Spatiotemporal distribution and fluctuation of radiocesium in Tokyo Bay in the five years following the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident

A monitoring survey was conducted from August 2011 to July 2016 of the spatiotemporal distribution in the 400 km2 area of the northern part of Tokyo Bay and in rivers flowing into it of radiocesium released from the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident. The average inventory in the river mouth (10 km2) was 131 kBq⋅m-2 and 0.73 kBq⋅m-2 in the central bay (330 km2) as the decay corrected value on March 16, 2011. Most of the radiocesium that flowed into Tokyo Bay originated in the northeastern section of the Tokyo metropolitan area, where the highest precipitation zone of 137Cs in soil was almost the same level as that in Fukushima City, then flowed into and was deposited in the Old-Edogawa River estuary, deep in Tokyo Bay. The highest precipitation of radiocesium measured in the high contaminated zone was 460 kBq⋅m-2. The inventory in sediment off the estuary of Old-Edogawa was 20.1 kBq⋅m-2 in August 2011 immediately after the accident, but it increased to 104 kBq⋅m-2 in July 2016. However, the radiocesium diffused minimally in sediments in the central area of Tokyo Bay in the five years following the FDNPP accident. The flux of radiocesium off the estuary decreased slightly immediately after the accident and conformed almost exactly to the values predicted based on its radioactive decay. Contrarily, the inventory of radiocesium in the sediment has increased. It was estimated that of the 8.33 TBq precipitated from the atmosphere in the catchment regions of the rivers Edogawa and Old-Edogawa, 1.31 TBq migrated through rivers and was deposited in the sediments of the Old-Edogawa estuary by July 2016. Currently, 0.25 TBq⋅yr-1 of radiocesium continues to flow into the deep parts of Tokyo Bay.


Introduction
Tokyo Bay is a closed bay that extends 70 km from north to south and 20 km from east to west, covers a total area of 1,380 km 2  Pacific Ocean by a 7 km wide strait at its south end. The average retention time of seawater varies seasonally but is reported to be approximately 31 days [1]. Central Tokyo is located on the west side of the bay, which is surrounded by a zone of large cities that forms the heart of Japan, and has a total population of 38 million. The catchment basins of rivers flowing into Tokyo Bay from the greater Tokyo region occupy a land area of 9,100 km 2 , and the total quantity of inflowing river water fluctuates greatly, but averages approximately 1.4 × 10 7 m 3 Áday -1 . The major rivers are the Edogawa, Old-Edogawa, Arakawa, Tamagawa, Sumidagawa, and Tsurumigawa. Even though Tokyo Bay is closed, its seawater flow is complex. In addition to tidal currents, permanent currents flow throughout the bay, and the surface water movement is dominated by circular drifts: clockwise in the winter and counterclockwise in the summer. The bottom water moves in the opposite direction to the flow of the surface water. The pelagic water from the Pacific Ocean flows north on the bottom inside the bay until it reaches the Bay's deepest section [1]. Aircraft monitoring of the 134+137 Cs precipitation was conducted by the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) [2], and the results were published by the Geospatial Information Authority of Japan (GSI) [3], showing that the catchment basin of Edogawa River was contaminated from 30 to 100 kBqÁm -2 by radiocesium discharged from the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident, but the radioactive contamination levels in the catchment basins of the Bay's other rivers were lower than those in the catchment basin of Edogawa. Radioactive materials precipitated on the ground surface in the greater Tokyo region are, as in the case of artificially discharged environmental contaminants, presumably carried by these rivers until they finally flow into Tokyo Bay.
Many reports outlining the FDNPP accident have already been released to the public [4][5][6][7]. However, many of these are analyses of the accident process, whereas few address the environmental radioactive contamination that was caused [2,[8][9][10]. In particular, the movement in the greater Tokyo region of radioactive contamination produced by the FDNPP accident has been insufficiently analyzed. Nevertheless, after the accident, high concentration radioactive plumes arrived in the greater Tokyo region, radionuclides washout with precipitation (rainfall) on March 16 and 22 in 2011 [11]. Clarifying the movement of environmental radioactive contaminants in the heavily populated greater Tokyo region is an important issue related to the problem of low dose exposure to large populations. Evaluation of the migration process of radiocesium from the Tokyo metropolitan area also is important from the viewpoint of reduction and decontamination of radioactive contamination in these areas. In our previous paper, the behavior of radioactive contaminants of the soil in the Tokyo metropolitan area was discussed [12]. It is estimated that 10 to 22% of the radiocesium precipitated in the surface soil and migrated to Tokyo Bay via rivers in the five years after the FDNPP accident.
