the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Unprecedented radioactive pollution in Spitsbergen air – first data as of 21st century
Abstract. The present study concerned the Arctic troposphere, providing an experimental database on nuclear aerosols that has improved considerably since 1999. The activity concentrations of 238Pu, 239+240Pu and 241Am were determined in the surface air of Hornsund, Spitsbergen, during 2007–2021. A multivariate approach was employed, incorporating meteorological data, 7Be, 137Cs, 210Pb records and isotopic ratios, to explain the transuranium dynamics of changes and provenance. 238Pu and 239+240Pu levels were comparable to recent observations from various locations, whereas 241Am exhibited particularly high values, with a maximum of 354 Bq/m3 detected in 2019. The highest activity concentrations of 6.61 nBq/m3 for 238Pu and 15.51 nBq/m3 for 239+240Pu identified in 2015 could be due to fly ash particles following wildfires in proximity to the Chernobyl zone. During the remaining period, 239+240Pu was correlated with seasonal processes, such as local resuspension and horizontal tropospheric transport of haze layers from remote areas. Similar mechanisms likely regulated a portion of 238Pu, but to a lesser extent. Unexpected single incidents of 237Np were encountered in 2013, 2014 and 2018. Furthermore, activity ratio evaluation demonstrated a frequent occurrence of 238Pu enrichment regarding known nuclear events. Unusual elevated levels of 241Am, 238Pu and 237Np were not associated with environmental processes; therefore, the possibility of recent anthropogenic emissions should be considered. Trajectory simulations performed in 2019 showed prominent transport pathways to the Hornsund from northern Eurasia.
- Preprint
(1884 KB) - Metadata XML
-
Supplement
(719 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1573', Anonymous Referee #1, 03 Jul 2025
-
CC1: 'Reply on RC1', Marek Kubicki, 04 Jul 2025
Thank you very much for the comments and suggestions described in lines 204 and 220. I agree with them. I would like to add that the filters are delivered for analysis conducted at NCBJ three times a year, i.e. in April, July and September, and this is related to the organization of transport from the Polish Polar Station in Hornsund. This causes losses in activity, especially Be7. Your suggestions will allow us to resume the discussion on the much more frequent shipment of filters.
My interests concern the electricity of the atmosphere and the research Global Electric Circuit of the Earth. The ability to detect Be7 concentration in a short time and with high accuracy is very much needed, because Be 7 ionizes the upper layers of the atmosphere, affecting changes in the electric current between the ionosphere and the earth.
We also made attempts to increase the time resolution of the measurement of radionuclide concentration using the NA-J gamma detector placed directly under the measuring filter in order to obtain results directly at the station.
Thank you very much for your attention to this important measurement aspect. I can assure you that we are working on better solutions than the current ones.
Marek Kubicki, co-author
Citation: https://doi.org/10.5194/egusphere-2025-1573-CC1 -
AC1: 'Reply on RC1', Anna Cwanek, 18 Feb 2026
We would like to express our sincere gratitude for your thorough review of our manuscript and your constructive feedback, highlighting areas that require refinement and offering valuable suggestions for improvement. Please find our responses and explanations to the points raised in the review below.
COMMENT 1
Line 130 The formulation “the most radiotoxic elements” is not sufficiently precise. While it is true that Pu and Am isotopes are among the most radiotoxic substances on a per-mass basis when incorporated into the human body, their activity concentrations in the environment are typically very low. As a result, their actual impact on human health under natural environmental conditions is negligible, except in cases of nuclear accidents. It would be more accurate to state that they are “among the most radiotoxic elements.”
REPLY TO COMMENT 1
This is a very accurate remark which will help to clarify this part of the text. We will, of course, revise the phrase as suggested.
COMMENT 2
Line 204 The manuscript states that “the set of weekly air filters was delivered from the Polish Polar Station in Hornsund, Svalbard Archipelago, to the National Centre for Nuclear Research (NCBJ) in Świerk, Poland.” However, the text does not clearly indicate whether the filters were actually sent to Poland every week, or whether shipments coincided with the rotation of polar station personnel and supply missions, which typically occur only a few times a year. This information is particularly important for short-lived radionuclides such as 7Be, as significant delays between sample collection and analysis can result in considerable activity losses and may affect the accuracy of the results.
