Sediments as sinks and sources of marine radionuclides: Implications for their use as ocean tracers

Marine Pollution Bulletin 194 (2023) 115316 Available online 28 July 2023 0025-326X/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Sediments as sinks and sources of marine radionuclides: Implications for their use as ocean tracers Raúl Peri´ a˜ nez a , * , Unai Abascal-Ruíz b , Jos´ e María L´ opez-Guti´ errez c , María Villa-Alfageme b a Dpt Física Aplicada I, ETSIA, Universidad de Sevilla, Ctra Utrera km 1, Sevilla, Spain b Dpt Física Aplicada II, ETSIE, Universidad de Sevilla, Avda Reina Mercedes s/n, Sevilla, Spain c Dpt Física Aplicada I, EPS, Universidad de Sevilla, Spain ARTICLE INFO Keywords: Transport model 236-Uranium 129-Iodine 137-Caesium sediments Atlantic Ocean ABSTRACT A Lagrangian transport model for the North Atlantic has been applied to simulate the historical releases of 137 Cs, 129 I and 236 U from the European nuclear fuel reprocessing plants. Advection by currents, mixing and decay are included, as radionuclide interactions between water, sediments and suspended matter. The model was validated comparing predictions with measured radionuclide concentrations in water and sediments in several areas. 129 I and 236 U signals entering the Arctic Ocean have been compared with the input terms: the 236 U signal is distorted, but the 129 I signal preserves its shape. In the first moments after the releases, the sediments act as sinks for 236 U, but not significantly for 129 I and ultimately they become sources of 236 U to the open sea. This results in a weaker correlation between input and output signals for 236 U than for 129 I. The same effects as for 236 U have been found for 137 Cs signal into the Arctic. 1. Introduction Radionuclide releases from the European nuclear fuel reprocessing plants (NFRPs) Sellafield, UK (SF) and La Hague, France (LH) have been used to trace water mass transport in the North Atlantic Ocean. In particular, 236 U has been used in the last years as a useful and powerful tracer due to its solubility, mainly in combination with 129 I (Shao et al., 2022; Casacuberta et al., 2016, 2018; Castrillejo et al., 2017), under the assumption that both radionuclides present an essentially perfect conservative behaviour in sea water (i.e., they remain completely dissolved without any interaction with suspended matter particles and sediments). The combined use of both radionuclides allows, for example, to determine circulation timescales (Wefing et al., 2021), water mass composition (Wefing et al., 2022) and transport pathways (Wefing et al., 2019). The use of these radionuclides as oceanic tracers requires to examine their geochemical properties when they are released from the reprocessing plants in their particular chemical forms. 129 I and 137 Cs properties in seawater have been evaluated and reviewed extensively [e.g. Inomata and Aoyama, 2022; Qiao et al., 2020 for 137 Cs and Hou et al., 2009 for 129 I]. However, only recently the 236 U geochemical behaviour has been addressed (Castrillejo et al., 2020; Villa et al., 2018). A numerical Lagrangian model which simulates the transport of 137 Cs (Peri´ a˜ nez et al., 2016), 129 I (Villa et al., 2015) and 236 U (Peri´ a˜ nez et al., 2018) in the North Atlantic Ocean, including shelf seas, was developed to simultaneously describe the pathways of the historical releases from SF and LH in the ocean. In Villa et al. (2015) 129 I was modeled as a conservative radionuclide in water, according to its geochemical behaviour described in the available bibliography, and the modeled results agreed well with measurements of dissolved 129 I in the Atlantic Ocean. For 137 Cs, in contrast, Peri´ a˜ nez et al. (2016) included in the model the interactions of 137 Cs with the particulate phase (suspended matter and bed sediments). The model successfully quantified its dissolution from the contaminated coastal sediments and its subsequent transport (Peri´ a˜ nez et al., 2016). It also predicted that a non-negligible amount of 137 Cs would be present in seawater, even after the drastic reduction in the releases of 137 Cs in mid 1980’s. When the model was applied to 236 U in Peri´ a˜ nez et al. (2018), considering a perfectly conservative behaviour, i.e. interactions with sediments were neglected, the simulated concentrations in surface waters of the North Sea were overestimated in relation to 236 U measurements in the area. However, the incorporation of water/sediment interactions to the model allowed to reproduce satisfactorily the measured concentrations of dissolved 236 U. Furthermore, the study predicted that when discharged, a third of the released 236 U would be * Corresponding author. E-mail address: [email protected] (R. Peri´ a˜ nez). Contents lists available at ScienceDirect Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul https://doi.org/10.1016/j.marpolbul.2023.115316 Received 29 March 2023; Received in revised form 10 July 2023; Accepted 17 July 2023 Marine Pollution Bulletin 194 (2023) 115316 2 initially trapped in bed sediments of the coastal areas close to the NFRPs. These radionuclides would be released back to the water column in subsequent years; thus, it was found the 236 U signal entering the Arctic Ocean was distorted with respect to the release signal from the NFRPs. Yet, further investigations are required to better