Fukushima-derived radionuclides in the ocean and biota off Japan

Abstract

The loss of power and subsequent overheating, meltdowns, and hydrogen explosions at the Fukushima Dai-ichi nuclear power plants (NPPs) resulted in airborne fallout over land and the ocean that peaked in mid-March 2011 (1–3). In addition to atmospheric fallout, waters used to cool the reactors that subsequently leaked from NPP buildings provided a direct path for contamination of the oceans (4). Concentrations at the NPP ocean discharge channels peaked in early April at more than 50 million times preexisting ocean levels of ¹³⁷Cs (5). Though considerable data have been released in Japanese reports regarding the concentration of selected radionuclides in the air, soil, and coastal discharge sites, large uncertainties remain, including even the magnitude of total atmospheric releases (6) and direct discharges (7). There is also little information on radionuclide distributions offshore to help assess contamination and transport in the North Pacific and for independent confirmation of whether the levels are of human health concern.

We report here on a comprehensive analysis of radionuclide concentrations off Japan in surface and subsurface waters as well as for mesopelagic fish and plankton collected from the R/V Ka'imikai-o-Kanaloa at 50 stations June 4–18, 2011 (Fig. S1). We used standard oceanographic methods to collect water and biota followed by laboratory processing and detection of marine radionuclides using high-purity germanium spectroscopy (Methods and SI Methods). This report describes the presence of the gamma-emitting isotopes ¹³⁴Cs (t_1/2 = 2.07 y), ¹³⁷Cs (t_1/2 = 30.07 y), and ^110mAg (t_1/2 = 250 d). With their shorter half-lives, any ¹³⁴Cs and ^110mAg seen in our samples could only be derived from the 2011 Fukushima NPP releases.

Cesium is a highly seawater soluble radionuclide whose primary source to the ocean before March 2011 has been from weapons testing in the 1960s, with lesser amounts from Chernobyl fallout in 1986 and intentional discharges such as from European nuclear fuel reprocessing facilities at Cap de la Hague (France) and Sellafield (United Kingdom) (8). Before 2011, ¹³⁷Cs levels off Japan were ∼1–2 Bq⋅m⁻³ (9) (1 Bq = 1 disintegration per second).

Results and Discussion

In June 2011, surface-water concentrations of ¹³⁴Cs were highly elevated in near-shore areas, although not necessarily at stations closest to the Fukushima Dai-ichi NPPs (Fig. 1A and Table S1). The highest ¹³⁴Cs activities, 3,900 Bq⋅m⁻³, were associated with a semipermanent eddy, seen here in the surface drifter data centered on 37°N 142.5°E (Fig. 2B). Activities up to 325 Bq⋅m⁻³ were found more than 600 km from the NPPs (Table S1). The ¹³⁴Cs:¹³⁷Cs ratio was close to 1 in all samples clearly indicating a Fukushima NPP source (5) (calculated end-member ratio of 0.975; Fig. S2); the exception to this are the southernmost stations, where we see ¹³⁷Cs at preexisting levels of 1–2 Bq⋅m⁻³ and no detectable ¹³⁴Cs. In our samples, < 0.1% of total Cs was caught on a 1-μm filter, consistent with its more soluble nature (SI Methods).

Open in a separate window

Fig. 1.

(A) Concentrations of ¹³⁴Cs in surface water (Bq⋅m⁻³), (B) ¹³⁴Cs in biological samples (Bq⋅kg⁻¹ dry weight), and (C) ^110mAg in biological samples (Bq⋅kg⁻¹ dry weight). Biological samples were separated into mixed zooplankton, crustaceans, and fish as indicated by legend (SI Discussion). Red star is location of Fukushima NPPs. Gray shaded area in A shows approximate position of Kuroshio Current during the cruise (Fig. S1). Dashed lines in A are areas used to calculate near-shore and offshore Cs inventories.

Open in a separate window

Fig. 2.

(A) Tracks of surface drifters released at the sampling stations, from time of release through October 26, 2011. Trajectories are color coded: blue drifters traveled north of the Kuroshio, red were advected east in the Kuroshio, green went south of the Kuroshio, and magenta stayed close to the coast. (B) Expanded view of coastal region showing more clearly the near-shore eddies and coastal transport.

It is evident that the Kuroshio Current forms a southern boundary for the transport of these Fukushima-derived radionuclides (at least in the western North Pacific), because samples at this boundary or to the south had ¹³⁴Cs activities <3 Bq⋅m⁻³ (our detection limit for ¹³⁴Cs was 1.5 Bq⋅m⁻³). This finding is supported by the tracks of our surface drifters, most of which traveled east along the Kuroshio or northeast (Fig. 2A, blue and red lines). However, several drifters moved westward toward the coast, with one drifter transiting along the coast to the south of our study area (Fig. 2B, magenta lines), suggesting a potential pathway for contaminated water. Although we have no Cs measurements this close to shore, this is consistent as well with the coastal model of Tsumune et al. (4).

