Introduction
Despite being a weak beta emitter, with a long physical half-life of 5730 years, 14C contributes to a significant part of the collective effective dose to the public from globally dispersed long-lived radionuclides emitted from the nuclear power industry (UNSCEAR 2016). Carbon is easily taken up by all living organisms, resulting in a certain amount of 14C, e.g. in food. Hence, it is important to monitor airborne emissions as well as waterborne discharges of this radionuclide from the nuclear power industry.
The current regulations in Sweden regarding monitoring of radioactive emissions during normal operation of nuclear power reactors require, as far as possible and reasonable, nuclide specific measurements of airborne emissions and waterborne discharge of several radionuclides, including 14C (SSMFS 2021:6). Airborne emissions of 14C from Swedish nuclear power reactors are currently routinely monitored with the aid of commercial stack air samplers, followed by analysis with liquid scintillation counting (IAEA 2004). To our knowledge, only a few measurements have been performed on process water and liquid discharges (Magnusson Reference Magnusson2007; Magnusson et al. Reference Magnusson, Aronsson, Lundgren and Stenström2008a, Reference Magnusson, Stenström and Aronsson2008b). These pointed to significantly lower liquid than airborne discharge rates, as also generally described by the IAEA (IAEA 2004). However, the excess of 14C in the terrestrial and marine environment does not only depend on the amounts of 14C in the gaseous and liquid waste streams. Airborne gaseous effluents of 14C, released through high stacks, are effectively diluted before reaching biota at ground level. For liquid 14C discharges the dilution is substantially less before reaching marine biota. Hence, biota in the local marine environment of a nuclear power plant (NPP) is relatively more exposed to discharged liquid 14C, than biota in the terrestrial environment is exposed to gaseous 14C releases.
In a previous study (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022), we investigated 14C in samples of brown algae (Fucus spp.) from numerous sites along the Swedish coast in 2020. Fucus spp. are often used as bioindicators in environmental studies, as they are stationary (attached to rocks at the seabed) and concentrate many types of radioactive and non-radioactive pollutants (Mattsson et al. Reference Mattsson, Eriksson Stenström and Pédehontaa-Hiaa2022). For 14C, Fucus spp. reflect the current level of 14C of Dissolved Inorganic Carbon (DIC) in the water. In the previous study (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022), Fucus spp. collected close to Ringhals NPP had up to ∼25% excess of 14C compared to that at nearby reference sites. This is a higher excess than what is normally found in the terrestrial environment of light-water cooled and moderated reactors such as the Ringhals NPP (Stenström et al. Reference Stenström, Skog, Nilsson, Hellborg, Leide-Svegborn, Georgiadou and Mattsson2010). Due to the lack of previous measurements of 14C in the marine environment of Ringhals NPP, we have investigated the possibility of retrieving such information by using the annual structures in blue mussels (Mytilus edulis), which are common in the waters along the Swedish west coast.
Mytilus edulis shells form annual structures as they grow (Figure 1). While there is conflicting information as to whether these structures form when the mussels are spawning, during summer, or due to slower growth during winter, there is a consensus that M. edulis shells form one band per year (Kautsky Reference Kautsky1982). Observations by Seed (Reference Seed1969) strongly indicate that in M. edulis, the annual structures form during winter.
(a). Terms used to denote parts of Mytilus edulis. (b) Illustration of the layers of a M. edulis shell where the mantle, a part of the soft tissue, is also included. The dark lines in the periostracum show the direction of the annual structures, while the bright parts show the direction of fiber growth. Based on figure 2 in Checa ( Reference Checa 2018 ), recreated with permission from the author.
The 14C in the M. edulis soft tissue reflects the amount of 14C in its current food (the turnover time typically varies between 100 to 300 days depending on season (Smaal and Vonck Reference Smaal and Vonck1997)). In the shells, 14C mainly reflects DIC in the surrounding water during the whole growth period (Tierney Reference Tierney2017). Mytilus edulis are usually attached to a specific location by proteinaceous byssus threads, but can actively move using their foot, and also occasionally voluntarily detach from a location and drift with a current or become involuntarily moved by waves (Ljunghager Reference Ljunghager2017). This semi-sessile characteristic lifestyle may partly limit their value as bioindicator (which preferably should be stationary).
There have been several past studies which have taken samples from annual structures in bivalves to study 14C, however most were done on the species Arctica islandica to study the bomb pulse (Dunca et al. Reference Dunca, Mutvei, Göransson, Mörth, Schöne, Whitehouse, Elfman and Baden2009; Schöne et al. Reference Schöne, Fiebig, Pfeiffer, Gleβ, Hickson, Johnson, Dreyer and Oschmann2005; Scourse et al. Reference Scourse, Wanamaker, Weidman, Heinemeier, Reimer, Butler, Witbaard and Richardson2012). There have also been some studies where M. edulis shells have been used to investigate 14C emissions from NPPs, but these studies have analyzed the entire shells and not the annual structures (Castrillejo et al. Reference Castrillejo, Witbaard, Richardson, Dekker, Welte, Wacker and Christl2020; Tierney Reference Tierney2017). To the best of our knowledge, no studies have been published where M. edulis shells have been divided into annual samples for retrospective analysis of 14C in the marine environment near NPPs.