understand the geochemical behaviour of 236 U in the vicinity of the nuclear reprocessing plants and to support and evaluate these findings. The purpose of this work is to underpin the preliminary conclusions described in Peri´ a˜ nez et al. (2018) relative to the role played by sediments in 236 U transport. This will be done obtaining additional validation of the model: water/sediment interactions have been included in the 129 I simulations for the first time; then measured and calculated 129 I, 236 U and 137 Cs concentrations have been compared in sediment samples of the Celtic and Irish Seas. Finally the input of 129 I, 236 U and 137 Cs into the Arctic Ocean have been calculated [only 236 U flux was calculated in Peri´ a˜ nez et al., 2018], which allowed to quantify the differences for these radionuclides between the input terms from the NFRPs and the output signals to the Arctic Ocean. The model is described briefly in Section 2.1 and sampling of sediment cores and measurement methods in Section 2.2. Results are presented and discussed in Section 3. 2. Methods 2.1. Description of the model A Lagrangian transport model was used. It includes advection due to three-dimensional currents, three-dimensional mixing due to turbulence, radioactive decay and exchanges of radionuclides between water and sediments. Exchanges of radionuclides with suspended particulate matter (SPM) and deposition of these particles on the seabed are also described. Water/sediment and water/SPM exchanges are described using a dynamic model based on kinetic transfer coefficients; thus the partition of radionuclides between water and the solid phases may be out from equilibrium conditions. Adsorption and desorption processes are described by the corresponding kinetic rates k1 and k2. Mixing, decay and radionuclide interactions between water and the solid phases (sediments and SPM) are solved using a stochastic method due to the Lagrangian nature of the model. A steady SPM distribution over the model domain was created, as SPM concentrations must be known to evaluate transfer of radionuclides to this phase (details are given in Peri´ a˜ nez et al., 2016). Vertical particle settling velocity is described by Stoke’s law. Water circulation was obtained from JAMSTEC 1 (Japan Agency for Marine-Earth Science and Technology) OFES 2 (Ocean global circulation model For the Earth Simulator) global ocean model. Monthly circulation data was used, as described in Villa et al. (2015). Results of this ocean model were compared with data in several regions of the Earth (including the North Atlantic) in Masumoto et al. (2004). The domain considered in this work expands from 50◦W to 25◦E in longitude and from 45.1◦N to 75.1◦N in latitude (Fig. 1). The model horizontal resolution is 0.1◦with 54 vertical levels. The desorption kinetic rate was fixed as k2=1.16 ×10−5 s −1 for the three radionuclides, which was obtained from some experiments described in Nyffeler et al. (1984). This value has been used in many previous published modelling works [see review in Peri´ a˜ nez et al., 2019b]. The adsorption rate k1, associated to the particle affinity of an element, is deduced from the desorption rate k2 and the radionuclide distribution coefficient kd for ocean waters as described, for instance, in Peri´ a˜ nez et al. (2016, 2018, 2019b). Essentially, the relation holding between these magnitudes is: kd=1 m k1 k2 (1) where m is the sediment concentration. The kd values have been compiled by IAEA (2004) and are presented in Table 1 for the considered elements. This way, the coefficient k1 describes the different geochemical behaviour of each radionuclide. Additional details of the model are included elsewhere (Peri´ a˜ nez et al., 2016, 2018; Villa et al., 2015 and references therein). In addition, the model was applied to simulate releases from Fukushima in the Pacific Ocean (Peri´ a˜ nez et al., 2015, 2019a; Abascal-Ruíz, 2021) and has also participated in model intercomparison exercises organized by the International Atomic Energy Agency (IAEA, 2019). 2.2. Sampling and measurement Two sediment cores from the Celtic Sea (Fig. 1) were collected in 2015 during the DY030 sampling campaign in the new RRS Discovery cruise included in the Shelf Sea Biogeochemistry (SSB), supported by the Natural Environment Research Council of the United Kingdom (Thompson et al., 2017). Core A (51◦13 ′ 59 ″ N; 6◦7 ′ 59 ″ W) was sampled on May the 11th and Core I (50◦36 ′ 0 ″ N; 7◦5 ′ 59 ″ W), five days later, on May the 16th. Core A is 25 cm long and Core I length is 21 cm. Both cores (10 cm diameter) were sliced into segments of different thicknesses: 0.5 cm segments in the most superficial part (top 2 cm), 1 cm segments in the middle (2–5 cm), and 2 cm segments in the deepest layers to the bottom of the core. Radionuclide profiles for both cores are presented in Abascal-Ruíz (2021), Abascal-Ruíz et al. (2023) and L´ erida-Toro et al. (2022). 129 I and 236 U were measured by AMS using the 1 MV Tandem AMS system (HVEE) in Centro Nacional de Aceleradores (CNA), Sevilla, Spain (G´ omez-Guzm´ an et al., 2012; Chamizo et al., 2015). 137 Cs measurement has proceeded in Centro de Investigaci´ on, Tecnología e Innovaci´ on of Sevilla (CITIUS) by gamma spectrometry using an HPGe detector system (Hurtado et al., 2007). 