Devon Ice Cap and Austfonna

As a reference, the radioactivity profiles of an ice core from 152 a glacier with negative (−23°C) 10 m temperature (Devon Ice Cap) will be compared with profiles from Austfonna, a sub-polar glacier. In cold glaciers the 10 m temperature is negative, closely representing the mean annual temperature. In the accumulation area of sub-polar glaciers, the 10 m temperature is generally about 0°C or slightly negative. It is negative in the ablation area.

For the Devon Ice Cap ice core, there is excellent agreement between total-beta ¹³⁷Cs and ³H measurements, the 1963 peak of artificial radioactivity occurring at almost the same depth (Fig. 3a and b). The 1963 horizon was generally assigned to depths where there was a clear peak of radio-activity for either ³H or ¹³⁷Cs. For at least two ice cores (Aus87 and Ves81) the identification of the 1963 horizon needs special attention. This 1963 date was fixed in the following way: after the large atmospheric nuclear tests in 1961 and 1962 there was a moratorium, with no other tests for some years. Even if there is not a clear maximum, one can consider that when the radioactivity, after ranging at a given level, falls to a low value at shallower depth, then the 1963 horizon is quite close to this depth.

Fig. 3. Radioactivity at time of fall-out (¹³⁷Cs and ³H), and at time of measurement (total beta and ²¹⁰Pb, respectively, 1986 and 1993) vs depth for Devon Ice Cap (a, b) and Austfonna (c, d).³H measurements from Devon Island (thin line) are not continuous.The arrow indicates the equivalent 1954 computed year.

The ³H profile is not continuous, and extrapolated values were included for estimating the fall-out. Thus the exact location of the 1963 maximum is estimated to occur at depths of 8.18–8.32 m w.e.(Table 2). The ¹³⁷Cs profile also shows the 1959 peak. As will be shown later for other ice cores (Figs 4–6), the beginning of atmospheric fall-out in 1954 is well marked only for the Devon ice core (Reference Koerner and TaniguchiKoerner and Taniguchi,1976). The initial increase of ¹³⁷Cs compared to ³H has already been pointed out by Reference Koide, Michel, Goldberg, Herron and LangwayKoide and others (1982) and Reference Holdsworth, Pourchet, Prantl and MeyerhofHoldsworth and others (1984).

Fig. 4. Radioactivity at time of fall-out (¹³⁷Cs and ³H), and at time of measurement (total beta and ²¹⁰Pb, respectively, 1988 and 1993) vs depth for three ice cores from Austfonna (a–d). The arrow indicates the equivalent 1954 computed year.

For the 1985 Austfonna ice core, the radioactivity profiles (Fig. 3c and d) include total beta,¹³⁷Cs, ²¹⁰Pb and ³H. For the ¹³⁷Cs profile, a clear maximum occurs at 9.88–10.76 m w.e. depth, indicating the 1963 maximum. This is in close agreement with the 1963 ³H maximum, which is, however, slightly shallower than the ¹³⁷Cs maximum. The ³H atoms are constituents of the water molecule and, compared to the ¹³⁷Cs particulates, much of the radioactivity (atoms) contributing to the ³H maximum was not propagated downwards. The ¹³⁷Cs profile shows another maximum at about 13.5 m w.e., possibly representing the 1959 fall-out (Fig. 3c).

