**1. Introduction**

Large amounts of radionuclides, such as cesium-134 (134Cs), cesium-137 (137Cs), and iodine-131 ( 131I), were dispersed into the terrestrial and aquatic environments as a result of an accident at the Fukushima Daiichi Nuclear Power Plant (FDNPP) of the Tokyo Electric Power Company (TEPCO) in March 2011. Atmospheric release of strontium-90 (90Sr) in March 2011 was two to four orders of magnitude lower than that of 137Cs on the basis of an analysis of highly contaminated soils (<1.1 Bq g<sup>−</sup>1) and vegetation (0.026–1.1 Bq g−1) collected from a contaminated area in Japan [1]. These 90Sr/ 137Cs activity ratios were much lower than the ratio for the estimated nuclear fuel compositions (90Sr/ 137Cs = 0.74) found in the reactor obtained by the ORIGEN2 code [2]. Atmospheric 90Sr release (0.01–0.14 PBq [3]) was estimated at less than 0.027% of the total amount in the nuclear fuel (5.2 <sup>×</sup> <sup>10</sup><sup>2</sup> PBq [2]) at FDNPP reactor units 1, 2, and 3. Most of the 90Sr remained in the reactor, although some of it had dissolved in stagnant water in the reactor and turbine buildings. Observed 90Sr and 137Cs concentrations in the stagnant water were 140 MBq L−<sup>1</sup> and 2.8 GBq L−1, respectively, on 27 March 2011 [4]. Hence, 137Cs concentrations were 20 times higher than 90Sr concentrations, and 1.6% of the 90Sr core inventory was dissolved into stagnant water [2], which was the most likely candidate for pollution to the ocean. 90Sr in seawater could be a useful tracer specific to the radionuclide contaminants directly released from the FDNPP into the ocean.

Analytical results of the stagnant water sampled from a turbine building in February 2012 indicated that 137Cs activity decreased to 240 MBq L−1, while 90Sr concentration remained high (170 MBq L−1) [5]. Highly contaminated stagnant water was decontaminated and stored in storage tanks on the FDNPP site. Some decontaminated water was transferred into reactors for cooling purposes after distillation or reverse osmosis processes. Before 2015, the decontamination system was optimized to remove Cs; hence, the treated water had significantly higher 90Sr activity (150 MBq L−1) than 137Cs activity (3.9 kBq L<sup>−</sup>1) [5]. This treated water in the storage tanks was a potential source for 90Sr contamination in the environment.

In observation wells between the reactor buildings and the harbor, groundwater was also monitored by TEPCO after a leakage event of contaminated water in December 2012 [4]. In particular, 90Sr activity in groundwater in the wells near the seawater intake for reactor units 1 and 2 were significantly higher than the 137Cs activity (e.g., 90Sr: 5 <sup>×</sup> <sup>10</sup><sup>6</sup> Bq L<sup>−</sup>1; 137Cs: 2.1 <sup>×</sup> <sup>10</sup><sup>2</sup> Bq L−<sup>1</sup> at the no. 1–2 wells on 5 July 2013 [6]). The 90Sr-enriched groundwater might have resulted from leakage of the decontamination system or from stagnant water. Due to these existing contamination sources, it is necessary to observe the 90Sr behavior in the aquatic environment near the FDNPP.

Kanda [7] indicated that continuous release of 137Cs from the FDNPP harbor to the ocean was occurring in 2012 based on time series seawater monitoring data. Due to the high 90Sr/ 137Cs activity ratio in the stagnant water, 90Sr release from the FDNPP should also be evaluated. TEPCO has continued seawater monitoring for 90Sr, 134,137Cs, and other radionuclides near the FDNPP [3,6]. However, only a few 90Sr data were obtained within small areas, particularly after 2012 (Figure 1).

This limited monitoring cannot evaluate how much 90Sr was released or its impact on the coastal environment and open ocean.

Time series seawater monitoring by TEPCO of 90Sr near the FDNPP was infrequent compared to that for radiocesium [3,6]. Povinec et al. [3] showed that the 90Sr/ 137Cs ratio in seawater at a monitoring point near FDNPP increased gradually from 0.01 to 1 between April 2011 and February 2012 (Figure 1), which clearly related with decontamination of stagnant water. The transient increase of 90Sr in seawater at the T2 site observed in December 2011 could reflect the leakage event from the 137Cs decontamination system [4]. After 2012, the 90Sr/ 137Cs ratio remained at a constant value around 0.5 at the T2 site with large variability. 90Sr/ 137Cs activity ratios in stagnant water have varied depending on the decontamination of 134, 137Cs. The agreement between the temporal variation of 90Sr/ 137Cs activity ratio and decontamination of the stagnant water supported the idea that the most probable candidate was the continuous release from reactor buildings of the FDNPP.