3. Results and discussion The predicted and measured radionuclide concentrations are first presented in Section 3.1. Then, 129 I, 236 U and 137 Cs signals into the Arctic Ocean are evaluated and discussed in Section 3.2. 3.1. 129 I, 137 Cs and 236 U concentrations in European seas The temporal frames of the simulations depend on the availability of release data from Sellafield and La Hague reprocessing plants: 1952–2011 for 137 Cs, 1952–2013 for 236 U and 1966–2012 for 129 I. The 129 I simulated concentrations (including water/sediment interactions) are compared with the results from Villa et al. (2015), where the sediment fraction was not included. Fig. 2 shows averaged 129 I concentrations in the North Sea surface water obtained from the new simulation compared with previous results from Villa et al. (2015) and with observations from the North Sea waters (Michel et al., 2012; Christl et al., 2015a). When the interactions with sediments were included, 129 I concentration in water was only slightly lower than that obtained without sediment interactions. This result is coherent with the weak interaction of 129 I with sediments, since it presents the lowest kd of the three considered radionuclides (Table 1). Both sets of calculations agree with 129 I measured concentrations in sea water, as shown in Fig. 2 (the uncertainties associated to measurements are larger than the differences between both simulations). As mentioned in the introduction, in the case of 236 U there are significant differences (above one order of magnitude) in model results if water/sediment interactions are included or not (Peri´ a˜ nez et al., 2018). Two sediment cores were collected in the Celtic Sea as explained in 1 www.jamstec.go.jp. 2 www.jamstec.go.jp/ofes/ofes.html. R. Peri´ a˜ nez et al. Marine Pollution Bulletin 194 (2023) 115316 3 Section 2.2. Fig. 3 compares measurements of 129 I, 137 Cs and 236 U in the surface layer of the cores (top 5 cm, see below) with simulation results. Water/sediment interactions were also included for 129 I to estimate its concentration in seabed sediments. In order to compare simulation results with measurements, the measured concentrations in the first 5 cm of the core were integrated, since the model considers a 5 cm thick sediment layer. Simulated and measured ratios 129 I/ 236 U, 129 I/ 137 Cs and 137 Cs/ 236 U are also included in Fig. 3. Fig. 3 shows that simulations and measurements are generally in good agreement. The simulation for 137 Cs provides the best results in comparison with measurements. In contrast, 236 U simulation results are slightly overestimated whereas 129 I is underestimated in the two cores. Consequently, 129 I/ 236 U ratio from the simulation is underestimated by slightly more than one order of magnitude. However, overall, the results show that the model is able to describe the adsorption of the three radionuclides by the sediments, in a position which is hundreds of km far from the radionuclide source location, SF NFRP. Note that previous Fig. 1. Model domain. Surface circulation for January 2008 is shown as an example. Only one of each 25 vectors is drawn for more clarity. EGC: East Greenland Current, NAC: North Atlantic Current (Gulf Stream extension), NCC: Norwegian Coastal Current. SF and LH indicate Sellafield and La Hague reprocessing plants and the red open circle the position where the sediment cores mentioned in Section 2.2 were sampled. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Table 1 Dimensionless distribution coefficients in the ocean recommended by IAEA (2004) for the considered elements, which are the values used in this work. Cs I U kd 4×103 7×101 1×103 1970 1980 1990 2000 2010 Year 0 2 4 6 8 10 12 129-I ( 10 6 at/kg) 10 4 With exchanges Without exchanges Fig. 2. Measured (points) and calculated (lines) 129 I concentrations in surface water of the North Sea for the simulations with and without water/sediment interactions. Measurements correspond to the calculated arithmetic mean concentrations and 1 σ standard deviations over the North Sea in 2005 (from data in Michel et al., 2012) and 2009 (from data in Christl et al., 2015a). Fig. 3. Calculated (lines) and measured (points) concentrations of 137 Cs, 129 I and 236 U and their ratios in the two cores of the Celtic Sea. The uncertainty in measurements is within the size of the points. Only one simulation result is shown for each radionuclide because both cores are within the same grid cell of the model. R. Peri´ a˜ nez et al. Marine Pollution Bulletin 194 (2023) 115316 4 results indicate that radionuclides from LH mainly travel to the east in the English Channel, so it is assumed that most radionuclides in the Celtic Sea are coming from SF (Villa et al., 2015; Peri´ a˜ nez et al., 2016, 2018). The discrepancies in 129 I results might be due to the fact that the model only includes the liquid releases from Sellafield and La Hague and does not incorporate 129 I atmospheric emissions from the NFRPs as a secondary source to the ocean (Kadowaki et al., 2020). In Abascal-Ruíz et al. (2023) this contribution was estimated and in the Celtic Sea area 1–8 % of the 129 I concentration in the sediment was assigned to the atmospheric deposition from Sellafield and La Hague. This is in agreement with previous estimations obtained in similar locations (Stewart and Wilkins, 1985). Other reasons for the underestimation could be attributed to uncertainties in the kd, since it