Radioecological and geochemical peculiarities of cryoconite on Novaya Zemlya glaciers

Data for all analysed radionuclides are presented in the “Supplementary Material”. Cryoconite samples were collected on Nalli Glacier (Supplementary Fig. S1) on Sept. 25, 2017 (samples 1701–1714) and on Sept. 10, 2018 (samples 1801–1814) at 28 spots (Fig. 2, Supplementary Table S1). Gamma spectrometric analysis of samples showed the presence of anthropogenic radionuclides ¹³⁷Cs, ²⁴¹Am, and ²⁰⁷Bi. All quoted radioactivity values were recalculated for the sampling date, except those for ²⁴¹Am since the concentration of the parent ²⁴¹Pu isotope is unknown. However, for this isotope, the correction for decay is negligible. The activity of ¹³⁷Cs reached 8093 (± 69) Bq kg⁻¹ of dry weight, that of ²⁴¹Am reached 58.3 (± 2.3) Bq kg⁻¹ and that of ²⁰⁷Bi reached 6.3 (± 0.6) Bq kg⁻¹. The natural radionuclides ²¹⁰Pb and ⁷Be were also present in all samples. The activity of ²¹⁰Pb varied in the range of 1280–9750 Bq kg⁻¹. In addition, in the investigated samples, a significant amount of short-lived cosmogenic radionuclide ⁷Be was found, whose specific activity reached 2418 (± 76) Bq kg⁻¹ (Fig. 3, Supplementary Table S2). To evaluate the contribution of atmospheric components to the total ²¹⁰Pb activity, ²²⁶Ra activity was determined and found to be 17–27 Bq kg⁻¹ (Supplementary Table S2). Based on the ²¹⁰Pb/²²⁶Ra ratio, we conclude that more than 98% of ²¹⁰Pb was of atmospheric provenance.

Anthropogenic radionuclides

A specific radioactivity of ¹³⁷Cs exceeding 8000 Bq kg⁻¹ is the record for Arctic cryoconite. It is known that levels of ¹³⁷Cs activity in cryoconite can reach even more extreme values of 140 kBq⁶ and even 223 kBq¹³. However, the data in Wilflinger et al.¹³ were recalculated for 01.05.1986, and those in Tieber et al.⁶ were recalculated for 03.10.2006. At present, both values would be approximately 100 kBq each. In addition, it should be noted that these data were obtained from Schladming Glacier and Stubacher Sonnblick Glacier, which are located in a relatively restricted area in the European Alps that had the highest density (reaching 1480 kBq m⁻²) of radioactive fallout from the Chernobyl accident in 1986²⁷. Both glaciers (47° 28′ N and 47° 07′ N) are located in the latitudinal belt with the highest density of global fallout in the Northern Hemisphere, exceeding this parameter by ~ 13 times compared to Arctic latitudes 70–80° N at which the Nalli Glacier is situated²⁸. Accordingly, based on generally accepted latitudinal zonation, we can assume that with the exception of spots directly affected by radiation accidents and/or nuclear tests²⁰, the ¹³⁷Cs activity levels in cryoconite from Arctic glaciers should generally reflect the global fallout level. This hypothesis is supported by literature data and eleven additional samples collected by us on Svalbard glaciers Jotufonna, Vestre Grønfjordbreen and Fridtjovbreen in the framework of another project (Table 1). Cryoconite samples from Svalbard glaciers were added to this work to better interpret our results (Supplementary Table S3).

Three other studies of ¹³⁷Cs in Northern Hemisphere cryoconite may be mentioned. On Russell Glacier in Greenland (67° 09′ N) the maximum ¹³⁷Cs activity was 123 (± 93) Bq kg⁻¹³⁰, on Castle Creek Glacier located 2670 km south of ‘our’ latitude belt (53° 02′ N) in Canada—3969 (± 149) Bq kg⁻¹³¹, and on Isfallsglaciären (67°, 54′ N) in Arctic Sweden—4533 (± 149) Bq kg⁻¹¹⁴. A comparison of our results with radiocesium activity data given in Table 1 shows that the determined maximum activity level of ¹³⁷Cs of 8100 Bq kg⁻¹ in sample 1813, as well as at three nearby locations from 5665 to 7125 Bq kg⁻¹, is currently the highest recorded for Arctic glaciers. The highest activity detected in one sample for Svalbard cryoconite on Werenskiold Glacier (no exact coordinates are given) is 4500 Bq kg⁻¹¹¹. Taking this value as the maximum for the global deposition level, and considering the difference in sampling time, it is two times lower than that at Nalli Glacier. Accordingly, at least four samples (1811, 1812, 1813, and 1814) of Nalli Glacier cryoconite (Supplementary Table S2) contain not only a global fallout contribution but also an additional contribution, which, according to our assumptions, is now being released from the contaminated layer that formed due to local fallout in the accumulation zone from nuclear tests.

