Observation of Dispersion in the Japanese Coastal Area of Released 90Sr, 134Cs, and 137Cs from the Fukushima Daiichi Nuclear Power Plant to the Sea in 2013

The March 2011 earthquake and tsunami resulted in significant damage to the Fukushima Daiichi Nuclear Power Plant (FDNPP) and the subsequent release of radionuclides into the ocean. Here, we investigated the spatial distribution of strontium-90 (90Sr) and cesium-134/cesium-137 (134, 137Cs) in surface seawater of the coastal region near the FDNPP. In the coastal region, 90Sr activity was high, from 0.89 to 29.13 mBq L−1, with detectable FDNPP site-derived 134Cs. This indicated that release of 90Sr from the power plant was ongoing even in May 2013, as was that of 134Cs and 137Cs. 90Sr activities measured at open ocean sites corresponded to background derived from atmospheric nuclear weapons testing fallout. The FDNPP site-derived 90Sr/137Cs activity ratios in seawater were much higher than those in the direct discharge event in March 2011, in river input, and in seabed sediment; those ratios showed large variability, ranging from 0.16 to 0.64 despite a short sampling period. This FDNPP site-derived 90Sr/137Cs activity ratio suggests that these radionuclides were mainly derived from stagnant water in the reactor and turbine buildings of the FDNPP, while a different source with a low 90Sr/137Cs ratio could contribute to and produce the temporal variability of the 90Sr/137Cs ratio in coastal water. We estimated the release rate of 90Sr from the power plant as 9.6 ± 6.1 GBq day−1 in May 2013 on the basis of the relationship between 90Sr and 137Cs activity (90Sr/137Cs = 0.66 ± 0.05) and 137Cs release rate.


Introduction
Large amounts of radionuclides, such as cesium-134 ( 134 Cs), cesium-137 ( 137 Cs), and iodine-131 ( 131 I), were dispersed into the terrestrial and aquatic environments as a result of an accident at the was dissolved into stagnant water [2], which was the most likely candidate for pollution to the ocean. 90 Sr 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 137 Cs activity decreased to 240 MBq L −1 , while 90 Sr 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 90 Sr activity (150 MBq L −1 ) than 137 Cs activity (3.9 kBq L −1 ) [5]. This treated water in the storage tanks was a potential source for 90 Sr 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, 90 Sr activity in groundwater in the wells near the seawater intake for reactor units 1 and 2 were significantly higher than the 137 Cs activity (e.g., 90 Sr: 5 × 10 6 Bq L −1 ; 137 Cs: 2.1 × 10 2 Bq L −1 at the no. 1-2 wells on 5 July 2013 [6]). The 90 Sr-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 90 Sr behavior in the aquatic environment near the FDNPP.
Kanda [7] indicated that continuous release of 137 Cs from the FDNPP harbor to the ocean was occurring in 2012 based on time series seawater monitoring data. Due to the high 90 Sr/ 137 Cs activity ratio in the stagnant water, 90 Sr release from the FDNPP should also be evaluated. TEPCO has continued seawater monitoring for 90 Sr,134,137 Cs, and other radionuclides near the FDNPP [3,6]. However, only a few 90 Sr data were obtained within small areas, particularly after 2012 ( Figure 1). This limited monitoring cannot evaluate how much 90 Sr was released or its impact on the coastal environment and open ocean.
Time series seawater monitoring by TEPCO of 90 Sr near the FDNPP was infrequent compared to that for radiocesium [3,6]. Povinec et al. [3] showed that the 90 Sr/ 137 Cs 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 90 Sr in seawater at the T2 site observed in December 2011 could reflect the leakage event from the 137 Cs decontamination system [4]. After 2012, the 90 Sr/ 137 Cs ratio remained at a constant value around 0.5 at the T2 site with large variability. 90 Sr/ 137 Cs activity ratios in stagnant water have varied depending on the decontamination of 134,137 Cs. The agreement between the temporal variation of 90 Sr/ 137 Cs 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 137 Cs in seawater and biota after the accident has been well documented [7][8][9][10][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 90 Sr derived from  It is still necessary to investigate the amount of released 90 Sr, including its subsequent dispersion from the FDNPP site to the ocean. Simultaneous determinations of 90 Sr and 137 Cs in seawater are important for monitoring the release of radionuclides from the reactor buildings and contaminated water from the storage tanks. By comparing 90 Sr behavior with that of 137 Cs 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 90 Sr, 134 Cs, and 137 Cs in 2013 and evaluate the continuous release of radionuclides from the FDNPP to the ocean based on the comprehensive analysis of seawater.