is given in IAEA (2004) as a combination for the different iodine species, information not fully detailed in such document. The differences between simulated and measured 236 U concentrations points out the complex geochemical behaviour of uranium when it is released from Sellafield NFRP. The provided kd corresponds to a combination of two main oxidation states: U(VI), highly soluble in water, and reduced U(IV), relatively insoluble. The oxidation state of uranium when it is released from Sellafield pipeline is not known. However, uranium compounds which are insoluble (e.g. U 3 O 8 , UO 2 , U metal) and partially soluble (UO 3 ) are produced and partially discharged (Gregory et al., 2013; IAEA, 2007) after the chemical treatment in the NFRPs. It is possible that a certain amount of 236 U is initially released to the coastal water in insoluble form, U(IV). This could be contributing to the overestimation of 236 U in the simulation. Simulated and measured 129 I and 236 U concentrations have also been compared for seabed sediments of the Irish Sea to have an additional test of the model. Simulated concentrations were averaged over the sea, as in the case of Fig. 2 for the North Sea. Measurements have been compiled from Lee et al. (2008), Marsden et al. (2006), Pham et al. (2008) and Povinec et al. (2002). Experimental results correspond to the inventory in the first 5 cm of sediments, which correspond to approximately 7 years (Marsden et al., 2006; Morris et al., 2001). Fig. 4 shows the experimental and simulation results, where the horizontal uncertainties correspond to the seven year age of the 5 cm layer. There is a reasonable agreement between simulated and measured results. These results confirm the role that sediments play as sink of 236 U in the Irish Sea, in contrast to previous publications showing that the primary form of 236 U in open sea water is, ultimately, soluble U(VI) (Anderson et al., 1989). Christl et al. (2017) described 236 U conservative behaviour, i.e. in a U(VI) oxidation state, in fifty open sea water samples measured in the North sea that displayed a clear correlation between 236 U/ 238 U ratio and 236 U concentration. However, those results correspond to open areas of the North Sea and little is known about the geochemistry of 236 U in the transect from the pipelines (both SF and LH releases travel to the North Sea) to these more open waters. In any case, care should be taken when using such type of relationship between any arbitrary ratio (x/u)and x since spurious correlations result, as clearly pointed out by Stenstr¨ om-Khalili and Håkanson (2009), for instance. Anderson et al. (1989) states that, despite its un-reactive behaviour in the open ocean, uranium might be removed from seawater to sediments of anoxic marine basins. There are several biogeochemical processes described by Anderson et al. (1989) which could cause the increase of uranium in the sediment phase. However, chemical reduction of soluble U(VI) to insoluble U(IV) and subsequent scavenging or interaction with organic sinking particles are not found to be a determining cause. Anderson et al. (1989) found that the main mechanism that precipitates uranium in the sediment happens when dissolved U(VI) is diffused into anoxic sediments where it is reduced and precipitated. Furthermore, note that, similarly to the coastal precipitation of releases from acid mine drainage (Villa et al., 2011), a partial co-precipitation of uranium in an insoluble compound might also occur when the acid releases from the pipeline encounter coastal seawater. The relatively good agreement between 137 Cs, 236 U and 129 I measurements and simulations suggests that the model is successfully describing water/sediment interactions and that sediments play a role in retaining part of the radioactive releases of, mainly, 236 U and 137 Cs. The deviation of 236 U simulation results in the Celtic Sea core might be associated to small discrepancies in the estimation of the average kd (IAEA, 2004), due to variations in the oxidation state of uranium in the coastal area before reaching the open sea and to uncertainties in its oxidation state when it is released from the NFRPs, as mentioned above. Further analysis and measurements are required to better understand the reduction process from U(VI) to U(IV) and determine its causes. In the next section we analyze how water/sediment interactions affect the radionuclide signal out of the Irish and North Sea into the Arctic Ocean and the role of sediments not only as sink, but as a delayed source of radionuclides as well. 3.2. 129 I, 236 U and 137 Cs input signals into the Arctic Ocean The fluxes of 129 I, 236 U and 137 Cs into the Arctic Ocean have been calculated as explained in detail in Peri´ a˜ nez et al. (2018), and results are presented in Fig. 5 for 129 I and 236 U. Essentially, the calculation is done considering that the flux is I=F⋅C, where C is the radionuclide concentration in the Norwegian Coastal Current and F=1.5×107 m 3 /s is the mean flow of such current, obtained from the applied ocean model. These fluxes are compared to the radionuclide releases from Sellafield and La Hague, which are shown in Fig. 6. 129 I releases were compiled from L´ opez-Guti´ errez et al. (2004), Areva (2014) and Sellafield (2014). 