The total-beta profile (Fig. 3d) cannot be used for dating. The clear maximum corresponds to fall-out that occurred well before 1963, as measured from ¹³⁷Cs. It is close to the bottom of the ice core and does not represent any artificial radioactivity. This maximum corresponds to a very high level of ²¹⁰Pb, as shown on the profile (Fig. 3d). Very similar disturbed total-beta profiles were also measured for Austfonna 1987 (Fig. 4d) and Vestfonna 1981 (Fig. 5c). This ²¹⁰Pb increase is not supported by long-lived parents ( ²³⁸U, ²²⁶Ra). The total-beta values involve beta activity from ²¹⁰Pb, but also the accompanying alpha and beta activities from the ²¹⁰Pb daughters (²¹⁰Bi and ²¹⁰Po) and to a lesser extent from radioactive dust particles (²²⁶Ra and ²¹⁴Pb from ²³⁸U parent). Such a large ²¹⁰Pb increase also occurs in all ice cores retrieved from Austfonna (1985, 1987 and 1998), except for the 1999 ice core. Although the ¹³⁷Cs and ³H profiles provide valuable dating information at Austfonna and for ten other ice-core locations (Reference PinglotPinglot and others, 1994), the ²¹⁰Pb and total-beta profiles reveal a strong scavenging process due to melting and infiltration for these particulate-composed elements, as for ionic species (Reference Goto-Azuma, Enomoto, Takahashi, Kobayashi, Kameda and WatanabeGoto-Azuma and others, 1993). Chronologies based on the radioactive decay of a given isotope (e.g. ²¹⁰Pb) are generally corrupted by post-depositional processes.

Fig. 5. Radioactivity at time of fall-out (¹³⁷Cs and ³H), and at time of measurement (total beta and ²¹⁰Pb, respectively, 1982 and 1993) vs depth for three ice cores fromVestfonna (a–c).The arrow indicates the equivalent 1954 computed year.

Along the ¹³⁷Cs profiles, the 1963 maximum is located at greater depth than in the ³H profiles (Fig. 4a and b, Table 2). For three ice cores from Austfonna, the 1963 depth difference ranges from 0–0.72 m w.e. corresponding to 0–1.6 years of net accumulation, while the atmospheric signals indicate that both ¹³⁷Cs and ³H fall-out peaks occurred at the same time (1963; Fig. 2). This confirms that the ¹³⁷Cs fall-out experiences downward migration due to melting and percolation. Total-beta and ²¹⁰Pb profiles at Aus87 (Fig. 4d) and Aus85 (Fig. 3d) are disturbed. This demonstrates that the absolute dating of ice cores in this sub-polar ice cap depends onthe isotope measured: the ³H profile gives a better chronology than ¹³⁷Cs.

Vestfonna

Samples from two ice cores from Vestfonna were also analyzed for ¹³⁷Cs and ³H (Fig. 5a and b).

For the 1981 Vestfonna ice core (Fig. 5b), there is a clear deformation of the original atmospheric signal both for ¹³⁷Cs and ³H. The tentative depth given for the 1963 peak is based on the ³H profile as the ¹³⁷Cs profile shows a strongly disturbed signal. The depth difference ranges from 0.60–0.73 m w.e. and corresponds to about 1.7 and 1.9 years of net accumulation for Vestfonna 1981 and 1995, respectively (Table 2). Note that the summit of Vestfonna (580–600 m a.s.l.) is lower than Austfonna (758–783 m a.s.l.), so higher melting is more likely at Vestfonna.

Lomonosovfonna

The same analyses were conducted for two ice cores from Lomonosovfonna, drilled in 1997 (Lom97) and 2000 (Lom00) (Table 1), located about 150 m from each other (Fig. 6a and b). For the1997 core, there are similar trends for ¹³⁷Cs and ³H and the 1963 peak is well-defined. The detailed ³H profile (preliminary, as all samples are not yet analyzed) was determined from the analysis of 5 cm resolution samples.

Fig. 6. Radioactivity at time of fall-out (¹³⁷Cs and ³H), and at time of measurement (total beta and ²¹⁰Pb, respectively, 1998 and 2000) vs depth for two ice cores from Lomonosovfonna (a, b) and for an ice core from Akademii Nauk ice cap (c). The arrow indicates the equivalent 1954 computed year.

For the core drilled in 2000, the ³H profile reflects the atmospheric signal while the ¹³⁷Cs profile shows large temporal variations.The 1963 layer at Lom00 (9.05 m w.e.) is not as deep as the Lom97 ice core (12.98 m w.e.) (Table 2). As discussed below, the low ³H fall-out at Lom00 most probably indicates that the original precipitation is low, as is the MANMB. The ²¹⁰Pb profile (Fig. 6b) at Lom97 indicates a general decrease of activity with depth. However, dating of this ice core from ²¹⁰Pb decay does not reproduce the 1963 horizon.