REPLY TO COMMENT 2
The filters are sent to NCBJ for analysis three times a year, in April, July and September. This is related to the organisation of transport from the Polish Polar Station in Hornsund. Therefore, it was necessary to recalibrate data from Hornsund due to radionuclide decay. As the manuscript explained, correction factors were used for all the gamma emitters because of decay during sampling, decay from the end of sampling to the start of measurement, and decay during measurement. We are aware that the measurement uncertainty for Be-7 is higher, but this does not imply that the result is erroneous. Specifically, the loss of activity in Be-7 does not affect the reliable determination and identification of the annual variability of this radionuclide. At this time, we have no other way to take gamma measurements.
COMMENT 3
Line 220 Due to low concentrations, the samples were aggregated on a quarterly or semi-annual basis. While this approach is technically understandable, it significantly limits the ability to detect short-term incidents and to link them to specific meteorological events or emission episodes. I recommend that the authors explicitly acknowledge this limitation in the manuscript.
REPLY TO COMMENT 3
Thank you for this accurate observation. Indeed, a database based on quarterly results suffers from low resolution. The reason for our decision was explained in the Materials and Methods section. However, we will emphasise this problem more strongly in the discussion of the results and conclusions.
COMMENT 4
Line 279 While the application of Tukey’s interquartile range (IQR) method with a conservative multiplier (k = 3) is a well-established and robust approach for detecting statistical outliers, it may have certain limitations when applied to highly dispersed and temporally aggregated environmental data, such as quarterly-averaged radionuclide concentrations. In such cases, individual extreme values often reflect true episodic events (e.g., sudden contaminant influx, wildfire transport, or atmospheric resuspension episodes) rather than errors or measurement artifacts. The IQR method, which does not consider temporal context or trends, may misclassify these scientifically relevant episodes as statistical outliers and overlook underlying regime shifts or changes in baseline levels. Therefore, it could be beneficial to consider supplementing the IQR analysis with additional time series methods—which are well suited to identifying genuine shifts or episodic changes in environmental data.
REPLY TO COMMENT 4
We appreciate the suggestion to use additional time series analysis methods. In our study, we used the Tukey IQR method (k = 3) to identify outliers when preparing the dataset for Spearman rank correlation calculations. Removing outliers at this stage was a technical step motivated by the need to reduce the influence of a small number of very high values on the correlation results, given the relatively small number of observations in the quarterly-aggregated time series. We would like to emphasise that labelling a point as an 'outlier' in the statistical analysis did not mean excluding it from scientific interpretation. All such cases were included in the discussion section, analysed in the context of meteorological conditions and possible sources of contamination, and presented in Figures 3–4 as potentially important environmental episodes.
COMMENT 5
Line 319 I suggest that Table 1 should also include information on the measurement uncertainty, particularly if the filters were analyzed after a significant storage period. If there was a substantial delay between sample collection and measurement, the decay correction and its associated uncertainty could have a notable impact on the final results—especially for radionuclides with shorter half-lives
REPLY TO COMMENT 5
Of course, we understand the importance of measurement uncertainties accompanying the values of the parameters studied. All uncertainties in the raw data (gamma and alpha emitters) are presented in the open-access database linked to the manuscript: https://doi.org/10.48733/no6.25.015. We would therefore prefer to avoid adding this information to Table 1, which presents purely statistical results. In addition, we propose adding information about where detailed results and uncertainties can be found in the Materials and Methods section.
COMMENT 6
Line 345 While the general patterns of 210Pb and 7Be in the lower atmosphere are similar, it is worth noting that, in the case of 210Pb, local emissions related to fuel combustion at the station during winter may also contribute to the elevated concentrations observed in this period
REPLY TO COMMENT 6
Certainly, heating the station affects the local Pb-210 concentration in the air. However, the heating season is quite long, covering particularly in the first and fourth quarters, so it is difficult to clearly link the fuel combustion at the station to the maximum Pb-210 in the first quarter.
COMMENT 7
Line 355 The outliers result from the application of Tukey’s method, as noted in the comment described in line 279.
REPLY TO COMMENT 7
Yes, and this method revealed outliers that could be related to incidental events, such as the Fukushima accident. These outliers are distinct from typical variability, which is determined by natural phenomena.
COMMENT 8
Line 569 Additionally, military activities in Ukraine in 2022 may have contributed to a temporary increase in airborne 137Cs, as observed across Europe.