Surface Current System off 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 [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]. 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 90 Sr and radiocesium analyses were carried out separately on land.
We conducted 134 Cs and 137 Cs 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-

Materials and Methods
Seawater samples for the analysis of 90 Sr, 134 Cs, and 137 Cs 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 90 Sr and radiocesium analyses were carried out separately on land.
We conducted 134 Cs and 137 Cs analyses based on Aoyama et al. [21]. First, 20 L of a filtered seawater sample was acidified to pH 1.6 with HNO 3 . 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 134 Cs and 661.7 keV for 137 Cs, at the Nihon University in Tokyo. Since the detector was slightly contaminated by atmospherically released 134 Cs and 137 Cs 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 134 Cs and 137 Cs were 0.5 mBq L −1 and 0.4 mBq L −1 , respectively.
For 90 Sr analysis, we added 150 g of (NH 4 ) 2 C 2 O 4 H 2 O 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 90 Sr and yttrium-90 ( 90 Y) (>2 weeks), 90 Y with stable Y carrier (0.1 mg) was "milked" from the 90 Sr 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.

Results
Activities of 90 Sr, 134 Cs, and 137 Cs in surface seawater samples collected in May 2013 are summarized in Table 1. Mean 90 Sr activity of 0.80 ± 0.11 mBq L −1 at offshore sites (S1, S2, S3, and N01) was slightly lower than the estimated value (1.0 ± 0.1 mBq L −1 [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. 134 Cs 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 137 Cs and 90 Sr, 134 Cs has a relatively short half-life (2.06 years compared to 30.17 years for 137 Cs and 28.8 years for 90 Sr).
High 90 Sr 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 90 Sr activity (29.13 ± 0.35 mBq L −1 ) was found at AN7, approximately 16 km south of the FDNPP, with 22.4 ± 0.6 mBq L −1 for 134 Cs activity and 44.7 ± 0.4 mBq L −1 for 137 Cs activity. At the NP-2 site closest to the FDNPP (5 km offshore) in this sampling campaign, we also found high 90 Sr activity (21.81 ± 0.28 mBq L −1 ). Furthermore, at S12 off Iwaki City, 57 km south of the FDNPP, relatively high 90 Sr activity (9.86 ± 0.22 mBq L −1 ) was found. Distributions of radiocesium activities in surface seawater showed similar trends to those of 90 Sr. The maximum radiocesium activities were obtained at AN7; in particular, 137 Cs activities ranged from 1.4 mBq L −1 at S2 to 44.7 mBq L −1 there. In the coastal region, 134 Cs activities were in agreement with 137 Cs activities corrected to 11 March 2011, which indicated that this radiocesium was derived from the Fukushima accident ( 134 Cs/ 137 Cs = 0.99 ± 0.03 [11]).

Dispersion of High 90 Sr and 134, 137 Cs Activity Plume
The high activities of 90 Sr and 137 Cs in the coastal region ( Figure 3) can be explained by the release at the time of the FDNPP accident and the physical processes that later occurred in the ocean. In the coastal region, high 90 Sr activity seawater samples with high 134 Cs and 137 Cs activities were mainly collected from south of the FDNPP (NP2, and AN7) to off Iwaki (S12 and S6), which reflects the southward transport of seawater along the Fukushima coast by the coastal currents. Higher 90 Sr activity in seawater (>8 mBq L −1 ) was found where the salinity was 33.23-33.33 psu and 11.95-12.41 psu. In the coastal region, no clear correlation between 90 Sr and salinity or 90 Sr and temperature was observed.

Results
Activities of 90 Sr, 134 Cs, and 137 Cs in surface seawater samples collected in May 2013 are summarized in Table 1. Mean 90 Sr activity of 0.80 ± 0.11 mBq L −1 at offshore sites (S1, S2, S3, and N01) was slightly lower than the estimated value (1.0 ± 0.1 mBq L −1 [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. 134 Cs 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 137 Cs and 90 Sr, 134 Cs has a relatively short half-life (2.06 years compared to 30.17 years for 137 Cs and 28.8 years for 90 Sr).