On the other hand, the periodic growth patterns of bivalve shells, including M. edulis, has been used for analysis of δ13C, δ18O and various elements such as Ba and Ca and applied to provide ecological and environmental data (Gillikin et al. Reference Gillikin, Lorrain, Bouillon, Willenz and Dehairs2006; Gröcke and Gillikin Reference Gröcke and Gillikin2008; Surge and Schöne Reference Surge and Schöne2013). Parameters such as water temperature and precipitation can be revealed with sub-annual resolution by using advanced micro-growth extraction methods, such as microdrilling, micromilling and laser ablation techniques (Spötl and Mattey Reference Spötl and Mattey2006). The carbon material of long-lived bivalves also has potential in environmental baseline monitoring (Steinhardt et al. Reference Steinhardt, Butler, Carroll and Hartley2016). E.g., records of metal pollution may be revealed in the shell annual structures (Steinhardt et al. Reference Steinhardt, Butler, Carroll and Hartley2016; Zuykov and Schindler Reference Zuykov and Schindler2019).
The aim of this study was to investigate the feasibility of M. edulis shells as bioindicators of past 14C discharges as DIC from nuclear power plants. A comparison between the 14C values in Fucus spp. and in annual structures of M. edulis shells was performed to evaluate how reliable the method using the annual bands of M. edulis shells is for estimating DIC at a specific location and time. The use of samples from the 1970s instead of contemporary samples is justified by the fact that the bomb pulse curve was steeper then. This provides a higher annual resolution in F14C compared to later years. An important requirement for the method developed, was that it should be possible to execute with the equipment already available at our laboratory. Furthermore, we wanted the method to be as simple as possible, to ensure it would be easy to reproduce. Micro-growth extraction methods, such as micromilling (Wurster et al. Reference Wurster, Patterson and Cheatham1999) were for these reasons not possible.
Materials and methods
Mytilus edulis shell annual structures
We have implemented a method to extract 14C from the fibrous layer of the annual structures of shells of M. edulis collected at the Swedish coast. Samples from Ringhals NPP as well as from a remote reference site (Båteviken close to the border of Norway) have been analyzed and compared to existing 14C data in Fucus spp. from the same sites. We have also analyzed shells collected at Särdal on the Swedish west-coast in 1978 and compared the data to 14C values in Fucus spp. from the same site and years 1972–1978 (Mattsson et al. Reference Mattsson, Eriksson Stenström and Pédehontaa-Hiaa2022).
The annual structures in the nacre and the fibrous layer of a shell of M. edulis (Figure 1) form in different directions as compared to the curve of the shell valves (Checa Reference Checa2018). The structure in the fibrous layer has thinner, threadlike fibers of calcite, while the nacre has a more brick-like structure of aragonite (Checa Reference Checa2018). As a result of this, a sample containing both layers will have materials from more than one year. The two layers therefore need to be separated to ensure that the samples only contain material from one specific year. The fibrous layer was chosen, due to it being easier to divide according to its annual structures. A further discussion of this decision is found in Bjarheim (Reference Bjarheim2024).
After removing the soft tissue from the M. edulis shells using a spoon and/or spatula, the annual structure of the shells was identified prior to dividing the shells into samples accordingly. The most widely used method for identifying the annual structures, thin sectioning (e.g. Clark Reference Clark, Donald and Lutz1980; Haag and Commens-Carson Reference Haag and Commens-Carson2008; Neves and Moyer Reference Neves and Moyer1988), was avoided due to the carbon contamination risk from the use of epoxy. Instead, acetate peels were made to allow for the study of the internal structures of the shells. Acetate peels require less manual work and sample material compared to thin sections (Kennish Reference Kennish, Rhoads and Lutz1980; Lutz Reference Lutz1976). An acetate peel is produced by first cutting the shell along the axis of maximal growth, the cut edge is then sanded using a fine grit, before acetone is poured along the cut. Lastly the shell is placed on a cellulose film leaving an imprint of the edge. A full step by step description of the method used within this project can be found in Bjarheim (Reference Bjarheim2024).
It was decided to divide the M. edulis shells according to their external annual structures, as this was practically the most achievable at this early stage of method development. The external annual structures were identified with the aid of a strong lamp. However, this method makes it difficult to differentiate between real annual structures and other disturbances. For example, a grain of sand trapped within the shell will result in the appearance of a darker area that can be mistaken for an annual band (Neves and Moyer Reference Neves and Moyer1988). Another issue is that the outer layer of the shell slowly erodes over time, predominately in the umbo region, thus removing the external annual bands. The external annual structures were compared with acetate peels made from the same shell, to identify false annual structures, as these will be less prominent internally.
To further ensure that age estimates of the M. edulis individuals were valid, the size of the M. edulis individuals was compared with growth curves from a nearby location (Dunca and Boalt Reference Dunca and Boalt2011). Growth curves are based on the assumption that a species grows at a certain rate and reaches a maximum size. By then measuring the age and size of several specimens a graph showing the expected size at a certain age, for a specific species, can be made (Seed Reference Seed, Rhoads and Lutz1980). However, as noted by Seed (Reference Seed, Rhoads and Lutz1980), individuals of M. edulis differ significantly in size depending on their growth conditions. Consequently, it is important to use relevant growth curves based on specimens found close to the sampling sites under investigation. It is also important to note that the M. edulis within a single settlement can vary in size even if they have the same age (Seed Reference Seed1969), altogether making a large sample size preferable when making growth curves.