Analysis of the distribution of ¹³⁷Cs, ²⁴¹Am and ²⁰⁷Bi activities in cryoconite relative to the hypsometric levels of point locations on the glacier surface leads to their division into three groups, termed zones A, B₁, and B₂ (Fig. 4). The first group of cryoconite holes consists of densely clustered points up to 210 m a.s.l., where ¹³⁷Cs activity varies from 58 to 436 Bq kg⁻¹ and ²⁴¹Am from 1.3 to 9.1 Bq kg⁻¹ (A in Fig. 4a/b). No ²⁰⁷Bi is detected. Cryoconite holes located between 240 and 345 m a.s.l. (B₁ in Fig. 4) compose the second group. In this group, the radioactivity of ¹³⁷Cs ranges from 2667 to 4659 Bq kg⁻¹, and ²⁴¹Am is in the range of 12.5–28.1 Bq kg⁻¹ (B₁ in Fig. 4a/b); these values are quite consistent with data from the Waldemar Glacier². In this group, ²⁰⁷Bi appears with specific activities ranging from 2.2 to 4.4 Bq kg⁻¹ (B₁ in Fig. 4c). It should be noted that only a few works address ²⁰⁷Bi in the natural environment, and this study is the first to show its presence in Arctic cryoconite. The origin of this radionuclide is still the subject of debate. Noshkin et al.³² showed the absence of a direct link between the yield of an explosive device and the amount of ²⁰⁷Bi released. Based on soil data in the Faroe Islands Aarkrog et al. suggested that a high-power thermonuclear bomb at the Novaya Zemlya test site on October 30, 1961, and other so-called “clean” thermonuclear weapons are the most likely sources of ²⁰⁷Bi³³.

Two samples 1710 and 1802 located at altitudes of 210–220 m a.s.l., form an intermediate zone between the groups mentioned above. On the glacier itself, this area is characterized by pronounced ridges and grooves with directions normal to the glacier flow direction; additionally the surface slope changes: in zone A the surface is 3°–3.5° steeper. The environment containing the cryoconite holes in zone A is more dynamic with larger amounts of meltwater and more intense flows. In the lower part of the glacier, the holes are less stable and contain smaller amounts of organics. Consequently, zone A is less suitable for the accumulation of radionuclides. A third group of cryoconite holes is located above 350 m a.s.l. (B₂ in Fig. 4). Here, ¹³⁷Cs activity reaches 8100 Bq kg⁻¹ (B₂ in Fig. 4a), ²⁴¹Am—58 Bq kg⁻¹ (B₂ in Fig. 4b), and ²⁰⁷Bi—6.3 Bq kg⁻¹ (B₂ in Fig. 4c). Apparently, the cryoconite samples in this group are enriched with “additional” radionuclides ¹³⁷Cs, ²⁴¹Am and ²⁰⁷Bi relative to the B₁ zone and are the leading frontal part of the radioactive material released from the radiation-contaminated layer.

