Oceanic dispersion of Fukushima‐derived radioactive cesium: a review
Abstract
This review summarizes the more than 70 papers published during the 4 years since the Fukushima Dai‐ichi nuclear power plant accident that occurred on 11 March 2011, and details the radioactive cesium dispersion pattern in the North Pacific and adjacent seas. The total amount of Fukushima‐derived radioactive cesium released into the North Pacific via atmospheric deposition and direct release, spatial and temporal changes in the Pacific coast around the accident site, and the concentration levels of radioactive cesium around the Japanese Islands, not only the Pacific coast but also in adjacent seas, such as Japan Sea, East China Sea are summarized. Based on observational data mostly obtained during 2 years since the accident, and simulation results, oceanic dispersion of radioactive cesium in the entire area of the North Pacific is described. The Fukushima‐derived radioactive cesium dispersed eastward as surface water and extended to the eastern side of the North Pacific in 2014, and was also observed via a southward intrusion to subsurface waters as Subtropical Mode Water and Central Mode Water. The radioactive cesium movement related to mode water is important in terms of the circulation of cesium into the ocean interior. Some new technologies and techniques concerning emergency monitoring of radioactivity in the ocean environment are also reported, the effectiveness of which has been demonstrated by use in relation to the Fukushima accident.
Introduction
On 11 March 2011, the Great East Japan Earthquake (Mw 9.0) occurred at the plate boundary off the coast of Tohoku, northeastern Japan. A huge tsunami was generated and caused 15 729 fatalities and 4539 missing in the Hokkaido, Tohoku and Kanto regions (The National Police Agency, as of 24 August 2011). Preliminary surveys reported tsunami waves with run‐up heights exceeding 30 m (Mori et al., 2011). The tsunami also hit the Fukushima Dai‐ichi Nuclear Power Plant (FNPP) sites located at 37˚25'N, 141˚02'E, and a loss of electric power at FNPP resulted in overheated reactors and hydrogen explosions. Radioactive materials were then released into the ocean through atmospheric fallout (such as aerosols and precipitation) and as direct releases (controlled releases related to safety issues at FNPP) as well as uncontrolled leaking of the heavily contaminated coolant water (Buesseler et al., 2011; Chino et al., 2011; Takemura et al., 2011). This accidental release of anthropogenic radionuclides (mostly iodine‐131, cesium‐134 and ‐137; 131I, 134Cs and 137Cs) resulted in severe elevations of these radionuclides in fisheries products in the coastal areas of Fukushima and adjacent prefectures (Buesseler, 2012; Yoshida and Kanda, 2012; Wada et al., 2013; Nakata and Sugisaki, 2015). Owing to its relatively long half‐life (2.07 years for 134Cs and 30.07 years for 137Cs), the evaluation of these radioactive Cs isotopes in the marine environment is important for addressing risks to both marine ecosystems and public health through consumption of fisheries products. Generally, cesium is a conservative element and mostly occurs in the dissolved phase in the marine environment. The concentration of radioactive cesium in marine organisms is strongly affected by its concentration in the surrounding seawater. Actually, temporal changes in radioactive Cs concentrations of many pelagic fish species in the near coastal area off Fukushima and adjacent prefectures were associated with those in seawater after the FNPP accident (e.g., Wada et al., 2013; Takagi et al., 2015; Morita et al., unpublished data). Kaeriyama et al. (2015) and Morita et al. unpublished data revealed the time‐lagged temporal changes in radioactive Cs in organisms (zooplankton and Pacific saury) and seawater under non‐steady‐state conditions after the FNPP accident, and showed that the concentration ratios in these organisms had been elevated when compared with those before the FNPP accident. With regard to zooplankton, Baumann et al. (2015) discussed the possible uptake of Fukushima‐derived radioactive Cs from phytoplankton dominated suspended particles. As a consequence, radioactive Cs would be transferred to the higher trophic level not only via surrounding seawater but also by prey‐predator interactions in the pelagic ecosystem. Shigenobu et al. (2014) reported the radioactive Cs concentrations of fat greenling (Hexagrammos otakii) caught off the coast of Fukushima Prefecture, and reported two outlier specimens caught in August 2012 and May 2013 which had ambiguously high 137Cs concentrations of more than 1000 Bq/kg‐wet. Probability analysis indicated that the two outlier fat greenlings had migrated from the port of FNPP. In the port of FNPP, extremely high 137Cs concentrations were reported from Japanese rockfish (Sebastes cheni), brown hakeling (Physiculus maximowiczi) and fat greenling (H. otakii) caught during January and February 2013 (Fujimoto et al., 2015). The maximum concentration of 137Cs (129 kBq/kg‐wet) was detected from fat greenlings. Wada et al. (2013) with the corrigendum (Wada et al., 2014) summarized the monitoring results of radioactive Cs concentrations in fisheries products from Fukushima Prefecture and revealed time‐series trends. Clear trends include a slower decrease of radioactive Cs in demersal fish compared to pelagic fish as well as spatial heterogeneity; specimens sampled in the area south of FNPP tended to have higher concentrations of radioactive Cs than those caught in the area north of FNPP. Sohtome et al. (2014) reported the time‐course trends in concentration of radioactive Cs in invertebrates in the coastal benthic food web near the FNPP. The difference in decreasing trends observed within the organisms and the concentrations of radioactive Cs in some of the sea urchins (Echinocardium cordatum and Glyptocidaris crenularis) were clearly affected by the contaminated sediments taken into their digestive tract.
