Introduction

Atmospheric chemistry shows that iodine plays a significant role in the depletion of ozone and aerosol particles for cloud nucleation1,2. Thus, considerable attention has been paid to understanding sources of inventory and distribution of iodine in natural environments. Oceans represent the main source of iodine to the Earth's surface environments and it is apparent that tracing the chemical species of iodine in ocean water provides further clue for incorporation in the atmosphere. As most of iodine atmospheric interaction is strongly linked to releases from ocean surface water, identifying iodine species in this compartment of the ocean is becoming important. Iodate is the predominant species in oxic ocean water with a comparable iodide concentration especially in the tropical surface waters. Data from the Atlantic and Pacific Oceans reveal occurrence of minimum iodate concentrations along the equatorial regions that increase polar wards3,4.

The addition of anthropogenic long-lived radioactive isotope of iodine, 129I (T1/2 = 15.7 Myr), to the marine environment provided another interesting tracer that can be used to shed further light on the 127I distribution and its geochemical behavior in ocean surface water. Huge amount of 129I has been released to the environment as a result of human nuclear activities since the 1940s. The nuclear fuel reprocessing facilities (NRFs) at Sellafield (UK) and La Hague (France), have together released >6000 kg of 129I both to the atmosphere and marine environment until 2010. The discharge rate of 129I from these two facilities increased and remained high after 1990 and peaked in 1997, but declined slightly during the past 10 years. Although concerns about 129I hazardous environmental impacts are there, the releases from Sellafield and La Hague also provided an ideal source point oceanographic tracer to investigate transport and circulation of water masses in specific areas.

Several studies show that the discharges of 129I from the English Channel and the Irish Sea are transported into the North Atlantic via the Norwegian Coastal Current (NCC) and the Arctic Ocean5,6,7. A twofold increase of 129I concentration was observed in the central Arctic since 19967 and has been used to trace Polar Water transport into the Atlantic Ocean8. However, most of these studies have focused on the Nordic Seas and the Arctic region and there is little information about concentration and/or inventory of 129I in other parts of the North Atlantic Ocean. He et al.9 summarized all available 129I data in the ocean's surface water and pointed out that there are many shortcomings in the data sets which need to be completed in future work.

Speciation analyses of 129I represent a new powerful tool for gaining detailed information on identifying redox rate of I−/IO3− pair in environmental conditions as well as sources and exchanges of seawater masses. This potential has been successfully applied in the North Sea and the Baltic Sea10,11,12. However, at present we have no idea on distribution and conversion patterns of 129I species in the open sea environment. In addition, to our best knowledge, there is no clear conclusions on whether 129I that originated from Sellafield and La Hague has influenced the middle latitudes of the North Atlantic Ocean or even more remote regions (i.e. South Atlantic Ocean and Southern Ocean), though evidence revealed elevated 129I/127I ratio in the Deep Western Boundary Current13,14. Orre et al.15 simulated the spread of 129I contamination from the west of the North Atlantic Ocean to the east over a 10 years period using Ocean General Circulation Model (OGCM). They concluded a possible transport of high 129I-labelled surface water from the Labrador Sea that can influence the eastern part of the Atlantic Ocean. This hypothesis has, however, not yet been supported and confirmed by experimental observations. In the study presented here, we provide first data on iodine isotopes (129I and 127I) and their species (iodate and iodide) in the northeastern Atlantic Ocean (Fig. 1), aiming at identifying effects of environmental conditions, sources and pathways on 129I distribution in this region.

Figure 1
figure 1

General bathymetric chart of the investigated area in the North Atlantic Ocean.

(a) Sampling transect of 129I along the northeastern Atlantic Ocean. Sampling locations are expressed as blank dots. Two European nuclear reprocessing facilities (NRFs) Sellafield and La Hague, are highlighted with stars. (b) General scheme of surface water circulation in the region of Bay of Biscay and (c) the Gulf of Cadiz. NAC = North Atlantic Current, IPC = Iberian Poleward Current, PC = Portuguese Current, ACN = Azores Current, northern branch, ACS = Azores Current, southern branch and CC = Canary Current. The bathymetry and coastal lines were created using Ocean Data View (ODV 4.5.3).

