Introduction

The radioactive fission product 90Sr has a sufficient half-life (T1/2 = 28.8 y, Emax = 546 keV) to be detected for a long time period after it has appeared in the environment. Due to chemical similarity to calcium, 90Sr is uptaken and easily accumulated by various objects of animate and inanimate nature. The radiotoxic effect of 90Sr is additionally increased by progeny—90Y (T1/2 = 64 h, Emax=2285 keV)—, which exists in equilibrium with 90Sr for most environmental samples. Therefore radiostrontium may be considered as a highly hazardous anthropogenic radionuclide.

The worldwide presence of 90Sr, to at least measurable extent, has been caused intentionally or accidentally since the 1940s, as a consequence of various human nuclear activities. For the High Arctic terrestrial environment, the most significant release is associated with the stratospheric global fallout (GF) [1, 2, 48]; i.e., far-reaching radioactive contamination injected into the atmosphere during nuclear weapons testing (1945–1980). The AMAP group estimated that the total inventory of 90Sr from the global fallout was around 21.9 PBq (decay corrected to 1995) over the land north of 60° N [1]. One of the largest yield test sites was located in the Arctic on the island of Novaya Zemlya. From 1955 to 1990 this territory was used by the Soviet Union for 88 atmospheric, 39 underground and 3 underwater nuclear explosions [23]. Despite an extensive nuclear programme, major tests were carried out at high altitudes, minimizing local and enhancing global fallout [48].

Chernobyl NPP’s accident has become an important source of 90Sr after 1986. Due to the different mode of the explosion, than during weapons testing, radiostrontium dispersion occurred in a relatively small range reaching mainly immediate vicinity of the Chernobyl. More specifically, the bomb derived deposition of 90Sr was lower (about 30‒35%) than 137Cs, whereas in the Chernobyl fallout the fraction of 90Sr was much lower, approximately in the range of 1–10% of 137Cs, depending on the distance from accident zone [26]. In total, the released activity of 90Sr from the Chernobyl NPP’s failure exceeds slightly 1% of the weapons test fallout [26]. The Arctic archipelago of Svalbard is generally believed to be relatively unaffected by the post-Chernobyl radiostrontium [38].

Of lesser concern, but still worthy of consideration nonetheless, is the possible contribution of the sea-to-land transfer of marine contaminants. The main radioactive discharges into the Arctic Ocean derived from the nuclear fuel reprocessing plants located in Western Europe (Cap de La Hague, France, and Sellafield, UK) and from nuclear facilities in Russia (Mayak PA; SCC, Tomsk; MCIC, Zheleznogorsk) placed within catchments of the Ob and the Yenisey rivers—major freshwater inflows of the Arctic seas [22, 25]. Furthermore, dumped nuclear waste, sunken nuclear submarines in the Arctic Ocean comprise other potential sources of 90Sr, once any leakage occurs [18, 40, 45].

The original distribution of radionuclides stored in the ecosystem changes over time. For the Arctic region, these changes may be, particularly, related to soil erosion, snow or marine aerosols, sea currents, an ice pack or iceberg drift and, finally, deposition of contaminated faeces (guano) by animals (predominantly birds) [11]. Different physicochemical forms of radionuclides (i.e., radionuclide species) affect their transport/mobility and bioavailability [41]. The modern issue is that the Arctic’s climate and weather patterns are changing most rapidly [4]. Rising air, surface and ocean temperatures are accelerating the decline of ice (including glaciers) and snow cover, as well as inducing direct and indirect effects on the interconnected physical, chemical and biological systems in the Arctic. Thus, global warming may be a significant factor triggering or enhancing both spatial and temporal variations in the distribution of contaminants trapped in the polar region [2, 3].

To identify and verify the levels, trends as well as dispersion and migration pathways of the contaminants accumulated or released by various elements of ecosystems, only selected environmental media are being commonly used. In recent years cryoconite holes have begun drawing attention among matrices known to be efficient in accumulating of radioisotopes and typically adopted for radioecological monitoring, such as lichens, mosses, peats, soils, sea and lake sediments. Cryoconite is a small size repository of a high level of pollutants on the glacier surfaces [9, 46]. These aggregates of mineral and organic components are associated with biological consortia. They decrease an ice-albedo and are responsible for the formation of water-filled holes. Research on heavy metals and radionuclides in different environmental media showed that cryoconite absorbs most of the analysed elements in high concentration [9, 42]. The hypothesis by Łokas et al. [30, 32] assumes that such material might be transferred into the proglacial zones while glaciers retreating, causing secondary contamination of the initial soils.