¹³⁷Cs and ¹³⁴Cs in biota ranged from below detection to 56 Bq⋅kBq⋅kgg⁻¹ dry weight (Fig. 1B and Table 1), depending primarily on the sampling location but also on the biological composition of the sample. As in the water, ¹³⁴Cs:¹³⁷Cs ratios were essentially 1. The only other artificial gamma-emitting radionuclide detected consistently in biota was ^110mAg (Fig. 1C), which was detected in zooplankton samples with concentrations up to 23.6 Bq⋅kg⁻¹ dry weight, but was undetectable in all three fish samples (Table 1). Overall, Ag is known to be highly particle-reactive in seawater and greatly concentrates in both phytoplankton, at the base of the food chain, and zooplankton (10), consistent with the data in Table 1. The ^110mAg concentrations in plankton samples north of the Kuroshio were lowest in samples more distant from Japan. The source of the ^110mAg measured in zooplankton is unclear; no direct releases to the coastal ocean have been reported, but soils collected near the power plants did show trace levels of ^110mAg (11). In zooplankton, ¹³⁴Cs:^110mAg ratios ranged up to 4.7, with most values 0.8–2. Gelatinous zooplankton, collected at one station, had about half the radioactivity of crustacean zooplankton at that same station for either Cs isotope or ^110mAg.

Concentration factors—essentially the degree of radionuclide enrichment in biota relative to ambient water—are used to evaluate the radiological risks associated with seafood consumption and were determined by dividing the radionuclide concentration in biota by the dissolved concentration in surface water. Median values were 44 for ¹³⁷Cs and 36 for ¹³⁴Cs, comparable to the recommended International Atomic Energy Agency (IAEA) value of 40 for zooplankton (10). Concentration factors for ^110mAg were not calculated because dissolved concentrations could not be measured using our seawater method, which was specific to Cs isotopes only.

The median concentration of ¹³⁷Cs in micronektonic fish (secondary consumers) was less than that in zooplankton (primary and secondary consumers) and about 150-fold below the Japanese legal limit for fish of 500 Bq⋅kg⁻¹ wet weight. Compared with oceans that were most heavily impacted by Chernobyl, our ¹³⁷Cs values in fish are equivalent to those found in the Black Sea, but less than in the Baltic Sea, where fish concentrations varied with locations (12). The combined gamma radioactivity of ¹³⁴Cs, ¹³⁷Cs, and ^110mAg in biota ranged from 1% to 54% (typically 10–30%) of total radioactivity attributable to naturally occurring radionuclides, of which ⁴⁰K contributed >99% of the total (Table 1).

The surface ocean Cs data leads to several key conclusions regarding the impact of the Fukushima Dai-ichi NPPs on the ocean. First, Fukushima-derived ¹³⁴Cs and ¹³⁷Cs were measured at up to 1,000× higher activities than existed previously and were found throughout a 150,000-km² area of the Pacific Ocean off Japan. Second, though elevated, substantial dilution had occurred between the discharge channels at the Fukushima Dai-ichi NPPs, where ¹³⁷Cs activities averaged 33,000 Bq⋅m⁻³ in June (5) and our closest samples 30 km offshore, which were on average 50× lower (600–800 Bq⋅m⁻³). Our Cs activities are thus consistent with the Japanese reporting in June “below detection” for ¹³⁴Cs and ¹³⁷Cs 30 km off-shore [Japanese Ministry of Education, Sports, Science and Technology (MEXT)], because their methods had a higher detection limit of 10,000 Bq⋅m⁻³ (5). As a result of ocean stirring and mixing processes, all ¹³⁷Cs activities offshore have been reduced to well below the Japanese regulatory limits for the ocean (90,000 Bq⋅m⁻³) and are below the levels of the most abundant natural radionuclide in the ocean, ⁴⁰K (∼12,000 Bq⋅m⁻³).

We measured ¹³⁴Cs activities decreasing with depth, and as with the surface data, activities in the subsurface are generally higher near-shore than offshore (Fig. 3). Prior studies show penetration of 1960s fallout radionuclides, including ¹³⁷Cs, down to 1,000 m in this area (9), but in the short time here since its release, Fukushima-derived Cs does not generally penetrate below ∼100–200 m. Over time, we expect to see deeper penetration and lateral mixing for Cs and other soluble Fukushima isotopes, providing a means to study the rates of vertical and horizontal mixing processes in the Pacific.

Open in a separate window

Fig. 3.

Profiles of ¹³⁴Cs (Bq⋅m⁻³) vs. depth for offshore (gray solid lines; mean, gray circles) and near-shore stations (black dashed lines; mean, black circles).

Cesium profiles are essential to calculate ocean contamination for comparison with land, and to calculate the total Cs inventory in our study area for comparison with the widely varying source estimates. When integrated vs. depth, aerial inventories of ¹³⁷Cs at different locations in the near-shore area range from 7,000 to 80,000 Bq⋅m⁻², decreasing to 200–14,000 Bq⋅m⁻² in the offshore stations (areas denoted in Fig. 1A; Methods). Preexisting ¹³⁷Cs inventories in these waters would be ∼200–300 Bq⋅m⁻² (down to 200 m) (9). It is important to note that our highest values are still 40× lower than the highest estimated aerial burdens on land within 30 km of the NPPs (3 million Bq⋅m^{-2 137}Cs; ref. 13, http://radioactivity.mext.go.jp/en/1750/2011/08/1750_083014.pdf).

