Introduction
Numerical models have been developed to describe the ice flow in temperate glaciers and ice sheets. Dating of ice is an important tool for validating these models. A reliable time-scale is also required in order to interpret the palaeoclimatic records derived from ice cores. The CO2 component of the enclosed air contains a radioactive 14C clock which starts when snow is transformed into ice and when air is sealed off from the atmosphere. Measurement of the residual 14C activity reveals the date of the ice. Unlike methods that count the annual varying parameters, such as δ 18O or conductivity. 14C dating does not depend on the observability of layers. This may be useful, particularly for dating ice from ablation zones.
The 14C activity in ice originates, however, not only from the trapping of atmospheric CO2, but also from the in situ production of 14C in the ice due to nuclear reactions by cosmic rays (Reference Fireman and NorrisFireman and Norris. 1982). After their formation, hot 14C atoms become oxidized as they slow down. During this process. 14CO2 and 14CO may be formed at a constant ratio Reference RoesslerRoessler, 1988). Because the atmospheric 14CO concentration is negligible, 14CO in ice is ascribed exclusively to in situ production and can be used to determine the amount of in situ 14CO2 on the basis of a known production ratio of in situ l4CO2 and 14CO (Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994).
After correcting the amount of 14CO2 in ice for the amount of in situ 14CO2 the ice can be dated.
The production of 14C decreases exponentially with depth due to the attenuation of cosmic rays in ice. Since surface ice is generally exposed to erosion, production depends not only on the intensity of cosmic rays but also on the exposure time and thus the ablation rate. The ablation rate ran thus be determined from the total amount of in situ products in the ice (Reference Lal, Jull, Donahue, Burtner and NishiizumiLal and others, 1990). A similar relation is expected between in situ production and the accumulation rate at the time of ice formation (Reference Lal and JullLal and Jull, 1990; Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994).
In this report we analyze shallow ice cores obtained from an Antarctic blue-ice area in Dronning Maud Land Both 14CO2 and 14CO measurements will be presented. Alter correction for the in situ 14C component the ages of the ice cores will be deduced. In addition we will discuss how the in situ 14C can provide a belter understanding of ablation and accumulation rates.
Field Description
Figure 1 shows schematically the blue-ice area at Scharffenbergbotnen (74°34′ S, 11° 03′ W). The coring locations are indicated. Cores 5. 6 and 15, with lengths of 3, 3 and 10 m. respectively, were used in this investigation. The cirque-like basin is downstream from the Heimefrontfjella nunataks, western Dronning Maud Land, East Antarctica. Inside the valley a depression has formed, where ablation occurs. The elevation here is about 1150 m, but the westerly ice divides are located 80–100 m higher. In addition to the ice that enters Scharffenbergbotnen from the ice divides, there is a small ice flow which enters the valley along a 400 m icefall in the east (Reference JonssonJonsson, 1992). The mean annual temperature in the centre of the depression is estimated at −18°C for the year 1988 (personal communication from R. Bintanja and S. Jonsson, 1994). Melt features have been observed only around the eastern part of the blue-ice area, at the bottom of the depression (Reference JonssonJonsson, 1992). The drilling locations
Fig. 1. View of the Scharffenbergbotnen basin. Indicated are the coring sites, from which cores 5, 6 and 15 are used in this work. For comparison, part of the Swedish stake net is marked
chosen were outside this area. In accordance with the Swedish Antarctic Research Programme (SWEDARP), a number of stakes were planted by Swedish scientists both inside and outside the basin for mass-balance and ice-movement studies. Measurements have been carried out since the field season 1988, and ablation rates have been reported over a 4a period (Reference Näslund, Melander and CarlssonNäslund, 1992).
Method
Ice sampling
The ice samples were handled carefully to prevent any contamination. No drilling fluids were used, and immediately after drilling the ice samples were sealed in polythene bags. The ice cores were kept below −12°C during transport by sledge, boat and truck. The ice samples were transported immediately in order to prevent the risk of post-core in situ contamination.
Before extraction, about 1 cm of the outer parts of the ice samples and a number of cracks in the samples were removed with a band-saw in a cold room at −25°C. Before the sample was loaded in the extraction vessel the outer parts were removed once more with a sharp knife under an argon atmosphere.
Experimental
A dry-extraction technique was chosen for extracting the gases from the ice (Reference RoijenVan Roijen, 1995). Dry extraction prevents dissolution of particulates or salts and minimises contamination from outgassnig (Reference Moor and StaufferMoor and Stauffer 1984). We expect in our set-up an efficiency comparable to that of the similar systems of Reference Moor and StaufferMoor and Stauffer (1984) and Reference NakazawaNakazawa and others (1993), which was found to be 90% and 98%, respectively.
After loading, the outer parts of the ice sample were evaporated for at least 5 h at −30°C. Under vacuum conditions, rotating razor blades driven by an external motor milled samples weighing approximately 3 kg for 20 min. The ice samples were chipped into small particles approximately 2 mm long and 0.2 mm in diameter. The stainless-steel walls of the 341 volume were kept at −30°C during operation to minimise outgassing. The total desorption of CO2 was measured and found to be less than 0.5/μlh−1 (STP).