As established so far, the level of contamination in Spitsbergen appears to be lower than in temperate zones [10, 1315, 44]. This is a simple consequence of the fact, that major human activities (e.g., production and release of radionuclides) took place outside of the polar region. Although, Svalbard archipelago, as the Arctic region in general, remains particularly vulnerable to radioactive contamination. Due to large organic matrix content, the High Arctic environment may be considered as a ’sponge‘ for captured particles with significant retention and accumulation [12].

This study was dedicated to determining the 90Sr content as well as tracking its behaviour in cryoconite, proglacial soils and peat profiles from Spitsbergen, Svalbard archipelago. The specific goals were to verify the 90Sr mobility relative to other radionuclides (e.g., 137Cs, Pu isotopes) and the possible transfer of pollutants from melting glacier to proglacial zones. Because of the limited knowledge in this area, the article provides scarce data on radiostrontium in various terrestrial ecosystems of Spitsbergen.

Material and methods

Study area and sampling

Svalbard (70° N, 20° E) is the Arctic archipelago placed north of the Arctic Circle between Greenland and Franz Josef Land. The meteorological conditions are temporally and spatially diverse due to the location at the confluence zone between cold and dry polar air masses from the north and warm and humid air masses from the Atlantic currents to the southwest [34]. Therefore, Svalbard has an extremely sensitive climate among other regions of the world [39]. The total glaciated area on Svalbard is 34,560 km2 (57% of the entire territory). There are six major islands, of which Spitsbergen is the largest and the most alpine one, having small cirque glaciers, extensive ice fields and valley glaciers.

Sampling was done during seven field campaigns into the different regions of the Spitsbergen within 10 years (from 2005 to 2015). The research material comprised 49 cryoconite samples, 3 proglacial soil profiles, 1 tundra soil profile and 3 peat profiles. The cryoconite was collected from each sampling site, whereas the soil and peat profiles were derived only from the Werenskiold area. The details of scientific expeditions into the Werenskiold, Kafføyra and Kongsfjord areas and the description of the fieldworks were presented elsewhere [28, 29, 32, 33, 43]. Samples taken from the other study areas (i.e., Calypsobyen and Sørkapp Land), as a part of a wider scientific project, will be characterised in a future paper. Herein we consider only selected aspects, that are common for all mentioned above Spitsbergen regions and research material. The locations of study sites are shown in Fig. 1 and the short sampling description is provided in Table 1.

Fig. 1
figure 1

The map of the Arctic archipelago of Svalbard with tagged study sites (black squares) and satellite pictures of the explored glaciers

Table 1 The details of the sampling positions

Experimental

Strontium fraction was extracted from the sample matrix after gamma radiation measurement (137Cs) by applying the sequential radiochemical treatment [35, 49]. As a result, the Pu, Th, Sr, Am and U fractions were obtained. The detailed information regarding the laboratory works was described by Łokas et al. [28, 31].

Detection of beta particles emitted by 90Sr and 90Y was based on the electron-scintillation conversion method [36]. The maximum beta energies of both nuclides are as high as follows: 0.54 MeV and 2.28 MeV, respectively. At least 2 weeks after separation, when the radioactive equilibrium between 90Sr and 90Y was reached, Sr fraction was mixed with the liquid organic scintillator followed by measurements on Wallac 1414-003 spectrometer. With the typical counting time of 30,000–45,000 s, the detection limit averaged at 5 mBq per sample.

Calculation

We computed activity concentrations C (Bq/kg) and total inventories I (Bq/m2) for 90Sr as well as isotopic ratios for selected pairs of radionuclides. These parameters were estimated with the aid of commonly known, simple equations [27]. Results were presented for the dry weight. The decay correction to the reference date 2020-01-01 was applied.

To infer the contamination origin, obtained isotopic ratios were compared to the signature typical for the global fallout (Table 2).

Table 2 Reference isotopic ratios of activity concentrations for the global fallout (decay corrected to 2020-01-01)

The law of propagation of uncertainty and the standard deviation of the mean [47] were used to estimate the errors associated with calculated parameters (only 1σ was taken).