This study was a continuous time-series analysis of the distribution and fluctuation of radiocesium in sediments and waters in Tokyo Bay and in the rivers flowing into Tokyo Bay starting in August 2011, immediately after the FDNPP accident. Based on the results, the process of the movement from the land and deposition in Tokyo Bay of radiocesium that was precipitated in the greater Tokyo region via the FDNPP accident were evaluated. However, before the FDNPP accident the Chernobyl accident and the Three Mile Island (TMI) accident affected many affecting people. TMI was located about 150 km west of Washington DC but because it avoided the destruction of the pressure vessel, the emission of radioactive nuclides to the environment is only rare gas nuclides, and the release amount of 131 I is estimated to be about 0.5 TBq [13,14]. In the case of the Chernobyl accident, Kiev City was located 130 km south, and 4 million residents lived in that metropolitan area. It was reported at the Chernobyl Forum by the IAEA in 2006 [15] that radioactive plume flew to Kyiv City by the north wind on May 1, 1986 immediately after the accident. However, regulations on information disclosure were made by the Soviet government at the time, and the actual state and dynamics of radioactive contamination in Kiev City are still hardly understood even now. Of course, do not forget the radioactive contamination by the atomic bombs of Hiroshima and Nagasaki. The results of our investigation on the environmental dynamics of 60 years after radiation exposure in Nagasaki has already been reported [16][17][18]. From such a viewpoint, we think that the FDNPP accident was the first time that an urban region as densely populated as Tokyo was contaminated by radioactive material over a wide area.
Studies on the behavior of radiocesium in an urban environment have been performed through simulations [19], but fluctuation in this radionuclide's spatiotemporal distribution has not been monitored nor analyzed over wide areas for long periods. Furthermore, the behavior of cesium as an alkaline element is often complicated and unknown in the estuary where seawater and river water mix [20,21]. In this study, the important roles that Tokyo Bay and rivers flowing into it play in the movement of radiocesium contaminants and the transport and accumulation mechanisms of such in the greater Tokyo region have been clarified.
The Nuclear Regulation Authority of the Japanese Government (NRA) has monitored the radioactive contamination derived from the FDNPP accident in the surface sediment of Tokyo Bay since June 2013 [22]. The Japan Coast Guard (JCG) has also been measuring the radioactive contamination of surface sediments in Tokyo Bay since 1981 [23]. On the other hand, survey results of radiocesium contamination in the Tokyo Bay area immediately after the accident have been published [24]. However, since their monitoring is limited spatiotemporally, it is insufficient to evaluate the dynamics of radioactive contamination throughout the environments of Tokyo Bay.

Material and methods
Sediment and water were sampled in Tokyo Bay and in the rivers in its catchment basin. The locations are shown in Fig 1. Sampling was performed at the same points one to seven times during the study period, which ran from August 20, 2011, until July 12, 2016. Sediment samples were collected at 77 points in Tokyo Bay, 10 points in Edogawa River, and 6 points in Sakagawa River. Of the sediment samples collected, 68 were core sediments and 142 surface sediments. Sediment cores were sampled at Point S1 (Fig 1), where Sakagawa flows into Edogawa, to evaluate the role of Sakagawa in the process of transporting radiocesium. To compare with sediment, soil samples were also collected from the 14 points shown in Fig 1B. Sediment core sampling was done using an acrylic pipe with a diameter of 10 cm and length of 100 cm. The core samples were collected by a diver pushing the pipe into the seabed by hand. Core samples of 20 to 80 cm length were obtained. Surface sediment specimens were sampled from a boat using an Ekman-Birge bottom sampler. Then, on the boat, after the sediments were collected, the material was inserted into an acrylic pipe with a diameter of 5 cm and length of 10 cm to obtain samples a top 5 cm sediment. Most of soil and sediment samples consisted of silt and sand with a particle size of 2 mm or less. However, more pebbles, plant pieces, shell fragments, etc. were removed with tweezers. Grain size sorting by sieve was not done. The sediments were pushed out of the pipes, cut into 1 or 2 cm thick slices in the depth direction, and then thermally dried to a constant weight in a 60˚C oven to remove the water from the sediments. The dried samples were pulverized in an agate mortar, and then the radioactivity of the samples was measured. Water samples were obtained from the surface of the water by lowering buckets from boats. Divers obtained bottom water from about 1 m above the seabed. Without filtering suspended materials out of the water, the radiocesium in 20 L of sample water was concentrated using an ammonium phosphomolybdate (AMP) method [25]. After standing overnight, the AMP precipitate was filtrated and collected on a membrane filter   (pore size 0.8 μm), then the radioactivity of the dried AMP precipitate was measured. In this way, it was confirmed in a preliminary experiment that the ionic and suspended radiocesium in the sample water can be recovered quantitatively.