The behavior of 137Cs in seawater and biota after the accident has been well documented [7–11]. High-density sampling of surface seawater to determine radiocesium activity [12] has been carried out. Kumamoto et al. [13] reported detailed vertical distributions of Fukushima-derived radiocesium along the 149 ◦E meridian in the western North Pacific. However, distributions of 90Sr derived from the FDNPP in the sea have been studied to a significantly lesser extent [3,6,14–17]. Castrillejo et al. [17]

suggested that continuous release of 90Sr from the FDNPP was occurring in September 2013 based on simultaneous observations of 90Sr and 137Cs. The estimated release rate of 90Sr was 2.3–8.5 GBq day−1, which was 2–3 orders of magnitude larger than river inputs.

**Figure 1.** Temporal variations of strontium-90/cesium-137 (90Sr/ 137Cs) activity ratio in seawater from monitoring sites T1 and T2 (T2-1) in the Fukushima Daiichi Nuclear Power Plant (FDNPP) site [6]. T1 and T2 (T2-1) sites are located north and south of the discharge channel of the FDNPP, respectively.

It is still necessary to investigate the amount of released 90Sr, including its subsequent dispersion from the FDNPP site to the ocean. Simultaneous determinations of 90Sr and 137Cs in seawater are important for monitoring the release of radionuclides from the reactor buildings and contaminated water from the storage tanks. By comparing 90Sr behavior with that of 137Cs in the ocean, we studied the input source to the sea and the environmental migration processes of both radionuclides, such as fluvial input, desorption from sediment, and atmospheric deposition. Accumulating environmental data and understanding the dispersion to the coastal and open oceans are necessary to respond to any accidental release during decommissioning of the FDNPP—work that will require more than 30 years. Our aim in this study is to determine the distributions of 90Sr, 134Cs, and 137Cs in 2013 and evaluate the continuous release of radionuclides from the FDNPP to the ocean based on the comprehensive analysis of seawater.

#### **2. Surface Current System o**ff **Fukushima Coast**

The Kuroshio and Oyashio currents are generated in the mixed region around 36 ◦N off the Ibaraki Prefecture coast in the subject area (Figure 2a,b). The warm (16.5–22.0 ◦C) and saline (34.4–34.8 psu) Kuroshio flows northeastward off the Boso Peninsula. The Oyashio current, off the Fukushima Prefecture coast, intrudes southward into the mixing region. The southward intrusion (9.5–10.5 ◦C, salinity 33.4–33.8 psu) reaches 36.5 ◦N, 141.3 ◦E, and is called the First Branch of the Oyashio [18]. The coastal currents near Fukushima Prefecture are variable on a time scale different from those of the Kuroshio and Oyashio currents. Coastal water is at a higher temperature (10.8–12.7 ◦C) and lower salinity (33.2–33.4 psu) relative to the first branch of the Oyashio current. Current meter observations made between 1971 and 1981 [19] indicated that the along-shore (north–south component) currents were dominant in this coastal area. The direction of the currents varied approximately every 3−4 days because of changes in the synoptic-scale wind fields [19]. The spread of radionuclides from the direct-release event in April 2011 depended on the coastal current system. Model simulation of directly released Cs employed the Regional Ocean Modeling System (ROMS), which indicated that the plume was southwardly advected to the coastal region [20].

**Figure 2.** Maps showing sampling locations. (**a**) Surface temperature and (**b**) salinity had two boundary currents (dashed arrows), the warm northeastwardly Kuroshio current south of the Boso Peninsula and the cold southerly first branch of the Oyashio current off the Fukushima coast. (**c**,**d**) Sampling locations are marked by red circles and located near the coast of Fukushima Prefecture.

#### **3. Materials and Methods**

Seawater samples for the analysis of 90Sr, 134Cs, and 137Cs were obtained during the UM13-5 cruise from 14 to 23 May 2013 undertaken by the RTV *Umitaka*−*Maru* of the Tokyo University of Fisheries, Japan. Seawater sampling sites were located in the offshore region in the first branch of the Oyashio current and the coastal region near Fukushima Prefecture (Figure 2c,d). Most of the coastal sites were south of the FDNPP and close to Iwaki city. The closest observation site to the FDNPP was NP-2, located approximately 6 km east of it. During the sampling period, most of the influence from the FDNPP could be detected in the region associated with the southerly coastal current.

During the cruise, surface seawater samples were collected by an underway sampling system, whose inlet was located on the bottom of the ship, at a depth 5 m below the surface. Collected samples were filtered through a 0.5 μm pore polypropylene cartridge filter (TCW-05N-PPS, Advantec, Tokyo, Japan). Filtered water samples were stored in 20 L polyethylene bags and 90Sr and radiocesium analyses were carried out separately on land.

