3.1 Determination of radioactive layers

The Chernobyl accident occurred on 26 April 1986 and incoming radioactivity was noted in Svalbard by

Reference Pourchet, Pinglot and GascardPourchet and others (1986). On Austfonna (Nordaustlandet), the first snowfall containing radioactive elements occurred on 4 May and was followed by precipitation events on 10, 11, 17, 18 and 25 May. Precipitation occurring on 10 and 11 May was highly contaminated. In Ny-Ålesund, the first radioactivity was recorded on 13 May in fresh snow which fell with a southern wind. Thereafter, precipitation occurred on 20 May and 3 and 4 June but was not analysed. Meteorological records at Ny-Ålesund indicate that except for 4d, the period 10 May–6 June was almost completely cloud-covered. It is therefore likely that radioactive snowfalls occurred in the Kongsvegen and Sveabreen basins during this period.

The first ice core (K in Fig. 1), to a depth of 20.65 m, was taken in 1989 at an altitude of 639 m on Kongsvegen, and nine ice cores (A, Β … I in Figure 1; Tables 1 and 2) were taken in April 1990 on Kongsvegen and Sveabreen. The ice cores were obtained in the accumulation area to depths between 7 and 12 m and one in the upper part of the ablation area to a depth of only 5 m. Each ice core was cut into sections approximately 15 cm long (almost ten samples for one accumulation year), with an average mass of 300–400 g. After density measurement, these samples were melted and filtered according to the method described by Reference Delmas and PourchetDelmas and Pourchet (1977). The radioactive active species were extracted on to ion exchange filters in the laboratory (Reference Pinglot and PourchetPinglot and Pourchet, 1979, Reference Pinglot and Pourchet1989). The total β radioactivitiy was measured on individual filters and ¹³⁷Cs was detected using γ spectrometry on composited samples.

Table 1. ¹³⁷ Cs activities and deposited amounts vs depths at all stations. In column 6, layer A corresponds to the ablation surface from 1989 and is equivalent to the 1988–89 winter-snow deposit, layers Β and C correspond to the ablation surfaces from 1986 and 1962, respectively

Table 2. Mean values of measured and estimated net balances in m of water equivalent. The altitudes in italics in column 2 are the estimated mean altitudes between 1964 and 1990. The methods which have been used for measurement or estimation are noted in column 6: “direct” = direct method by stake reading and density measurement, “geodetic” = lowering of the surface from topographic work in the ablation area, “index” = from detected layers with ³⁷ Cs activity to the 1989 ablation surface

3.2 Measurement of total β radioactivity

The measurement of total β radioactivity was carried out on samples of all ice cores. This measurement includes but does not discriminate between the total β radiation emitted by natural (cosmogenic and terrestrial) and artificial isotopes (fission products from atmospheric nuclear tests, accidents in power stations, etc.). Some years after their emission, the main artificial sources of β radiation are ⁹⁰Sr and ¹³⁷Cs, the latter emitting also γ radiation.

The total β radioactivity profiles (see Fig. 2) do not show clearly characteristic peaks which can be absolutely attributed to the fall-out of Chernobyl elements. This is due to the fact that the natural total β radioactivity is about 10 times the artificial radioactivity. Between the surface and about 2.5 m depth at stations A, B, C, F and I, the winter snow shows a total β radioactivity level constantly higher than 200 dph kg⁻¹. This activity is due to terrestrial and cosmogenic radio-isotopes, mostly of short half-lives, from ²³⁸U and ²³²Th series, as well as ⁴⁰K, etc. Below this winter deposit (noted A in column 6 of Table 1), the profiles from the previously mentioned stations clearly exhibit individual peaks which may correspond to different summer layers, because it is during this period that the local outcrops are free of snow and surface dust is exposed to aeolian transport and also because the radio-isotopes may become concentrated near the ablation surface.

Fig. 2. Total β activity (thin lines) and ¹³⁷Cs activity (thick lines) versus depth, at all stations in accumulation areas

3.3 Measurement of the γ radioactivity

To determine the 1986 Chernobyl reference level, γ spectrometry has been carried out on all ice cores. The samples were analysed in groups because this technique of measurement is less sensitive than total β counting. The re-grouping was made around the previously described individual β peaks.