Measurement of radioactivity
Radionuclides in the samples were quantified by connecting a 4096-multichannel pulse height analyzer (Lab Equipment, MCA600) to a low energy HPGe detector (ORTEC, LO-AX/30P) shielded in lead 10 cm thick, sealing the specimens inside a plastic container with a diameter of 5.5 cm and depth of 2.0 cm, then measuring them via γ-ray spectrometry. The Ge detector calculated the geometric efficiency relative to the sample weight using the American NIST (National Institute of Standards and Technology) Environmental Radioactivity Standards, SRM 4350B (River Sediment) and SRM 4354 (Freshwater Lake Sediment), and the efficiency was corrected to within a range of 2 to 30 g of the sample weight [26]. The measurement time was set so that the counting error would be less than 5% according to the radioactive intensity of the samples. 134 Cs (605 keV) and 137 Cs (662 keV) were quantified in this study. A 134 Cs solution with known concentration was used to correct the sum peak effect for 134 Cs counting. The detection limits of 134 Cs and 137 Cs under appropriate conditions were 0.6 BqÁkg -1 in sediment samples and 0.3 mBqÁL -1 in water samples. Radiocesium activity was indicated by the values per sampling day, but was corrected for radioactive decay to the value on March 16, 2011, as necessary. In that case, it is denoted as "corrected activity."

Measurements of heavy metals and particle size distributions in the sediments
The heavy metals in the sediments were measured via an XRF method (Rigaku, ZSX-Primus I) using the NIST SRM 1646 (Estuarine Sediment) as the standard sample. Sample measured were made from cellulose powder pressed into 4 cm diameter aluminum ring 0.4 tonÁcm -2 , and then 1.2 g of powdered sample was placed on the disk and repressed at 1.6 tonÁcm -2 . The correction of matrix effect was achieved by X-ray intensity ratio of peak to back ground for each element [27]. Mercury in the sediments was measured via a heating-vaporization atomic absorption spectrometry (Hiranuma, HG-300). The particle size distribution of the wet sediment samples was measured using a laser diffractometer (Shimadzu, SALD-3000) with a measurement range of particle size 0.05 to 3000 μm. Dispersion of sedimentary particles was carried out via ultrasonic irradiation (Shimadzu, SUS-200, 42 kHz) using sodium hexametaphosphate as a dispersant. In this paper, the particle size obtained is presented as the volume-based average particle diameter.

Spatiotemporal distribution of radiocesium in Tokyo Bay sediment
All measured values obtained in this study are shown in S1-S4 Tables of the supporting information file. Geographic coordinates of sampling points are also shown in S5  Table) shows the spatial distribution of the 134+137 Cs activity (total value of 134 Cs and 137 Cs) in the surface layer of the sediment, from 0 to 5 cm depth. When multiple measurements were done at the same point, the 134+137 Cs activity was evaluated based on the value in a weighted average with the counting error. The deviation of the weighted average approximated according to the law of uncertainty propagation.