A common equation for such a growth curve, known as the von Bertalanffy growth equation, is
$${L_t} = {L_\infty }\left( {1 - {e^{ - kt}}} \right) + {L_0},$$
where
${L_\infty }$
is the maximum attainable length,
$k$
is a growth constant,
$t$
is the time, and
${L_0}$
is the length when the first annual band is formed (t = 0 in Figure 3, we later in manuscript denote this as age 1 year) (Dunca and Boalt Reference Dunca and Boalt2011). Values of
${L_\infty }$
and
$k$
can vary between sample sites. The software OriginPro 2024 (OriginLab, Northampton, Mass., USA) was used to fit the von Bertalanffy growth equations to plots of the length of the M. edulis shells as a function of the age of the mussel. The growth curves of the M. edulis individuals were compared to length data of M. edulis from Kullen (in Kattegat on the Swedish west coast, 56.3N, 12.4E), reported by Dunca and Boalt (Reference Dunca and Boalt2011). Additionally, data for M. edulis shells collected in the Baltic Sea (Högby fyr, 57.1N, 17.0E), with lower salinity than on the Swedish west coast, were used for size comparisons between different conditions (data from Dunca and Boalt (Reference Dunca and Boalt2011)). This method of combining acetate peels and growth curves for determination of the age of bivalves has also been used when studying other shells in the Mytilus family (Millstein and O'Clair Reference Millstein and O'Clair2001).
(a) Map of southern Sweden including sampling sites and nuclear power plants. Sk. – Skagerrak, K. – Kattegat. Ocean currents are indicated (surface water movement in orange and bottom water movement in blue) (adapted from Qiao et al. (Reference Qiao, Andersson and Nielsen2020) and Qiao et al. (Reference Qiao, Zhang, Steier, Hain, Hou, Vartti, Henderson, Eriksson and Aldahan2021)). (b) Enlarged map of area surrounding Ringhals NPP. Arrows indicate inlet and outlet of cooling water. Maps: Schlitzer, Reiner, Ocean Data View, https://odv.awi.de, 2025.
Length of the six shells from Mytilus edulis analyzed (two individuals from each of the sites Båteviken (site 1), Bua (site 11) and Särdal (site 15) on the Swedish west coast) as a function of assessed age of the M. edulis individuals. The von Bertalanffy growth fit applied to the data is similar to that of age and length data of M. edulis from another site at the Swedish west coast, Kullen (data from Dunca and Boalt (Reference Dunca and Boalt2011)). The Högby M. edulis shell data (Dunca and Boalt (Reference Dunca and Boalt2011)), from M. edulis in the Baltic Sea, demonstrate the reduction in size of M. edulis in the Baltic Sea due to lower salinity.
Sites
The relevant sampling sites along the Swedish west coast, and the location of the Swedish NPP Ringhals, are shown in Figure 2a. The site numbers 1 (Båteviken, 58.95N, 11.13E), 11 (Bua, at Ringhals NPP; 57.24N, 12.10E) and 15 (Särdal, 56.76N, 12.63E) are the same as in Eriksson Stenström and Mattsson (Reference Eriksson Stenström and Mattsson2022). Båteviken is located close to the border to Norway at about 220 km NNW of Ringhals NPP. At Bua (at Ringhals NPP), sampling was performed at two locations close to Krogstadsudde Lighthouse (57.24N, 12.10E): Fucus vesiculosus was collected ∼70 m east of the lighthouse, while M. edulis was collected ∼90 m south-west of the lighthouse (see Figure 2b). Further information summarizing some relevant aspects on airborne as well as liquid 14C discharges from Ringhals NPP can be found in Eriksson Stenström and Mattsson (Reference Eriksson Stenström and Mattsson2022).
Särdal is situated about 65 km SSW of Ringhals NPP. The Särdal site is a reference site, which since 1967 has been used for regular sampling of mainly Fucus spp. for investigations of radionuclide concentrations and metals (see Mattsson et al. (Reference Mattsson, Eriksson Stenström and Pédehontaa-Hiaa2022) and references therein). Two additional sites (not shown in Figure 2a, but located close to Båteviken), 52 Svinesund (59.10N, 11.27E) and 53 Saltbacken (59.08N and 11.23E) located at Ringdalsfjorden by the Norwegian boarder, were added to rule out contamination, especially at Båteviken, by discharges from the nuclear research reactor in Halden, Norway. Halden is located about 10 km further into the fjord from Saltbacken and about 10 km from Svinesund.
Sampling
Sampling of Fucus spp. was performed as described in Eriksson Stenström and Mattsson (Reference Eriksson Stenström and Mattsson2022), and M. edulis were collected in a similar manner by entering the water on foot. The Fucus spp. samples collected were constantly submerged in water to avoid uptake of atmospheric CO2 (Cook et al. Reference Cook, MacKenzie, Muir, Mackie and Gulliver2004). Several individuals of M. edulis and F. vesiculosus were retrieved at Båteviken on the 12th of October 2022. Fucus vesiculosus was also sampled at Båteviken at four other occasions: in spring 2020 (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022), autumn 2020, and spring and late summer in 2023. Mytlius edulis was collected at Bua by Ringhals NPP on the 12th of May 2023. At Bua, F. vesiculosus was sampled in spring (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022) and autumn 2020, in autumn 2022 and in spring and late summer 2023. The M. edulis and F. vesiculosus samples from Särdal were collected already on the 24th of December 1978, and have been stored in dried form in a biobank since the time of collection. Furthermore, F. vesiculosus was sampled on the 17th of May 2024 at Båteviken, Saltbacken and Svinesund. Water temperature and salinity were measured at the time of sampling of all marine samples (see Eriksson Stenström and Mattsson (Reference Eriksson Stenström and Mattsson2022)). Samples of grass, representing the terrestrial environment, were collected at Båteviken on the 12th of October 2022 and at Bua the following day.