High 90 Sr 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 90 Sr activity (29.13 ± 0.35 mBq L −1 ) was found  Both activities decreased rapidly from NP2 toward the eastern sites (S6, NP1, and NP3), which indicate that the eastward dispersion was limited because of the effect of the southern coastal current during this sampling period. Compared with the pre-Fukushima accident, offshore 90 Sr activities in the north Pacific Ocean (1.0 ± 0.1) [3] and the activities measured in May 2013 indicate that the influences of the Fukushima-derived 90 Sr on open ocean sites in the mixed region between the Oyashio and Kuroshio currents were negligibly small, as were those of 134 Cs and 137 Cs. However, if any accidental releases from the FDNPP site were to occur during the decommissioning of the reactors, coastal areas could be exposed to a high activity plume.
In September 2017, low 90 Sr (1.0−1.8 mBq L −1 ) and high 137 Cs (9−43 mBq L −1 ) activities were obtained at low salinity (4-28 psu) in groundwater and beach seawater samples from Sendai Bay, located north of the FDNPP [17]. 90 Sr/ 137 Cs ratios ranged from 0.036 to 0.19. The 137 Cs activity in low salinity samples was affected by atmospheric fallout from the FDNPP accident that was deposited on land, while 90 Sr activity was not sensitive to terrestrial input. Therefore, the relationship between 90 Sr and 137 Cs can be a useful indicator for river input.
The mouth of the Ukedo River is located between collection sites NP2 and AN6. 137 Cs activity at NP2 (39.0 mBq L −1 ) was more than five times higher than that at AN6 (6.9 mBq L −1 ), though their salinities were comparable (33.24 and 33.26 psu). Dissolved 137 Cs activity in the Ukedo River, which drains a highly contaminated area, ranged from 200 to 1100 Bq L −1 in August and November 2012 [6]. 90 Sr activity in the Ukedo River was not available but the reported 90 Sr/ 137 Cs ratio for river water in the Fukushima Prefecture [6] was less than 0.04. The contribution to 90 Sr activity in seawater by input from the Ukedo River should be minor.

90 Sr/ 137 Cs Activity Ratios Derived from the FDNPP Accident
The 90 Sr/ 137 Cs activity ratios in seawater are different according to timing of any release or leakage event (e.g., direct discharge event from late March to early April 2011 [3,15]). Since Sr and Cs are highly soluble in seawater, the 90 Sr/ 137 Cs activity ratio depends on the source, which could be a useful tracer for the source. The most possible source of 90 Sr and 137 Cs is stagnant water in the reactor building of unit 2. The 90 Sr/ 137 Cs ratio of open ocean seawater [3], seawater monitoring data near the FDNPP [6], stagnant water [4], atmospheric input [1], seabed sediment [24], and river water [6] are summarized with our data in Figure 4.
The 90 Sr activity of 0.80 ± 0.11 mBq L −1 was obtained at the offshore sites, S1, S2, S3, and N01 (Table 2). To evaluate FDNPP site-derived 90 Sr, measured 90 Sr activity was subtracted from this value as the background value for North Pacific seawater. The measured 134 Cs was a pure FDNPP site-derived component because of its short half-life (T 1/2 = 2.06 yr). The FDNPP site-derived 134 Cs/ 137 Cs ratio was reported to be 0.99 ± 0.03 [11] in March 2011. FDNPP site-derived 137 Cs was calculated on the basis of measured 134 Cs activity and the FDNPP site-derived 134 Cs/ 137 Cs ratio. The 90 Sr corr / 137 Cs corr ratio estimated from the slope of a linear regression fitting was 0.66 ± 0.05 in Figure 5. 90 Sr corr activities strongly correlated with those of 137 Cs corr (R 2 = 0.919), as described in similar contour maps of 90 Sr and 137 Cs (Figure 3). The high correlation between 90 Sr and 137 Cs indicates that 90 Sr and 137 Cs were derived from a common source. However, the low 90 Sr activity samples (<10 mBq L −1 ) showed larger variability in 90 Sr/ 137 Cs activity ratio (0.34 ± 0.14) relative to those for high 90 Sr activity samples (>10 mBq L −1 : ratio of 0.56 ± 0.08). If multiple sources to seawater exist, such as stagnant water, storage water, and groundwater, contributions from each source could yield temporal and spatial variations. To distinguish these components, detailed 134,137 Cs and 90 Sr distributions should be investigated. Castrillejo et al. [17] found a short-term transition of 90 Sr/ 137 Cs ratio from 0.14 to 0.36 and an abrupt increase in 137 Cs activity in the vicinity of the FDNPP (observation site St. 1 (or NP0)) in September 2013. 90 Sr and 137 Cs release from the FDNPP could be related to the tidal cycle and weather conditions, which caused a temporal variation of the released 90 Sr/ 137 Cs ratio from the FDNPP site. Fukushima Prefecture [6] was less than 0.04. The contribution to 90 Sr activity in seawater by input from the Ukedo River should be minor.