Full size image

Results

Distribution of total iodine and its species (127I− and 127IO3−)

The concentrations of total iodine (127I) and its species (127I− and 127IO3−) in the sampled surface water of the North Atlantic Ocean are shown in Fig. 2a and listed in Supplementary Table S1. In general, surface concentrations of total iodine and its species (iodate and iodide), show a fairly small variation in the samples from the break of Celtic Slope to the east of Madeira Archipelago. It seems that 127I and its species, iodate and iodide, show rather constant values in the surface water of the open sea, with mean concentrations of 0.39, 0.35 and 0.05 μM, respectively. The range of iodide is 0.02 to 0.08 μM and the lowest concentration of iodide occurred along the shelf break of the Celtic Sea (sample 1) while the two pronounced peaks (sample 6 and 14) appear in the north and middle parts of the Portugal coastal region. Unlike the iodide, the concentrations of total iodine is positively correlated with iodate (r2 = 0.5), which suggests rather similar distribution trend for both 127I and 127IO3−. The highest concentrations are in the area off the Cape St Vicente, whereas the minimum values occurred in the northeast of Madeira. More than 75% of 127I-iodate is observed in these surface waters, which is comparable to seawaters collected in any other ocean regions (Supplementary Table S2).

Figure 2
figure 2

Iodine isotopes (127I and 129I) and their species in the Atlantic seawaters.

(a) Concentration of total iodine (127I) and its species (127I− and 127IO3−) along the sampled transect in the northeastern Atlantic Ocean. (b) Variations of isotopic ratio of 129I/127I, 127I−/129I− and 127IO3−/129IO3− in the sampled transect along the northeastern Atlantic Ocean. (c) Concentration of 129I and its species (129I− and 129IO3−) along the sampled transect in the northeastern Atlantic Ocean.

Full size image

Distribution of 129I and its species (129I− and 129IO3−)

Data on 129I and its species (129I− and 129IO3−) in the surface water samples are listed in Supplementary Table S1. It is apparent that, as a sensitive radioactive tracer, the concentration of 129I (including 129I− and 129IO3−) shows a larger variation in the seawater compared with that of 127I (Fig. 2c). More than one order of magnitude difference in 129I concentration is shown in the studied region, with an average of 1.53 × 108 atoms/L. The highest concentration occurred in the middle of the Biscay Bay (12.67 × 108 atoms/L) and all the 129I data here are higher than 4.0 × 107 atoms/L. Five pronounced peaks of 129I (>2 × 108 atoms/L) are observed along the surface water transect. However, if these distinct peaks are not taken into consideration, concentration of 129I in the sampled transect is less than 1 × 108 atoms/L (mean 6.9 × 107 atoms/L) and also does not vary a lot in the open sea surface water. Despite nearly similar behavior for 129I− and 129IO3− in the sampled surface water, the concentrations of 129I− show a relatively wider range (0.07–5.73 × 108 atoms/L). Additionally, 129IO3− concentrations are normally higher than that of 129I− in the sampled region with a few exceptions. The isotopic ratio of 129I/127I has the same trend as with that of 129I and varies between 1.82 × 10−10 and 5.69 × 10−9 with an average value of 6.63 × 10−10 (Fig. 2b). Both 129I and 129I/127I values are 2–4 orders of magnitude higher than the natural levels in the ocean (estimated as ~105 atoms/L for 129I and ~10−12 for 129I/127I). Similarly, there are five peaks found at sampling locations 3, 8, 17, 24 and 33. The same pattern also appears for the ratios of 129I−/127I− and 129IO3−/127IO3− and that the atomic ratio of 129I/127I for iodide is significantly higher than that for iodate in the investigated region.