This paper focuses on the radioactive Cs in seawater and summarizes estimates of the total amount of released radioactive Cs from the FNPP site, spatio–temporal changes in the concentrations of 134Cs and 137Cs not only off the coast of Fukushima and adjacent prefectures, but also in the North Pacific, and adjacent seas such as Japan Sea, East China Sea, based on measurement results and simulation models published during 4 years since the FNPP accident.
Total Amount of FNPP–Released Radioactive Cesium
Information on the total amount of the FNPP‐released radioactive Cs into the North Pacific is critical information to enable effective monitoring and resource management. However, despite its importance, estimation of atmospheric deposition is complex due to lack of the observational data in the oceanic environment. The activity ratios of 134Cs/137Cs, decay corrected to March–April 2011, were reported to be almost 1.0 for the entire North Pacific (e.g., Buesseler et al., 2011, 2012; Kaeriyama et al., 2014). This ratio means an equivalent amount of 134Cs and 137Cs was released into the ocean. Under the limitation of data concerning not only the amount of radioactive Cs in aerosols but also on precipitation in the North Pacific, estimation of atmospheric deposition remains a source of considerable uncertainty (5–15 PBq of 134Cs and 137Cs; 1 PBq = 1015 Bq, Table 1). In contrast, the direct release of radioactive Cs (134Cs and 137Cs) into the ocean as uncontrolled leaking of the heavily contaminated coolant water is well estimated as approximating the value of 3.5 PBq, with the exception of Bailly du Bois et al. (2012) and Charette et al. (2013) (Table 1). Dietze and Kriest (2012) discussed the possible overestimates by Bailly du Bois et al. (2012) as a result of methodological issues. Charette et al. (2013) estimated the direct release inventory from the observational data of radioactive Cs with radium isotopes in May–June 2011, and no atmospheric deposition was assumed. Their estimates of direct releases may be included in the atmospheric deposition. Tsumune et al. (2012) clearly showed that direct releases started on 26 March 2011 using 131I/137Cs activity ratios, which varied much more before 26 March 2011 when the atmospheric deposition was the major source. The most recent estimations have revealed that 3–4 PBq of 134Cs and 137Cs were directly released into the ocean and 12–15 PBq of 134Cs and 137Cs were deposited on the surface seawater in the North Pacific (Aoyama et al., 2015a).
| Source | Direct release | Atmospheric deposition on ocean surface | Reference |
|---|---|---|---|
| Fukushima accident | 2.8–4.2 | 12–15 | Aoyama et al. (2015a) |
| 12–42 | Bailly du Bois et al. (2012) | ||
| 11–16 | Charette et al. (2013) | ||
| 2.3 | Dietze and Kriest (2012) | ||
| 2.9–4.3 | Tsumune et al. (2013) | ||
| 4 | 5 | Kawamura et al. (2011) | |
| 4.1–4.5 | 5.7–5.9 | Estournel et al. (2012) | |
| 3.5 | 7.6 | Kobayashi et al. (2013) | |
| Global fallout | 76 | Buesseler (2014)aa
based on Aarkrog (2003).
|
|
| Close‐in fallout | 28 | Buesseler (2014)aa
based on Aarkrog (2003).
|
|
- a based on Aarkrog (2003).