About 80/μl (STP) of 14C-free CO was added to the extraction vessel before milling, as a carrier for the extraction of 14CO from the ice. After milling, the released air was extracted with a molecular sieve at T = 70 K for 45 min. The H2O and CO2 were removed in traps cooled with a dry-ice-alcohol mixture and liquid nitrogen. The CO was oxidized to CO2 by flushing it over CuO at 620°C. Air and CO2 yields were measured with an uncertainty of about 1% using a capacitance manometer.
The CO2 samples were converted into graphite samples weighing about 20-50μg and analyzed with the Utrecht AMS facility (Reference Borg, Alderliesten, Houston, de Jong and van ZwolVan der Borg and others, 1987). A mean contamination-level value of 1.4 ± 0.5 pMC (percentage of modern carbon) was obtained for CO and CO2 samples of 70–140μl (Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994).
Results and Interpretation
14C depth profiles
Table 1 describes the samples from cores 5, 6 and 15, the CO2 yield and the CO carrier used. The size of each ice sample is about 2.5 kg, a compromise between the sample size and the depth resolution. The CO2 concentrations obtained are 14–18 μg kg−1 and are shown in Table 1. Air extracted from core 15 decreases from 106 ± 1 ml kg−1 at 10 m depth, to 93 ± 1 ml kg−1 at 1 m depth. Table 2 summarises the results of the 14C analysis. Figures 2 and 3 show the 14CO2 and 14CO profiles (in atoms g−1 of ice) vs depth for the 3 m cores 5 and 6, and Figure 4 for the 10 m core 15.
Table 1. Ice samples used for analysis
Table 2. 14CO and 14CO2 concentrations in Antarctic ice
Fig. 2. Result of the 14C measurements of core 5. The filted equation C0exp(−x/Λ) + C1, is shown for 14CO and 14CO2. Note the low value in the 14CO concentration of the surface sample, which is not included in the fits
Fig. 3. Result of the 14C measurements of core 6.
Fig. 4. Result of the 14C measurements of core 15.
The presence of 14CO clearly indicates in situ production of 14C. The exponential decrease is according to the attenuation length of cosmic rays in ice. The equation C(x) = C
0exp(−x/Λ)+C1 is fitted to the 14CO data. where C(x) is the concentration (atoms g−1) as a function of depth x (cm), Λ is the attenuation length of cosmic rays in ice (170 cm), C
0 is the production coefficient (atoms g−1) and C
1 is a constant (asymptotic) level (atoms g−1) (Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994). A similar fit is made to the CO2 data from which the coefficients [
] and [
] are derived. Table 3 shows the results for cores 5. 6 and 15. Also listed is the ratio [
], which represents the 14CO2/14CO ratio for in situ production. The deduced values of 3.5, 3.1 and 4.1, respectively, agree within error limit. The fits are drawn in Figures 2–4 to illustrate the results.
Table 3. Calculated 14C concentrations (which result from the fit), deduced ages and surface mass-balance data. For comparison the ablation rates obtained from stake readings (Reference Näslund, Melander and CarlssonNäslund, 1992) are added. For the accumulation rales it is assumed that the altitude of ice formation is 1240 m.a.s.l.
Clearly the 14CO constant level differs from zero, which indicates the presence of another in situ component. Such a component is to be expected for 14CO2 as well. Assuming an equal ratio for in situ production of these components, the in situ 14CO2 at the constant level is obtained from the 14CO2 value by applying the deduced ratio. Subtraction of the in situ 14CO2, thus obtained, from the constant-level 14CO2 results in the remaining natural 14CO2.
Calculation of ages, ablation and accumulation rates
The deduced natural 14CO2 concentrations are compared with the 100 pMC for modem CO2, in ice (Table 2). The 14C ages obtained are 10000 ± 3000 BP (Table 3). Only the decrease by radioactive decay (14C half-life of 5570a) is taken into account in this calculation. The ages obtained in this way do not depend on the efficiency of dry extraction of in situ products, as the correction takes this into account (Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994).
For a constant ablation rate C 0 = P 0(Λ + a/Λ) (Reference Lal, Jull, Donahue, Burtner and NishiizumiLal and others, 1990), where P 0 is the production rate of the total in situ 14C in the ablation zone (atoms g−1 a−1), a is the ablation rate (cm a −1, and λ is the decay constant (λ = 1.25 × 10−4a−1). Using the total production rate P0 = 45 atoms g−1a−1 (Reference Lal, Jull, Donahue, Burtner and NishiizumiLal and others, 1990 for Scharffenbergbotnen, the ablation rates calculated are about 10 cm a−1 (Table 3).
The ratio of the total in situ 14C and the natural component as found in the asymptotic values of the ice in the ablation region may still he equal to the ratio in the region of origin, as the decay does not affect the ratio. We calculate this ratio R for the three cores (Table 3). We use this value to obtain the accumulation rate s in the relation R =0.176 P 0/s , where s (cm w.e.a −1) is the accumulation rate, and. P 0 (atoms g−1a−1) is the production rate of in situ 14C in the accumulation zone (Lal and rates varying from 45 to 85 atoms g−1 a−1 can be derived. Calculated accumulation rates for cores 5, 6 and 15 are 7–20 cm w.e. a −1 for P0 =45 atoms g −1a −1(Table 3).