Distribution of 129I−/129IO3− and 127I−/127IO3−

The ratios of 129I−/129IO3− are generally higher than 127I−/127IO3− (Fig. 3c). This discrepancy between abundance of iodine species for 127I and 129I likely reflects different sources of the isotopes. Some 129I−/129IO3− values, however, are close to the corresponding 127I−/127IO3− levels which are expected to reflect a local effect. This may either be related to older water ventilated to the surface during winter, or the small-scale accelerated redox cycle induced by local chemical or biological characters in these waters. The 129I−/129IO3− value varies from 0.14 to 2.02 in the samples where the highest level occurs in the region between Madeira and Africa coast (sample 32). Most of the ratios for 129I−/129IO3− lie at 0.3–0.7 while the 127I−/127IO3− values are normally below 0.2. High values of 129I−/129IO3− (above 0.7) are thought to be directly linked to the large-scale circulation pattern as described in the next section. Similar to 129I and 129I/127I, there are five peaks observed for 129I−/129IO3− and the intensity of the peaks generally decreases towards the north. Unlike the 129I−/129IO3−, the ratios of iodide to iodate for 127I remain nearly constant in the sampled surface water transect. On average the value for 127I−/127IO3− is 0.14, which is a factor of five lower than that of 129I−/129IO3− (0.65).

Figure 3
figure 3

Ratios of Iodine species in the Atlantic seawaters.

(a) Variations of 127I−/(127IO3− + 127I−) as a function of latitude along the sampled transect in the northeastern Atlantic Ocean. Linear fit shows a significant increase of 127I− proportion southwards (r = −0.61; P < 0.005). (b) Negative correlation between 129I−/(129IO3− + 129I−) ratio and latitude towards south along the sampled transect of the northeastern Atlantic Ocean (r = −0.39; P < 0.05). Clusters in blue and red that separated by 40°N may attribute to different sources of 129I (Mediterranean source and English Channel source, respectively). (c) Variations of 129I−/129IO3− and 127I−/127IO3− along the sampled transect of the northeastern Atlantic Ocean. Colored zones represent 129I species in samples that is close to (<0.3, cyan), approaching (0.3–0.7, gray) and far from reaching (>0.7, white) the iodine redox equilibrium in seawaters.

Full size image

Discussion

It is well established that marine discharges from Sellafield and La Hague (NRFs) are the main source of 129I in the North Sea, Baltic Sea, Nordic Seas and the Arctic Ocean8,16,17. Whether the discharges have influenced the northeastern Atlantic Ocean and, if so, on what magnitude was not fully investigated. The natural 129I/127I of the pre-nuclear era is estimated at 1.5 × 10−12 based on analysis of marine sediment cores18 and consequently the inventory of natural 129I in the world ocean is calculated to be 100 kg using average iodine concentration of 0.4 μM from our Atlantic data presented here. About 90 kg of 129I were released to the environment by atmospheric nuclear weapons tests in the 1950s and 1960s which raised the 129I concentration to ~107 (atoms/L) and 129I/127I to ~10−10 (atoms/atoms)19 and account for about 10% of the concentration found in our samples. This finding indicates that the 129I in surface water of the northeastern Atlantic Ocean had been labeled by mainly marine and/or atmospheric discharges from the nuclear fuel reprocessing.