Sample preparation
The length and weight of the M. edulis individuals were measured (length 28-83 mm; typical weights from <1 g to a few g wet weight). For the 2022–2024 samples, the soft tissue was removed from the M. edulis shells, and frozen. Then the soft tissue was freeze-dried at – 55 °C for about 48 hours, before a sample was taken from the soft tissue of each shell and sent to the 14C laboratory at Lund University. The sample from Särdal consisted of 14 M. edulis individuals. The soft tissue was removed from the shells and both the shells and soft tissue contents were dried in an oven at 65°C. The dry weight/wet weight ratio of the shells and contents were 94.8% and 16.3% respectively. Shells from 2 of the M. edulis individuals were used for analysis and the soft tissue content represented all 14 individuals.
When shining through the shells with a strong lamp, the annual structures appear as dark bands, as explained by Kautsky (Reference Kautsky1982). To better see the dark bands, the periostracum (the thin outer shell layer, see Figure 1b) was carefully sanded away beforehand. Each annual band was marked, using a marker with carbon-free ink, to keep track of their locations for later stages of the sample preparation. The M. edulis samples, two shells from each site, were then divided both according to the annual structures and according to the shell layers. After the shell layers had been separated through heating (see below), the pieces became very brittle and difficult to handle. Therefore, samples were divided according to the annual structures in the fibrous layer, before separating the latter from the nacre.
The first step in dividing the shell samples according to the annual structures was to cut off the dorsal and ventral edges using a lapidary saw. Then cuts were made along the annual structures, which were previously marked. When cutting out the annual structures, a Dremel tool with a diamond tipped saw blade rotating on a slow speed was used. Separating the fibrous layer from the nacre was then accomplished by heating the cut samples. Each sample was placed in separate labelled beakers. After testing several different temperatures and durations, it was decided to heat the shell for ten minutes in an oven with air at 500°C, to achieve separation of the shell layers. Heating the shells for longer periods or at higher temperatures did not facilitate the separation, instead the shells became more brittle. What was left of the periostracum seemed to burn off at about 450°C, therefore a higher temperature than that was wanted. Furthermore, it was found that heating the shells for less than 10 minutes made the separation of the shell layers more difficult. The full process behind the method development, and a step-by-step instruction is found in Bjarheim (Reference Bjarheim2024).
After being heated, the cut shell samples from the fibrous layer were subjected to standard sample preparation techniques used at the Lund Radiocarbon Dating Laboratory. The samples were thoroughly cleaned by brushing, and then put in an ultrasonic bath. Next, the shells were treated with a weak HCl acid, 10 ml 37% HCl diluted with deionized water to 1000 ml, to detach any organic material as well as nacre that might have been left on the fibrous layer. Shell weight before and after acid treatment is included in the Supplementary materials. The amount of acid was proportional to the weight of the sample (for further information, see Appendix B in Bjarheim (Reference Bjarheim2024)). The solutions with the shell samples were put in an oven, with air at about 80°C, for three hours. Afterwards the shell samples were again thoroughly cleaned in an ultrasonic bath and then dried in a heating cabinet at 110°C.
Fucus spp. samples were dried in an oven at 70°C, ground and homogenized via mixing, prior to being cleaned with an acid-alkali-acid (AAA) pretreatment to remove carbonates and organic acids (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022). Grass samples were dried in air.
14C analysis
The samples were analyzed using single-stage accelerator mass spectrometry (SSAMS) at the Radiocarbon Dating Laboratory at Lund University. Prior to graphitization in an AGE system (Wacker et al. Reference Wacker, Němec and Bourquin2010), CO2 was liberated from the pretreated shell samples using phosphoric acid and Fucus spp. samples were combusted in an elemental analyzer. The resulting amount of graphite was ∼1 mg. The procedure for 14C measurement is described e.g. in Skog (Reference Skog2007) and Skog et al. (Reference Skog, Rundgren and Sköld2010). The background arising from sample preparation and the accelerator system was determined using IAEA-C1 marble. The standards used were IAEA-C7, IAEA-C8, SRM 4990B (OxI), and SRM 4990C (OxII). Raw data were analyzed at the Lund Radiocarbon Dating Laboratory. Results were expressed in terms of F14C (Eriksson Stenström et al. Reference Eriksson Stenström, Skog, Georgiadou, Genberg and Johansson2011; Reimer et al. Reference Reimer, Brown and Reimer2004). Reported uncertainties correspond to 1 σ.
Results and discussion
F14C of all samples analyzed are presented in the Supplementary Material.
Growth curves
The estimated age of the M. edulis individuals that were subjected to shell analysis (2 from each site) were 6 and 4 years at Båteviken, 9 and 8 years at Bua, and 7 and 6 years at Särdal. The lengths of the mussels were between 47 mm and 83 mm (Figure 3). Figure 3 also includes data from Kullen on the Swedish west coast, obtained from Dunca and Boalt (Reference Dunca and Boalt2011). Kullen is located ∼310 km from Båteviken, ∼110 km from Bua, and ∼44 km from Särdal and is the closest site for which a growth curve for M. edulis was found in the literature.
The similarity in the fitted growth curves for the Kullen site and for the six M. edulis individuals from the sites Båteviken, Bua and Särdal demonstrates comparable growth conditions for M. edulis along the Swedish west coast. This conformity also strengthens the age estimates of our samples. Furthermore, Figure 3 highlights the size differences that can be observed within M. edulis. As shown, M. edulis individuals from Högby fyr, at the island Öland off the Swedish east coast, are much smaller than those from the west coast, due to the lower salinity of the Baltic Sea compared to the Kattegat and Skagerrak on the west coast. Methodological aspects and uncertainties regarding the age of M. edulis samples are further discussed below.