90 Sr/ 137 Cs Activity Ratios Derived from the FDNPP Accident
The 90 Sr/ 137 Cs activity ratios in seawater are different according to timing of any release or leakage event (e.g., direct discharge event from late March to early April 2011 [3,15]). Since Sr and Cs are highly soluble in seawater, the 90 Sr/ 137 Cs activity ratio depends on the source, which could be a useful tracer for the source. The most possible source of 90 Sr and 137 Cs is stagnant water in the reactor building of unit 2. The 90 Sr/ 137 Cs ratio of open ocean seawater [3], seawater monitoring data near the FDNPP [6], stagnant water [4], atmospheric input [1], seabed sediment [24], and river water [6] are summarized with our data in Figure 4. . The 90 Sr/ 137 Cs activity ratios in the environment and the FDNPP site. The 90 Sr/ 137 Cs activity ratio in seawater near the FDNPP was consistent with those from the monitoring points, T1 and T2 [12]. Soil [4] and sediment [24] samples had extremely low 90 Sr/ 137 Cs activity ratios.  The 90 Sr/ 137 Cs activity ratio in seawater near the FDNPP was consistent with those from the monitoring points, T1 and T2 [12]. Soil [4] and sediment [24] samples had extremely low 90 Sr/ 137 Cs activity ratios.
The slope of a linear regression fitting (0.66) was similar to the reported 90 Sr/ 137 Cs activity ratios in stagnant water of 0.78 and 0.88, respectively, in July and August of 2013 [5]. The stagnant water samples were collected from the sampling line behind the mixing point of water from each reactor building [4,5] (Figure 5). These radionuclides were thought to mainly be derived from the reactor building of unit 2 on the basis of the initial data for stagnant water in the unit 2 turbine building [2], which was severely damaged. The 90 Sr/ 137 Cs activity ratio in stagnant water varied depending on the decontamination of 134, 137 Cs, and gradually increased from the direct-release event in March 2011 (0.0256 ± 0.0006 [15]). The 90 Sr/ 137 Cs activity ratio of seawater in this study is slightly lower than that of stagnant water, although the most possible source candidate is the continuous release of stagnant water from the FDNPP.
The discrepancy between our data and monitoring data at T1 (1.25 ± 0.71) implies multiple sources exist at the FDNPP site. The higher 90 Sr/ 137 Cs at the T1 site could reflect a contribution from 90 Sr-rich groundwater. Groundwater around the reactor buildings had a 90 Sr/ 137 Cs activity ratio (2.4 × 10 4 [5]) that was 5 orders of magnitude higher than the seawater value observed in this study. The decontaminated water in storage tanks in the FDNPP was also observed to have high 90 Sr/ 137 Cs activity ratios.   [6]. Core inventory (0.74) was estimated by the ORIGEN2 code [2]. Shaded areas were averaged monitoring data at T1 (1.25 ± 0.71) and T2-1 (0.31 ± 0.14) sites from January 2013 to October 2013.
The 90 Sr activity of 0.80 ± 0.11 mBq L −1 was obtained at the offshore sites, S1, S2, S3, and N01 (Table 2). To evaluate FDNPP site-derived 90 Sr, measured 90 Sr activity was subtracted from this value as the background value for North Pacific seawater. The measured 134 Cs was a pure FDNPP sitederived component because of its short half-life (T1/2 = 2.06 yr). The FDNPP site-derived 134 Cs/ 137 Cs ratio was reported to be 0.99 ± 0.03 [11] in March 2011. FDNPP site-derived 137 Cs was calculated on the basis of measured 134 Cs activity and the FDNPP site-derived 134 Cs/ 137 Cs ratio.