Atmospheric 129I emission from many NRFs is a possible source to the North Atlantic Ocean. Hanford nuclear fuel reprocessing plant located in the west coast of USA has released about 260 kg 129I into the atmosphere during its operation20. Meanwhile, the major European NRFs (i.e Marcoule (France), Sellafield (UK) and La Hague (France)), have combined release of 400 kg 129I into the atmosphere until 200720 and estimation of total atmospheric discharges from major Russian facilities (Mayak, Seversk and Zheleznogorsk) are about 210 kg21. A large portion of atmospheric 129I may attach to aerosols and be washed out and deposited in the vicinity of the point sources while long-distance transport of 129I atoms will be continuously removed by precipitation along their journey. Concentration 129I in precipitation and surface soil in Europe together with the fallout pattern of 137Cs after Chernobyl accident suggest that more than 90% of the radionuclides are deposited within 1500 km of the nuclear power plant22,23,24. Thus the contribution from Hanford reprocessing plant (~4000 km away from east coast of America) to the Atlantic Ocean must be negligible and the same applies for the atmospheric releases from the Russian facilities. Furthermore, precipitation collected in Europe suggests that only tiny portion (<1%) of 129I atmospheric releases occurs in the atmosphere which again supports insignificant effect of gaseous emission from the European NRFs25. Thus for the 129I inventory, the influence of direct marine discharges and later transport by ocean currents system play a major role for the concentration in the surface water of the investigated area. The 129I concentration and 129I/127I value in samples north of Cape Finisterre (Spain) analyzed here also confirm the influence from north, where a gradually decreasing pattern is documented southward. Therefore, the 129I level in the north part of the investigated region exhibited a combined contribution from both Sellafield and La Hague and may likely be advection effect further south.

Except the two high 129I concentrations in the north part of our transect, three peaks are also found in water samples south of 40°N (Fig. 2). These waters exhibit relatively high 129I concentration (1.8–2.6 × 108 atoms/L) and high 129I/127I values (8.7–11.5 × 10−10 atoms/atoms), which are about one order of magnitude higher than the level of atmospheric nuclear weapons test (~10−10) and 4–6 times higher than surrounding waters. These three isolated peaks do not seem linked to waters from the north due to the general pattern of meridional water movement with respect to the Portugal Current system in this region26. This is also confirmed by speciation analysis of 129I that shows, iodide as the main species, in the southern three peaks, whereas in samples 3 and 8 the iodate species dominates. Accordingly, waters with high 129I concentration south of 40°N clearly imply a source other than the influence of Sellafield and La Hague from the north. In addition to the two NRFs mentioned above, the Marcoule facility (France), which is located on the banks of Rhone river, directly released 45 kg of liquid 129I to the river and 145 kg of 129I to the atmosphere during its operation until 199720. Considering 50% of atmospheric 129I deposited in the vicinity of the plant and assume all 129I has eventually injected into the Mediterranean Sea through continental runoff, about 120 kg of 129I would be in the Mediterranean Sea. As a result, the average concentration of 129I is estimated at 1.46 × 108 atoms/L on the assumption that the 129I is homogeneously mixed in the Mediterranean water. The large amount of 129I released by Marcoule and the relatively long residence time of iodine (~1800 y) compared with the water turnover time (~1000 y) in the Mediterranean Sea, point out the Mediterranean water as a source of 129I to the North Atlantic Ocean. Seawater and algae samples taken from the Mediterranean Sea show a comparable 129I level with our samples that show high 129I level in the south part of the transect. This similarity suggests a Mediterranean source in these waters27,28. To our best knowledge, there is no 129I speciation data in the Mediterranean Sea, but data from the English Channel (close to the La Hague facility) indicate high 129I−/129IO3− values10 and it is reasonable to conclude high 129I−/129IO3− values for Mediterranean seawater. Moreover, high iodide concentration is normally found in the coasts, bays, estuaries and semi-enclosed basin10,29,30. Therefore, our transect seawater samples south of 40°N that are characterized by high 129I and 129I−/129IO3− may represent direct influence from Mediterranean water plume.

