The Baltic Sea, a significant brackish water ecosystem, faces pollution from various anthropogenic radionuclides. Identifying and quantifying these sources is crucial for assessing ecological and human health risks. While sources like global fallout from nuclear weapons testing and discharges from European reprocessing plants (La Hague and Sellafield) contribute to radionuclide levels in the Baltic, unidentified sources remain a concern. Uranium-236 (²³⁶U), a significant byproduct of nuclear reactors, acts as a sensitive tracer for such releases. However, distinguishing reactor-derived ²³⁶U from global fallout ²³⁶U is challenging due to the abundance of natural ²³⁸U and similar ²³⁶U/²³⁸U ratios. This study addresses this challenge by employing a novel multi-isotope approach involving ²³³U, a radionuclide with different production mechanisms in nuclear reactors and nuclear weapons testing, leading to distinct ²³³U/²³⁶U ratios. This ratio serves as a powerful fingerprint to differentiate the sources of ²³⁶U. Additionally, the study incorporates ¹²⁹I, another anthropogenic radionuclide, to refine source identification and tracing transport pathways. The integration of these isotopes provides a robust methodology to uncover previously unknown sources of radionuclide contamination within the Baltic Sea ecosystem and improve our understanding of the impact of past and ongoing nuclear activities on the marine environment.

Previous research has established the presence of reactor ²³⁶U in the North Atlantic and Arctic oceans, primarily attributed to releases from La Hague and Sellafield reprocessing plants. Global fallout from atmospheric nuclear weapons testing also contributes significantly to the ²³⁶U inventory in the environment. However, distinguishing between these sources based solely on ²³⁶U/²³⁸U ratios is problematic. Recent advancements in accelerator mass spectrometry (AMS) have enabled accurate measurements of ²³³U at environmentally relevant concentrations. The distinct ²³³U/²³⁶U ratios in fallout (high) and reactor discharges (low) provide a powerful tool for source discrimination. Studies in the Irish Sea have demonstrated the utility of this ratio in identifying a dominant reactor signal from Sellafield. Combining ²³⁶U with other radionuclides, particularly ¹²⁹I, enhances source tracing capabilities, with successful applications in the subpolar North Atlantic. This existing research sets the stage for the current study's investigation into the Baltic Sea, a region characterized by unique hydrological conditions and a complex history of nuclear activities.

This study involved the collection of water and sediment samples from various locations in the Baltic Sea and surrounding waters between 2011 and 2016. Samples were collected from five geographical regions: Kattegat-Skagerrak, Danish Straits, South Baltic Sea, Middle Baltic Sea, and North Baltic Sea. The majority of water samples were from the surface (0–5 m), with some deep-water samples and one lake water sample from Lake Mälaren. Sediment samples were also collected to assess isotope accumulation trends. The concentration of ²³⁸U and ¹²⁷I in seawater was determined using ICP-MS (Inductively Coupled Plasma Mass Spectrometry) after appropriate sample preparation and dilution. A radiochemical method was employed to separate ²³³U and ²³⁶U from seawater samples prior to analysis by AMS. This involved preconcentration through Fe(OH)₃ co-precipitation followed by separation on a UTEVA resin column. The purified uranium fractions were then measured using accelerator mass spectrometry (AMS) at the Vienna Environmental Research Accelerator (VERA) facility to determine ²³³U/²³⁶U and ²³⁶U/²³⁸U atomic ratios. For sediment samples, the samples were ashed, leached with aqua regia, and processed similarly for AMS analysis. The ¹²⁹I concentration and ¹²⁹I/¹²⁷I atomic ratios were measured using AMS at the Uppsala University Tandem Laboratory. This analysis was conducted on samples from a subset of the campaign, primarily from 2015. A binary mixing model was utilized to evaluate source contributions of ²³⁶U, considering end members representing North Sea water and freshwater inputs. This model was applied to analyze the behavior of both the ²³⁶U/²³⁸U and ²³⁶U/¹²⁹I atomic ratios to identify the origin of excess ²³⁶U in the Baltic. The ²³³U/²³⁶U atomic ratio also provided crucial insights for disentangling the sources of ²³⁶U by analyzing deviations from the binary mixing models.

The study revealed a previously unknown source of reactor ²³⁶U in the Baltic Sea, evident in the high ²³⁶U concentrations and ²³⁶U/²³⁸U atomic ratios, particularly in the central and northern parts of the sea. The spatial distribution of ²³⁶U shows a pattern not solely explained by inputs from La Hague and Sellafield reprocessing plants or global fallout. The ²³³U/²³⁶U atomic ratios, ranging from (0.14–0.87) × 10⁻², provide further evidence for this additional source. Samples from the central Baltic Sea exhibit higher ²³³U/²³⁶U ratios compared to the expected fallout signature. Binary mixing analysis using ²³⁶U/²³⁸U and ²³⁶U/¹²⁹I atomic ratios, plotted against salinity, indicates that approximately 90% of the ²³⁶U in the central Baltic Sea originates from a local source lacking significant ¹²⁹I. A more detailed binary mixing model using ²³³U/²³⁶U ratios suggests that roughly two-thirds of the anthropogenic uranium in the middle and north Baltic regions comes from a local source characterized by a low ²³³U/²³⁶U atomic ratio – characteristic of thermal nuclear reactor ²³⁶U. This additional reactor ²³⁶U source contributes approximately 200 ± 47 g to the Baltic Sea. Sediment samples from the Studsvik area, near a Swedish nuclear research facility, show extremely high ²³⁶U concentrations (three orders of magnitude higher than other Baltic sediments), providing strong evidence for this location as the likely source of the additional ²³⁶U. The ²³³U/²³⁶U atomic ratio in the Studsvik sediment (0.36±0.05) × 10⁻² confirms a reactor origin. While the specific amount of ²³⁶U released from Studsvik is not well-documented, the high levels in the sediments strongly support this as the source.

The findings demonstrate the presence of a previously unrecognized source of reactor-derived ²³⁶U in the Baltic Sea, highlighting the limitations of relying solely on ²³⁶U/²³⁸U ratios for source apportionment. The multi-isotope approach, particularly the use of ²³³U/²³⁶U ratios, proves crucial in distinguishing between different sources of ²³⁶U. The identified source, likely linked to past waste discharges or accidental releases from the Studsvik nuclear research facility, emphasizes the potential for undetected or underreported radioactive releases in the marine environment. This study underscores the importance of comprehensive monitoring and improved record-keeping of nuclear facilities to accurately assess and mitigate risks associated with anthropogenic radionuclides in marine ecosystems. While the radiological risk from the current ²³⁶U levels is deemed negligible, the potential for future release of other co-located radionuclides needs to be further investigated and monitored, especially given the potential impact of climate and environmental change.

This study successfully identified a previously unknown source of reactor ²³⁶U in the Baltic Sea using a novel multi-isotope fingerprinting technique. The high concentrations of ²³⁶U in sediments near the Studsvik nuclear research facility strongly implicate this site as the primary source. While the current radiological risk is low, this discovery underscores the need for enhanced monitoring and transparency in nuclear waste management practices. Future research should focus on investigating the potential for other radionuclide releases from the same source and the impact of environmental change on the resuspension and remobilization of previously deposited radioactive materials.

The study's analysis is based on a limited number of samples spanning several years. The interpretation relies on a binary mixing model, which simplifies the complex hydrodynamics of the Baltic Sea. The study focused primarily on surface water samples, potentially missing information about the vertical distribution of radionuclides. Further, the precise quantification of the ²³⁶U release from Studsvik is hampered by incomplete historical records.