Akademii Nauk ice cap

In order to extend our study of the spatio-temporal variations of the MANMB and the global mass budget to other glaciers and ice caps in the high Arctic, ¹³⁷Cs and ²¹⁰Pb profiles were also measured in an ice core from Akademii Nauk ice cap (Fig. 6c). The 1963 peak for ¹³⁷Cs is well-defined between 15.55 and 16.73 m w.e. This is in agreement with the 1956 Bezymianni volcanic layer studied by Reference FritzscheFritzsche and others (2002). Previously, Vaikmäe and others (1980) analyzed ³H at the nearby Vavilov ice cap and concluded that the corresponding profile was disturbed. The 1954 computed year of ¹³⁷Cs fall-out does not correspond to this absolute date. Therefore it could be concluded that melting also occurs at the summit of Akademii Nauk ice cap, which is supported by positive air temperatures recorded by an automatic weather station here in summer 2000 (Reference Hagen and MelvoldHagen and Melvold, 2001). Reference Vaikmäe, Punning, Romanov and BarkovVaikmäe and others (1981) even reported temperature of +10−15°C at Severnaya Zemlya.

Chronology

The ice cores from the Arctic area have been dated using radioactivity measurements. The most accurate detection of the 1963 peak of artificial radioactivity is obtained from ³H using the proportional-counting technique, and then from ¹³⁷Cs. This is due to the higher resolution of samples subjected to ³H analysis, 0.04 and 0.07 m w.e. for the Lom97 and Dev98 ice cores, respectively (Table 2). Gamma spectrometry of ¹³⁷Cs needs samples with higher mass and length, compared to ³H analysis. ³H dates are also better than ¹³⁷Cs because the 1963 horizons are much stronger for ³H and largely remain at the original fall-out depth.There is generally a stronger migration of ¹³⁷Cs compared to ³H, as shown below (Fig. 7). The depth differences for the 1963 peak, based on ¹³⁷Cs and ³H, are indicated in Table 2, together with the corresponding age differences. Minimum and maximum depths correspond to the length of samples representing the 1963 layer. While this layer is well defined for ³H, the ¹³⁷Cs layer corresponding to 1963 was in some cases estimated to be just before the decrease of activity following the nuclear-test moratorium. However, for each ice core, the depth differences for 1963 is lower than the length of the corresponding ³H or ¹³⁷Cs sample (Table 2).

Fig. 7. ¹³⁷Cs and ³H fall-out ratio from atmospheric nuclear tests, determined from 1963 mean time of fall-out and absolute date of snow layers, respectively.

The differences in dating obtained using ¹³⁷Cs and ³H are close to zero for several ice cores and may be not significant for the other ice cores.The maximum difference may extend up to 2 years at Vestfonna and slightly opposite (−0.8 year) for Devon Ice Cap.This may be due to the discontinuous ³H sampling, with one sample in the 1963 layer that was not analyzed. The total-beta activity profile was used to estimate the probable depth of the 1963 layer for ³H (i.e. 8.18–8.32 m w.e.).

Mean annual net mass balances (MANMB)

All ice cores were retrieved from the summits of the studied ice caps with very low horizontal velocities and terrain slopes. The thinning effect of deeper ice layers (down to 40 m in this study) is negligible, and we can use the chronology results to determine the MANMB.

The MANMB values range from 0.17 (Fin94) to 0.47 m w.e. a⁻¹ (Aus85 and Sno92) (Table 3) from 1963 to the drilling dates (1981–2000). Apart from a MANMB value determined for Devon Ice Cap by Reference Koerner and TaniguchiKoerner and Taniguchi (1976), earlier MANMB values determined for other Arctic locations do not agree with our determinations, due to difficulties in interpreting the stratigraphy and total-beta profiles. This is particularly true for ice cores from Nordaustlandet (Vestfonna and Austfonna), where MANMBs were previously misinterpreted , as shown by Reference Pinglot, Hagen, Melvold, Eiken and VincentPinglot and others, 2001.