We conducted 134Cs and 137Cs analyses based on Aoyama et al. [21]. First, 20 L of a filtered seawater sample was acidified to pH 1.6 with HNO3. Next, 0.26 g of CsCl was added and the solution was adsorbed on ammonium phosphomolybdate (AMP) [21]. Then, AMP was collected by filtering through a 0.45 μm pore mixed cellulose esters membrane (A045A047A, Advantec, Tokyo, Japan). After drying the AMP/Cs compound, gamma rays were counted for 80,000–200,000 s with a lead-shielded HPGe detector (EGPC 250-P 15, EURISYS MEASURES, NV, USA), 604.7 keV for 134Cs and 661.7 keV for 137Cs, at the Nihon University in Tokyo. Since the detector was slightly contaminated by atmospherically released 134Cs and 137Cs at the time of the accident of the FDNPP, the background was determined before and after this counting period and subtracted from the detected signals for seawater samples. Cs yield was determined gravimetrically based on AMP weight. The typical minimum detectable concentrations (MDCs) of 134Cs and 137Cs were 0.5 mBq L−<sup>1</sup> and 0.4 mBq L<sup>−</sup>1, respectively.

For 90Sr analysis, we added 150 g of (NH4)2C2O4 H2O to 20 L of filtered seawater and shook the solution vigorously. Sr was precipitated with Ca oxalate. Oxalate precipitate was decomposed to carbonate at 550 ◦C in a muffle oven. Then, the precipitate was dissolved in HCl and diluted to about 200 mL with Milli-Q water. A small portion of sample solution was used for determination of stable Sr yield by ICP-OES (SPECTROBLUE TI, SPECTRO Analytical Instruments GmbH, Kleve, Germany). After secular equilibrium between 90Sr and yttrium-90 (90Y) (>2 weeks), 90Y with stable Y carrier (0.1 mg) was "milked" from the 90Sr by precipitating the Fe hydroxide and purified by solid phase extraction using DGA Resin (DN1ML-R50-S) purchased from Eichrom Technologies, LLC. (IL, USA). Detailed chemical separation and beta counting procedures are described elsewhere [22,23]. Beta particles were counted by a low background 2π gas flow proportional counter (LB−4200, Canberra, NV, USA) during 120 min intervals for more than 20 h. Typical Sr and Y yields were 82 ± 9 % and 95 ± 5 %, respectively.

#### **4. Results**

Activities of 90Sr, 134Cs, and 137Cs in surface seawater samples collected in May 2013 are summarized in Table 1. Mean 90Sr activity of 0.80 <sup>±</sup> 0.11 mBq L−<sup>1</sup> at offshore sites (S1, S2, S3, and N01) was slightly lower than the estimated value (1.0 <sup>±</sup> 0.1 mBq L−<sup>1</sup> [3]) based on long-term monitoring for surface water of the western North Pacific. Around sites S2, S3, and N01, cool surface water (9.4–10.4 ◦C) from the southerly first branch of the Oyashio current was present. 134Cs activities were lower than the MDC (<0.5 mBq L−1) at the offshore sites. In this study, we used values obtained at the offshore sites as the background level originating from atmospheric nuclear weapons testing. Compared to 137Cs and 90Sr, 134Cs has a relatively short half-life (2.06 years compared to 30.17 years for 137Cs and 28.8 years for 90Sr).

High 90Sr activities were observed along the coastal region with higher temperatures (10.86–12.89 ◦C) and higher salinity (33.23–33.35 psu). The highest 90Sr activity (29.13 <sup>±</sup> 0.35 mBq L<sup>−</sup>1) was found at AN7, approximately 16 km south of the FDNPP, with 22.4 <sup>±</sup> 0.6 mBq L−<sup>1</sup> for 134Cs activity and 44.7 <sup>±</sup> 0.4 mBq L−<sup>1</sup> for 137Cs activity. At the NP-2 site closest to the FDNPP (5 km offshore) in this sampling campaign, we also found high 90Sr activity (21.81 <sup>±</sup> 0.28 mBq L−1). Furthermore, at S12 off Iwaki City, 57 km south of the FDNPP, relatively high 90Sr activity (9.86 <sup>±</sup> 0.22 mBq L−1) was found. Distributions of radiocesium activities in surface seawater showed similar trends to those of 90Sr. The maximum radiocesium activities were obtained at AN7; in particular, 137Cs activities ranged from 1.4 mBq L−<sup>1</sup> at S2 to 44.7 mBq L−<sup>1</sup> there. In the coastal region, 134Cs activities were in agreement with 137Cs activities corrected to 11 March 2011, which indicated that this radiocesium was derived from the Fukushima accident (134Cs/ 137Cs <sup>=</sup> 0.99 <sup>±</sup> 0.03 [11]).



#### **5. Discussion**