The Chernobyl evidence is based on our measurements of ¹³⁷Cs and ¹³⁴Cs. The spectrometry of these elements was carried out using a high-purity germanium diode of Ν type. This detector has a relative efficiency of 20%. It is made of material free of radioactivity and located at the centre of nuclear shielding in a special building, free from all parasitic radioactivities such as cosmic rays, radon emanations, γ rays from building materials, etc.

Whilst the detection of ¹³⁴Cs confirms the local Chernobyl fall-out, the work has been focused on the ¹³⁷Cs activity (see Table 1 and Fig. 2). From the snow surface down to a definite depth, the caesium content is zero or insignificant (below detection level, which is about 10 dph kg⁻¹or 3 × 10⁻³ Bq kg⁻¹ at 95% confidence level). Thereafter, for every ice core, the caesium level exceeds the minimum detection level. A positive identification of ¹³⁷Cs indicates the presence of snow contaminated by Chernobyl even though, due to percolation, the maximum activity may be located lower down. This assessment agrees with previous work by Arnbach and others (1989).

Ice core I (726 m a.s.l. and 12 m depth), collected at Kongsvegpasset, provides the best reference data on the Chernobyl layer of May 1986 and allows the determination of total deposition of ¹³⁷Cs. Even at this station, which is the highest point on the glacier, some impurities have migrated downward into deeper layers by percolation, and ¹³⁷Cs is found 2–3 m below the reference level of May 1986. The total ^l37Cs fall-out measured is 20 Bq m ⁻², which is similar to fall-outs measured in Greenland (Reference Pourchet, Pinglot and GascardPourchet and others, 1986; Reference Davidson, Harrington, Stephenson, Monaghan, Pudykiewicz and SchellDavidson and others, 1987; Reference DibbDibb, 1989) ranging from 10 to 20 Bq m⁻² in the northernmost part (between 77° and 81° N) and at the highest location (Summit station). These values are, however, much lower than the fall-outs measured in Alpine glaciers (Arnbach and others, 1987; Reference Haeberli, Gäggeler, Baltensperger, Jost and SchottererHaeberli and others, 1988). At all other stations, the ice cores do not penetrate deep enough and therefore we have only partial fall-out results. Ice core H at 390 m a.s.l., in the ablation area, shows a high concentration of radioactive elements at the glacier surface. At this location, the successive radioactive layers deposited since the first atmospheric tests are concentrated at the ice surface.

Ice core Κ (639 m a.s.l. and 20.65 m deep) was taken in 1989 mainly for use as a test; thus, the radioactive profile is not continuous and samples between 5.90 and 12.95 m were not analysed. Nevertheless, its γ profile shows two steps of radioactivity at 14.82 and 16.74 m; then, a further step from 18.75 to 20.65 m which reaches a ¹³⁷Cs activity of 80 dph kg⁻¹. This activity is around 8 times the activity of the samples located between 12 and 15 m, and may be attributed to the last important nuclear tests conducted in 1961 and 1962, especially by the U.S.S.R. in Novaya Zemlya (75° N, 55° E). Comparison with total β profiles obtained in Nordaustlandet (Reference Vaykmyae and PunningVaykmyae and Punning, 1989) and at Station Central in Greenland (Reference Holdsworth, Pourchet, Prantl and MeyerhofHoldsworth and others, 1984) (used here as references), suggests that the samples from 18.65 m depth correspond to the year 1962. Radioactive profiles from an ice core of 24 m collected on top of Kongsvegen in April 1992 (Reference Pinglot and PourchetPinglot and others, 1992) and from an ice core of 26.80 m depth, collected nearby at the same time, have recently confirmed this conclusion.

3.4 Kongsvegen balance from the direct glaciological method

The mass balance of Kongsvegen has also been obtained since 1986–87 by stake measurements, snow-depth sounding profiles and density measurements in pits along the glacier. There are nine stakes from the summit at 726 m a.s.l. down to 198 m (S9, S8 … S1 in Figure 1; stake altitudes in Table 2). Soundings were made every 200 m. At the end of the ablation period, density of the firn was determined in two pits, one at the summit and one around 620 m a.s.l. Net accumulation close to the equilibrium line consists only of superimposed ice.

As mentioned previously, the calving of Kongsvegen is very small (less than 2% of the ablation by melting). Nevertheless, this calving is taken into account in the results in Table 2. The mean annual balance in the 3 years 1986–89 is positive (0.11 m w.eq.; Table 3), mainly due to a higher positive value in 1987. Reference Hagen and LiestelHagen and Liestol (1990) have measured the balances of Broggerbreen, a small glacier in the vicinity of Kongsvegen, since 1967 and their result is also presented in Table 3. There is consistency in the change of the balance of both glaciers.