Särdal reference samples
The F14C data from the Särdal site is graphically summarized in Figure 4a, also including atmospheric 14C data from the ICOS station Jungfraujoch in Switzerland (3580 m a.s.l) (Conen et al. Reference Conen, Emmenegger, Leuenberger, Steger and Steinbacher2019; Emmenegger et al. Reference Emmenegger, Leuenberger and Steinbacher2024; Hammer and Levin Reference Hammer and Levin2017; Levin and Kromer Reference Levin and Kromer2004; and Levin et al. Reference Levin, Kromer and Hammer2013) and a modeled estimation of the global average F14C levels in the upper layer of the ocean (Reimer et al. Reference Reimer, Baillie, Bard, Bayliss, Beck, Blackwell, Ramsey, Buck, Burr and Edwards2009). The entire Särdal Fucus spp. data series (Mattsson et al. Reference Mattsson, Eriksson Stenström, Pédehontaa-Hiaa, Bernhardsson, Jönsson, López-Gutiérrez, Lérida-Toro and Chamizo Calvo2025; Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022) is included in Figure 4a to demonstrate the influence of long-range transport of 14C from the spent nuclear fuel reprocessing plants La Hague (France) and Sellafield (United Kingdom). The increase in F14C in Fucus spp. in Särdal in the late 1990s correspond to increased liquid 14C discharges, mainly from La Hague, reaching the Swedish west coast after about 2 years (see Mattsson et al. (Reference Mattsson, Eriksson Stenström, Pédehontaa-Hiaa, Bernhardsson, Jönsson, López-Gutiérrez, Lérida-Toro and Chamizo Calvo2025)). The general trend in Fucus spp. from Swedish west coast waters is that F14C increases towards the north (Eriksson Stenström and Mattson Reference Eriksson Stenström and Mattsson2022). As previously reported (Eriksson Stenström and Mattsson Reference Eriksson Stenström and Mattsson2022), the increase in F14C correlates positively with the increase in salinity towards the north of the Swedish west coast due to inflow of water of high salinity from the Atlantic, also carrying 14C originating mainly from La Hague, but also from Sellafield.
F14C in annual structures of the fibrous layer of two individuals of Mytilus edulis collected at Särdal (56.76N, 12.63E) in December 1978, average F14C in soft tissue of several individuals of M. edulis from the same site and sampling occasion, and F14C in Fucus spp. from the same site collected between 1972 and 1979 (Mattsson et al. Reference Mattsson, Eriksson Stenström, Pédehontaa-Hiaa, Bernhardsson, Jönsson, López-Gutiérrez, Lérida-Toro and Chamizo Calvo2025). The average formation season is plotted as summer for the M. edulis shell structures (dark ring forming in winter). The Fucus spp. data is plotted at the collection date. Atmospheric data are for Central European clean air CO2 (Jungfraujoch, Switzerland) from Emmenegger et al. (Reference Emmenegger, Leuenberger and Steinbacher2024), Conen et al. (Reference Conen, Emmenegger, Leuenberger, Steger and Steinbacher2019), Hammer and Levin (Reference Hammer and Levin2017), Levin and Kromer (Reference Levin and Kromer2004) and Levin et al. (Reference Levin, Kromer and Hammer2013). The dashed curve shows the modelled global marine surface mixed-layer bomb pulse for the period 1950 to 1996 (Reimer et al. (Reference Reimer, Baillie, Bard, Bayliss, Beck, Blackwell, Ramsey, Buck, Burr and Edwards2009), and values after this were extrapolated based on linear regression of marine data from 1987 to 1996. Uncertainty bars represent the analytical uncertainties (1 σ).
Enlargement of years 1972 to 1980 from Figure 4a. Uncertainty bars represent the analytical uncertainties (1 σ).
The detailed results of the 14C analysis of M. edulis collected at Särdal in 1978, and of Fucus spp. from the same site, are shown in Figure 4b. F14C of one of the shells (shell 2 in Figure 4) agrees well with F14C in Fucus spp. For the other shell (shell 1), F14C is generally lower (up to 5.2%) in the shell annual structures than in Fucus spp., in particular for the structures from 1972–1975. The lowered F14C in the four oldest annual structures of shell 1, compared to the more aligned F14C in shell 2 and Fucus spp., may indicate that this individual has moved from another habitat with lower F14C in DIC. For example, the specimen may originate from a place closer to the mouth of one of the three smaller creeks that open in the area and brings 14C-depleted carbon into coastal waters (Lougheed et al. Reference Lougheed, Filipsson and Snowball2013), resulting in lower F14C in DIC at the creek mouth. F14C in the M. edulis soft tissue (F14C = 1.258 ± 0.006, average from several individuals, reflecting F14C in the dietary carbon of mussels) is identical to F14C in the annual band from 1978 in shell 2 (F14C = 1.258 ± 0.006), and overlapping within 1 σ to F14C in the annual band from 1978 in shell 1 (F14C = 1.267 ± 0.005). Hence, dietary carbon of M. edulis (represented by the soft tissue) and DIC (represented by the newer parts of the shell) appear to have been in equilibrium at Särdal at the time of sampling.