We inferred the inventory and flux of radiocesium accumulation in Tokyo Bay sediment from the catchment basin owing to the FDNPP accident for the five years studied. Table 1 shows the inventory, flux, and 134 Cs/ 137 Cs activity ratio of radiocesium in the sediments collected from the Edogawa water system and Tokyo Bay. The 134 Cs/ 137 Cs activity ratio in 117 sediment samples (Table 1, S6 Table), with counting error of the radioactivity measurements within 5%, was 1.006 ± 0.003 (weighted average value), which conforms to the 134 Cs/ 137 Cs ratio of radiocesium discharged by the FDNPP accident [28][29][30].
The vertical distribution of radiocesium in Tokyo Bay sediment was investigated via the core samples. Fig 3 (S1 Table) shows the vertical distribution of the 134+137 Cs activities from the Old-Edogawa mouth (Point 02) to the center of the bay (Point J) in 2014 and 2015. The highest activity of radiocesium was 1070 BqÁkg -1 , detected in the 35 cm depth layer of Point 02, which is near the river mouth. However, for Point J in the center of the bay, the peak of   Table). The activity of radiocesium was high in the 5 cm layer in August 2011, immediately after the FDNPP accident, and it peaked at 547 BqÁkg -1 in   It is shown that sedimentary materials with high water content and low apparent density are deposited in the estuary area during flooding [31]. Furthermore, the radiocesium derived from the FDNPP accident was detected as approximately 350 BqÁkg -1 in the 70 cm depth layer, collected on July 2016. This value in April 2012 when correction of radioactive decay was 560 BqÁkg -1 . Furthermore, It is suggested that the concentration peak of radiocesium about five years after the accident is buried while being retained in the sediment.

Radiocesium activity in water samples around Tokyo Bay
It was assumed that radiocesium flows into Tokyo Bay through the rivers. Thus, the 134+137 Cs activity in water from the estuaries of the major rivers that flow into Tokyo Bay was analyzed.
The results are shown in Table 2. The 134+137 Cs activity in water standardized for March 16, 2011, ranged from 4.4 to 178 mBqÁL -1 , and in the estuary of the Old-Edogawa, it was higher, by 20 mBqÁL -1 or more. In particular, higher radiocesium activity (23 to 178 mBqÁL -1 ) was detected in the water at the confluence of Sakagawa and Edogawa (Point S1). In the estuaries of Sumidagawa and Tamagawa, on the other hand, it was lower than 10 mBqÁL -1 . This  conforms to the radiocesium distribution in the sediments of these points. Surface and bottom seawater was sampled from the Old-Edogawa estuary to the center of the bay, but the ratio of the 134+137 Cs activity in the surface and bottom water was 2.05 ± 0.91 (n = 15), excluding Point J in the center of the bay, showing that the activity was approximately twofold higher in the surface water than in the bottom water. This means that the radiocesium flowed into Tokyo Bay via river water. Moreover, the radioactivity ratio of 134 Cs/ 137 Cs was 0.893 ± 0.025 (n = 51). It was thought that the water still contained 1 to 2 mBqÁL -1 of 137 Cs as background in Japan, owing to global fallout [32].

Spatiotemporal distribution of radiocesium in the river sediment of Old-Edogawa, Edogawa, and Sakagawa
In the Sakagawa (Fig 2), which converges with the middle Edogawa, the weighted average 134+137 Cs activity was 3630 ± 11 BqÁkg -1 (n = 11), whereas in Edogawa, it was 170 ± 2 BqÁkg -1 (n = 20) upstream from Sakagawa and 473 ± 3 BqÁkg -1 (n = 14) downstream from Sakagawa. As the results of the aircraft monitoring [2] in Figs 1 and 2 clearly show, the forested zone in Gunma Prefecture upstream on Tonegawa River was also contaminated with radiocesium at 60 to 300 kBqÁm -2 . However, upstream in Edogawa after it diverges from Tonegawa, the 134+137 Cs activity was low in both the sediment and water at Point E4. Therefore, it considered that the radiocesium-contaminated area in upstream Tonegawa is not an important radiocesium source supplying Tokyo Bay.
The Matsudo Weir is installed about 100 m downstream from Point S1 and is normally closed to adjust the flow rate of the Sakagawa water. This means that contaminated suspended materials flowing down from the Sakagawa catchment basin is deposited at the point where the flow velocity is low. However, this weir was opened during the Kanto-Tohoku heavy rainfall event from September 9 to 11, 2015 [33,34], and the sediment deposited upstream of the weir flowed out into the confluence with Edogawa.