Evaluation of the analytical quality was performed by the analysis of IAEA reference materials (soil 375, moss-soil 447) using the same procedure as for the sample set. These results were presented by Łokas et al. [28, 31].

Results and discussion

Activity concentration of 90Sr in cryoconite

The highest individual and mean activity concentrations were observed for cryoconite from Werenskiold (Table 3), that is the most southern study site along the west shoreline of Spitsbergen. Moreover, the widest range of results was also noted in Werenskiold (Fig. 2a). In contrast, the most southern Sørkapp Land, placed on the east coastline of Spitsbergen, together with Calypsobyen and Kongsfjord, characterised the lowest mean activity of 90Sr for cryoconite (Table 3). For the latter regions rather high consistency of result ranges was observed (Fig. 2a).

Fig. 2
figure 2

The box-whisker plots illustrating the ranges of 90Sr activity concentrations (a) and dispersions of the isotopic ratios of 137Cs/90Sr (b), and 239+240Pu/90Sr (c) for the investigated Spitsbergen sites

Table 3 The mean, min and max values of 90Sr activity concentration for examined Spitsbergen sites

The data by Łokas et al. [31] revealed the average activity concentration of 90Sr at 57 ± 8 Bq/kg for cryoconite collected from Hans glacier, Hornsund fjord (SW Spitsbergen). Our results were slightly lower, although represented the same order of magnitude for cryoconite from Kongsfjord and other studied regions, except Werenskiold. Therefore, one might claim, that the level of radiostrontium was relatively similar for both projects.

However, analysing Table 3; Fig. 2a, the considerable difference between results for cryoconite as well as soil and peat profiles derived from Werenskiold could be noted. Activity concentrations for the latter were even two orders of magnitude lower than in cryoconite samples. This observation corresponds to the previous findings. Namely, cryoconite granules are considered as highly rich in natural (unsupported 210Pb) and artificial radioisotopes, significantly more radioactive than other environmental matrices routinely used in radioecological monitoring [9].

Activity concentration and inventory of 90Sr for soil and peat

Peat profiles, collected from the strandflat adjacent to the proglacial zone of Werenskiold, were named as P3, P5 and P6 [29]. Proglacial soil profiles were coded as GA, GB, GC—according to decreasing distance from the Werenskiold glacier terminus—, whereas H1 represented the only one tundra soil core [32]. To better interpret 90Sr results, we included also information on 137Cs and 239+240Pu activity concentrations for each profile, already analysed and published elsewhere [29, 32].

Table 3 provides the means and ranges of 90Sr activity concentration and Table 4 presents 90Sr inventory for each core. Generally, the inventories varied between 128.1 ± 4.9 Bq/m2 (P3) and 1082 ± 104 Bq/m2 (GC). The UNSCEAR report [48] estimated the deposition of 137Cs over the globe due to weapons testing, distinguishing different Earth’s regions. Assuming 1.5 is the isotopic ratio of 137Cs/90Sr, deposition of radiostrontium derived from the global fallout, should have reached approximately 0.63 kBq/m2 for 70‒80° N latitude belt (decay corrected to the year of 2020). The above assessment provided a reference point, which was relatively compatible with inventories for P5, H1, GA and GB. This might suggest the global fallout as prevailing source of radiostrontium for the Werenskiold area, but the more profound analysis is necessary to infer the origin.

Table 4 Inventory of 90Sr for peat and soil profiles collected from Werenskiold, Spitsbergen

90Sr inventory discrepancies between peat profiles P5 and P3, P6 were, most likely, caused by the lithological heterogeneity of the examined samples. Upper intervals of P3 and P6 consisted of fresh moss-peat, whereas lower layers were strongly humified. In contrast, the bottom part of the profile P5 contained considerable amounts of mineral matter as expressed by the lower loss on ignition (LOI) values [29]. Such conclusion was additionally confirmed by analogous behaviour of 137Cs for mentioned peat profiles [29].