The zonal distribution of cryoconite into three groups becomes more prominent when considering the ²⁴¹Am/¹³⁷Cs (Fig. 5a), ²⁰⁷Bi/¹³⁷Cs (Fig. 5b) and, especially, ²⁰⁷Bi/²⁴¹Am (Fig. 5c) ratios. The ²⁴¹Am/¹³⁷Cs ratio (Fig. 5a) is used to distinguish Chernobyl fallout from global levels³⁴. In the Chernobyl accident relatively small amounts of transuranic elements were released into the atmosphere, mostly in the form of relatively large ‘hot’ particles that fell in a relatively small area around the reactor³⁵. A significant fraction of volatile radionuclides, including ¹³⁷Cs, were released in aerosol form and distributed worldwide^36,37. ²⁴¹Am is not produced directly in nuclear explosions; instead, it results from ²⁴¹Pu decay. There are few data on the presence of the latter in natural objects because its measurement is complicated. The activity of ²⁴¹Pu delivered to the environment is more difficult to estimate than that of fission fragments since its production in any given explosion depends both on the explosion yield and on the initial isotopic composition of the charge. It has been estimated³⁴ that atmospheric nuclear explosions have released 142 PBq of ²⁴¹Pu into the environment. Most of it fell out between 1961 and 1964, and by now, it has almost all decayed. The decay of ²⁴¹Pu leads to an increase in ²⁴¹Am activity, which will pass through a maximum in 2036³⁴. Both ²⁴¹Am and ²⁴¹Pu are firmly bound to suspended matter, so their fractionation after fallout is unlikely.

A recent paper³⁸ gives a global fallout ratio of ²⁴¹Am/^239+240Pu = 0.37, as recalculated for 2018. The ¹³⁷Cs/^239+240Pu ratio was determined in several papers^39,40. Recalculated for 2018, it is ≈ 24. Based on these sources, we estimate that the global activity ratio ¹³⁷Cs/²⁴¹Am ≈ 0.015. Analysis of 26 samples of cryoconite and alpine soils gave the following ratio ²⁰⁷Bi/¹³⁷Cs_(global) = (1.70 ± 0.12) × 10^–341. In the five samples from Svalbard (our unpublished data), the average ²⁰⁷Bi/¹³⁷Cs ratio was (1.7 ± 0.4) × 10^–3, which correlates well with⁴¹ (Fig. 5b). In the cryoconite of Novaya Zemlya, however, this ratio is noticeably lower, (9.2 ± 1.6) × 10^–4 (n = 13), which is closer to the ratio given in Ref.³³. For the Stubacher and Hallstatter alpine glaciers, the values of ²⁰⁷Bi/¹³⁷Cs_(global) are (1.00 ± 0.37) × 10^–3 and (2.75 ± 1.24) × 10^–3, respectively¹³. Note that the ratio of ²⁰⁷Bi/¹³⁷Cs in natural objects may not correspond to the original composition of the fallout but may result from secondary fractionation processes⁴¹. This may explain some of the variation in the values obtained in different regions. It is natural to assume that fractionation should increase the ²⁰⁷Bi/¹³⁷Cs ratio since ¹³⁷Cs is a more mobile radionuclide than ²⁰⁷Bi. According to Aarkrog³³, the super high-power explosion on Novaya Zemlya was the main global source of ²⁰⁷Bi. This assumption is indirectly supported by the higher activity of ²⁰⁷Bi in Novaya Zemlya than in Svalbard. However, our results do not permit us to make firm conclusions about the origin of this nuclide. Variations in the ²⁰⁷Bi/¹³⁷Cs ratio in different regions may result both from the existence of several sources (e.g., high-yield thermonuclear explosions) and from the markedly different geochemical behaviour of these elements. More detailed studies are required to unambiguously identify the origin of artificial radionuclides on Novaya Zemlya involving other matrices such as peat, soils, lichens, and mosses.

Natural radionuclides

Cryoconite holes on a glacier surface may survive for years⁴², accumulating precipitated matter, radionuclides of various origins in particular, which are transported into cryoconite via meltwater. For example, high concentrations of ²¹⁰Pb (T_1/2 = 22.2 years) and ⁷Be (T_1/2 = 53.2 days) have been reported in cryoconite^{1,2,8,9,10,11,12,14}. ²¹⁰Pb results from Rn decay, which, in turn, is produced from Ra present in rocks and soils in trace amounts. Cosmogenic ⁷Be is produced via the interaction of cosmic rays with the upper troposphere and stratosphere. Both nuclides tend to fix in particulate matter, but markedly different half-lives permit the use of them to estimate the residence time of cryoconite holes¹³. Almost all our samples contain both isotopes. In contrast to other radionuclides, the spatial distribution of ⁷Be does not show marked peculiarities or zoning (Fig. 6b). Its activity is in the range of 38 (± 10)–1264 (± 45) Bq kg⁻¹, with the exception of sample 1802, which contains 2418 (± 76) Bq kg⁻¹. The old-inactive cryoconite, having no contact with meltwater, is devoid of ⁷Be⁹. The presence of ⁷Be implies that on the date of sampling the cryoconite was “active”, i.e., was accumulating radionuclides. Consequently, cryoconite holes in Nalli Glacier accumulate radionuclides independently of their location on the glacier. This highlights that the supra-glacial environment of Nalli Glacier is dynamic from a hydrologic point of view, with a network of supraglacial channels capable of spreading radionuclides in cryoconite across the entire glacier.