Saline and dense Mediterranean Outflow Water (MOW) that passes through the Strait of Gibraltar descends as Mediterranean Undercurrent in the Gulf of Cadiz along the continental slope and interacts with local bottom topography and spreads towards the Azores, the Canary Islands and the coast of Portugal31. Two major Mediterranean cores with maxima in the temperature and salinity can be well detected in the depth of 800 and 1200 m32. This feature is also shown by two peaks of 129I/127I ratio in a water profile taken from the northeastern Atlantic Ocean33. The dense Mediterranean Water typically settles at depth of 800–1200 m in the North Atlantic Ocean, which is physically separated by a salinity-minimum layer of Subpolar Mode Water (S < 36) above34. The isolated Mediterranean layer should not have direct influence on the surface and the North Atlantic Central Water (NACW) that control the surface circulation. Thus, those high 129I samples (>2 × 108 atoms/L) in the south of our transect seem to originate from the Gulf of Cadiz and Portugal coast rather than local upwelling effect in the open sea. Intensive in situ hydrographic data and sea surface temperature (SST) satellite observations, as well as numerical models reveal that the eastward Azores Current, combined with equatorial Portuguese Current that flows parallel to the west Iberia coast, enters the Gulf of Cadiz along the continental slope towards the Strait of Gibraltar. The easternmost part of this anticyclonic circulation eventually feeds the Canary Current (Fig. 1c). This general feature of circulation pattern was thought to be weakened in winter and enhanced in summer35,36,37. Mediterranean Outflow Water (MOW) overflows the sills of Gibraltar in a shallow depth (~200 m) and intense local vertical mixing and entrainment processes taken place at the mouth of the strait are even capable to induce and maintain the eastward surface Azores Current38,39. Mauritzen et al.40 hypothesized a cross-isopycnal MOW towards the surface in the vicinity of the Strait of Gibraltar, which they referred to ‘detrainment’ process and later confirmed by a rotating tank experiment. Thus the high 129I concentration may occur in the surface water as a result of Mediterranean water upwelled in this turbulent and instable fluid condition within the eastern Gulf of Cadiz. In addition to the double-core structure of the MOW, a shallow layer is identified through salinity and temperature field at depths between 400 m and 700 m36,41, or depicted in the geostrophic velocity field42,43 along the upper continental slope of the Gulf of Cadiz and off the southern half of the Iberian western coast41,44. Therefore, upwarping of this Mediterranean shallow core, in association of coastal upwelling that is especially occurring off the Capes St. Vicente and St. Maria, by favored westerlies during summer45, probably bring the Mediterranean-labeled 129I from mid-depth to the surface. Accordingly, we suggest the surface water in location 24 and 25 may reflect the propagation of a branch of this onshore water westwards towards the open ocean after leaving Cape St. Vicente when easterlies dominated (Fig. 4). Seasonal and regional variations of wind system in the transition zone between westerlies and trade wind highly influences the surface circulation pattern in most parts of our sampling transect. Based on buoy and hydrographic data, large-scale movement of upper water between subpolar and subtropical gyres is week, flowing slowly towards east and south46. However, a poleward surface current has been reported off the west coast of Portugal during winter when the wind blowing northward47. Although the origin and variability of this slope-flow current is still unclear, it may be connected with the coastal counter flow in the western of Cadiz, which flow anticyclonically around Cape St. Vicente and thus bearing high 129I signal from MOW48. When this water reaches the Cape Espichel, local coastal morphology, shelf/slope bathymetry as well as propagation of eddies make the 129I plume extends seaward, which is reflected in our sample 17. The other two high-129I samples, which occur in the middle of Madeira and African coast (location 32 and 33), however, are likely related to the onshore flow of Canary Current, which with one branch separated from the northern Morocco coast37, transports 129I from the Strait of Gibraltar to further west around 14°W (Fig. 4).

Figure 4
figure 4

Concentrations of 129I (108 atoms/L) along the transect and suggested surface 129I pathways from the English Channel and the Strait of Gibraltar (dashed arrows).

Red regions represent major coastal upwelling and solid blue lines show the spread pathways of deep Mediterranean water. Major nuclear reprocessing facilities (NRFs) are marked as stars. The original map was constructed by a free software Ocean Data View (ODV 4.5.3).