Table 3. Compared mean net balances of Kongsvegen and Breggerbreen

4.1 Definition of the periods

Fall-out of Chernobyl elements occurred in May 1986. Following this event, there was negligible precipitation until the beginning of the ablation period which started in June. In Ny-Ålesund, the atmospheric temperature rose by 5°C up to 1 June and only 14 mm of rain was recorded during this month. Thus, the Chernobyl horizon corresponds to the ablation surface from 1986. Nuclear tests in 1962 in Novaya Zemlya occurred between August and December. In the present work, we link this horizon to the 1962 ablation surface, even though the upper part of this horizon could be included within the snow layer from the 1962–63 winter. The introduced uncertainty is small as the calculations will be carried out using a period of 26 years.

The sampling used in the radioactivity measurements was conducted in the spring (April and May 1989 and 1990). To obtain net-balance measurements corresponding to budget years, the current winter snow depth was measured at every pit in order to determine the 1988 and 1989 ablation surfaces. The winter snow depths, which are given in Table 1 (A in column 6), are not taken into account in the balance of the following periods of 3 years (1986–89) and 26 years (1962–88). Therefore, in Table 2 and Figure 5, the periods given correspond to 3 or 26 budget years.

4.2 The 1986–89 balance

There is consistency between the net-balance results of both the direct measurements and using the Chernobyl layer as shown in Figure 3. At the two locations where sampling was made coincidentally with stake measurements (i.e. stations F and I and stake 9, station A and stake D2) both results are identical or very close.

Fig. 3. Mean net balance on Kongsvegen and Sveabreen. The heavy line shows the mean net balance of Kongsvegen during the period 1986–89 using the direct glaciological method. S1, S2... S9, D1 and D2 are ablation stakes on the glacier surfaces. The dotted line shows the average for the 26years’ period 1964–90, based on geodetic work in the ablation area of Kongsvegen connected to the mean balance estimated at station Κ by radioactive analysis for the 26years’ period 1962–88. Inset is shown the range of net-accumulation values during the period 1986–89 obtained from the Chernobyl layer. The letters refer to the ice cores described in the text: (I) and (F) are located at Kongsvegpasset in the common accumulation area of the two glaciers; B, A and G with heavy lines and C, D and Ε with dotted lines are located on Kongsvegen and Sveabreen, respectively

Results from ice cores located between stakes complete the data set and demonstrate that the maximum of net accumulation extends in a zone 2.5 km long from the top of Kongvegen and 7 km long from the top of Sveabreen. The net accumulation on Sveabreen is higher than on Kongsvegen (see Fig. 3) and the ELA is around 520 m a.s.l. for Kongsvegen and 450 m a.s.l. or a little lower for Sveabreen. Stratigraphie profiles at station G on Kongsvegen (562 m a.s.l.) indicate that the total net accumulation is composed of superimposed ice, while at station C on Sveabreen (538 m a.s.l.) the accumulation comprised both firn and ice lenses down to a depth of 6.80 m. This asymmetry of the two accumulation areas has also been observed by Reference Korolev, Sin’kevich and TarusovKorolev and others (1988). The maximum of net accumulation reflects the winter-balance distribution (see Fig. 4), especially on Sveabreen. Meteorological records at Ny-Ålesund and Longyearbyen, the two closest stations, indicate that snow precipitation occurs mainly with the winter prevailing winds from the southeast sector. This may explain the high rate of snow deposition on the windward Sveabreen and its higher net balance in the accumulation area.

For the period 1986-89, the balance for Kongsvegen was slightly positive (bn = 0.11 m w.eq.) and the ELA at 520 m a.s.l. (see Fig. 3). This ELA corresponds to an accumulation-area ratio of 0.67. On Sveabreen, for the same period, an ELA of 450 m a.s.l. corresponds to an accumulation-area ratio of 0.6 but, because of the higher accumulation value on Sveabreen, the balance of this glacier for the 3 analysed years would also have been very close to the equilibrium when the calving is ignored.