In contrast, ²¹⁰Pb behaves differently. The specific activity of this nuclide varies in the range of 1283 (± 31)–9748 (± 274) Bq kg⁻¹ (Supplementary Table S2). Similar to anthropogenic radionuclides, it clearly correlates with hypsometric position: spots with lower activity (1200–3000 Bq kg⁻¹) mostly lie below 210 m a.s.l. (Fig. 6a). This behaviour likely indicates that cryoconite in the upper part of the glacier is older than in the lower part, possibly due to removal of the latter with meltwater and rain. The lower part of the glacier is, in general, more dynamic than the upper part. Cryoconite holes above 220 m a.s.l. appear to be more stable and survive several seasons, accumulating significant amounts of ²¹⁰Pb.

Phase composition and selected properties of cryoconite samples

The X-ray diffraction patterns of the three studied samples are virtually identical (Fig. 7a). Semiquantitative phase analysis gives 40% quartz and 17% albite, and the rest is represented by micas (chlorite, muscovite, biotite, and phlogopite). These results are consistent with data on cryoconites from other polar regions^43,44. Note that the sample preparation complicates the quantitative analysis of micas due to possible texturing.

Infrared spectra (Fig. 7b) are dominated by micas (e.g., phlogopite): bands are due to Si–O and Si–O–Si (Al, Mg) between 690 and 1200 cm⁻¹⁴⁵, Si–O stretch overtones are within 1650–1870 cm⁻¹, and aliphatic C–H bonds (2850 and 2920 cm⁻¹) and OH groups (2900–3650 cm⁻¹) are present. In contrast to the XRD results, the samples demonstrate considerable scatter in the abundance of various species. This clearly results from a much smaller beam size (here—100 µm) than in the XRD case (approx. 1 cm²). Since cryoconites consist of a mixture of minerals and biological films, the results of micro-IR spectroscopy are very sensitive to the selection of the studied grains. To obtain sample-averaged information, preparation of KBr-based pellets is required. However, in our view, microspectroscopy is an extremely valuable tool for cryoconite research, since after proper sample preparation, it allows examination of different types of granules (e.g., those suggested by Takeuchi et al.⁴⁶) and interaction of organic matter with individual minerals.

An important supplement to the precise study of radionuclides was the investigation of spatial distribution of radioactivity in cryoconite samples using digital autoradiography. Autoradiography of a large set of samples from various locations (Novaya Zemlya and Svalbard) gave similar results—a rather uniform distribution of weak levels of activity. However, in some samples, “hot spots” with activity approximately two times higher than that of the rest of the sample were observed (Fig. 7c). The splitting method was applied to extract some “hot particles” for subsequent gamma spectroscopy, but the activity of separated mineral grains was extremely low, ruling out the presence of radioactive particles from historic nuclear explosions. Some information about the origin of these “hot spots” was provided by comparison of autoradiographs with µ-XRF mapping. In the studied cryoconite samples, the “hot spots” are associated either with small (10–20 µm) Zr-rich grains (likely zircon) or with Fe-containing particles, most likely Fe oxides. Zircon often contains an admixture of uranium and/or thorium, and thus observation using autoradiography is expected. Explanation of the radioactivity of Fe-containing grains is less certain, but sorption of Pu and some other radionuclides on Fe oxides may be important in a wide range of environmental conditions⁴⁷.