Full size image

Iodate is found to be predominated species in ocean waters, however, since some biological-mediated processes may be active on iodate reduction, significant amounts of iodide have also been observed in the surface water of the open sea. By combining available inorganic iodine species data in the different ocean regions, total iodine concentration is fairly constant in the world ocean (~0.45 μM) whereas 127I−/127IO3− ratio shows a meridional variation (Supplementary Fig. S1). This feature implies reduction of iodate in the tropical areas, though how fast this conversion is in the open sea is still unclear. Our results (iodide/total iodine) also show a 50% increase of iodide from 50 to 30°N for both 127I and 129I (Fig. 3a, b). It seems that the water samples show a different behavior in terms of 129I− proportion at the two sides of latitude 40°N (Fig. 3b). This could be explained either by different sources of 129I (English Channel and Mediterranean Sea) or different degrees of biological activity in each cluster. Marked discrepancies of iodide/iodate ratio for 127I and 129I (Fig. 3c) suggest that the 129I speciation system is not kinetically and thermally at equilibrium in the investigated area. Thus, given the transit time between source and sampled region, speciation of 129I can be used as an ideal temporal and spatial tracer to distinguish 129I sources as well as to evaluate iodine redox processes in the ocean.

Compared to iodate, iodide is a thermodynamically unfavorable species under oxic conditions. Thus theoretically the spontaneous oxidation of iodide should occur in the ocean surface until dynamic equilibrium has been achieved for the iodide-iodate couple. However, kinetic barrier prevents this mechanism and natural oxidation process in ocean waters is thought to be rather sluggish once newly iodide is formed10,49. Regarding available ratio of 129I−/(129I− + 129IO3−), the decreasing trend from the source of 129I indeed suggests iodide oxidation occurred in ocean surface water as the water is transported (Fig. 5 and Supplementary Table S3). Considering a constant level of 129I speciation in the source water and 6.5-years transport of 129I plume from the northern North Sea (Location B, Fig. 5) to the Nansen Basin in the Arctic Ocean (Location C, Fig. 5)50, 7.3–10.3% (mean 8.9%) of iodide is annually oxidized as the 129I-contaminant water masses move. Seawater transit times from the northern North Sea to the Denmark Strait and the interior of Labrador Sea are estimated at 3.5–6.5 and 5.5–8.5 years, respectively51,52,53. The 129I−/(129I− + 129IO3−) values in the surface water of the Denmark Strait and the Labrador Sea are 0.19–0.42 (mean 0.32) and 0.07–0.34 (mean 0.22) when analogous annual oxidation rate is applied. Regarding species for 127I, concentration of iodide seems rather homogenous and remains at low value in the seawaters (<0.01 μM) at high and middle latitude. Similarly, the average ratio of 127I−/(127I− + 127IO3−) remains fairly constant (0.13) both in the Arctic Ocean and the North Atlantic Ocean (Supplementary Table S3), which suggests an equilibration for iodine species in these waters. Thus, if we simply assume there is no upwelling or vertical mixing, as well as no additional redox processes occurring during the transport of 129I plume in the open ocean, it takes 7.5–10.6y (mean 8.7 y) for 129I−/(129I− + 129IO3−) to reach the about 0.13 seawater equilibrium ratio as they transport from the northern North Sea (Location B, Fig. 5). Given 2 years transport from the Labrador Sea to the northeastern Atlantic region, via the Azores Current54,55, it will take about 7.5–10.5 y for 129I to be transported from the northern North Sea, via the Greenland Sea and the Labrador Sea, to the northeastern Atlantic Ocean. In addition, even much longer time is estimated (>10 y) for 129I in the upper 200 m from the Labrador Sea to reach our sampling transect using Ocean General Circulation Model15. This indicates that the ratio of 129I−/(129I− + 129IO3−) in our sampled transect should be in a state of geochemically balanced value (~0.13). However, this feature is not the fact in collected water samples as the observed 129I−/(129I− + 129IO3−) ratio varies at 0.12–0.41 (mean 0.30), which is far from reaching their thermodynamic equilibrium. This suggests that a dominated source of 129I in the northeastern Atlantic Ocean (sampled transect) from the Nordic Seas and the Labrador Sea seems doubtful. We suggest that 129I is mainly originated from outflow in the English Channel and possibly the Irish Sea rather than transport from the northwestern Atlantic. Transit time of 2.6–3.7 y (mean 3.0 y) was calculated for water transport from the English Channel to the northeastern Atlantic Ocean, via the Bay of Biscay, based on the same iodine oxidation rate used in the open ocean. This hypothesis is quite possible because of the relatively stagnant circulation in this region56.