REPLY TO COMMENT 8
This is generally possible, but 2022 is outside the studied sampling period. Air filter samples were taken between 2007 and 2021. We intended to include nuclear events that took place before and during the mentioned period.
Citation: https://doi.org/10.5194/egusphere-2025-1573-AC1
-
CC1: 'Reply on RC1', Marek Kubicki, 04 Jul 2025
-
CC2: 'Comment on egusphere-2025-1573', Krystyna Kozioł, 21 Oct 2025
The manuscript „Unprecedented radioactive pollution in Spitsbergen air – first data as of 21st century” presents an exceptionally long time series, which is a record of atmospheric radionuclide concentrations. This allows for tracing unusual events in the 21st century, where man-made actinides occur in the Arctic, likely due to long-range atmospheric transport (however, other possibilities are also considered in the study). The temporal resolution represented by this study is unavailable to environmental studies of sediment or ice cores, which makes it a positively anticipated contribution to the field of environmental pollution studies.
Perhaps the only other medium which offers similar temporal resolution is fresh snow, on which I will draw in particular in this comment, since there are studies published on snowfall records of other types of pollution, falling within the commented study period and somewhat corroborating the presented interpretations. Fresh snow has been sampled in the very interesting year 2019 in Hornsund, within two studies. Cappelletti et al. (2023) noted a very high 14C activity incindent, having occurred in April 2019, and attributed it to a long-range transport event bringing nuclear waste products to southern Svalbard from Europe, which they supported with scanning electron microscopy (SEM) analysis of the particle composition in the sampled snow. The latter components have been also found on filters exposed in an active air sampler collecting short-term samples at the very same environmental hut roof that Cwanek et al. (discussed manuscript) used for their study site, which were then analysed by SEM (Pawlak et al., 2024), and attributed to anthropogenic activities, without a clear link to nuclear waste. Pawlak et al. (2024) also noticed extraordinarily high concentrations of orgnochlorine persistent organic pollutants (POPs) in the atmospheric air of Hornsund in April/May 2019 (as compared to the Zeppelin station in Svalbard with a long-term record), while Pawlak et al. (2023) showed through a comparison of spatially distributed fresh snowfall collected during April/May of 2019 and 2021, that 2019 snowfalls were under a strong influence of long-range atmospheric transport of pollution, while in 2021, the same area was mostly exposed to air masses with an Arctic origin and a very low pollution level.
The Authors make an interesting speculation about the smoke detector incorrect disposal as a source, which is not entirely excluded as there are smoke detectors used at the PPS Hornsund and the non-hazardous waste is partly disposed of there in an incinerator. However, the station ensures proper handling of dangerous waste and the removal of such items into facilities outside of Hornsund. Furthermore, it is unlikely that repeated events could all be attributed to this source of 241Am, since its level appears to have been generally elevated in the Hornsund area. Moreover, spatially distributed evidence from snow studies suggests a notable long-range transport event affecting both atmospheric air and precipitation concentrations in the Hornsund area in spring 2019. With the prevailing eastern winds, the upper part of Hans glacier is unlikely to receive air connected to the exhaust of the Polish Polar Station’s incinerator.
Further support is provided by air mass trajectory analysis in the manuscript. While air mass trajectory usually needs to be interpreted in synchrony with other environmental data, due to its oftentimes ambigous intepretation on its own, the authors have made that effort by considering potential sources by their radioisotope ratio characteristics. Furthermore, the aforementioned papers listing extraordinary events in the 2019 season indicate a special case of long-range transported pollution occurring in that year.
To conclude, I believe the example of the unusual long-range atmospheric transport events in the 2019 spring season in Hornsund shows that a time series of air concentrations of pollutants at this location is a highly anticipated record for the scientific community studing environmental pollution in the Arctic and the effects of long-range atmospheric pollution transport.