Radioactive Cesium in the North Pacific before the FNPP Accident
Before the FNPP accident, the largest source of 137Cs was global fallout from nuclear weapons testing. The synoptic dataset on anthropogenic radionuclides in the North Pacific has been collated with the geochemical and hydrographic measurements of projects such as the GEOSECS program (Craig and Turekian, 1976). Bowen et al. (1980) summarized the vertical and horizontal distributions and inventory of anthropogenic radionuclides such as 137Cs and plutonium isotopes (239,240Pu) in the North Pacific which were derived from atmospheric nuclear weapons testing. The global fallout resulted in 220 PBq of 137Cs deposited in the North Pacific (Aarkrog, 2003), when decay corrected for 2011 just before the FNPP accident, 76 PBq of global fallout 137Cs had occurred in the North Pacific (Buesseler, 2014). Another fallout unique to the North Pacific was close‐in fallout from nuclear testing conducted on the islands of Bikini and Eniwetok in the 1950s. The close‐in fallout 137Cs accounts for an additional 28 PBq in the North Pacific in 2011 (Buesseler, 2014). As a consequence, FNPP‐derived 137Cs of 15–19 PBq were added to the global fallout and close‐in fallout 137Cs of 104 PBq in 2011 in the North Pacific. In the surface seawater, 137Cs levels showed a typical geographic distribution high in the mid‐latitude region due to the global and close‐in fallout and low in the equatorial Pacific during the 1950s and 2000s (Hirose and Aoyama, 2003). As a result of the dilution process associated with water movement and mixing, the concentration of 137Cs in the surface seawater had decreased with time and the horizontal distribution of 137Cs in surface seawater had been relatively homogenous over the entire North Pacific in the 2000s (Aarkrog, 2003; Hirose and Aoyama, 2003; Hong et al., 2004). In the 2000s, the 137Cs concentration in the surface water was almost at a level of 1.5–2.0 Bq m−3 (Hirose and Aoyama, 2003; Povinec et al., 2004). The effective half‐life of 137Cs was reported as 12.7 ± 1.4 years over the entire North Pacific (Hong et al., 2004). In the area off the coastal region of Japan, Kasamatsu and Inatomi (1998) reported the effective environmental half‐life of 137Cs as 18.7 years. In the mid‐latitude area in the North Pacific, the subarctic and subtropical gyres are located (Fig. 1a) (Kida et al., 2015). The winter convection of the surface water is another pathway of radioactive Cs, which subducts the surface water into the ocean interior. The intrusion of 137Cs with mode waters (Subtropical Mode Water, STMW and Central Mode Water, CMW) was observed in the early 2000s (Aoyama et al., 2008). These intruded water with elevated 137Cs levels into the ocean interior and were transported southward with water movement. Including the movement of 137Cs trapped in the mode waters, the transport process of 137Cs in the northern hemisphere to the southern hemisphere was studied with very low‐level measurement techniques of 137Cs in seawater [such as modification of the ammonium phosphomolybdate (AMP) co‐precipitation method and removal of radioactive potassium from the AMP/Cs compounds; Levy et al., 2011;] and simulation models, just before the FNPP accident (Nakano et al., 2010; Aoyama et al., 2011; Povinec et al., 2011; Sanchez‐Cabeza et al., 2011; Tsumune et al., 2011).

Fukushima‐Derived Radioactive Cesium in Near Coastal Area Off FNPP
Tokyo Electric Power Company Holdings, Inc. has carried out measurements of radioactive Cs in seawater at station T1 around the south discharge channel since 21 March 2011 (TEPCO; http://www.tepco.co.jp/en/nu/fukushima-np/f1/smp/index-e.html). From 23 March 2011, the sea area monitoring by the Japanese Government had begun in the coastal area near the FNPP site. The monitoring data of the Japanese Government obtained during the monitoring program have been published on the website by Nuclear Regulation Agency (NRA; http://radioactivity.nsr.go.jp/en/), previously by the Ministry of Education, Culture, Sports, Science and Technology (MEXT). Oikawa et al. (2013) modified monitoring results of MEXT by the addition of error information, re‐analysis of “not detectable (< 6000–9000 Bq m−3)” samples using the ammonium phosphomolybdate (AMP) co‐precipitation method (e.g., Aoyama et al., 2000). The oceanic conditions around the coastal area near the FNPP site, the mean and tidal currents were generally weak (Kubota et al., 1981). The offshore area of FNPP site is located in the Kuroshio‐Oyashio transition area, which is the region of the confluence of the two wind‐driven boundary currents of the North Pacific (Fig. 1). The Kuroshio, which is the subtropical gyre transporting warm, saline waters along the south coast of Japan and then eastward as its extension (Kuroshio Extension; KE), and the Oyashio, which is the subarctic gyre transporting, cold, less saline water southward (Yasuda, 2003). Under such oceanic conditions, the Fukushima‐derived radioactive Cs was considered to disperse southward along the coast, and then the strong eastward current, KE transports the Fukushima‐derived radioactive Cs eastward to the North Pacific (Fig. 1b). Figure 2 shows temporal changes of 137Cs in seawater off the coast near the FNPP site (Aoyama et al., 2012, 2015b; Kaeriyama, 2015; Kaeriyama et al., 2015; Kaeriyama, this study; NRA). The most remarkable changes in the concentration of 134Cs and 137Cs in the surface seawater were observed during the first six months of 2011 (Oikawa et al., 2013). Just off the FNPP site the 134Cs and 137Cs