The MANMBs of Lom97 and Lom00 ice cores (1250 m a.s.l.) are 0.36 and 0.23 m w.e. a⁻¹, respectively. These are much lower than the previous 0.82 m w.e. a⁻¹ value determined for a 1000 m a.s.l. site (Reference Gordiyenko, Kotlyakov, Punning and VaikmäeGordiyenko and others, 1981). Earlier studies (Reference Zagorodnov, Klementyev, Nikiforov, Nikolaëv, Savatyugin and SasunkevichZagorodnov and others, 1990) reported MANMB values at the summit of Akademii Nauk ice cap ranging from 0.20–0.30 m w.e. a⁻¹ compared to our determination of 0.45 m w.e. a⁻¹.

¹³⁷Cs and ³H fall-out and melt indexes

Both ¹³⁷Cs and ³H fall-out from atmospheric nuclear tests (1954–74) over Arctic glaciers have been determined (Table 3). The transport mechanisms (residence time) of ³H and ¹³⁷Cs through the atmosphere are almost the same (Reference Pourchet and PinglotPourchet and Pinglot, 1979), so they do not affect the location of the 1963 peak in the cores. Assuming a constant MANMB at each ice-core location, as already demonstrated for the period from 1963 to the date of the drillings (Reference PinglotPinglot and others, 1999), the ¹³⁷Cs and ³H activity profiles have been decay-corrected, taking into account the precise date of the snow layers. This will enable us to compute the fall-out of both isotopes in two ways. The hypothesis is that the exact 1963 peak did not propagate downward. Even if 1963 is not the exact mean date of atmospheric nuclear tests fall-out, the melt index studied is representative of the strength of melting and percolation processes.

In the first correction for decay, we assume that the mean date of fall-out for all samples is 1963. This is a reasonable assumption, given that the total period of atmospheric nuclear testing extends equally before and after 1963. For the second correction, each sample radioactivity value is corrected to take into account the absolute date of the snow layers. From the date of drilling, dating is given by the equivalent depth (m w.e.) divided by the MANMB, as previously determined. Both corrections take into account the date of all measurements, extending from 1983–2000 (Table 3). The depth of the layer corresponding to the year1954 (first significant increase of artificial radioactivity in the Arctic) was then determined and included on all profiles (Figs 3–6).

There is a general disagreement between the 1954 computed equivalent depth and the first increase of isotopes. The ¹³⁷Cs profiles apparently propagated to greater depths than the 1954 equivalent depth. This clearly demonstrates the post-depositional processes due to melting and percolation during successive summers. Periods of melting during warmer summers in the early 1950s have been reported (Reference Fisher and KoernerFisher and Koerner, 1994; Reference Førland, Hanssen-Bauer and NordliFørland and others, 1997). This is also in accordance with the ²¹⁰Pb profile (Fig. 3d), showing a major activity peak at greater depths, as has been detected for other Austfonna ice cores (Reference PinglotPinglot and others, 1994). ¹³⁷Cs and ³H fall-out will be overestimated as a result of the overestimation of the age of deeper samples. Figure 7 indicates both the ¹³⁷Cs and ³H fall-out ratio, as determined either from a decay correction of fall-out computed for 1963 or for the absolute date of each snow layer.

This fall-out ratio represents a melt-index equivalent. Instead of a fall-out ratio ideally equal to 1 for cold glaciers and ice caps, the present fall-out ratios are close to 1.2 for ice cores from Lomonosovfonna (Lom97), Devon Ice Cap (Dev98) and Snøfjella (Sno92) (Table 3). For the other cores studied, the respective fall-out ratio increases, ranging up to about 2 for lower-altitude cores from Vestfonna (Ves81) and Finsterwalderbreen (Fin94), at 580 and 668 m a.s.l., respectively.

The fall-out ratio is generally higher for ¹³⁷Cs than for ³H (five cores), except for Ves95 and Lom00. This melt-index study also demonstrates that percolation is less important for ³H than ¹³⁷Cs. As the ¹³⁷Cs profile at Lom00 core reveals large variations, perhaps the corresponding corrected fallouts were not determined accurately. This demonstrates that for a given ice-core location, the study of the fall-out of radioactivity can be used to determine the relative strength of the melting and percolation processes. There is evidence that cores from cold glaciers are less subjected to these processes. However, all cores from sub-polar glaciers experience large post-depositional processes, as described by Reference KoernerKoerner (1997) and Tarussov (1992).