The core sampled on April 29, 2014 (Fig 5A, S3 Table) shows a record of the vertical distribution of radiocesium from before the FDNPP accident. If it is presumed that the 38 to 40 cm layer in which 134 Cs was detected is the layer was deposited immediately after the accident, the rate of deposition of this sediment is 1.0 cmÁmonth -1 . Thus, the sediment layer deposited in October 2011, when decontamination work started in Kashiwa City [35], which is in the catchment basin of Sakagawa, is the sediment layer from 32 to 34 cm. This decontamination work was completed in December 2012, which corresponds to the layer from 14 to 16 cm. For this reason, the vertical distribution of 134+137 Cs, shown in Fig 5A, presumably is sediment with a record of contaminated soil from the high contamination zone accompanied by a discharge of contaminated sludge caused by the decontamination work in Kashiwa City. Some of the decontamination wastewater that does not receive treatment flows into Sakagawa, and it is highly likely that this contaminated sludge was deposited at Point S1. In a core sampled at the same point on November 13, 2015, three months after the Kanto-Tohoku heavy rainfall event, the high contamination sediment layer had disappeared, but new contaminated sediment had been deposited (Fig 5B). The sedimentation rate at Point S1 after the flood was 2.2 cmÁmonth -1 , suggesting that the sedimentation environment had changed under the effects of the flood. Another core sampled in July 2016 showed a sedimentation rate of 2.2 cmÁmonth -1 , suggesting that, similarly to the core after the flood (Fig 5B), the 134+137 Cs activity was a constant value of about 2,000 BqÁkg -1 (Fig 5C). This implies that contaminated soil has been flowing constantly into Sakagawa from the high contamination zone around Kashiwa, and thus is a supply source of radiocesium that has accumulated in the Old-Edogawa estuary in Tokyo Bay.

Importance of Sakagawa River as a radiocesium source supplying Tokyo Bay
As shown in Fig 2, the radiocesium activity of the sediments in the Sakagawa, Edogawa and Old-Edogawa rivers is higher than that in the Edogawa upstream. Furthermore, it is suggested from Temporal changes in the estimated flux of 134+137 Cs to the sediment at Point S1. The radioactive decay curve of 134+137 Cs was calculated assuming that the activities of 134 Cs and 137 Cs were equal immediately after the FDNPP accident [28][29][30].
Kashiwa City are being transported to Tokyo Bay through these rivers. In other words, the origin of radiocesium contamination in Tokyo Bay is polluted soil in the high contaminated zone, and it can be thought that the Edogawa water system plays an important role in its transportation. Fig 6 (S7 Table) shows the flux of the 134+137 Cs under the sediments at Point S1 calculated from Fig 5 (S3 Table). The estimated flux immediately after the accident was about 1.0 kBqÁ m -2 Áday -1 and had decreased to about 0.5 kBqÁm -2 Áday -1 by July 2016 . Fig 6 also shows the flux estimated for each the sampling period of each sediment core, and the flux decay corrected based on the value of March 16, 2011. For both, the radioactive decay curve of 134+137 Cs conforms closely, and from this fact as well, it was assumed that a constant supply of radiocesium currently continues at Point S1.

Contamination via global fallout of 137 Cs in Tokyo Bay sediment before the FDNPP accident
Compared with the water samples, the radiocesium activity in the present soil and sediment samples before the FDNPP accident is extremely low; therefore, we did not consider the background in this study. Global fallout via atmospheric nuclear testing contributed 137 Cs to the environment before the FDNPP accident. Precipitation of the global fallout 137 Cs from the atmosphere was 6.56 kBqÁm -2 from 1954 to 1985 (before the Chernobyl accident), in Chiba City, located in east of Tokyo Bay (S8 Table) [36]. The inventory of the global fallout 137 Cs in the sediment of central Tokyo Bay was presumed from 0.37 to 0.51 kBqÁm -2 [37]. However, it is possible these values are underestimates, as the cores were too short. The global fallout 137 Cs in the initial stage probably was not measured. On the other hand, we also found a record of the global fallout 137 Cs in the core collected from Points 36 and J in the center of Tokyo Bay. The inventories were 0.80 and 0.71 kBqÁm -2 , respectively, for Points 36 and J (S8 Table). From our data, it was estimated that 11 to 12% of the global fallout 137 Cs precipitation observed in Chiba City from 1954 to 1985 was accumulated in the Tokyo Bay sediment. However, as shown in Table 2, it is presumed from the 134 Cs/ 137 Cs activity ratio that 137 Cs due to the global fallout and the fallout from the Chernobyl accident is significant in river water and seawater in the Tokyo Bay water system.