Fig. 4. Relationship between winter accumulation and altitude on Kongsvegen (Ko) and Sveabreen (Sv) in May 1991 (upper curves) and May 1990 (lower curves)

4.3 The 1962–88 balance

The net balance on Kongsvegen can be estimated using the 1962 radioactive layer in the accumulation area combined with the surface-topographic survey carried out in 1991 in the ablation area. The accumulation value is for the 26 years 1962–88 and the ablation values are for the 26 years 1964–90. In the accumulation area of Kongsvegen, at station Κ (639 m a.s.l.), the 1962 level is located between 18.75 and 20.65 m depths. That corresponds, when the snow from the winter 1988–89 is not taken into account, to a minimum total accumulation of 11.28 m or a maximum of 12.93 m of water equivalent. Annual mean accumulation is then between 0.43 and 0.50 m w.eq. for the 26 years of the period 1962–88, i.e. slightly lower than the mean net accumulation for the years 1986–89 obtained from ice core A (between 0.45 and 0.52 m w.eq.) or by direct annual measurements (0.48 m w.eq.).

In the ablation area, the net-balance values can be estimated from geodetic work for the period 1964—90. Reference PillewizerPillewizer (1967) carried out a topographic survey on the tongue of Kongsvegen in 1962 and 1964. He constructed a map at a scale of 1:50 000 with a contour interval of 10 m covering the entire ablation area of the glacier. Positions and altitudes of the ten stakes (S1 … S9 and D2) located on the central axis of the glacier were surveyed in spring 1991, and the altitude of the ablation surface in 1990 was determined. In the ablation area, the net ablation corresponds to a lowering of the surface multiplied by 0.9, the mean ice density. The ice flow is slow and almost constant along the area. Reference MelvoldMelvold (1992) has measured movements between 0.72 and 0.81 cm d ⁻¹ for stakes 1–6. Owing to a fairly constant slope of 2.5% and a constant velocity, the unknown vertical flow component may be neglected. At each stake, the change in altitude of the ice surface between 1964 and 1990 has been calculated and counted as the net ablation for the 26 years. The results (see Fig. 3) are the mean annual values for the 26 years plotted versus the mean altitudes for the period 1964–90. The result is remarkably parallel to the curve obtained from direct measurement during the period 1986–89. Long-term stability of mass-balance gradient has been noted by various authors (Reference Meier and TangbomMeier and Tangborn, 1965) and has been used as the basis of a linear model (Reference LliboutryLliboutry, 1974). From the values obtained, an annual net ablation of −34.70 × 10⁶ m³ w.eq. can be calculated for the 26 years’ period 1964–90. In the accumulation area, the total accumulation must be estimated from the single-range value in the core at point Κ (between 11.28 and 12.93 m w.eq.). From the maximum and minimum values at this point, extrapolation of the mean accumulation curve obtained by direct measurements in the years 1986–89 and plotted in Figure 3, gives an accumulation between 22.2 and 26.7 × 10⁶ m³ w.eq. That corresponds to an approximate mean net balance of between −0.13 and −0.08 m w.eq. and an ELA around 550–560 m a.s.l.

In the accumulation area of Sveabreen for the same 26 year period, extrapolation gives a total mean net accumulation of between 31.9 and 44.8 × 10⁶ m³ w.eq., i.e. much greater than on Kongsvegen. The ELA was around 480 m and the accumulation-area ratio only 0.52. The ablation and calving rates are not known. Hence, in spite of the fact that the net accumulation on Sveabreen is higher than on Kongsvegen, Sveabreen more than likely had a negative balance.

In Svalbard, the change in net balance from year to year is well correlated with climatic parameters and is consistent between all surveyed glaciers (Reference Hagen and LiestelHagen and Liestøl, 1990; Reference Lefauconnier and HagenLefauconnier and Hagen, 1990). The present results may be compared with the change in balance from Broggerbreen (see Table 3) which has been surveyed since 1967. During the 2 years 1987–89 for which the Kongsvegen balance has been the same as the estimated mean net balance for the 26 year period (−0.10 m w.eq.), the Broggerbreen balance was lower (although very slightly) than its mean value since 1967. We may conclude that Kongsvegen had effectively a mean net balance near −0.10 m w.eq., i.e. not far from equilibrium. The comparison also confirms, for the recent 3 year period, a more positive balance than during the longer period. Such a situation may have been caused by the slight change in climate towards a positive balance that has been detected for Broggerbreen since 1967 (Reference Lefauconnier and HagenLefauconnier and Hagen, 1990).