236 U releases were adapted from Christl et al. (2015b), as explained in detail in Peri´ a˜ nez et al. (2018). The black lines indicate the total (Sellafield plus La Hague) releases. If a radionuclide is perfectly conservative in water, in both open and coastal waters, a correlation between the input signal (radionuclide release into the ocean) and the output signal (measured concentration somewhere downstream), is expected; with a time-delay related to the transit time between the release location and the measurement location. In our case the output is the flux to the Arctic Ocean, displayed in Fig. 5, since these fluxes are proportional to the 129 I and 236 U concentrations in the Norwegian Coastal Current, and they represent the export of the radionuclides to the Arctic Ocean, as explained above. Fig. 4. Calculated (lines) and measured (points) [Lee et al., 2008; Marsden et al., 2006; Pham et al., 2008; Povinec et al., 2002] concentrations of 236 U and 129 I in bed sediments of the Irish Sea. The horizontal uncertainties correspond to the seven years age of the 5 cm layer and the uncertainties in the measured concentrations are within the size of the points. Note that the calculated concentrations correspond to averages over the Irish Sea, while measurements for specific points are provided. R. Peri´ a˜ nez et al. Marine Pollution Bulletin 194 (2023) 115316 5 A comparison between Fig. 6 and Fig. 5 shows that the shapes of the release and export curves in the case of 129 I are fairly similar. However, this is not the case for 236 U, where the peaks in releases in the 80 s and 90 s are almost completely lost, leading to a relatively smooth export flux signal. Indeed, the export signal for this radionuclide is quite similar to that of 129 I (compare both panels in Fig. 5). This weaker correlation for 236 U signals can be quantified in Fig. 7, where the best linear correlations between export signal (to the Arctic) and release signal (from NFRPs) are presented. Best correlations are obtained with delay times of 7 years and 6 years for 236 U and 129 I respectively. Transit time for 236 U is longer than for 129 I. This is in agreement with a less conservative behaviour of uranium than iodine in water since the mobility of a radionuclide, and hence its transit time between two locations, decreases as its kd increases [this can be clearly seen, for instance, in the results presented in Peri´ a˜ nez (2021, 2022). Finally, the more conservative behaviour of 129 I in relation to 236 U can quantified through the correlation between signals. Thus, 129 I signals present a significantly better correlation (r2=0.9563) than the 236 U signals (r2=0.7827). Our model is able to correctly simulate the transport of 236 U and 129 I in the North Atlantic, predicting that, according to simulated and measured concentrations in water and sediments, 129 I is a more conservative tracer in both coastal and open seas. The potential sedimentation of uranium near the coast should be considered when 236 U is used as a tracer of ocean circulation, since its interactions with seabed sediments cannot be totally neglected. It is important to consider, when using the 236 U signal from the NFRPs into the Arctic, that the interactions with the sediment phase reduce the correlations between release and export (flux into the Arctic) signals, since 236 U is deposited in sediments and later it is slowly released back to the water column. We present here evidences that sediments would behave as a long-term delayed source of 236 U. This effect is already known to occur in the case of 137 Cs in the Irish Sea after the reduction of Sellafield releases (Hunt et al., 2013; Ray et al., 2020 among many others) and also predicted for Fukushima releases in the Pacific Ocean coastal waters after the 2011 accident (Peri´ a˜ nez et al., 2021; Peri´ a˜ nez, 2022). We have calculated the export to the Arctic also for 137 Cs and results confirm what is outlined above. A comparison between releases from SF and LH NFRPs and the export to the Arctic for this radionuclide is presented in Fig. 8. The peak in releases (which are dominated by SF NFRP) in the 70 s is clearly detected in the flux some years later. However, although SF releases abruptly decrease in the 80 s in several orders of magnitude, this quick reduction is not so apparent in the flux, which diminishes slower than releases do. This is most likely due to the remobilisation of 137 Cs from previously contaminated sediments, now acting as a delayed radionuclide source. The best correlation between release and flux (obtained with a 8 year delay time) is presented in Fig. 9. This correlation is weaker than for 129 I and similar to that obtained for 236 U (Fig. 7). Actually, the correlation for 137 Cs is 1970 1980 1990 2000 2010 0 2 4 6 atoms / year 10 27 129-I fl ux 1950 1960 1970 1980 1990 2000 2010 Year 0 2 4 6 atoms / year 10 24 236-U flux Fig. 5. Calculated fluxes of 236 U and 129 I into the Arctic Ocean. 1960 1970 1980 1990 2000 2010 0 100 200 300 4 00 129-I ( kg/yr ) Sellafield La Hague Total 1960 1970 1980 1990 2000 2010 Year 0 1 2 3 4 236-U (kg/yr) Sellafield La Hague Total Fig. 6. Releases of 236 U and 129 I from Sellafield and La Hague. 0 200 400 Release (kg/y) 0 2 4 6 Export ( atoms/year ) 10 27 129-I r²=0.9563 024 Release (kg/y) 0 2 4 610 24 236-U r²=0.7827 Fig. 7. Correlations between the exports to the Arctic Ocean and releases from Sellafield and La Hague. Fig. 8. 