concentrations reached a maximum in the middle of April 2011 of up to 108 Bq m−3 and after that rapidly decreased (Fig. 2b; stations T1, T12, T18 and Iwasawa). Aoyama et al. (2012) reported the results of the weekly seawater monitoring of 134Cs and 137Cs collected at Hasaki, 180 km south of the FNPP site. The maximum radioactive Cs concentration at Hasaki was observed in June 2011 of up to 2000 Bq m−3, a delay of two months from the corresponding maximum value observed near the FNPP site. This delay in the timing of maximum values observed between the FNPP site, and Hasaki may be because of the anticyclonic eddy restricting the southward transport of Fukushima‐derived radioactive Cs (Aoyama et al., 2012). The radioactive Cs in Hasaki decreased with time, and an apparent half‐life of radioactive Cs was reported as 60 days (Aoyama et al., 2012). Kaeriyama et al. (2015) reported the radioactive Cs off Fukushima and adjacent prefectures (mostly within the area of 36–38˚N, 141–142˚E) during July 2011 and December 2013. In July 2011, the maximum concentration of 137Cs, 1800 Bq m−3 was observed and then rapidly decreased until December 2013 with an apparent half‐life of 85 days (Kaeriyama et al., 2015). Considerable variations were also observed within each sampling location of up to two orders of magnitude. The heterogeneity in the spatial distribution of radioactive Cs to the offshore may be due to the complex interactions of water masses, affected by the subarctic Oyashio and subtropical KE, which resulted in the creation of mesoscale eddies. In the area north of the FNPP site, time‐course monitoring of radioactive Cs had been conducted in Sendai Bay, 100 km north of the FNPP site (Fig. 2a). The maximum concentration of 137Cs of 2700 Bq m−3 was observed in June 2011, and also decreased with time (Kaeriyama et al., 2015). The apparent half‐life of 137Cs in Sendai Bay was estimated as 120 days (Kaeriyama et al., 2015). Fisheries Research Agency, Japan had begun weekly monitoring of radioactive Cs in seawater at Onahama station 35 km south from the FNPP site from May 2012 (Kaeriyama, 2015; Kaeriyama, this study). The concentration of 137Cs and its temporal changes are roughly comparable with those of adjacent stations of NRA monitoring (T12 and T18 in Fig. 2b). The concentrations of 137Cs had reached nearly background level (~ 2 Bq m−3) in late 2012 at offshore stations (36–38˚N, 141–142˚E), by late 2014 in Sendai Bay, and Hasaki (Fig. 2b). At the FNPP site (T1 in Fig. 2b), the 137Cs concentration has remained three orders of magnitude higher than the background level. In coastal areas such as T12, T18 and Onahama stations, also slightly higher 137Cs levels remain 3 or 4 years after the FNPP accident (Fig. 2b). The possible sources of radioactive Cs to these coastal areas (T1, T12, T18 and Onahama in Fig. 2) have been discussed such as the continuous leakage from the port of FNPP site (Kanda, 2013), and river discharge (Nagao et al., 2013; Kakehi et al., 2016), after the peak duration of heavily contaminated water release during the end of March from May 2011 (Tsumune et al., 2012). Continuous monitoring should be conducted to clarify how these continuous sources affect the concentration level of radioactive Cs in the coastal area around the FNPP site.

Inoue et al. (2014) discussed the subsurface peaks of Fukushima‐derived radioactive Cs observed based on radium isotope data in July 2012 along the northern Sanriku coast (Fig. 1b), further north from Sendai Bay. The subsurface peaks of radioactive Cs are considered to result from the atmospheric deposition of Fukushima‐derived radiocesium during March and April 2011 and lateral penetration of low‐level radioactive Cs water from the Japan Sea and/or Oyashio region during the following spring season. Inoue et al. (2012a) reported the radioactive Cs level in northern Sanriku and the Tsugaru Strait (39˚40’–41˚30'N; 250–450 km north of the FNPP site). In May 2011, 134Cs was detected in the range of 2 to 3 Bq m−3, the source of which was considered as atmospheric deposition from the FNPP accident. A rapid decrease in 134Cs concentration by 25–45% was observed in June 2011 due to the dilution process of water movement. In terms of the effects of direct release from the FNPP site, Kofuji and Inoue (2013) discussed the delivery mechanism of Fukushima‐derived radioactive Cs in the northern Sanriku area. They reported the rapid increase and decrease of radioactive Cs during June–October 2011 off the northern Sanriku coast (39–41˚N). Such a temporally variable change is considered to be associated with the low‐level radioactive Cs water from the Japan Sea with the clockwise gyre from the Tsugaru warm current and Oyashio region which was little affected by Fukushima‐derived radioactive Cs. These temporal changes in the concentration of Fukushima‐derived radioactive Cs were strongly affected by the direct release from the FNPP site. The regional ocean model with only the direct release source is well simulated with these spatiotemporal changes in radioactive Cs in the coastal area off Fukushima and adjacent prefectures (Masumoto et al., 2012; Miyazawa et al., 2012, 2013; Perianez et al., 2012; Tsumune et al., 2012, 2013; Rypina et al., 2013). Although the regional ocean models with atmospheric deposition have large uncertainties (see Table 3.2 and Fig. 3.2 in Science Council of Japan, 2014), the difficulty in precise estimations of the atmospheric deposition onto the coastal area off Fukushima and adjacent prefectures is still unresolved (Science Council of Japan, 2014).