Båteviken
Figure 5 displays shell F14C data and F14C data from F. vesiculosus collected at Site 1 Båteviken at the northernmost Swedish west coast. The trend in F14C in the terrestrial environment in southernmost Sweden, obtained previously from analysis of tree rings and vegetation (Eriksson Stenström et al. Reference Eriksson Stenström, Skog, Bernhardsson, Mattsson, Nielsen, Rundgren, Muscheler, Linderson, Pédehontaa-Hiaa and Rääf2022; Bernhardsson et al. Reference Bernhardsson, Eriksson Stenström, Jönsson, Nilsson and Pédehontaa-Hiaa2023) is included, as well as F14C in terrestrial grass samples from Båteviken and from Bua (the latter close to Ringhals NPP). F14C in the terrestrial environment at both sites follows the trend line from southern Sweden, which also agrees well with central European data (Eriksson Stenström et al. Reference Eriksson Stenström, Skog, Bernhardsson, Mattsson, Nielsen, Rundgren, Muscheler, Linderson, Pédehontaa-Hiaa and Rääf2022; Bernhardsson et al. Reference Bernhardsson, Eriksson Stenström, Jönsson, Nilsson and Pédehontaa-Hiaa2023). Soft tissue from 3 M. edulis individuals collected in October 2022 has an average F14C of 1.058 ± 0.002, which is consistent with F14C from the annual band for 2022 for Mussel 2 (1.054 ± 0.005). As for Särdal, the dietary carbon of M. edulis at Båteviken appear to be in equilibrium with DIC (represented by the outer band(s) of the shell).
F14C in annual structures from shells of Mytilus edulis, in soft tissue of M. edulis and in Fucus vesiculosus collected at Site 1 Båteviken (about 200 km NNE of Ringhals NPP). The average formation season is plotted as summer for the M. edulis shell structures (dark ring forming in winter). The F14C F. vesiculosus data is plotted at the collection date. F14C in the terrestrial environment is represented by grass samples from the Båteviken (Site 1) and Bua (Site 11) (plotted at collection date). A trendline of F14C in the terrestrial environment of southern Sweden is also included (data from (Eriksson Stenström et al. Reference Eriksson Stenström, Skog, Bernhardsson, Mattsson, Nielsen, Rundgren, Muscheler, Linderson, Pédehontaa-Hiaa and Rääf2022) and from (Bernhardsson et al. Reference Bernhardsson, Eriksson Stenström, Jönsson, Nilsson and Pédehontaa-Hiaa2023)). Uncertainty bars represent the analytical uncertainties (1 σ). The right y axis displays annual liquid discharges from La Hague (France) and Sellafield (United Kingdom) (OSPAR 2012-2022).
As demonstrated in Figure 5, the marine environment (shells from M. edulis as well as Fucus spp.) at Båteviken has higher F14C in DIC than atmospheric CO2 at the same site (represented by F14C in grass samples). For Båteviken, as for Särdal (Figure 4a), the main cause of this excess is believed to be long-range transport of 14C from La Hague and Sellafield (Eriksson Stenström and Mattson Reference Eriksson Stenström and Mattsson2022; Mattsson et al. Reference Mattsson, Eriksson Stenström, Pédehontaa-Hiaa, Bernhardsson, Jönsson, López-Gutiérrez, Lérida-Toro and Chamizo Calvo2025), but also other sources may occur, such as a contribution from Ringhals NPP (see below).
The peak in the shell data for 2020 may possibly be attributed to a peak in liquid discharges reported from La Hague in the year before (2019), see Figure 5 (OSPAR Commission Reference OSPAR2024). We have previously estimated the transportation time from La Hague to Särdal (site 15) to about 2 years, hence the transport time to Båteviken should be less. At the Särdal site, we have previously also estimated a dilution factor of about 40 for La Hague and a dilution factor of 200 for Sellafield (transport time of about 4 years). Taking these estimates into account for Båteviken, the La Hague discharges would completely dominate compared to the discharges from Sellafield for the M. edulis individuals analyzed in the present study.
The Fucus spp. samples collected on the 17th of May 2024 at Båteviken, Saltbacken and Svinesund, to investigate the potential impact of the nuclear research reactor in Halden, Norway, showed that F14C decreased upstream in the fjord towards the reactor (F14C = 1.070±0.005 at Båteviken, 1.058±0.005 at Saltbacken and 1.038±0.005 at Svinesund). As salinity decreased linearly with the F14C data upstream in the fjord (see Supplementary Material), discharges from the nuclear research reactor in Halden were ruled out as a significant source of 14C on the Swedish west coast.
Bua at Ringhals NPP
The M. edulis (shell annual structures and soft tissue) and F. vesiculosus data for samples from Bua, located close to Ringhals NPP, is clearly influenced by water-borne discharges from the NPP (see Figure 6). Temporal variations in F14C are expected at the site. The mixing conditions of surface water of low salinity from the river Viskan (entering Kattegat <10 km south of the NPP), of low-salinity water from the Baltic Sea (in the surface water, usually flowing north) and of water with higher salinity from the North Atlantic (deeper water, usually flowing south) changes with time. Furthermore, local weather conditions (mainly wind) may result in frequent changes in the direction of the cooling water plume (also carrying the liquid radioactive discharges) from the NPP (Skoglund and Peterson Reference Skoglund and Peterson1988). The liquid 14C release rates from the NPP may also change over time.
F14C in annual structures from shells of Mytilus edulis, in soft tissue of M. edulis and in Fucus vesiculosus collected at Site 11 Bua (close to Ringhals NPP). The average formation season is plotted as summer for the M. edulis shell structures (dark ring forming in winter). The F. vesiculosus and M. edulis soft tissue data are plotted at the collection date. F14C in the terrestrial environment is represented by grass samples from the Bua (Site 11) and Båteviken (Site 1) (plotted at collection date). A trendline of F14C in the terrestrial environment of southern Sweden is also included (data from (Eriksson Stenström et al. Reference Eriksson Stenström, Skog, Bernhardsson, Mattsson, Nielsen, Rundgren, Muscheler, Linderson, Pédehontaa-Hiaa and Rääf2022) and from (Bernhardsson et al. Reference Bernhardsson, Eriksson Stenström, Jönsson, Nilsson and Pédehontaa-Hiaa2023)). Uncertainty bars represent the analytical uncertainties (1 σ).