Activity concentrations were higher in the case of the peats (Table 3), but total inventories turned out to be greater for the soil profiles (Table 4). The main reason for such inversion could be related to notably larger surface mass density (kg/m2) for soil layers and entire cores; the mass data was presented by Łokas et al. [29, 32]. On the other hand, peat accumulates pollutants more efficiently than soil (hence higher activity concentrations), which could somewhat compensate lower bulk density of peats for inventory estimation. The analysis of 137Cs inventories in peat [29] and soil profiles [32] by Łokas et al., revealed large differences between tundra samples (soil and peat) and proglacial material (soil). Namely, results for initial soils GA, GB and GC reached significantly higher levels of 137Cs in comparison with profiles H1, P3, P5 and P6, whereas for the latter ones 137Cs inventories were rather less varied. Therefore, we cannot linkage dissimilarities of activity concentrations and total inventories of 90Sr between examined peat and soil profiles with LOI values only. The problem seems to be more complex and affected by other factors.

Further interpretation of the data set requires understanding the mode of pollutant deposition for initial soil as well as tundra soil and peat profiles. Generally, the tundra is rather exposed for fresh fallout (i.e., suspended airborne particles, aerosols), which might be captured and accumulated almost continuously, in accordance with temporal flux variations of the given radioisotope. (Other processes of radionuclide supplies could also take place, however, atmospheric deposition is assumed as dominating). If so, such media should record, layer-by-layer, the history of the fresh radioactive fallout. In contrast, the proglacial zone was covered by the ice cap for most of the time, thus the ground was unavailable for suspended airborne particles. It is only when the glacier begins to melt that proglacial soils open up to the direct atmospheric fallout. Therefore, the radionuclide content within recent initial soils is derived mostly from melted ice, as the pollutants are deposited on its surface (e.g., in form of cryoconite). This mode of contaminant pathway throughout the ecosystem may be called secondary transport. No temporal changes of the radioactive fallout (i.e., changes with the core depth) are preserved correctly in such samples (unless they would become uncovered early enough to record fresh air deposition).

Following Fig. 3, the diversity of vertical radionuclide variations for proglacial and non-proglacial cores is evident. Tundra profiles P3, P5, P6 and H1 preserved temporal fallout oscillations over the Spitsbergen territory, but only for 137Cs and 239+240Pu. The observed peaks of 137Cs and 239+240Pu have been already linked to the Partial Nuclear Test Ban Treaty in 1963 [29, 32]; till then the nuclear weapons testing was the most intensive. The position of the maximum within the given profile depended mainly on the mass accumulation rate, which apparently varied from profile to profile. The initial soils GB and GC appeared to contain 137Cs and 239+240Pu from secondary sources only. However, the soil GA, collected from the longest distance from the glacier front position, had to be exposed early enough to record the maximum of the global fallout.

Fig. 3
figure 3

The plots present activity concentrations of 90Sr, 137Cs and 239+240Pu versus core depths in soil and peat profiles from Werenskiold (a—tundra, b—proglacial zone)

In contrast, vertical variations of the 90Sr activity were distinctly inconsistent with 137Cs and 239+240Pu deposition records (Fig. 3). Even if the maximum of 90Sr activity occurred (e.g., for P6), it was shifted relative to 137Cs and 239+240Pu peaks. Overall, 90Sr profiles seemed to be strongly disturbed.

We claim, that the above described discrepancy resulted from enhanced mobility of 90Sr compared to 137Cs and 239+240Pu. This phenomenon has been observed and reported by many researchers for a variety of soil types and environmental conditions since the 1950s [5, 6, 14, 19,20,21, 24]. For the Arctic tundra regions, the presence of thick lichen-moss layers on the ground surface prevents the substantial penetration of deposited radionuclides into the deeper layers of the vegetation-soil carpet. However, Sr can migrate faster, than Cs and Pu, from the upper to the lower moss-lichen sections and further down into the soil surface [16].

Environmental mobility of radiostrontium relative to radiocaesium, and particularly to plutonium, could be more intense when radionuclides occur in the ionic form attached to aerosols (e.g., atmospheric nuclear tests’ origin). For the places contaminated by local and on-site fallout, no principle differences in depth distributions for anthropogenic radionuclides in soil were found even 30–40 years after the deposition [7, 17]. This is explained by the presence of radiostrontium and other fission or activation products in the form associated with small fuel particles that are rather stable for natural weathering and chemical treatment [37].