Geochemical characteristics of cryoconite

A summary table of the results of all analytical measurements is attached in Supplementary Table S4. The distribution of concentration variability of a large number of trace elements, as well as several major (macro-) elements, showed clear patterns. The main reference for the normalization of most trace elements was chosen as their generalized average concentrations (Clarke value) in the upper continental crust (UCC)⁴⁸. Macroelements and their ratios are usually reported in weight percentages without normalization. By plotting the distribution of the values obtained for the elements with contrasting behaviour as a function of the altitude of the samples, it is possible to confidently distinguish the two groups. The first group consists of lithophile elements with higher contents in the lower part of the glacier (Fig. 8), which corresponds to the zone A in Figs. 2 and 11. The second group comprises chalcophile elements and W concentrated in the upper part of the sampled area (Fig. 9), i.e., in zone B in Figs. 2 and 11. Similar to radionuclides, we observe a clear distinction between the two altitudinal zones of the glacier at altitudes of 200–220 m a.s.l. (termed an “inversion band”). The established altitude-dependent distribution of concentrations of elements in cryoconite follows general geochemical patterns⁴⁹. The rare earth elements (REEs), Li, Rb, Be, Sc, U, and Th (Fig. 8) are geochemically inert lithophiles. The chalcophiles Bi, Ag, Sn, Sb, Pb, Cd, and Cu (Fig. 9), in turn, are chemically active and mobile under certain geological and geochemical conditions. Accordingly, they are much more affected by human perturbations than lithophiles. Tungsten, which possesses mixed siderophile and lithophile properties, shows a clear relationship with Cu (Fig. 9).

Concentrations of chalcophiles and W in the lower part of the glacier (below 210–220 m a.s.l.) have normalized values close to unity, i.e., are at the level of UCC clarkes. Above the “inversion band”, their concentrations gradually increase, and above approximately 350 m a.s.l., another step is visible (Fig. 9). The largest effect is observed for Bi and Ag, showing a 30-fold enrichment with respect to UCC; Sn and Sb show a 15-fold increase. This anomalous jump is observed within zone B above the boundary separating zones B₁ and B₂ (Figs. 2 and 11) at the same sampling points where maximum activities of anthropogenic radionuclides are established. The concentrations of some macroelements (C, S, P, Al, and Si) and trace elements are distinctly different in the lower and upper parts of the studied section of the glacier (Fig. 10).

Geochemical analysis reveals the following trends. In the lower part of the glacier (zone A), the contents of U, Th, Be, Sc, Bi, Ag, Pb, Cd, W, Cu, and every individual REE (in Fig. 8, they are grouped into light and heavy REEs: LREEs and HREEs) are close to the UCC Clarke values (see Supplementary Table S4). Only Li, Rb and Sc are slightly enriched. In the upper zone of the glacier (zone B), the contents of REEs, Be, U, and Th are somewhat lower than their clarkes; Li, Rb and Sb are slightly below the reference. However, in the zone B, the concentrations of Pb, Cd, W, Cu, and, in particular, of Bi, Ag, Sn, Sb, are markedly higher than the reference. Concentration of the mentioned elements reached maximum values in the topmost analysed glacier zone (B₂). These elements often form polymetallic ores, but the nearest large-scale mining and processing plants are 1100 and 1300 km away (Norilsk and Pechenga, respectively). More likely, polymetallic ores of Novaya Zemlya⁵¹ are the source of these elements. In particular, in 2001 one of the largest Zn–Pb–Ag-containing deposits in Russia was discovered just ~ 80 km south of the Sukhoy Nos test ground (A in Fig. 1)⁵². High exposure (> 95%) of corresponding rocks and numerous outcrops likely promoted entrapment of these elements into explosion clouds, and their subsequent precipitation with radionuclides. This feature of the geological structure of the area explains the extremely high enrichment of surface waters in elements such as Zn, Pb, Sr, Ni, As, Cr, Co, Se, Te, Cd, W, Cu, Sb, and Sn; for many of them, the excess reaches 10-fold with respect to the Clrake values⁵¹. This hypothesis is supported by obvious correlations between the concentrations of Bi, Ag, Sn, Sb, Pb, Cd, W, and Cu and the activity of anthropogenic radionuclides ¹³⁷Cs, ²⁴¹Am and ²⁰⁷Bi. This relationship is obviously related to the simultaneous release of elements and radionuclides from the contaminated ice layer and their entrapment in cryoconite holes.

Source: Ecology - nature.com