Cited references:
Cappelletti, D., Ežerinskis, Ž., Šapolaitė, J., Bučinskas, L., Luks, B., Nawrot, A., Larose, C., Tuccella, P., Gallet, J.C., Crocchianti, S., Bruschi, F., Moroni, B., Spolaor, A., 2023. Long-range transport and deposition on the Arctic snowpack of nuclear contaminated particulate matter. J. Hazard. Mater. 452, 131317. https://doi.org/10.1016/j.jhazmat.2023.131317
Pawlak, F., Koziol, K., Frankowski, M., Nowicki, Ł., Marlin, C., Sulej-Suchomska, A.M., Polkowska, Ż., 2023. Sea spray as a secondary source of chlorinated persistent organic pollutants? - Conclusions from a comparison of seven fresh snowfall events in 2019 and 2021. Sci. Total Environ. 891. https://doi.org/10.1016/j.scitotenv.2023.164357
Pawlak, F., Koziol, K., Wilczyńska-Michalik, W., Worosz, M., Michalik, M., Lehmann-Konera, S., Polkowska, Ż., 2024. Characteristics of Anthropogenic Pollution in the Atmospheric Air of South-Western Svalbard (Hornsund, Spring 2019). Water (Switzerland) 16. https://doi.org/10.3390/w16111486
Citation: https://doi.org/10.5194/egusphere-2025-1573-CC2 -
AC3: 'Reply on CC2', Anna Cwanek, 18 Feb 2026
We would like to express our sincere appreciation for your insightful commentary. It is a valuable source of knowledge, with numerous insightful remarks and observations. Of particular interest are the 2019 results in Hornsund, which revealed elevated levels of C-14, organic POPs and clear evidence of long-range atmospheric transport. Even though the primary focus of our research lies in other types of pollutants, the observed convergence of incidents in the air filters or snowfall in Hornsund confirms that this area was affected by an intense stream of aerosols from distant locations. The information regarding hazardous waste management at the station has been found to be extremely useful. The insights provided in this commentary will undoubtedly contribute to further refinement of our discussion on the outcomes.
Citation: https://doi.org/10.5194/egusphere-2025-1573-AC3
-
AC3: 'Reply on CC2', Anna Cwanek, 18 Feb 2026
-
RC2: 'Comment on egusphere-2025-1573', Anonymous Referee #2, 06 Jan 2026
Ref. No.: EGUsphere-2025-1573
This paper measures the activity concentrations of 238Pu, 239+240Pu, and 241Am in the surface air of Spitsbergen, from 2007 to 2021, and uses a multivariate method to analyze them in conjunction with meteorological data and the activity concentrations of 7Be, 210Pb, and 137Cs monitored during the same period. It demonstrates that these records and isotope ratios can be used to explain the dynamic changes in transuranium elements and provide one of the best methods for identifying and distinguishing resuspended or newly released radioactive isotopes in the atmosphere. This study is very meaningful, and I recommend that it be accepted for publication after minor revisions.- There are Table S1, S2,,,,S7 and Fig. S1,,,,S5 are cited in the text, but cannot be found in paper?
- Starting from line 488, the author states "strong correlations were obtained...". In fact, I cannot see these correlations in Figure 6, nor can I find these strong correlations in Table 4, except for the strong correlation between 239+240Pu and 241Am in the third quarter.
- It is recommended that the 238Pu/239+240Pu and 241Am/239+240Pu ratios obtained from other studies be added to Figure 7 and represented with different colors or shapes for easy comparison.
Citation: https://doi.org/10.5194/egusphere-2025-1573-RC2 -
AC2: 'Reply on RC2', Anna Cwanek, 18 Feb 2026
We would like to thank you very much for reviewing our manuscript, providing us with useful suggestions for changes and highlighting elements that require refinement. Our responses and explanations to the review are provided below.
COMMENT 1
There are Table S1, S2,,,,S7 and Fig. S1,,,,S5 are cited in the text, but cannot be found in paper?
REPLY TO COMMENT 1
Yes, the mentioned tables and figures are cited in the text. Tables and figures numbered with the letter S are located in a separate file called ‘Supplementary material’. We will add this information to the main text to make it clearer.
COMMENT 2
Starting from line 488, the author states "strong correlations were obtained...". In fact, I cannot see these correlations in Figure 6, nor can I find these strong correlations in Table 4, except for the strong correlation between 239+240Pu and 241Am in the third quarter.
REPLY TO COMMENT 2
While Figure 4 attempts to illustrate the results of descriptive statistics in quarters, Table 4 presents the results of cross-correlations in quarters for the radionuclides studied. As stated in the manuscript, strong correlations are highlighted in bold and marked with an asterisk *. We have verified that this table is correctly presented in the manuscript PDF file posted by the EGUsphere. For better clarity, we will add information on how strong correlations are marked in the description of Table 4.