137 Cs releases from the NFRPs and calculated exports to the Arctic Ocean for this radionuclide. Releases are compiled from Areva (2014), Sellafield (2014) and Povinec et al. (2003). R. Peri´ a˜ nez et al. Marine Pollution Bulletin 194 (2023) 115316 6 slightly worse than for 236 U, in agreement with the slightly higher kd of Cs than of U (Table 1). Sakaguchi et al. (2012) detected 236 U in seabed sediments of the Japan Sea at water depths larger than 2500 m. Sakaguchi et al. (2012) argue that this fact could compromise the use of this radionuclide as a conservative tracer, depending on the relative importance of 236 U deposition on sediments with respect to the inventory in the water column. In fact, the deposited inventory could be neglected in comparison with the dissolved inventory in these deep waters of the Japan Sea, but this may not be the general situation. Adsorption of radionuclides in sediments is higher in shallow waters (where direct adsorption from water is higher than in deep waters) and can also be in coastal waters (where suspended matter concentrations are generally larger due to river supply and coastal runoff) and thus deposition of radionuclides is enhanced (Villa et al., 2011; Marsden et al., 2006). The European Shelf Seas seem to be a clear example of these cases, where a significant fraction of the released 236 U (and also 137 Cs) is trapped in the seabed. 4. Conclusions A Lagrangian model which simulates the transport of radionuclides in the North Atlantic has been applied to 137 Cs, 129 I and 236 U. The model includes advection by three dimensional currents, three dimensional turbulent mixing, radioactive decay and interactions of radionuclides between water, bed sediments and suspended matter. The last three processes are solved using stochastic techniques. Water circulation was obtained from a global ocean circulation model. The model, applied to simulate the transport of historical releases of such radionuclides from the European NFRPs, produces radionuclide concentrations generally in agreement with measurements in water and sediments. The main conclusion of this work is that 236 U interactions with seabed sediments cannot be totally neglected, at least in shallow waters (e.g., the European Shelf Seas, next to the NFRPs; but this effect is likely to occur in shallow regions over the world), if this radionuclide is used as an ocean tracer. These interactions reduce the correlation between release and export signals at two locations since particle reactive radionuclides are fixed in sediments and they are later released back to the water column. Thus, the sediments turn into an additional source of 236 U. The same conclusions are found in the case of 137 Cs, although this radionuclide has been extensively studied in the past. Interactions with sediments are much weaker in the case of 129 I and consequently good release-export correlations are found; thus this radionuclide seems a suitable ocean tracer in the region under study. CRediT authorship contribution statement Raúl Peri´ a˜ nez: Conceptualization, Methodology, Formal analysis, Writing – original draft. Unai Abascal-Ruíz: Methodology, Writing – review & editing. Jos´ e María L´ opez-Guti´ errez: Funding acquisition, Supervision. María Villa-Alfageme: Funding acquisition, Supervision, Methodology, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. Acknowledgement This work was partially supported by project TRACECARBON (PAIDI-Junta de Andalucía-FEDER), P20_01217. References Abascal-Ruíz, U., 2021. Uso reciente de 129 I y 137 Cs como trazadores en oceanografía: estudios de dos casos. Master Thesis. Universidad de Sevilla (in Spanish). Abascal-Ruíz, U., L´ opez-Guti´ errez, J.M., L´ erida-Toro, V., Chamizo, E., Klar, J., Hicks, N., Casacuberta, N., Christl, M., Hurtado, S., Villa-Alfageme, M., 2023. Transport and Accumulation of Artificial Radionuclides in a Marine Core From the Celtic Sea. In preparation. Anderson, R.F., Fleisher, Martin Q., LeHuray, Anne P., 1989. Concentration, oxidation state, and particulate flux of uranium in the Black Sea. Geochim. Cosmochim. Acta 53, 2215–2224. Areva, 2014. http://www.areva.com/EN/operations-2315/cumulative-release-re sults-report-for-the-areva-la-hague-plant.html. Casacuberta, N., Masqu´ e, P., Henderson, G., Bauch, D., Vockenhuber, C., Daraoui, A., Walther, C., Synal, H., Christl, M., Rutgers-van-der-Loeff, M., 2016. First 236 U data from the Arctic Ocean and use of 236 U/ 238 U and 129 I/ 236 U as a new dual tracer. Earth Planet. Sci. Lett. 440, 127–134. Casacuberta, N., Christl, M., Vockenhuber, C., Wefing, A.M., Wacker, L., Masqu´ e, P., Synal, H.A., van der Loeff, M.R., 2018. Tracing the three Atlantic branches entering the Arctic Ocean with 129 I and 236 U. J. Geophys. Res. Oceans 123, 6909–6921. Castrillejo, M., Casacuberta, N., Christl, M., Garcia-Orellana, J., Vockenhuber, C., Synal, H.A., Masque, P., 2017. Anthropogenic 236 U and 129 I in the Mediterranean Sea: first comprehensive distribution and constrain of their sources. Sci. Total Environ. 593–594, 745–759. Castrillejo, M., Witbaard, R., Casacuberta, N., Richardson, C.A., Dekker, R., Synal, H.A., Christl, M., 2020. Unravelling 5 decades of anthropogenic 236 U discharge from nuclear reprocessing plants. Sci. Total Environ. 