Radioactive Cesium in Adjacent Seas
In contrast to the coastal and offshore areas around the FNPP site in the North Pacific, the elevation of 137Cs in seawater was almost negligible or only at trace levels in adjacent seas, such as the Japan Sea, East China Sea and the Bering Sea even just after the FNPP accident (Fig. 3a). Inoue et al. (2012b) reported the 134Cs and 137Cs levels in June 2011 in Japan Sea. Trace levels of 134Cs (~1.5 Bq m−3) were detected in the northeastern Japan Sea with 137Cs levels of 2.0–2.8 Bq m−3, higher than those in the southwestern Japan Sea (Fig. 3a). They concluded that Fukushima‐derived atmospheric deposition of radioactive Cs was observed only in the northeastern area of Japan Sea. Ramzaev et al. (2014) reported the background level 137Cs at two stations (1.6 and 1.7 Bq m−3) in the northern part of Japan Sea in April 2011, and 1.4–3.6 Bq m−3 just south of the Kuril Islands, Oyashio region during April and May 2011. 134Cs at a level of 0.3–2.6 Bq m−3 was observed in seawater samples collected in 2011 near the Kuril Islands and Kamchatka in the Oyashio region. They also reported the 137Cs level in August–September 2012 as 1.0–1.7 Bq m−3 in the northern part of the Japan Sea, near the Kuril Islands and Kamchatka in the Oyashio region, and 1.1 Bq m−3 at one station in the Okhotsk Sea. Inoue et al. (2012c) also reported a small amount of Fukushima‐derived radioactive Cs in the Okhotsk Sea. Kaeriyama (2015) reported only the background level 137Cs detection in the Japan Sea and the East China Sea during 2011 and 2012. Kim et al. (2012) reported the 137Cs level around the coastal area of Korea, in the Japan Sea and the Yellow Sea during April and July 2011 (Fig. 3a, b). The range of detected 137Cs concentrations in surface seawater was 1.2 to 2.6 Bq m−3, without 134Cs detection. In the East China Sea, the South China Sea, and the Yellow Sea, the 137Cs concentration ranged from 0.8 to 1.4 Bq m−3 during April and June 2011 (Fig. 3a, Wu et al., 2012). Wu et al. (2012) concluded that the Fukushima‐derived radioactive Cs input through atmospheric deposition into the East China Sea was negligible. In coastal areas of Indonesia, Suseno and Prihatiningsih (2014) reported 137Cs in the seawater collected during June 2011 and June 2013. The 137Cs concentration in seawater was 0.1–0.3 Bq m−3 and concluded that the background level was due to global fallout and that Fukushima‐derived radioactive Cs had not been reached to the coastal areas of Indonesia.

Surface Dispersion of Radioactive Cesium in the North Pacific
Over a broad area of the North Pacific, the concentration of Fukushima‐derived radioactive Cs in surface seawater has been already reported (e.g., Buesseler et al., 2012; Honda et al., 2012; Inoue et al., 2012a; Aoyama et al., 2013a,b, 2015b; Charette et al., 2013; Kaeriyama et al., 2013, 2014, 2015; Kamenik et al., 2013; Kitamura et al., 2013; Kumamoto et al., 2013, 2014, 2015a,b; Povinec et al., 2013a,b; Ramzaev et al., 2014; Smith et al., 2014; Kaeriyama, 2015; Men et al., 2015; Yoshida et al., 2015; Yu et al., 2015). Figure 3 shows the 137Cs concentration in surface seawater during March–June 2011 (Fig. 3a), during July–December 2011 (Fig. 3b), during January–June 2012 (Fig. 3c) and during July–December 2012 (Fig. 3d) in the North Pacific and adjacent seas, the data were compiled from the references listed above. Briefly, the surface dispersion of the Fukushima‐derived 137Cs can be described as three phases; first atmospheric deposition resulted in a very patchy distribution over a broad area of the North Pacific and was immediately diluted, in the second phase, the direct released plume of Fukushima‐derived radioactive Cs (> 10 Bq m−3 of 137Cs) travelled eastward in the mid‐latitude, thirdly with the weakening with eastward dispersion of KE flow in the central part of the North Pacific, the north‐south dispersion was enhanced, and the concentration of 137Cs decreased markedly.
During the first phase (Fig. 3a; March–June 2011), the high concentration of 137Cs (>150 Bq m−3) was observed not only offshore of the FNPP site (Buesseler et al., 2012; Oikawa et al., 2013), but also in the subtropical area (Honda et al., 2012), central North Pacific, and in the eastern North Pacific (Aoyama et al., 2013a,b). A considerably high 137Cs concentration of 196 Bq m−3 was observed at 43˚N 180˚ in May 2011 reported by Aoyama et al. (2013b) (Fig. 3a). These patchy distributions of 137Cs are considered as a result of atmospheric deposition soon after the FNPP accident, and the model prediction of the atmospheric deposition also supported this scenario (e.g., Kawamura et al., 2011; Honda et al., 2012; Kobayashi et al., 2013; Tsumune et al., 2013). The relatively high concentrations of 137Cs (> 10 Bq m−3) were also observed in the subtropical region, south of KE during October–November 2011 (Kaeriyama et al., 2013). In terms of surface dispersion, the KE is considered to act as a barrier against the southward dispersion of Fukushima‐derived 137Cs in the North Pacific. The Fukushima‐derived 137Cs in the subtropical region may be transported via atmospheric deposition into this area with less dilution because of the weaker water exchange processes than those in the transition region, and also associated with mesoscale eddies that separated from KE entraining high 137Cs concentration water from the transition area to the subtropical region. The heterogeneity in the horizontal distribution of 137Cs around mesoscale eddies in both areas north and south of KE are discussed not only from observational data (Fig. 4; Kaeriyama et al., 2013) but also a high‐resolution oceanic model (Prants et al., 2011).