The expected temporal variations in F14C are clearly reflected in the F. vesiculosus samples. Also the annual structures of the M. edulis individuals show significant changes over the years. For the youngest 3 annual structures (2022, 2021 and 2020, see Figure 6), the two M. edulis shells analyzed display similar (but not identical) values in F14C (shell 2 is 13%, 3% and 6% higher than shell 3). For most of the earlier years, differences in F14C are pronounced for the two shells. The M. edulis individuals may, voluntarily or involuntarily, have changed habitat, or have had different temporal growth patterns. It should be noted that the F. vesiculosus data and the M. edulis data are not completely comparable, as the samples were collected about 150 m apart. Nonetheless, both sites are about 1.2 km from the outlet of cooling water from the NPP.
Another possibility is that the acid pretreatment may result in unrepresentative samples. Varying degrees of material may be removed from the surface and edges of the shell during acid pretreatment. The edges of the cut shell annual pieces represent the winter months. Hence, material from the winter months may be totally lost during the acid treatment. Additionally, if F14C in DIC in the surrounding water varies over the year, depending on fluctuating 14C discharge rates, water currents and dispersion patterns, the removed material has a potential to give rise to an unrepresentative annual average F14C value. The degree of shell material removal (ARF, acid removal fraction) varied between 6% and 29% for the cut annual samples of the two shells from Bua (see Supplementary material). For each year, the absolute difference between the ARFs (ΔARF = abs(ARFshell 2−ARFshell 3)) and the absolute difference between their F14C values (ΔF14C = abs(F14Cshell 2−F14Cshell 3)) have been calculated (see Supplementary material). The largest difference in F14C for the cut annual samples of two shells from Bua is for 2017 (ΔF14C = 0.536±0.009). This year also has the largest difference in weight removal (ΔARF = 23%). The opposite is valid for 2021: equal weight removals (ΔARF = 0%) correspond to a low ΔF14C (0.042±0.008). The data is not conclusive but should be seen as an indication that the acid treatment may skew the results. However, the acid treatment procedure may not be necessary when preceded by the heating step. For most of the years with a large disagreement between samples, one of the shell pieces weighed less than 80 mg, therefore the sample weight could also be a factor of uncertainty. Despite this, since sample weight is related to the annual shell growth it cannot be controlled.
One further observation is that F14C in soft tissue of one the three individuals of M. edulis is significantly higher than the other two, despite being collected at the same site. Changes of habitat, or possibly different dietary conditions, may be the reason. This observation demonstrates that it is suitable to analyze not only one individual but several from the same site (individually or in a pooled sample).
It should also be noted that Ringhals NPP has had two more operational reactors apart from the two pressurized water reactors (PWRs) in current operation. One reactor was closed in 2019 (a PWR) and the other in 2020 (a Boiling Water Reactor, BWR), which implies a reduction in operationally discharged 14C.
As a final note on the data from Bua, the grass samples from this site show no elevated levels of 14C in the terrestrial environment. The site is, however, not in the main downwind direction from the stacks of the NPP. Furthermore, the two operational reactors at Ringhals NPP are PWRs, which are known to release 14C to air mainly as hydrocarbons (thus not directly accessible to vegetation by photosynthesis) (IAEA 2004).
Methodological aspects and estimations of Mytilus edulis ages
In the determination of the age of mussels, growth curves are valuable as a complementary tool to the counting of annual structures. Ideally, individual growth curves from each sampling site would be preferred to ensure that size comparisons are made for as identical growth conditions as possible. As previously mentioned, growth curves assume that there is a maximum size that a species cannot surpass, it is however debated whether this is actually true for M. edulis (Seed Reference Seed, Rhoads and Lutz1980). Seed (Reference Seed, Rhoads and Lutz1980) states that while there is some evidence for asymptotic growth, old mussels from a site with poor conditions will start growing more rapidly if they are placed in more favorable conditions. In the current study, growth curves have mainly been used as a further validation for the total age of the M. edulis individuals collected. Notably, the growth curves were not used to find the location of specific annual structures. Our estimates agree fairly well with the growth curve for M. edulis from Kullen, implying that the true age is indeed close to the age assessed from the identification of annual structures.
The largest challenge when locating the annual structures in the shell of M. edulis is to discern between true annual structures and other disturbance lines. As mentioned above, if a small particle is trapped within the shell it will appear as a darker section, which can resemble an annual band (Neves and Moyer Reference Neves and Moyer1988). Moreover, in contrast, Neves and Moyer (Reference Neves and Moyer1988) noted that annual structures may be missing due to erosion of the fibrous layer. While most erosion will occur in the older region of the shell, by the umbo, particles are more likely to get trapped within the younger regions, towards the posterior edge of the shell. From this it would be natural to assume that the lowest uncertainties about the location of the annual structures would be in the middle region. However, in the acetate peels made during this project the middle region of the shell were often difficult to analyze visually. Hence, in this pilot study, the uncertainties related to the locations of annual structures are most prominent in the middle regions of the M. edulis shells. The quality of the acetate peels could most likely be improved by adjusting and refining the technique specifically for preparation of M. edulis samples.