717, 137094. Chamizo, E., L´ opez-Lora, M., Villa, M., Casacuberta, N., L´ opez-Guti´ errez, J.M., Pham, M. K., 2015. Analysis of 236 U and plutonium isotopes, 239,240 Pu, on the 1MV AMS system at the Centro Nacional de Aceleradores, as a potential tool in oceanography. Nucl. Inst. Methods Phys. Res. B 361, 535–540. Christl, M., Casacuberta, N., Lachner, J., Maxeiner, S., Vockenhuber, C., Synal, H.A., Goroncy, I., Herrmann, J., Daraoui, A., Walther, C., Michel, R., 2015a. Status of 236 U analysis at ETH Zurich and the distribution of 236 U and 129 I in the North Sea in 2009. Nucl. Inst. Methods Phys. Res. B 361, 510–516. Christl, M., Casacuberta, N., Vockenhuber, C., Els´ asser, C., Bailly du Bois, P., Herrmann, J., Synal, H.A., 2015b. Reconstruction of the 236 U input function for the Northeast Atlantic Ocean: implications for 129 I/ 236 U and 236 U/ 238 U-based tracer ages. J. Geophys. Res. Oceans 120, 7282–7299. Christl, M., Casacuberta, N., Lachner, J., Herrmann, J., Synal, H., 2017. Environ. Sci. Technol. 51, 12146–12153. G´ omez-Guzm´ an, J.M., L´ opez-Guti´ errez, J.M., Pinto-G´ omez, A.R., Holm, E., 2012. 129 I measurements on the 1MV AMS facility at the Centro Nacional de Aceleradores (CNA, Spain). Appl. Radiat. Isot. 70 (1), 263–268. Gregory, C., Liljenzin, J.O., Rydberg, J., Ekberg, C., 2013. The nuclear fuel cycle. In: Radiochemistry and Nuclear Chemistry, Fourth edition, pp. 685–751. 0 0.5 1 1.5 2 Release (kg/y) 0 2 4 6 8 10 12 14 16 18 Export (at/year) 10 23 137- C s r²=0.6892 Fig. 9. Correlation between the exports to the Arctic Ocean (output signal) and releases from Sellafield and La Hague (input signal) for 137 Cs. R. Peri´ a˜ nez et al. Marine Pollution Bulletin 194 (2023) 115316 7 Hou, X., Hansen, V., Aldahan, A., Possnert, G., Lind, O.C., Lujaniene, G., 2009. A review on speciation of iodine-129 in the environmental and biological samples. Anal. Chim. Acta 632, 181–196. Hunt, J., Leonard, K., Hughes, L., 2013. Artificial radionuclides in the Irish Sea from Sellafield: remobilisation revisited. J. Radiol. Prot. 33, 261–279. Hurtado, S., Villa, M., Manj´ on, G., García-Tenorio, R., 2007. A self-sufficient and general method for self-absorption correction in gamma-ray spectrometry using GEANT4. Nucl. Instrum. Methods Phys. Res. Sect. A 580 (1), 234–237. IAEA, 2004. Sediment Distribution Coefficients and Concentration Factors for Biota in the Marine Environment. In: Technical Reports Series, vol. 422. Vienna. IAEA, 2007. Management of Reprocessed Uranium Current Status and Future Prospects. In: IAEA-TECDOC-1529. Vienna. IAEA, 2019. Modelling of Marine Dispersion and Transfer of Radionuclides Accidentally Released from Land Based Facilities. In: IAEA-TECDOC-1876. Vienna. Inomata, Y., Aoyama, M., 2022. Evaluating the transport of surface seawater from 1956 to 2021 using 137 Cs deposited in the global ocean as a chemical tracer. Earth Syst. Sci. Data Discuss. 1–108. Kadowaki, M., Terada, H., Nagai, H., 2020. Global budget of atmospheric 129 I during 2007–2010 estimated by a chemical transport model: GEARN-FDM. Atmos. Environ. 8, 100098. Lee, S.H., Povinec, P.P., Wyse, E., Hotchkis, M.A.C., 2008. Ultra-low-level determination of 236 U in IAEA marine reference materials by ICPMS and AMS. Appl. Radiat. Isot. 66 (6–7), 823–828. L´ erida-Toro, V., Abascal, U., Villa, M., Klar, J., Hicks, N., L´ opez-Guti´ errez, J.M., 2022. 129 I in sediment cores from the Celtic Sea by AMS through a microwave digestion process. Nucl. Instrum. Methods Phys. Res., Sect. B 529, 61–67. L´ opez-Guti´ errez, J.M., García-Le´ on, M., Schnabel, Ch., Suter, M., Synal, H.A., Szidat, S., García-Tenorio, R., 2004. Relative influence of 129 I sources in a sediment core from the Kattegat area. Sci. Total Environ. 323, 195–210. Marsden, O.J., Abrahamsen, L., Bryan, N.D., Day, J.P., Fifield, L.K., Gent, C., Goodall, P. S., Morris, K., Livens, F.R., 2006. Transport and accumulation of actinide elements in the near-shore environment: field and modelling studies. Sedimentology 53 (1), 237–248. Masumoto, Y., Sasaki, H., Kagimoto, T., Komori, N., Ishida, A., Sasai, Y., Miyama, T., Motoi, T., Mitsudera, H., Takahashi, K., Sakuma, H., Yagamata, T., 2004. A fifty-year eddy-resolving simulation of the World Ocean — preliminary outcomes of OFES (OGCM for the Earth Simulator). J. Earth Simul. 1, 36–56. Michel, R., Daraoui, A., Gorny, M., Jakob, D., Sachse, R., Tosch, L., Nies, H., Goroncy, I., Herrmann, J., Synal, H.A., Stocker, M., Alfimov, V., 2012. Iodine-129 and iodine-127 in European seawaters and in precipitation from Northern Germany. Sci. Total Environ. 419, 151–169. Morris, K., Bryan, N.D., Livens, F.R., 2001. Plutonium solubility in sediment pore waters. J. Environ. Radioact. 56, 259–267. Nyffeler, U.P., Li, Y.H., Santschi, P.H., 1984. A kinetic approach to describe trace element distribution between particles and solution in natural aquatic systems. Geochim. Cosmochim. Acta 48, 1513–1522. Peri´ a˜ nez, R., 2021. APERTRACK: a particle-tracking model to simulate radionuclide transport in the Arabian/Persian Gulf. Prog. Nucl. Energy 142, 103998. Peri´ a˜ nez, R., 2022. Viewpoint on the integration of geochemical processes into tracer transport models for the marine environment. Geosciences 12, 152. Peri´ a˜ nez, R., Brovchenko, I., Duffa, C., Jung, K.T., Kobayashi, T., Lamego, F., Maderich, V., Min, B.I., Nies, H., Osvath, I., Psaltaki, M., Suh, K.S., 2015. A new comparison of marine dispersion model performances for Fukushima Dai-ichi releases in the frame of IAEA MODARIA program. J. Environ. Radioact. 