The 90 Srcorr/ 137 Cscorr ratio estimated from the slope of a linear regression fitting was 0.66 ± 0.05 in Figure 5. 90 Srcorr activities strongly correlated with those of 137 Cscorr (R 2 = 0.919), as described in similar contour maps of 90 Sr and 137 Cs (Figure 3). The high correlation between 90 Sr and 137 Cs indicates that 90 Sr and 137 Cs were derived from a common source. However, the low 90 Sr activity samples (<10 mBq L −1 ) showed larger variability in 90 Sr/ 137 Cs activity ratio (0.34 ± 0.14) relative to those for high 90 Sr activity samples (>10 mBq L −1 : ratio of 0.56 ± 0.08). If multiple sources to seawater exist, such as stagnant water, storage water, and groundwater, contributions from each source could yield temporal and spatial variations. To distinguish these components, detailed 134,137 Cs and 90 Sr distributions should be investigated. Castrillejo et al. [17] found a short-term transition of 90 Sr/ 137 Cs ratio from 0.14 to 0.36 and an abrupt increase in 137 Cs activity in the vicinity of the FDNPP (observation site St. 1 (or NP0)) in September 2013. 90 Sr and 137 Cs release from the FDNPP could be related to the tidal cycle and weather conditions, which caused a temporal variation of the released 90 Sr/ 137 Cs ratio from the FDNPP site.
The slope of a linear regression fitting (0.66) was similar to the reported 90 Sr/ 137 Cs activity ratios in stagnant water of 0.78 and 0.88, respectively, in July and August of 2013 [5]. The stagnant water samples were collected from the sampling line behind the mixing point of water from each reactor building [4,5] (Figure 5). These radionuclides were thought to mainly be derived from the reactor  [6]. Core inventory (0.74) was estimated by the ORIGEN2 code [2]. Shaded areas were averaged monitoring data at T1 (1.25 ± 0.71) and T2-1 (0.31 ± 0.14) sites from January 2013 to October 2013. The fitted regression line had an x-intercept of 5.8 mBq L −1 . A very low 90 Sr/ 137 Cs ratio (0.16) was observed at S6 without a change in salinity. These results indicate that there is a missing source for the site with a low 90 Sr/ 137 Cs ratio. In the coastal region, salinities ranged from 33.2 to 33.3 psu and showed no correlation with activities of 90 Sr and 134, 137 Cs. Atmospherically deposited 90 Sr on land soil in March 2011 was at a lower level (<1.1 Bq g −1 ) than 137 Cs, where the 90 Sr/ 137 Cs activity ratio was considered to be 0.00008−0.017 [1] (Figure 5). Higher mobility of 90 Sr has been recognized, but 90 Sr activity in water of the Fukushima River was less than 4 mBq L −1 in 2012 [6]. 137 Cs activity ranged from 12 to 190 mBq L −1 , which yielded a low 90 Sr/ 137 Cs activity ratio of 0.01-0.04 [6]. Considering the 90 Sr/ 137 Cs activity ratio in seawater, riverine input from the land to the ocean was minor for 90 Sr, though dissolved 90 Sr activity in the Ukedo River was never reported.
A possible supply process for 137 Cs is the release from seabed sediments. Some amount of Cs could be scavenged by seabed sediments through adsorption onto particles, such as clay minerals [25][26][27] during the direct discharge event. The sedimentary 137 Cs inventory of 100-200 TBq represents only 1%-3% of the total discharge from the FDNPP to the Pacific Ocean in 2011 [28,29]. Approximately 80% of the total 137 Cs sedimentary inventory was found in coastal sediments at less than 150 m water depth [29]. The highest 90 Sr activity of 63 Bq kg-dry −1 in seabed sediments was observed near the south discharge gate of the FDNPP site (T2 monitoring point) in September 2011 [6]. Sedimentary 90 Sr/ 137 Cs activity ratios observed after the accident ranged from 0.001 to 0.08, which were lower than those in seawater. 137 Cs could be attributed to the direct discharge event in late March to April 2011 [20].