The best estimates for ¹³⁷Cs and ³H fall-out are determined from the 1963 decay-corrected values, as the decay correction of fall-out, applied for each snow-layer date, leads to overestimated values. As already specified, MANMB at each ice-core location is almost constant for periods extending from 1963–86 and from 1986 to the drilling dates (Reference PinglotPinglot and others, 1999). From this important feature for ice cores from Svalbard, ¹³⁷Cs and ³H fall-out is studied in relation to the MANMB for each core location (Fig. 7a and b), even if the all periods are not the same.

¹³⁷Cs fall-out from the nuclear tests for all studied cores spans from 224 (Ves81) to 549 Bq m⁻² (Sev99) (Fig. 8a). Altogether, the mean ¹³⁷Cs fall-out values for the 13 Arctic and 11 Svalbard ice cores are 355 Bq m⁻² (±95 Bq m⁻²) and 341Bq m⁻² (±81 Bq m⁻²), respectively. For four ice cores near the summit of Austfonna, the fall-out values extend from 269–434 Bq m⁻². Fall-out values are 310 and 549 Bq m⁻², respectively, for the Devon Ice Cap and Akademii Nauk ice cap cores (Table 3). The fall-out apparently tends to decrease from eastern to more western Arctic locations. This agrees with the ¹³⁷Cs fall-out in the Arctic predicted by Reference Wright, Howard, Strand, Nylén and SickelWright and others (1999) using a geographical information system to combine ¹³⁷Cs fall-out and precipitation data.

Fig. 8. ¹³⁷Cs (a) and ³H(b) fall-out from atmospheric nuclear tests vs MANMB for 13 high-Arctic ice cores.

The mean ³H fall-out for the same 1954–74 period is 4762 ± 1126 TU m for eight ice cores and spans from 2979–6505 TU m (Fig. 8b). For comparison, the ³H fall-out measured for the Isfjord Radio coastal station in Svalbard (IAEA, 1984) for the period 1961–76 is 2793 TU m. This is in accordance with the ³H data from ice cores, as the ³H fall-out at Isfjord Radio (mean annual precipitation is 0.444 m) was not recorded before 1961. The ³H fall-out increases with MANMB (r = 0.56), from a previous study in Greenland (Reference Merlivat, Ravoire, Vergnaud and LoriusMerlivat and others, 1973). A null ¹³⁷Cs fall-out corresponds to about a 1000 TU m ³H fall-out (Fig. 8b). This corresponds to the natural ³H fall-out (∼50 TU m a⁻¹) over 20 years, the period when artificial ³H was also deposited (1954–74).

For all ice cores, there should be proportionality between ¹³⁷Cs and ³H fall-out from atmospheric nuclear tests. This relationship prevails and exhibits a 0.69 correlation coefficient (Fig. 9). The ³H fall-out at Lomonosovfonna (Lom97 and Lom00), 3709 (estimated value, as some samples were not analyzed) and 2979 TU m, respectively, is low compared to the other ice cores (mean of 4913 TUm) (Fig. 8b).

Fig .9 ³H vs ³⁷¹Cs fall-out from atmospheric nuclear tests at eight Arctic ice-core locations.

¹³⁷Cs fall-out in the Lom97 and Lom00 ice cores, 343 and 279 Bq m⁻², respectively, is close to the mean value of Svalbard ice cores (341 Bq m⁻²), indicating that wind erosion may be low, but cannot explain the surprisingly low MANMB at the Lomonosovfonna sites at the summit of this ice cap. Analysis of three ice cores from the Lomonosovfonna dome did not show the presence of the Chernobyl signal, indicating that some wind scouring occurs at this location. However, the Chernobyl event was present in ice cores located at lower altitudes (1044 and 1173 m a.s.l.). A close examination at the relationship between ¹³⁷Cs and ³H fallout (Fig. 9) reveals that both the Lom97 and Lom00 ice cores experience about 25% lower ³H fall-out than the mean values. This may be due to snow and ³H sublimation, as described byReference Winther, Bruland, Sand, Killingtveit and MarechalWinther and others (1998).