Figure 5
figure 5

Scheme of surface water circulation in the North Atlantic Ocean and the variation of 129I−/(129IO3− + 129I−) (upper parenthesis) in the (A) English Channel, (B) the North Sea, (C) the Arctic Ocean, (D) the Labrador Sea and (E) the eastern North Atlantic Ocean.

NCC = Norwegian Coastal Current, NAC = North Atlantic Current, AC = Azores Current, CC = Canary Current, GS = Gulf Stream, EGC = East Greenland Current and IC = Irminger Current. Background map was created using Ocean Data View (ODV 4.5.3). Numbers in lower parenthesis refer to the sampling year and the bold italic number represent estimated 129I−/(129IO3− + 129I−) value.

Full size image

At present concentration levels in the northeastern Atlantic, 129I does not pose a risk to human health. However, because of its high mobility and long half life, as well as continuous release from NRFs, anthropogenic 129I can easily dispersed and redistributed in the Atlantic Ocean, thus affects marine ecosystem. Accumulation of 129I has been reported in the marine organisms such as algae, mussel and fish57,58. This seafood with elevated 129I can therefore find its way into food chain and eventually into the human, which makes 129I monitoring important. In addition, as algae blooming in coastal areas is linked to nutrient concentration associated with stratification and upwelling effects, it is possible to couple 129I distribution with respect to biotic processes and subsequent anoxic environment. Rather similar nutrient (e.g. phosphorus) geochemical behavior as that of iodine is suggested in seawater59.

Fish otolith has been used as environmental marker for fish lifetime and migration pathway60. Majority of elements deposited in otolith originate from water, with a minor fraction comes from food sources. 127I was already found in otolith61 and 129I has been detected in fish flesh reference IAEA58, then 129I may also be incorporated into the otolith. Thus, variations of iodine isotopes (129I and 129I) and their species in seawater are expected to be reflected in growth of otolith (e.g. Atlantic cod and Atlantic herring). This possibility can open a new interesting approach where growth and migration pathways of fish can be traced using the ratio of 129I/127I. For example, fishes that inhabit the English Channel, the Nordic Seas and the Arctic Ocean are expected to build up higher 129I/129I levels than those inhabit other parts of the Atlantic Ocean or even at different depths in the ocean or originate in a fresh water environment like salmon. Therefore, levels of 127I/129I and speciation analysis in seawater can add new dimensions in otolith chemistry to investigate fish migration pathways, 129I contamination history and marine environmental change. The same ideas could also be implied to other calcified structures, such as coral reef growth rings.

The data presented here are the first study on iodine isotopes (129I and 127I) and their species (iodate and iodide) in the water of the northeastern Atlantic Ocean. Generally, iodate is the dominant species of dissolved iodine for both 127I and 129I in the surface waters of the sampled northeastern Atlantic. Although relatively low level of 129I appears in the studied area at present, continuous 129I releases from La Hague and Sellafield make monitoring of this radionuclide valuable in the Atlantic Ocean since the concentration of 129I in the subpolar gyre tends to be increased by ocean circulation. This study also indicates that 129I and species could be used as an ideal oceanographic tracer not only in the Arctic and Nordic regions, but also in the North Atlantic Ocean.