The extremely low 90 Sr/ 137 Cs ratio indicates that the contribution of 90 Sr in seawater from the soil and seafloor sediments is less than that of 137 Cs, even if there is a higher mobility for Sr than for Cs in the soil and sediments.
Another possible low 90 Sr/ 137 Cs source is contaminated water that remained in a tunnel for pipes and cables, which were connected to the turbine buildings of units 2 and 3. Contaminated water in the turbine and reactor buildings was released into the ocean via the tunnel and cracks resulting from the earthquake and tsunami. During this direct release event, the 90 Sr/ 137 Cs ratio was very low 0.0256 ± 0.0006 [15] (Figure 5). 137 Cs activity at T1 reached 68 kBq L −1 [20]. After the direct release was stopped in early April 2011 by sealing cracks and the tunnel entrance, contaminated water could have been left in the tunnel until July 2015. Such highly contaminated water could be the source of the low 90 Sr/ 137 Cs ratio.
The 90 Sr/ 137 Cs activity ratio of 0.66 ± 0.05 observed in this study was higher than data at the monitoring point, T2-1, near the south discharge gate (0.31 ± 0.14) [6] from January to December 2013, but was lower than that at T1 near the north discharge gate (1.25 ± 0.71) [6] (Figure 5). A large variation of 90 Sr/ 137 Cs at T1 was observed, which might reflect the local input processes of 90 Sr and 137 Cs. A much higher 90 Sr/ 137 Cs activity ratio (e.g., 90 Sr activity of 7.5 Bq L −1 and 90 Sr/ 137 Cs activity ratio of 3.2 on 26 June 2013) was observed by TEPCO in the harbor [6] than the coastal region as observed in this study. The variation in 90 Sr/ 137 Cs activity ratios might reflect the spatial and temporal heterogeneities of released water.
As mentioned above, the similarity of the 90 Sr/ 137 Cs activity ratio between seawater and the stagnant water supported the idea that the most likely candidate was the continuous release from the reactor buildings of the FDNPP. Both high 90 Sr activity and 90 Sr/ 137 Cs activity ratio in the coastal region reflect the input of the stagnant water. Variability of 90 Sr/ 137 Cs activity ratios in seawater is an important indicator to understand the status of the release of contaminated water from the FDNPP. Unfortunately, the contribution of underground water near the reactor buildings, and released from sediments to the harbor water, could not be distinguished from the release of the reactor buildings on the basis of seawater obtained from outside of the harbor. More detailed temporal and spatial data in the harbor and for other radionuclides such as tritium ( 3 H) and iodine-129 ( 129 I) are necessary.

Estimation of 90 Sr Input to the Ocean from the FDNPP
The continuous release from the FDNPP was the main source to the Fukushima coast. The amount of 137 Cs released daily to the ocean was estimated to be from 8.1 GBq day −1 [7] to 30 GBq day −1 [8] in 2012 on the basis of simulation of the 137 Cs activities of seawater in the harbor and at the north discharge gate, respectively. We examined the amount of released 90 Sr based on that of 137 Cs in 2013 by using Equation (1): where N, C, and F represent release rate, activity, and conversion factor from activity to daily release rate of 137 Cs, respectively. For the estimation of daily released 137 Cs, the activities of 137 Cs at T2-1 300 m south of the south discharge gate were used (Table 3) [6]. Most of the 137 Cs activities were lower than the MDA (1.2-1.5 Bq L −1 ). To avoid overestimation of the averaged 137 Cs activity, we used only precise analysis data. The 137 Cs activity at the T2-1 site ranged from 0.14 to 0.98 Bq L −1 and showed considerable variation (mean value = 0.60 ± 0.35 Bq L −1 ). The conversion factors, F, from activity to daily release rate of 137 Cs were obtained on the basis of the amount of released 137 Cs in the direct release event of March 2011 and 137 Cs activities at the T2-1 site [7,20]. The conversion factor applied was 25.5 × 10 9 [20] for the T2-1 site. The resulting daily released amount of 90 Sr was 9.1 ± 6.1 GBq day −1 during our sampling campaign in May 2013. The observed 90 Sr activity in the coastal region was too low to disturb the ecological system and affect the background radiation dose, as mentioned above. Continuous release could increase the inventory of 90 Sr in the Pacific Ocean. If the constant release (9.1 GBq day −1 ) continued over the year, the annual release rate would be estimated at 3.3 TBq yr −1 , which is small relative to the inventory of 105 PBq in the ocean [3]. However, 90 Sr in seawater should be closely observed to detect any unexpected release from the nuclear reactor buildings and the contaminated water storage tanks. This estimation needs to assume a stable release rate from the single source. As discussed above, the low 90 Sr/ 137 Cs source contributed to seawater around the FDNPP. Therefore, this result could be overestimated.