Inventory and flux of radiocesium owing to the FDNPP accident in the sediment of Point D in the Old-Edogawa estuary
Based on the vertical distribution of radiocesium in the core sediment deposits at Point D, as shown in Fig 4, the changes over time in the activity, inventory, and flux of the 134+137 Cs in the sediment were analyzed. The results are shown in Fig 7 (S9 Table). The radioactive decay curve of the 134+137 Cs activity leads to the hypothesis that the 134 Cs/ 137 Cs activity ratio emitted during the accident was 1.0 [28][29][30]. The 134+137 Cs activity of surface sediment presumably reflects the level of contamination of the catchment basin, which is the supply source of radiocesium. Fig 7A shows the average radioactivity of 134+137 Cs in the sediment layers contaminated by radiocesium and in the top 5 cm of the sediments. In both the surface and in the contaminated layers, the maximum values appeared at the end of 2012, about two years after the accident. Afterwards, the 134+137 Cs activity of sediments decreased until its value matched the activity anticipated from the radioactive decay curve. Fig 7B shows the inventory and change over time of 134+137 Cs in sediments and the flux estimated from the inventory. The inventory immediately after the accident was about 20 kBqÁm -2 , but five years later in 2016, it had increased to about 100 kBqÁm -2 . Theoretically, by that time, the 134+137 Cs activity should have decayed to 53% of its value immediately after the accident, yet the inventory had greatly increased. The flux was 0.13 kBqÁm -2 Áday -1 immediately after the accident, but in 2016, it had decreased to 0.053 kBqÁm -2 Áday -1 (19 kBqÁm -2 Áyr -1 ). Beginning in 2014, when it is assumed that the inflow of radiocesium to Point D had become constant, the flux also conformed almost exactly to the values anticipated, based on the radioactive decay of 134+137 Cs. This suggests that, even now, radiocesium continues to constantly flow into Tokyo Bay.

Balance of the radiocesium flowing into Tokyo Bay from the Edogawa watershed
Based on the analytical results, the balance of radiocesium in the Edogawa water system and in Tokyo Bay sediment are shown in Table 3 (S2 Fig). Under the activity value standardized to that of March 16, 2011, the quantity of 134+137 Cs precipitated in the catchment basin of the Edogawa water system was 8.33 TBq per the results of aircraft monitoring [2]. Judging from the results of the analysis of the core samples, the average inventory of 134+137 Cs in Area X, which has an area of 10 km 2 and is located about 8 km southeast of Chiyoda-Ward in central Tokyo, was 131 kBqÁm -2 (n = 10), and the total inventory of 134+137 Cs was 1.31 TBq. In July 2016, about 70% of the radiocesium deposited in the study area of Tokyo Bay accumulated in Area X. Similarly, the average inventory in Area Y, which is 40 km 2 in area, was 5.52 kBqÁm -2 (n = 11), and the total inventory was 0.22 TBq. Assuming that all the radiocesium deposited in Area X was supplied from the Edogawa water system, about 16% of radiocesium precipitated in the catchment basin during the five years following the accident had moved to Area X. This value is larger than the 11 to 12% global fallout measured in the center of Tokyo Bay, but because Area X is located in the estuary, it is reasonable that it has a relatively large value. Furthermore, the contaminated particles that flowed out during the decontamination work in the highly contaminated zone also may have caused this value to be higher. The average flux in sediment of Area X for 5.4 years (from the FDNPP accident to July 2016) was 0.067 kBqÁm -2 Áday -1 (24.5 kBqÁm -2 Áyr -1 ), so 0.245 TBq of radiocesium continues to flow into Area X every year. This value is consistent with the value for July 2016 of the flux estimated at Point D, off the coast of the river mouth, as shown in Fig 7. The average radiocesium flux of Area X in July 2016 from the FDNPP accident is about 15 times the maximum value of 1.59 kBqÁm -2 Áyr -1 of the global fallout 137 Cs, found in 1963 in Chiba City [36].