The second phase (Fig. 3b; July–December 2011), the eastward dispersion of Fukushima‐derived 137Cs was reported mainly in the area north of KE (Aoyama et al., 2013b; Kaeriyama et al., 2013). In the mid‐latitude, around 40˚N, high concentrations of 137Cs more than 50 Bq m−3 were observed in the western–central North Pacific. From the temporal changes of the 134Cs concentration, Aoyama et al. (2013b) estimated the eastward speed of the Fukushima‐derived Cs plume as 8 cm sec−1 in the western North Pacific, this estimated speed was consistent with the one estimated by ARGO float data. Such an eastward dispersion of Fukushima‐derived radioactive Cs in the area north of KE was well simulated based on the global circulation models (Miyazawa et al., 2012; Tsumune et al., 2013; Kawamura et al., 2014).
After January 2012, concentrations of 137Cs off the east coast of the Japanese Islands, except for the coastal area of FNPP site markedly decreased due to the supply of waters with low or no Fukushima Cs content from the Kuroshio and Oyashio (Fig. 3c). Relatively high concentrations of 137Cs (5–30 Bq m−3) were observed in the area of 30–45˚N, 142˚E–170˚W. These 137Cs levels are considered mainly to be due to the direct release from the FNPP site and dispersion within the transition region, north of KE as mentioned above. Until December 2012, the main body of this Fukushima‐derived radioactive Cs was located in the central part of the North Pacific (Kaeriyama et al., 2013; Kumamoto et al., 2015b) (Fig. 3d). On the eastern side of North Pacific, Smith et al. (2014) reported the intrusion of Fukushima‐derived 137Cs offshore of the Canadian coast firstly in June 2012 at a trace level (0.2–0.4 Bq m−3) which was estimated from the data of 134Cs and 134Cs/137Cs ratio of Fukushima‐derived radioactive Cs, and in February 2014 at the maximum level of 137Cs as 2.1 Bq m−3, which includes background 137Cs and Fukushima‐derived 137Cs. Yoshida et al. (2015) reported the Fukushima‐derived radioactive Cs in the vicinity of the Californian coast in 2014. The first recording of Fukushima‐derived radioactive Cs in the eastern side of the North Pacific shows good agreement with the eastward speed of Fukushima‐derived radioactive Cs as 5 cm sec−1 in the central and eastern North Pacific estimated by Aoyama et al. (2015b). Some of the global circulation models (Behrens et al., 2012; Rypina et al., 2014) also closely simulated the results of Smith et al. (2014) and Yoshida et al. (2015). Although Fukushima‐derived 137Cs in seawater has dispersed across the North Pacific, the concentrations of 137Cs including the background 137Cs on the eastern side of the North Pacific were at a comparatively low level (~ 2.1 Bq m−3) because of the dilution process over the entire North Pacific. No Fukushima‐derived radioactive Cs (134Cs and 137Cs) in marine organisms has been observed from the eastern side of North Pacific. The exceptions were detection of 134Cs from samples of Pacific bluefin tuna, Thunnus orientalis collected from the coast of California in August 2011 (0.9–6.2 Bq kg dry−1; Madigan et al., 2012), and Pacific albacore, Thunnus alalunga, collected at the coast of Oregon and Washington during summer of 2012 (0.018–0.36 Bq kg‐wet−1; Neville et al., 2014) which were considered to have migrated from the high radioactive Cs area (the western North Pacific) to the eastern North Pacific during 2011.
Overall, the Fukushima‐derived radioactive Cs in the surface seawater was dispersed widely in the North Pacific. The concentrations of 137Cs were elevated to be more than 100 Bq m−3 in the western North Pacific during the 1st year from the accident, because of the dilution process with the water movement. However the levels were below 10 Bq m−3 during the 2nd year from the accident. The simulation models allowed the dispersion features to be re‐drawn for the Fukushima‐derived radioactive Cs in the surface seawater over the entire North Pacific.
Subsurface Intrusion of Radioactive Cesium with Mode Waters
Before the FNPP accident, Aoyama et al. (2008) found two sub‐surface peaks of 137Cs, which derived from nuclear weapons testing (hereafter ‘bomb‐derived 137Cs’), at the potential density ranges of 25.0–25.6 σθ and 26.0–26.6 σθ. These sub‐surface peaks of bomb‐derived 137Cs were observed in the Subtropical Mode Water (STMW) and Central Mode Water (CMW). Mode waters are characterized as vertically homogenous layers, i.e., pycnostads, covering a large geographical area, and thus are dominant components of the permanent pycnocline (Suga et al., 2008). The mechanisms forming the subsurface peaks of 137Cs with mode waters are considered as follows; winter cooling of surface seawater results in vertical mixing of the water column, the bomb‐derived 137Cs in the surface water entered the depth of > 100 m during the winter season, after that the elevation of surface temperature during spring–summer seasons, results in the formation of the thermocline which separates the surface water from the subsurface water. In the surface seawater, the dilution process results in a low and homogeneous distribution of bomb‐derived 137Cs. In contrast, the subsurface waters with bomb‐derived 137Cs were less diluted because of the vertically homogenous characteristics. More than 20 years from the nuclear weapons testing era (mostly 1950s–1960s, last test in 1980), Aoyama et al. (2008) successfully illustrated the subsurface intrusion of bomb‐derived 137Cs within the STMW and CMW in the North Pacific.
