Polar glaciers and ice sheets contain the precipitation of the last 10⁵ to 10⁶ a in an undisturbed stratigraphy and are therefore best suited to study variations of ¹⁰Be production and deposition processes during this time period.

Reference RaisbeckRaisbeck and others (1981) first reported an increased ¹⁰Be concentration during the last glaciation and an apparent correlation with δ¹⁸O in Antarctic ice samples. These results were confirmed by more detailed measurements on ice cores from Dye 3, Greenland, recovered in 1981 in the frame of the Greenland Ice Sheet Programme (GISP) which reveal the following results (Reference BeerBeer and others 1983): (i) the ¹⁰Be concentration in biannual ice samples of the period 1900 to 1977 shows variations which are correlated with the 11-a solar cycle, (ii) at the end of the last glaciation the ¹⁰Be concentration was higher by a factor of 2 or 3 than during the Holocene, and (iii) the ^l0Be variations during this period and the transition from glacial to postglacial time are correlated with δ¹⁸O values and the CO₂ content of air bubbles in ice.

To study in more detail the variability of the ¹⁰Be concentration and its correlation with other parameters we investigated the ice core from Dye 3, Greenland, at a depth from 1 860 to 1 890 m, which corresponds approximately to the time interval 30 to 40 ka BP (Reference DansgaardDansgaard and others 1982). This interval shows several strong variations of the δ¹⁸O values. The 30 m-long section of the 1ce core was cut into pieces of length 1 m, each thus corresponding to about 350 a. The first 3 cm of each piece were used to determine the CO₂ content (Reference Stauffer, Hofer, Oeschger, Schwander and SiegenthalerStauffer and others 1984). The rest (1 to 2 kg) was melted and a few ml of water were separated to measure the δ¹⁸O, ¹⁰Be was extracted in the form of Be0 as described elsewhere (Reference BeerBeer and others 1983). The ¹⁰Be measurements were performed using accelerator mass spectrometry (Reference WölfliWölfli and others 1983). Figure 1 shows the results of the measurements in units of 10^{4 10}Be atoms per gram of ice together with the δ¹⁸0 data as a function of depth. Large ¹⁰Be variations are observed. The lowest values coincide with the mean value of the period from 1900 to 1977 (0.9 ± 0.2), but the mean value (1.6 ± 0.4) is higher by a factor of 1.7 than during that period. There are several possible causes which could explain the ¹⁰Be variations: (i) changes of the production rate in the atmosphere due to modulation of the cosmic-ray flux by magnetic properties of the solar wind plasma, (ii) changes of the production rate due to variations in the intensity of the geomagnetic field, and (iii) changes of the precipitation rate or the atmospheric circulation and mixing processes which are responsible for the transport of ¹⁰Be from the atmosphere to the Earth’s surface.

Fig. 1 ¹⁰Be concentration in units of 10⁴ atoms per gram of ice in the Dye 3 deep ice core (upper part) and δ¹⁸O values in parts per thousand (lower part).

δ¹⁸O also exhibits strong variations between about −31 and −36 ‰ (Fig.1) indicating significant climatic fluctuations. The variations of ¹⁰Be and δ¹⁸O are, in general, parallel. In Figure 2 the δ¹⁸O values are plotted together with the ¹⁰Be results. The scale is chosen such that the extreme values of the two parameters fit. The correlation coefficient between ¹⁰Be and δ¹⁸O is −0.61 (calculated using equal weights for all data points) and 1s different from zero with a significance level of 0.975. This strongly suggests that there is a mechanism which causes similar variations of both parameters. From the possible causes of ¹⁰Be variations mentioned above, atmospheric processes also influence δ¹⁸O. One mechanism which would affect ¹⁰Be concentration at the same time as δ¹⁸O is suggested as follows. Low δ¹⁸O values correspond to periods with lower mean temperatures. During colder periods the amount of precipitable water in the atmosphere is smaller than during relatively warm periods. This leads to lower precipitation rates. If the precipitation rate is reduced over large areas and if the atmospheric production rate is assumed to be constant, then the ¹⁰Be concentration per unit mass of precipitation is increased (Reference JungeJunge 1977). Further insight into the mechanisms determining the ¹⁰Be concentration 1n precipitation can be expected from measurements of additional parameters such as concentrations of NO3, SO₄, etc.

Fig. 2 Comparison of the δ¹⁸O curve (dotted line) and the ¹⁰Be concentrations (solid line).

The correlation in the depth range from 1 880 to 1 890 m is very good. However, the number of ¹⁰Be values is relatively small. For the interval from 1 867 to 1 880 m, which is dominated by low δ¹⁸O values, the correlation is less pronounced. This means that the observed ¹⁰Be variations can be only partly explained by changes of a single parameter, such as the precipitation rate. Variations of the atmospheric production rate due to changing solar properties cannot be excluded. ¹⁴C measurements on tree rings show that cold periods, for instance the Little Ice Age during the seventeenth century, were accompanied by high atmospheric radiocarbon levels, suggesting that both the isotopic production rate and the climate might be influenced by the state of the sun (Reference Suess and OlssonSuess 1970).

From the results shown in Figures 1 and 2 we conclude that during the period 40 to 30 ka BP there were strong ¹⁰Be variations with amplitudes similar to those at the end of the last glaciation (10 ka BP). The ¹⁰Be variations are correlated with δ¹⁸O values, probably due to changes of the precipitation rate. However, to explain the entire variability, changes in other processes such as ¹⁰Be production and atmospheric circulation and mixing have to be considered. A more detailed discussion is given elsewhere (Oeschger and others in preparation).