Estimation of the sedimentation process of radiocesium in the Edogawa water system and Tokyo Bay
Model calculation was conducted and the process of the movement of radiocesium from the land to the aquatic system was simulated based on values observed in various environments [38][39][40][41]. Our study has shown that radiocesium precipitated in the northeastern part of the greater Tokyo region was accumulated through the rivers into the estuaries of deep in Tokyo Bay without diffusing to the center of the bay (Figs 2, 3 and 8A, S6 Table). deposited. The fine particles are not found in sediments at Points 02 and D of the estuary where seawater and river water mix. This suggests that coagulated deposition of colloidal particles by salting out did not occur. Since the average particle size of the sediment decreases from the estuary towards the offshore, it is shown that sedimentary substances flowing from the river are accumulating first from large particles as they diffuse through the sea water. It is well known that cesium cations is inserted between layers of 2:1:1 type clay minerals like vermiculite and strongly absorbs by ion exchange [42][43][44][45][46][47]. Therefore, radiocesium  [56][57][58], the relationship should obey the inverse square law. The vertical arrow in the Fig 8A indicates the range of the radiocesium background referred from the JCG report [23].
https://doi.org/10.1371/journal.pone.0193414.g008 precipitated from atmosphere on ground surfaces is absorbed and held by clay minerals in the soil. Recently, it has been confirmed directly that radiocesium released from the FDNPP is adsorbed by the clay mineral using a new technic such as EXAFS and NMR [48][49][50]. Moreover, fine particles with a particle size of 10 μm or less containing radiocesium at high activity are found in various places of eastern Japan [51][52][53][54][55], but we believe that washout due to rainfall plays an important role in precipitation on the ground even if radiocesium is scattered in the atmosphere as gaseous or fine particles. It can be hypothesized that in the greater Tokyo region, radiocesium precipitated on ground surfaces was absorbed by the soil and transported to the Tokyo Bay estuary by such a mechanism, underwent coagulating sedimentation via the salting out effect as it mixed with seawater [56,57], and then accumulated in the estuary sediment. However, such a mechanism cannot explain the difference in the sedimentation processes of radiocesium and that of heavy metals in Tokyo Bay (Fig 8, S6 Table, S3 Fig).
Assuming that the background lead concentration in the Tokyo Bay sediment older than 1800s is about 10 mgÁkg -1 [59], the anthropogenic lead concentration of the sediment was estimated. As shown in Fig 8B, a clear grain size effect between the concentrations of lead and the average particle size of the surface sediments in the Tokyo Bay water system has been confirmed. This indicates that the lead is selectively accumulated in sediments in the center of Tokyo Bay, where the average particle size is relatively small. This phenomenon occurs when anthropogenic lead absorbs onto the surface of suspended particles in water. But the slope is less than two. As many rivers flow into Tokyo Bay, this may be due to the dilution effect of suspended particles with low anthropogenic lead contamination. Moreover, as the particle size increases, it deviates slightly from the inverse square law. This is probably due to the relatively higher lead concentration of constituent components of sediments than anthropogenic lead in large particles. Such a tendency was also observed with zinc and mercury (S6 Table, S4 Fig). If it can be assumed that the deposited particle are spherical, the radiocesium activity should follow the inverse square law for the sediment particles [56][57][58]. However, a negative correlation between them is not observed.