The 1986 Chernobyl layer

¹³⁷Cs fall-out from the Chernobyl accident has been successfully detected in about 50 Svalbard ice cores (Reference Lefauconnier, Hagen, Pinglot and PourchetLefauconnier and others, 1994; Reference Pourchet, Lefauconnier, Pinglot and HagenPourchet and others, 1996; Reference PinglotPinglot and others, 1994, Reference Pinglot, Pourchet, Lefauconnier and Creseveur1997, Reference Pinglot1999, Reference Pinglot, Hagen, Melvold, Eiken and Vincent2001). More recently, we determined the Chernobyl fall-out for the Devon and Akademii Nauk ice caps. The ¹³⁷Cs fall-out varies from 0–47 Bq m⁻² for the Svalbard ice cores. Wind scouring has been determined to be the main factor explaining this variability, as the fall-out from Chernobyl only occurred during a few days (Reference Pinglot, Hagen, Melvold, Eiken and VincentPinglot and others, 2001). The fall-out on Devon Island is 5 Bq m⁻² for two ice cores. This value is of the same order as the Svalbard and Greenland values (Reference DibbDibb, 1989). On Severnaya Zemlya, the ¹³⁷Cs fall-out is about 1 Bq m⁻², as determined from a shallow ice core retrieved in 2000. This low value is close to the detection limit of our low-level gamma spectrometer. It would appear however to be valid, given that the corresponding mass balance for 1986–2000 is 0.52 m w.e. compared to 0.45 m w.e. for 1963–2000(Table 3).

Dating from ²¹⁰Pb profiles

Both ¹³⁷Cs and ²¹⁰Pb were simultaneously analyzed by gamma spectrometry. The ²¹⁰Pb profiles for Aus85, Aus87 and Ves81 ice cores (Figs 3d, 4d and 5c) indicate a strong ²¹⁰Pb increase between 18 and 24 m w.e. Samples from other ice cores from Austfonna,Vestfonna and Spitsbergen were also analyzed for ²¹⁰Pb. The corresponding ²¹⁰Pb profiles are included in Reference PinglotPinglot and others (1994). Other ice cores from Svalbard (Asg93 and Sno92) have also been analyzed for ²¹⁰Pb (Reference Suzuki and FujiiSuzuki and Fujii, 1992; Reference Suzuki, Ohta and WatanabeSuzuki and others, 1995). There is no exponential decay starting from the surface down, for any profile. The ²¹⁰Pb profile (Fig. 6b) for Lom97 indicates a general decrease of activity with depth. However, the dating of this ice core, based on ²¹⁰Pb decay, does not correctly position the 1963 radioactive layer from atmospheric nuclear tests.

For Akademii Nauk ice cap, a formerly determined ²¹⁰Pb profile (Reference Zagorodnov, Klementyev, Nikiforov, Nikolaëv, Savatyugin and SasunkevichZagorodnov and others, 1990) can be closely compared to the new profile (Fig. 6c). Instead of a regular decrease of the ²¹⁰Pb activity with depth, the activity is almost stable (and even increases with depth in the first 10 m) from the surface down to about 25 m w.e. Then the activity regularly decreases, as expected. This unusual feature was already reported by Reference Zagorodnov, Klementyev, Nikiforov, Nikolaëv, Savatyugin and SasunkevichZagorodnov and others (1990). At a depth of about 20 m w.e., there is even a high level of ²¹⁰Pb activity. The interpretation may be that ²¹⁰Pb particulates travel downward into the deeper firn layers due to summer melting and percolation. Once the firn layers reach a density close to 0.9, the migration of ²¹⁰Pb stops and its activity begins to decrease, as seen below 30 m depth. This indicates that dating based on the ²¹⁰Pb half-life (22 years) may be impossible for most ice cores from sub-polar glaciers.