The acetate peels revealed that if the M. edulis individuals were collected early in the growth season the youngest annual ring was easily missed when externally identifying the annual structures, as it had only grown a few millimeters at this time. Consequentially this led to the shell samples from the youngest part of some shells containing a narrow band of material from the current growth season in addition to the material from the preceding year of growth. The youngest ring in the samples from Bua, which were picked during spring, were originally estimated to be formed a year later than what they actually were. The youngest annual ring of M. edulis samples collected in May is hence assumed to be from the previous year. Practically this means that the data in Figure 6 is shifted backwards by a year, as compared to the year of sample collection. To avoid this issue in the future it would be preferable to collect M. edulis in the middle of winter, before a new growth ring starts forming. As the nacre forms annual layers parallel to the curvature of the shell, starting on the inner surface of the fibrous layer growing towards the mantle, all annual layers of the nacre can be seen in the umbo region. Furthermore, the nacre is less likely than the fibrous layer to be eroded, as the fibrous layer first has to be fully eroded for the nacre to become exposed. Due to this, annual markings are often easier to identify in the nacre compared to the fibrous layer (Lutz Reference Lutz1976). However, the markings in the nacre cannot be used to locate annual structures in the fibrous layer since, as previously mentioned, the annual markings in the two layers do not align. When studying the annual markings in the nacre, as depicted in the acetate peels, several of the M. edulis individuals appear to be between one and two years older than assumed. In conclusion, if a year was missed when dividing the shells according to their external structure, it is difficult to find where it should have been located in the fibrous layer, if the annual marking is only clearly defined in the nacre. In future studies, collecting more shell samples and averaging the result would be preferable.
When re-examining the acetate peels from the Bua 2 and Bua 3 M. edulis shells, 11 and 9 annual structures were seen respectively. For comparison, 9 annual structures for Bua 2 and 8 for Bua 3 were identified during the initial sample inspection and preparation. Unfortunately, the imprint of the fibrous layer in the acetate peels was not detailed enough to identify the location of the structures, resulting in it being impossible to tell which in what part of the shell an annual ring has been missed. This does, however, open up for the possibility that the samples assumed to be from 2017 and 2018 from the Bua 2 shell instead being from the same year as the 2016 and 2017 samples from the Bua 3 shell giving an even higher correlation between the sample data (see Figure 6). If, for instance, an annual structure in the young part of the Bua 2 shell has been missed, the later samples will be assumed to be a from a too early date, thus shifting samples by a year. Another possibility is that the aforementioned samples have been mixed during sample preparation. This is deemed to be unlikely as all samples were carefully labeled during each stage of the preparation.
The maximum age of M. edulis limits the time span that can be investigated. According to the Norwegian institute of marine research, M. edulis can live for 20 years and possibly even longer, and their shells can reach a length of over 10 cm depending on their growth conditions (Havforskningsinstituttet 2024). However, many M. edulis individuals do not survive past the first few years of life (McGrorty and Goss-Custard Reference McGrorty and Goss-Custard1993), and it seems reasonable to expect the age of most samples to be less than 20 years.
An alternative to explore for future studies would be to characterize the shell with high-resolution micromilling techniques, combining 14C analysis with measurement of δ13C and δ18O. The stable isotopes could be used to give information e.g. on water temperature, which could aid to define the annual growth pattern. Time-resolved 14C measurements could reveal discharge patterns from the NPP on a sub-annual basis.
Conclusions
We have evaluated the potential use of annual structures of M. edulis shells to estimate past liquid discharges of 14C in the form of DIC from Ringhals nuclear power plant on the Swedish west coast. The approach has proven to be promising but with some challenges, and a few aspects that need further attention.
One challenge is to identify and extract relevant growth structures from the shells. Annual structures appear in the fibrous layer of the shell—which is suitable for 14C analysis—as well as in the nacre, but as these do not align, the nacre needs to be removed. We have found that heating of the shells is efficient to separate the nacre from the fibrous layer. Another challenge, which we believe introduces the main uncertainty within the method, is connected to locating each annual ring within the shell. False annual structures, as well as absent annual structures, may occur in individual mussels. To minimize such effects, several shell samples from each location should be evaluated, instead of individual shells. The method could also be further developed by improving the quality of the acetate peels, and by making growth curves for each sampling location. A further simplification and improvement of the method would be to exclude the acid treatment after the heating step and after cutting the fibrous layer into annual samples.
An advantage of using M. edulis shells to retrieve past discharge data is that M. edulis is abundant in the area and that they can live up to 20 years. One inherent disadvantage of the approach is that M. edulis is semi-sessile, and some individuals may have moved between various habitats during their lifetime. In two shell samples collected at the reference site Särdal in 1978, we observed that F14C in one of the shells agreed well with F14C in Fucus spp. from the same site for the period 1974 to 1978. For the other Särdal shell, data agreed well for shell and Fucus spp. for 1976–1978, but not for 1972–1975. During these earlier years, this particular individual of M. edulis may have resided elsewhere. In two shells collected in 2022 from another reference site, Båteviken, F14C agreed well for both shells and Fucus spp. for years 2020–2022. The number of samples from reference sites analyzed within this study are too few to be able to estimate the extent of individuals of M. edulis changing habitat, and further investigations are needed. Using several M. edulis shells may circumvent the problem, given that changes in habitat are relatively rare.
In this first study of examining the potential of shells of M. edulis for retrospective analysis of 14C in the marine environment of Ringhals NPP, we have observed significantly higher F14C values in shells and soft tissue of M. edulis from the Ringhals area than from reference sites. Furthermore, F14C varies substantially between years for the Ringhals samples. Despite the uncertainties using shells from M. edulis in environmental studies related to liquid discharges from NPPs, we believe that the approach is valuable when other samples from the past are not available.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2025.10137
Acknowledgment
The project was funded by the Swedish Radiation Safety Authority (SSM), projects SSM2022-4035 and SSM2023-1044-13.
Open access