As shown in Fig 8, the average particle size in the soil of the high contamination zone was 20 to 100 μm and the radiocesium activity was in the range of 1000 to 35000 BqÁkg -1 . It is believed that this soil in the high contamination zone is the source of radiocesium contamination in Tokyo Bay. When it flows into Sakagawa, the average particle size is separated into 8 to 20 μm (4000 to 35000 BqÁkg -1 ) and 250 to 350 μm (1500 to 6000 BqÁkg -1 ) size classes. The larger particles settle in Sakagawa, and the smaller particles move to the confluence with Edogawa. From Sakagawa to Tokyo Bay, there seems to be a positive correlation in the process between the particle size and the radiocesium activity as opposed to the grain size effect. This suggests that the radiocesium is not adsorbed to suspended particles in water during transportation from Sakagawa to Tokyo Bay. The low radiocesium activity in Area Z in the central part of Tokyo Bay, where the sediment particles are smallest, is thought to be due to the dilution effect observed, the same as for the heavy metals. The relationship between radiocesium activity and particle size in Tokyo Bay is much different from the relationship observed in the estuary area of Abukuma River, which flows through the most contaminated zone owing to the FDNPP accident [60][61][62]. This is probably because the Abukuma River estuary faces the Pacific Ocean, hence particle size fractionation of the contaminated suspended particles due to tidal currents and waves is occurring. On the other hand, because Tokyo Bay is closed, it is not affected by the ocean waves. Therefore, the large particles contaminated with radiocesium are selectively deposited in the river mouth as the flow rate decreases, and small particles are transported to the center of the bay, where they settle. The distribution in Fig 8A (S3 Fig) can be explained by such a mechanism. When radiocesium adsorbed onto fine clay minerals flows from the river into Tokyo Bay, there is a possibility of coagulation and precipitation owing to the salting out effect of seawater [42,56,57], but from our results regarding this period, no such effect was found.

Conclusions
Changes in the radiocesium contamination of the Tokyo Bay sediment surrounded by the Tokyo metropolitan area were discussed for five years after the FDNPP accident. Most of radiocesium flowed from the high contaminated zone in the northeastern part of the Tokyo metropolitan area through the river into the Old-Edogawa estuary inner part of Tokyo Bay and was accumulated without diffusing to the center of the bay. The radioactivity increased from immediately after the accident and decreased following the theoretical radioactive decay after showing the maximum value of 547 BqÁkg -1 at end of 2012. The average inventory of radiocesium in the Old-Edogawa estuary increased to 104 kBqÁm -2 in July 2016. At that time about 70% of the radiocesium deposited in the study area of Tokyo Bay had accumulated in the Old-Edogawa sediment. On the other hand, the inventory of radiocesium in the central bay was 0.46 kBqÁm -2 .
Radiocesium may be transported to Tokyo Bay as contaminated soil particles through rivers such as Sakagawa, Edogawa, and Old-Edogawa from the high contaminated zone in the northeastern part of the Tokyo metropolitan area. Observation results suggested that the river plays an important role in transporting radiocesium from pollution sources. Furthermore, it is considered that the radiocesium precipitated from the atmosphere was adsorbed onto soil particles. The contaminated particles flow out into the river, but they selectively deposited and accumulated in the estuary of Tokyo Bay where the flow velocity of the river water decreases. Therefore, it was presumed that it did not diffusing to the center of the bay.
It is important to recognize that, as the present study has indicated, radiocesium from the FDNPP accident which was deposited in distant watersheds is still flowing into Tokyo Bay, and the consequences are not fully understood. Continued and careful long-term monitoring of these environmental radionuclides is warranted. In the other words, Tokyo Bay plays the role of a sink for radioactive contaminants discharged in the greater Tokyo region.
Supporting information S1 Fig. Evidence that flood sediment deposited at Point D. Coastal flood sediments have high water content and small particle size, so their apparent density is small. The cumulative mass of the 8 to 22 cm layer of the core collected in November 2015 immediately after the Kanto-Tohoku heavy rainfall event is clearly lower than that of the other cores. It can be thought that this is a trace of the flood sedimentary layer [31].  Table. Radiocesium activity, anthropogenic heavy metal concentration, and particle size in the surface sediment collected from the Tokyo Bay water system. a: X (Old-Edogawa estuary), Y (Off the Old-Edogawa estuary), Z (Center of Tokyo Bay), V (Tamagawa estuary), W (Sumidagawa estuary). b: Value on sampling date. c: Corrected values were corrected for radioactive decay on March 16, 2011. d: Weighted average value of the decay corrected radiocesium activity for multiple samples from the same site. e: Ratio for the decay corrected value. f: Value for samples whose counting error of the decay corrected total radiocesium activity was within 5%. The background concentration of heavy metal was assumed to be Zn 100 mgÁkg -1 , Hg 40 μgÁkg -1 , and Pb 10 mgÁkg -1 .