There has been considerable debate about both the total radionuclide releases and the extent of atmospheric fallout vs. direct discharges from the Fukushima NPPs. Total atmospheric releases of ¹³⁷Cs have been estimated by several Japanese groups and official sources to range from 10 to 13 PBq (1 PBq = 1 × 10¹⁵ Bq; Table S2), with peak discharges delivered March 15–25, 2011 (1, 2). From the 10-PBq total atmospheric source, Morino et al. (2) suggest that 22% fell on land in Japan and 10% (1 PBq) in their ocean domain (defined out to 145°E), with about 50% advecting farther east over the ocean. A more recent study by Stohl et al. (6) found a similar fraction on land, but suggested much higher total atmospheric releases for ¹³⁷Cs of 36 PBq (23–50 PBq).

Direct discharges were an additional radionuclide source to the ocean, dominated by a leak from Dai-ichi reactor 2 April 1–6, 2011. Similar to atmospheric fallout, there is a considerable range of release estimates, from 3.5 PBq (4) to as high as 22 PBq (Table S2). We measured a total inventory of Fukushima-derived ¹³⁷Cs of 1.9–2.1 PBq in our study area (Methods), which is lower because most of the ¹³⁷Cs delivered either as fallout in March or direct discharges in April would have been transported out of the study area by June.

To quantify this transport and assess if our data support the higher or lower release estimates, we modeled the spreading of contaminated waters by ocean currents (Methods and SI Discussion). The model reproduced the measured ¹³⁷Cs distribution in June, with the largest concentrations associated with a near-shore cyclonic eddy, and intermediate concentrations at stations closest to shore, resulting from later stages of direct discharges, and the southern edge of contaminated waters aligning with the Kuroshio (Fig. S3). The model predicts an inventory of ∼0.4 PBq (using the lowest atmospheric and direct discharge source estimates) to ∼2.0 PBq (using the highest estimates) in our study area in June (SI Discussion). Thus, if we include realistic transport, our measured inventory in June agrees better with what is predicted using the largest release estimates; this is important, but needs to be supported by additional data over larger areas of the North Pacific. However, at present, such data are extremely limited and insufficient to make a basin-wide inventory calculation. A more complex 3D ocean model would also be helpful to quantify transport processes and predict long-term concentration trends expected throughout the Pacific.

In terms of potential biological impacts, radiation doses in marine organisms are generally dominated by the naturally occurring radionuclides ²¹⁰Po (an alpha emitter) and ⁴⁰K, even when organisms are exposed to anthropogenic radioactivity discharged to coastal waters (14). To be comparable just to doses from ²¹⁰Po, ¹³⁷Cs levels in fish would need to range from 300 to 12,000 Bq⋅kg⁻¹ dry weight, some 1–3 orders of magnitude higher than what we observed ≥30 km off Japan. Thus, radiation risks of these isotopes to marine organisms and human consumers of seafood are well below those from natural radionuclides (SI Discussion). Further, we can calculate an external dose to humans if immersed/swimming in these waters of <0.01 μSv⋅d⁻¹ at ¹³⁴Cs and ¹³⁷Cs levels of 1,000 Bq⋅m⁻³, which is <0.3% of the average Japanese dose of about 4 μSv⋅d⁻¹ from all radioactivity sources. Finally, these levels are several orders of magnitude lower than those used in one study that assumed exposure to the most heavily impacted water discharged from the Fukushima NPPs to predict marked reproductive effects and possible mortality in marine biota (15).

Several other gamma-emitting fission products (e.g., ¹⁰⁶Ru, ¹⁰³Ru, ¹⁴⁴Ce, and ¹⁴¹Ce) were detected after Chernobyl in biota and sinking particles (16), but were not seen here in biota (or on filters), or by others in soils around the Fukushima NPP (11). These refractory isotopes were prevalent from Chernobyl due to the massive explosion and release of core materials in addition to radioactive gases and volatiles. The Fukushima accident was characterized by core overheating that led to the venting of radioactive gases, hydrogen explosions, and fires associated with spent fuel rods; this resulted in the preferential release of more volatile radionuclides, such as Cs, and gases to the atmosphere. However, the salt and fresh water used to cool the Fukushima NPPs and acidic conditions in the core provide possible pathways to mobilize refractory radionuclides from the core that may have subsequently been discharged to the ocean but have yet to be assessed. Ultimately, though the radionuclide levels of ¹³⁷Cs and ¹³⁴Cs offshore are currently low with respect to human health effects, any assessment of radiation dose should also consider long-term exposure if the NPP remains a continued source of radionuclides (5) and if, as has been reported, coastal sediments are contaminated with multiple radionuclides.

Methods

Details of sampling, analytical methods, transport model, and calculations of dose are provided in SI Discussion, with a brief summary below.

Supplementary Material

Acknowledgments

Footnotes

References