150, 247–269. Peri´ a˜ nez, R., Suh, Kyung-Suk, Min, Byung-Il, 2016. The behaviour of 137 Cs in the North Atlantic Ocean assessed from numerical modelling: releases from nuclear fuel reprocessing factories, redissolution from contaminated sediments and leakage from dumped nuclear wastes. Mar. Pollut. Bull. 113, 343–361. Peri´ a˜ nez, R., Suh, Kyung-Suk, Min, Byung-Il, Villa, M., 2018. The behaviour of 236 U in the North Atlantic Ocean assessed from numerical modelling: a new evaluation of the input function into the Arctic. Sci. Total Environ. 626, 255–263. Peri´ a˜ nez, R., Bezhenar, R., Brovchenko, I., Jung, K.T., Kamidara, Y., Kim, K.O., Kobayashi, T., Liptak, L., Maderich, V., Min, B.I., Suh, K.S., 2019a. Fukushima 137 Cs releases dispersion modelling over the Pacific Ocean. Comparisons of models with water, sediment and biota data. J. Environ. Radioact. 198, 50–63. Peri´ a˜ nez, R., Bezhenar, R., Brovchenko, I., Duffa, C., Iosjpe, M., Jung, K.T., Kobayashi, T., Liptak, L., Little, A., Maderich, V., Min, B.I., Nies, H., Osvath, I., Suh, K.S., de With, G., 2019b. Marine radionuclide transport modelling: recent developments, problems and challenges. Environ. Model. Softw. 122, 104523. Peri´ a˜ nez, R., Qiao, Fangli, Zhao, Chang, de With, Govert, Jung, Kyung-Tae, Sangmanee, Chalermrat, Wang, Guansuo, Xia, Changshui, Zhang, Min, 2021. Opening Fukushima floodgates: modelling 137 Cs impact in marine biota. Mar. Pollut. Bull. 170, 112645. Pham, M.K., Sanchez-Cabeza, J.A., Povinec, P.P., Andor, K., Arnold, D., Benmansour, M., Bikit, I., Carvalho, F.P., Dimitrova, K., Edrev, Z.H., Engeler, C., Fouche, F.J., Garcia- Orellana, J., Gasco, C., Gastaud, J., Gudelis, A., Hancock, G., 2008. A new Certified Reference Material for radionuclides in Irish sea sediment (IAEA-385). Appl. Radiat. Isot. 66 (11), 1711–1717. Povinec, P.P., Badie, C., Baeza, A., 2002. Certified reference material for radionuclides in seawater IAEA-381 (Irish Sea Water). J. Radioanal. Nucl. Chem. 251, 369–374. Povinec, P.P., Bailly du Bois, P., Kershaw, P.J., Nies, H., Scotto, P., 2003. Temporal and spatial trends in the distribution of 137 Cs in surface waters of Northern European Seas — a record of 40 years of investigations. Deep Sea Res. Pt. II 50, 2785–2801. Qiao, J., Andersson, K., Nielsen, S., 2020. A 40-year marine record of 137 Cs and 99 Tc transported into the Danish Straits: significance for oceanic tracer studies. Chemosphere 244, 125595. Ray, D., Leary, P., Livens, F., Gray, N., Morris, K., Law, K.A., Fuller, A.J., Abrahamsen- Mills, L., Howe, J., Tierney, K., Muir, G., Law, G.T.W., 2020. Controls on anthropogenic radionuclide distribution in the Sellafield-impacted Eastern Irish Sea. Sci. Total Environ. 743, 140765. Sakaguchi, A., Kadokura, A., Steier, P., Takahashi, Y., Shizuma, K., Hoshi, M., Nakakuki, T., Yamamoto, M., 2012. Uranium-236 as a new oceanic tracer: a first depth profile in the Japan Sea and comparison with caesium-137. Earth Planet. Sci. Lett. 333–334, 165–170. Sellafield, 2014. http://sustainability.sellafieldsites.com/environment/environment-pag e/annual-discharge-monitoring-reports/. Shao, Y., Zhang, J., Luo, M., Xu, D., Ma, L., 2022. A review of anthropogenic radionuclide 236 U: environmental application and analytical advances. J. Environ. Radioact. 251–252, 106944. Stenstr¨ om-Khalili, M.I., Håkanson, L., 2009. Variations and spurious correlations related to DIN, DIP, TN, TP, DIN/DIP and TN/TP in lakes, rivers, and marine systems. Open Mar. Biol. J. 3, 21–28. Stewart, S.P., Wilkins, B.T., 1985. Areal distribution of 129 I in west cumbrian solis. J. Environ. Radioact. 2, 175–182. Thompson, C.E.L., Silburn, B., Williams, M.E., et al., 2017. An approach for the identification of exemplar sites for scaling up targeted field observations of benthic biogeochemistry in heterogeneous environments. Biogeochemistry 135, 1–34. Villa, M., Manj´ on, G., Hurtado, S., García-Tenorio, R., 2011. Uranium pollution in an estuary affected by pyrite acid mine drainage and releases of naturally occurring radioactive materials. Mar. Pollut. Bull. 62, 1521–1529. Villa, M., L´ opez-Guti´ errez, J.M., Suh, Kyung-Suk, Min, Byung-Il, Peri´ a˜ nez, R., 2015. The behaviour of 129 I released from nuclear fuel reprocessing factories in the North Atlantic Ocean and transport to the Arctic assessed from numerical modelling. Mar. Pollut. Bull. 90, 15–24. Villa, M., Chamizo, E., Santos-Ar´ evalo, F.J., L´ opez-Guti´ errez, J.M., G´ omez-Martínez, I., Hurtado-Bermúdez, S., 2018. Natural and artificial radionuclides in a marine core. First results of 236 U in North Atlantic Ocean sediments. J. Environ. Radioact. 186, 152–160. Wefing, A.M., Christl, M., Vockenhuber, C., Rutgers van der Loeff, M., Casacuberta, N., 2019. Tracing Atlantic waters using 129 I and 236 U in the Fram Strait in 2016. J. Geophys. Res. Oceans 124, 882–896. Wefing, A.M., Casacuberta, N., Christl, M., Gruber, N., Smith, J.N., 2021. Circulation timescales of Atlantic Water in the Arctic Ocean determined from anthropogenic radionuclides. Ocean Sci. 17, 111–129. Wefing, A.M., Casacuberta, N., Christl, M., Dodd, P.A., 2022. Water mass composition in Fram Strait determined from the combination of 129 I and 236 U: changes between 2016, 2018, and 2019. Front. Mar. Sci. 9, 973507. R. Peri´ a˜ nez et al.