Multiple sources of 129I are suggested to be received in the northeastern Atlantic Ocean as revealed by the 129I speciation data. Among these sources, occurrence of Mediterranean seawater contribution is exposed as three narrow jets south of 40°N. Mediterranean water is commonly difficult to detect in the Atlantic surface water using salinity and temperature distribution. Here we show the potential implication of 129I and its species to label MOW in the Atlantic surface and even deeper water. As most of MOW descends within the Gulf of Cadiz, depth profiles are needed in the vicinity of the Strait of Gibraltar for further research.

Methods

Samples collection

Surface seawater samples (47.4–31.1°N, 7.6–14.5°W) were collected in the North Atlantic Ocean from October to November, 2010 (Fig. 1a). The sampling transect was done onboard the Ice Breaker Oden during the passage from Sweden to Ponta Arenas (Chile) as part of the 2010/2011 Antarctica expedition jointly funded by the Swedish Polar Research Secretariat and the US National Science Foundation. Sampling was performed using a Teflon intake direct surface water sampling system that has been tested for possible memory effect and contamination62. At the same time an automatic continuous measurements of parameters such as salinity, temperature and wind speed were conducted. Each 2-L sample was instantly filtered onboard through a 0.45 μm membrane (Sartorius AG, Gottingen, Germany) and filled in a clean polyethylene container under cold and dark conditions. All chemical reagents used were of analytical grade and all solutions were prepared using deionized water (18.2 MΩ·cm).

Iodine speciation analysis

A method for separation of iodine species (iodate and iodide) developed by Hou et al.63 was used for both 127I and 129I (Supplementary Fig. S2). The filtered seawater mixed with 250 Bq of 125I− (Amersham Pharmacia Biotech, Little Chalfout, Buckinghamshire, UK) was loaded to a column (Ø 1.0 × 20 cm) filled with prepared NO3− form of AG1- × 4 anion exchange resin (50–100 mesh, Bio-Rad laboratories, Richmond, CA). 30 ml deionized water and a 50 ml of 2 M NaNO3 were used to wash the column and the effluent and two washes were collected for solvent extraction of iodate. Iodide absorbed on the resin was eluted using 60 ml of 10% NaClO solution. One milliliter of each fraction (original seawater, iodate effluent and iodide eluate) was diluted with Cs internal standard (~200 ppb) and 1% NH3·H2O solution.

About 250 Bq of 125I− tracer and 0.1 ml 2 mg/ml of Woodward iodine (MICAL Specialty Chemicals, New Jersey, USA) were added to iodate effluent for chemical separation of 129I species. To convert all the iodine to iodide form, 5 ml of 3 mol/L HNO3 and 1.0 ml of 1 mol/L Na2S2O5 were added, to ensure a fast reduction. CHCl3 was used for solvent extraction of iodine. Iodine was first extracted to CHCl3 phase (as I2) by adding 2–5 ml of 1 mol/L NaNO2 solution and then back-extracted to water with Na2S2O5. This procedures was repeated several times until no I2 in CHCl3 phase was observed. The extraction of iodide from the eluate is much the same, the differences are that we use 1.0 mol/L NH2OHHCl solution to reduce iodate to molecular iodine and no 125I− was added during the first extraction. The analytical uncertainty is 5%–10%.

Iodine determination

The concentration of total iodine, iodate and iodide were determined using X-SeriesII ICP-MS (Thermal Electron Corporation) under hot plasma conditions with Xt interface. The detection limit calculated as 3SD of blanks was 0.02 ppb. 125I activity was measured using a NaI gamma detector (well type, Bicron) at channels 25–115 in 26–36 keV.

The dried iodine precipitation (AgI) was mixed with Nb powder in a mass ratio of 1∶2 and the mixture was transferred into a copper holder and pressed for the AMS measurement using a 5 MV Pelletron machine at the Tandem Laboratory, Uppsala University. The 129I/127I isotopic ratio in the 129I standard (diluted NIST SRM 4949C) is (1.1 ± 0.1) × 10−11 and the background of the AMS system is 4 × 10−14. The statistical error of the measurements was <7% (one standard deviation).