COMMENT 3
It is recommended that the 238Pu/239+240Pu and 241Am/239+240Pu ratios obtained from other studies be added to Figure 7 and represented with different colours or shapes for easy comparison.
REPLY TO COMMENT 3
Naturally, we understand the need for a visual comparison of the Hornsund activity ratio results with those from other stations. Such a comparison is presented in the 'Results and Discussion' section. Due to the infrequent analysis of Pu-238, 239 and 240, or Am-241, in air filters, comparative data is generally scarce. Furthermore, data from other stations often refers to a different sampling period than the one we studied and to filters that are aggregated differently (i.e. not quarterly). Therefore, we feel that adding further data points from other stations to Figure 7 would be inconsistent with the manuscript section that discusses the data comparison in more detail and provides contextual information. We sincerely hope that our points of view will be found convincing.
Citation: https://doi.org/10.5194/egusphere-2025-1573-AC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 943 | 255 | 45 | 1,243 | 78 | 25 | 43 |
- HTML: 943
- PDF: 255
- XML: 45
- Total: 1,243
- Supplement: 78
- BibTeX: 25
- EndNote: 43
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript entitled "Unprecedented radioactive pollution in Spitsbergen air – first data as of 21st century" presents a very interesting and well-written study. The research addresses an important topic in the field of environmental radioactivity and provides new, valuable data on the concentrations and sources of transuranic radionuclides and other artificial and natural radioisotopes in the Arctic atmosphere. The manuscript is clear, logically structured, and presents its methodology and results in a transparent way. I found the discussion to be insightful and supported by relevant literature. While I have a few minor comments and suggestions for improvement, overall, I believe the work is of high quality and recommend it for publication after minor revision.
Line 130 The formulation “the most radiotoxic elements” is not sufficiently precise. While it is true that Pu and Am isotopes are among the most radiotoxic substances on a per-mass basis when incorporated into the human body, their activity concentrations in the environment are typically very low. As a result, their actual impact on human health under natural environmental conditions is negligible, except in cases of nuclear accidents. It would be more accurate to state that they are “among the most radiotoxic elements.”
Line 204 The manuscript states that “the set of weekly air filters was delivered from the Polish Polar Station in Hornsund, Svalbard Archipelago, to the National Centre for Nuclear Research (NCBJ) in Świerk, Poland.” However, the text does not clearly indicate whether the filters were actually sent to Poland every week, or whether shipments coincided with the rotation of polar station personnel and supply missions, which typically occur only a few times a year. This information is particularly important for short-lived radionuclides such as 7Be, as significant delays between sample collection and analysis can result in considerable activity losses and may affect the accuracy of the results.
Line 220 Due to low concentrations, the samples were aggregated on a quarterly or semi-annual basis. While this approach is technically understandable, it significantly limits the ability to detect short-term incidents and to link them to specific meteorological events or emission episodes. I recommend that the authors explicitly acknowledge this limitation in the manuscript.
Line 279 While the application of Tukey’s interquartile range (IQR) method with a conservative multiplier (k = 3) is a well-established and robust approach for detecting statistical outliers, it may have certain limitations when applied to highly dispersed and temporally aggregated environmental data, such as quarterly-averaged radionuclide concentrations. In such cases, individual extreme values often reflect true episodic events (e.g., sudden contaminant influx, wildfire transport, or atmospheric resuspension episodes) rather than errors or measurement artifacts. The IQR method, which does not consider temporal context or trends, may misclassify these scientifically relevant episodes as statistical outliers and overlook underlying regime shifts or changes in baseline levels. Therefore, it could be beneficial to consider supplementing the IQR analysis with additional time series methods—which are well suited to identifying genuine shifts or episodic changes in environmental data.
Line 319 I suggest that Table 1 should also include information on the measurement uncertainty, particularly if the filters were analyzed after a significant storage period. If there was a substantial delay between sample collection and measurement, the decay correction and its associated uncertainty could have a notable impact on the final results—especially for radionuclides with shorter half-lives
Line 345 While the general patterns of 210Pb and 7Be in the lower atmosphere are similar, it is worth noting that, in the case of 210Pb, local emissions related to fuel combustion at the station during winter may also contribute to the elevated concentrations observed in this period
Line 355 The outliers result from the application of Tukey’s method, as noted in the comment described in line 279.
Line 569 Additionally, military activities in Ukraine in 2022 may have contributed to a temporary increase in airborne 137Cs, as observed across Europe.