After the FNPP accident, subsurface intrusion of Fukushima‐derived radioactive Cs associated with these mode waters has been reported (Kitamura et al., 2013; Kumamoto et al., 2013, 2014; Kaeriyama et al., 2014, 2016; Aoyama et al., 2015b; Men et al., 2015; Yoshida et al., 2015). Kitamura et al. (2013), Kumamoto et al. (2013, 2014) reported the subsurface peak of Fukushima‐derived radioactive Cs within STMW along 149˚E during December 2011–February 2012, in the subtropical region far south from the FNPP‐site and just south of KE. Their findings revealed that the Fukushima‐derived radioactive Cs had intruded soon after the accident, during late winter of 2011 (March–April 2011). In the southwest area of the FNPP site (12–30 ˚N, 135–138˚E), Kaeriyama et al. (2014) reported subsurface peaks of Fukushima‐derived radioactive Cs in STMW. They reported the seasonal change of 137Cs inventory within the 0–500 m depth and an increase in the inventory after the second winter season from the FNPP accident (December 2011–April 2012). They suggest that the Fukushima‐derived radioactive Cs in STMW had intruded in the subtropical area after the second winter season from the FNPP accident. Men et al. (2015) reported the Fukushima‐derived radioactive Cs (9.7 Bq m−3 of 134Cs) in STMW at 21˚30'N, 125˚E during May–June 2012. Kaeriyama et al. (2016) reported the Fukushima‐derived radioactive Cs during October–November 2012 in the 147˚E and 155˚E transects, areas where STMW formation occurs. Most of the Fukushima‐derived radioactive Cs occurred in the STMW and CMW rather than in the transition area during June–November 2012 (Fig. 5; Kaeriyama et al., 2016). They estimated the total amount of Fukushima‐derived radioactive Cs (134Cs) in the STMW in October 2012 as 4.2 ± 1.1 PBq decay corrected to April 2011, which accounts for 22–28% of the total deposition of Fukushima‐derived radioactive Cs (15–19 PBq, Aoyama et al., 2015a). Kumamoto et al. (2014) estimated the total amount of 134Cs in STMW during December 2011–February 2012 as 6.0 PBq, which intruded in the STMW during March and April 2011. Since STMW is lighter than CMW, it outcropped into the surface layer during the next winter season; furthermore, the volume of STMW varies year to year (Oka et al., 2015). The estimated value of 4.2 PBq of 134Cs in STMW by Kaeriyama et al. (2016) probably represents the results of two winter–spring periods of deep mixing (March–April 2011 and February–April 2012), as well as one period of outcropping in the surface water during the second winter–spring time interval (February–April 2012). In contrast, as CMW is denser than the STMW, CMW is less likely to outcrop and effectively traps the surface water contents into the ocean interior, when compared with the STMW (Oka et al., 2011). In the formation area of CMW, Aoyama et al. (2015b) reported the Fukushima‐derived radioactive Cs both in STMW and CMW during June 2012 along the 165˚E transect. Most of the Fukushima‐derived radioactive Cs (more than 80% of the water column inventory) had been distributed in the CMW at 165˚E transect. Kaeriyama et al. (2016) also reported the Fukushima‐derived radioactive Cs in the CMW at just north of KE along 155˚E during October–November 2012. Yoshida et al. (2015) reported the subsurface peak of Fukushima‐derived radioactive Cs around 300 m depth west of 175˚E along 30˚N during March–May 2013. Their observations revealed the subsurface penetration of radioactive Cs in the potential density range of CMW, the time scale for CMW to be transported to the observational area is too long (10 years) compared with actual observation timing (3 years from the accident). They discussed another possible pathway of this Fukushima‐derived radioactive Cs penetration to the subsurface water as North Pacific Intermediate Water (e.g., Talley, 1993). The subsurface penetration pathways of Fukushima‐derived radioactive Cs have not yet been clarified. These subsurface trapped radioactive Cs derived from the FNPP accident need to be monitored in the future, because of their relatively large portion of the total deposition onto the North Pacific as is discussed by Aoyama et al. (2015b) and Kaeriyama et al. (2016) (Fig. 5). The model simulation studies also reported the intrusion of the Fukushima‐derived radioactive Cs in the North Pacific (e.g., Nakano and Povinec, 2012; Rossi et al., 2013, 2014). Although discrepancies have been observed between observation (Kaeriyama et al., 2014) and simulation results (Nakano and Povinec, 2012) especially in the timing of the Fukushima‐derived radioactive Cs intrusion, the observational data are useful to validate the models (see Kaeriyama et al., 2014). Furthermore, Nakano and Povinec (2012) used an old version of the ocean general circulation model with a rough resolution and poor parameterization, their simulated results of the surface dispersion of Fukushima‐derived radioactive Cs showed large differences in the observed results, especially in the subarctic region. On the other hand, Rossi et al. (2013) had missed the source term value; the corrigendum was in Rossi et al. (2014). Although all concentration data in Rossi et al. (2013) need to be read as 1/10, their model well described the transport processes of Fukushima‐derived radioactive Cs such as the eastward dispersion in surface water and also the time‐course intrusion into the ocean interior. To develop a circulation model dealing with the Fukushima‐derived radioactive Cs in the North Pacific, updated information concerning Fukushima‐derived radioactive Cs as atmospheric deposition, the subsurface distribution will be essential in the future.