Discussion
Calculated ages are independent of the 14C extraction efficiency because the correction method used in this work takes the efficiency into account (Reference Roijen, Bintanja, van der Borg, van den Broebe, de Jong and Oerlemans.Van Roijen and others, 1994). Calculation of the ages depends on the fixed relation between the in situ production of 14CO2 and that of 14CO. The obtained values of 3.5, 3.1 and 4.1 agree with a reported value of 6 ± 3 (Reference RoesslerRoessler, 1988). The ages obtained ate supported by the CO2 concentrations of the ice samples. Concentrations of about 16 μg kg−1 have been reported for Interglacials whereas at glacial stages concentrations were only 10μg kg−1 (Reference WalVan de Wal and Others, 1994). The CO2 yields for cores 5, 6 and 15 show values of about 14–18 μg kg−1, indicating a Holocene age. The almost Pleistocene age for core 15 indicates that any slow oxidation of CO to CO2 in the long duration between accumulation and ablation is negligible. But even if oxidation of 14CO occurred, the derived ratio would be too small and this would result only in an under- estimate of the ages.The calculated ages indicate that the surface ice at Scharffenbergbotnen is relatively young. From ice-flow considerations, “zero-age” ice is expected to be found on the equilibrium line between accumulation and ablation zones, and the oldest ice is expected to be near the boundary formed by the two opposite flow regimes. The oldest ice is still of Holocene age. The increasing age along the transect formed by cores 5, 15 and 6 is not contradicted by the 14C data.
The ablation rates obtained are in agreement with field observations of about 10 cm a−1 (Reference Näslund, Melander and CarlssonNäslund, 1992). It has been pointed out that the ablation rate is related to the total amount of in situ 14C in ice. Although dry extraction is known to be nearly 100% efficient when applied to gases, its efficiency when applied to in situ 14CO2 and 14CO is still unknown. The similarity of ablation rates obtained from field observations and from 14C measurements indicates that in situ products have been extracted almost completely. In fact, 14C ablation rates tend to be even lower than those from field observations, possibly due to the 4 a average of the stake readings, which is shorter than the 20 a period average of the 14C data. This shows that a precise determination of the extraction efficiency should make it possible to determine mean ablation rates.
Although some loss of in situ 14C from the firn cannot be ruled out. Reference JullJull and others (1994) recently reported on high concentrations of 14C in two Antarctic ice cores, and their data support the approximately total retention of in situ 14C. In our calculation of accumulation rates we assume total retention.
The calculated accumulation rales are proportional to lhe production rates, which in tum are a function of the altitude at the ice origin. Calculation of accumulation rates requires knowledge of the altitude of the area where the ice was formed. The calculated ages indicate a nearby origin, since reported surface velocities in Scharffenbergbotnen are only about 1 m a−1 (Reference JonssonJonsson, 1992). The accumulation zone may be located on the westerly ice divides (altitude 1240 m). This view is supported by the air content data of core 15. Using the empirical relation between total gas content and altitude of the area where the ice is formed (Reference Raynaud and Lebel.Raynaud and Lebel, 1979), a maximum altitude of 1900 m can be derived assuming a 100% gas-extraction efficiency. The obtained accumulation rates of approximately 7–20 cm w.e.a−1 are in the same range as the value at the equilibrium line and a present-day value found 5 km from the valley, 20 cm w.e.a−1 (Reference Jonssonjonsson, 1992). It would be valuable to compare the information obtained about ages and ablation and accumulation rates with results obtained with an ice-flow model. This model should Focus on the local flow around Scharffenbergbotnen. Of course, this would mean extension of the measurements to parts of the blue-ice area that have lower ablation rates and younger ages.
Conclusions
It is shown that 14C ages can be determined in surface ice samples using the dry-extraction method. However, the correction that has to be made for the in situ component in 14CO2 exceeds the natural value. This means significant errors of about 25% in the final ages. Calculated ages are 10 000 ± 3000 BP; these values are supported by the CO2 concentrations of the ice samples and indicate that most of the surface ice of Scharffenbergbotnen is of Holocene age. The deduced ablation rates are in agreement with field observations, indicating an almost complete extraction for in situ produced 14C. The deduced ages, and measurements of the air yield. support a nearby accumulation zone as the area where the ice was formed, The area is presumably at the westerly ice divides.
Acknowledgements
The authors thank C, Alderliesten for his contribution to the AMS measurements, M.P. R. van den Broeke, R. Bintanja and M.J. Portanger for their drilling efforts and L.A. Conrads for organizing the drilling logistics. Discussions with L. Lindner have proved valuable. M. Hanegraaf is thanked for his contribution to the construction and K. C. Welten, R. Eisma and P. van Andel for their valuable assistance with the operation of the ice-milling device. financial support For this investigation was provided by the Netherlands Marine Research foundation (SOZ).