Taking into account the behaviour of radiostrontium in the tundra samples from Werenskiold (peats P3, P5, P6 and soil H1) as well as presented above theory regarding its mobility, the global fallout origin of 90Sr is highly possible.

Variation of 137Cs/90Sr and 239+240Pu/90Sr isotopic ratios

To recognize the potential sources of environmental radioactive contamination, isotopic signatures were examined. In general, both 137Cs/90Sr and 239+240Pu/90Sr varied over wide ranges for each studied region of Spitsbergen (Fig. 2b, c). For cryoconite and, to a minor scale, for initial soils the mean values of ratios were extremely higher than reference values of the global fallout (Fig. 2b, c; Table 2). The overall analysis of 137Cs/90Sr and 239+240Pu/90Sr results shed no light on the provenance of 90Sr in glacial and proglacial zones. The enormous excess of Cs and Pu relative to Sr could not be assigned for any known releases of radioactivity. On the other hand, tundra samples (peats and soil) from Werenskiold were characterised by the range and mean values rather consistent with the weapons tests (Fig. 2b, c; Table 2).

The previous investigation for the Arctic archipelago of Svalbard concluded that the terrestrial environment was primarily contaminated with bomb-derived Pu isotopes and 137Cs, and, to a lesser extent, by 137Cs originating from the Chernobyl accident in 1986 [15]. For initial soils and cryoconite unexplained deviations from the weapons testing were noted, but mainly for 238Pu/239+240Pu and 240Pu/239Pu activity and mass ratios, respectively [29, 31, 33].

The studies by Bossew et al. [8], Tieber et al. [46] and Łokas et al. [31] showed the strong depletion of Sr activity in cryoconite media. 90Sr/137Cs and 90Sr/239+240Pu activity ratios, received by Łokas et al. [31] for SW Spitsbergen, varied from 0.04 ± 0.01 to 0.90 ± 0.13 and from 2.2 ± 0.8 to 48.5 ± 18.6 with mean values of 0.26 ± 0.06 and 12.4 ± 4.7, respectively. Our results indicated even greater excess of Cs and Pu relative to Sr (Fig. 2b, c). Typically, Cs and Pu are more strongly bounded, Cs in clay minerals (omnipresent in cryoconite), Pu also in the organic matter, whereas high solubility of Sr in water leads to its leaching to a much larger scale. Therefore, we assumed, following Bossew et al. [8], that Sr has been effectively removed from the examined cryoconite samples. This would also explain the loss of Sr relative to Cs and Pu observed for initial soils from Werenskiold, since the radionuclides found in proglacial zones, most likely, derived from the cryoconite.

Following the literature data and the observed intensive mobility of 90Sr in the environment, the atmospheric weapons test fallout could be the major source of 90Sr in each studied matrices and the entire examined region. However, isotopic signatures did not support such a thesis for all samples and sites.

Conclusion

Currently, the greatest attention in radiostrontium survey is paid to the dosimetric consideration, somewhat neglecting other environmental monitoring aspects. The presented research attempt was to fill the lack of knowledge regarding 90Sr levels, transfer pathways and trends in various parts of the terrestrial environment of the High Arctic and continue radioecological monitoring in terms of 90Sr.

Our results pointed out the relatively low level of 90Sr activity concentrations in different environmental media of Spitsbergen. In general, the mean activities were higher for cryoconite than for peats and soils. The highest value of 90Sr concentration was received for cryoconite from Werenskiold. Moreover, the transfer of radiostrontium from retreating glacier to initial soils was noted, that confirms previous findings for Cs and Pu radioisotopes. The total inventory of 90Sr in proglacial zones corresponded to estimation for adequate latitude belt, except the soil profile collected from the closest site to the glacier terminus. However, this value was not significantly higher than the reference level. No hot spot was discovered in terms of 90Sr content in the entire region.

Analysis of the vertical changes of 90Sr in the tundra soil and peat profiles revealed its significant migration throughout the cores. Furthermore, the strong depletion of radiostrontium compared to Cs and Pu for cryoconite matrix was detected.

Due to the large dispersion of both 137Cs/90Sr and 239+240Pu/90Sr ratios, the 90Sr provenance could not be identified. Surprisingly, intensive mobility of 90Sr might indicate the global fallout origin, because such property is attributed to ionic form, in which 90Sr was released during nuclear weapons testing.