New Technologies and Techniques
After the FNPP accident, real‐time measurement techniques for radioactive Cs were developed (Caffrey et al., 2012; Pike et al., 2012). Caffrey et al. (2012) developed an onboard real‐time approximation system of the radioactive Cs concentration, enabling recording of occurrences of areas with high radioactive Cs concentrations. The system consists of a NaI(Tl) scintillation detector with a compact digital spectroscope connected to a seawater tank through which a continuous flow of surface seawater is entrained. The system was deployed off the FNPP site and detected 137Cs ranging from 800 to 3800 Bq m−3 in June 2011. Pike et al. (2012) modified the AMP (ammonium phosphomolybdate) resin analysis method; the AMP was incorporated with organic polymer polyacrylonitrile (PAN; Šebesta and Štefula, 1990). They also tested the PAN‐AMP resin method in June 2011 off the FNPP site, and samples with measured activities of Fukushima‐derived radioactive Cs ranging from a few Bq m−3 to >3000 Bq m−3 were monitored. The modified PAN‐AMP resin method enabled a reduction in time from sampling to analysis and thus allowing for large numbers of samples to be processed. Lutter et al. (2015) focused on metrological aspects of gamma‐ray spectrometry analysis with a trace level detection limit of radioactive Cs performed in underground laboratories. In terms of measurement of radioactive Cs, they concluded that the best detectors with a low‐background environment (e.g., underground laboratories), optimization of detection efficiency, and reduction of the potassium‐40 intrinsic activity is used for samples with very low activities and detectors with lower efficiencies and higher background are used for samples containing the highest activities.
Inomata et al. (2014) successfully estimated the 131I, 134Cs and 137Cs levels in the sea surface using data obtained from an aerial survey by U.S. Department of Energy National Nuclear Security Administration on 18 April 2011, soon after the Fukushima accident (Lyons and Colton, 2012). They analyzed the surface dispersion of radioactive Cs in seawater and estimated the areal inventory based on correlations between gamma‐ray peak count rates determined by the aerial survey and in‐situ activities of 131I, 134Cs and 137Cs in seawater samples. The estimated inventory of 137Cs 12 days after the Fukushima accident indicated that more than one‐third of the directly released 137Cs as uncontrolled leaking of the heavily contaminated coolant water remained in an offshore area less than 30 km from the FNPP site. These new technologies will enhance understanding the radioactive contamination levels soon after any unexpected accident such as the Fukushima accident, and also trace the released radionuclides into the ocean environment as far from the accidental site as low level because of the dilution process in the open ocean.
Concluding Remarks
During the more than 4 years from the FNPP accident, many papers that have dealt with an oceanic dispersion of Fukushima‐derived radioactive Cs have been published. In the coastal area near the FNPP site, a rapid decrease of Fukushima‐derived radioactive Cs in seawater was observed within a half year from the accident, although higher concentrations of radioactive Cs than the background levels measured before the accident are still detected. Continued monitoring is necessary, and the possible sources of the continued releases of Fukushima‐derived radioactive Cs into the coastal area around FNPP site should be clarified such as via river discharge. Such information is essential to discuss the temporal changes in radioactive Cs concentrations of pelagic organisms in the coastal area around FNPP site (e.g., Wada et al., 2013; Kaeriyama et al., 2015; Takagi et al., 2015; Takata et al., 2015; Miki et al., 2016; Morita et al., unpublished data). In the North Pacific, the Fukushima‐derived radioactive Cs in surface seawater has dispersed eastward in the mid‐latitude, mainly because of the wind‐driven surface currents, and has already been recorded on the eastern side of North Pacific. The subsurface intrusion and southward transport had also been reported as a result of intrusion and formation of mode waters. The overview of the dispersion pattern of Fukushima‐derived radioactive Cs has been well documented from observational data and model simulation studies during the 4 years since the accident. The dispersion pattern of Fukushima‐derived radioactive cesium will be clarified that the timing and total amount of Fukushima‐derived radioactive Cs intruded to the mode waters and their transport processes because some discrepancies observed between the observational data and model results. The heterogeneity in the surface distribution of Fukushima‐derived radioactive Cs around the mesoscale eddies observed in the year 2011 should be clarified by meso‐, sub‐mesoscale resolving ocean models. Furthermore, estimation of atmospheric deposition from the FNPP site into the North Pacific needs to be better understood to assess the effect of this accident as in terms of the ‘mass balance’ of Fukushima‐derived radioactive Cs released into the environment.
Acknowledgement
I would like to thank four anonymous reviewers for comments that improved the manuscript. This study was supported by the Fisheries Agency, Ministry of Agriculture, Forestry and Fisheries, Japan.