In this study, the 90 Sr/ 137 Cs activity ratio of 0.66 ± 0.05, which was influenced by continuous release from the FDNPP, was distinguished based on precise 90 Sr analysis. Buesseler et al. [11,30] suggested that the possible source of 137 Cs was not only continuous release from the FDNPP but also the input from subsurface groundwater [31], river water [32], and desorption from the marine sediments in the coastal region [29,30,33]. The environmental migration of 137 Cs through particulate and dissolved fluvial inputs, and remineralization from the sediments contaminated by direct discharge of stagnant water from 26 March to 6 April 2011, must also be taken into consideration. 90 Sr/ 137 Cs activity ratios could fluctuate according to the source in the FDNPP area and remobilization of 137 Cs in coastal water. In addition to monitoring for ongoing release from the reactor buildings and possible leakage of stored contaminated water in tanks, continuous measurement of 90 Sr is necessary for investigation of the migration of 137 Cs in the marine environment. A combination of other fission product nuclides, 129 I and 3 H activity, will provide precise information for the current status of leakages from stagnant water, groundwater, and stored water in tanks.

Estimation of Effective Dose Rate by Ingestion from Marine Products
90 Sr dispersion to the coastal area is the most serious issue for fisheries due to its radiotoxicity.
We estimated the dose impact to human health from marine products. The highest 90 Sr activity (29.13 mBq L −1 at AN7; Table 1) was comparable to typical levels for North Pacific surface seawater in the early 1960s during nuclear weapons testing [34]. Taking into consideration the processes in the food chain and the highest activity in the coastal water observed in this study (29.13 mBq L −1 at AN−7), we obtained Equation (2): where D is representative of the annual dose rate. C, CF, IR, and F are representative of the 90 Sr activity in seawater, the concentration factor from seawater to marine products (5-10 [35]), the intake rate of marine products (28.4 kg yr −1 [36]), and dose coefficient (2.8 × 10 −8 Sv/Bq [37]), respectively. It should be noted that these concentrations are quite small (0.23 µSv yr −1 ) compared with the International Commission on Radiological Protection (ICRP) limit of 1 mSv yr −1 for a member of the general public. Much higher 90 Sr activities were observed at monitoring points near the south (150−670 mBq L −1 ) and north (260−5800 mBq L −1 ) discharge gates [6]. Even this anomalously high 90 Sr activity (5800 mBq L −1 ) close to the FDNPP would contribute 46 µSv yr −1 to the annual effective dose rate by marine products.

Conclusions
90 Sr is useful as a tracer for continuous releases from the FDNPP site. We reported 90 Sr data in seawater along with 134 Cs and 137 Cs in samples collected in the coastal area off Fukushima Prefecture. Released 90 Sr was dispersed along the Fukushima coast, and the highest 90 Sr activity was 29.13 mBq L −1 at a sampling site 16 km south of the FDNPP. FDNPP site-derived 90 Sr/ 137 Cs ranged from 0.16 to 0.64 and the slope of a linear regression fitting of the relationship of Fukushima site-derived 90 Sr and 137 Cs was 0.66 ± 0.05, which was similar to the ratio of contaminated water in the FDNPP reactor and turbine buildings. These results suggest that the major contamination source is contaminated water in the FDNPP buildings. On the other hand, the l37 Cs-rich source could also affect seawater and cause temporal and spatial variations. The estimated release rate of 90 Sr (9.6 ± 6.1 GBq day −1 ) was small relative to the inventory of 90 Sr in the Pacific Ocean. Release of 90 Sr has been controlled by the water shielding wall between the reactor buildings and the harbor since 2015. However, our results imply that if any accidental release of radionuclides, including 90 Sr from the FDNPP, occurs during decommissioning of the reactors, the coastal area can be exposed to a high activity plume.