Transport and dispersion of tritium from the radioactive water of the Fukushima Daiichi nuclear plant

Marine Pollution Bulletin 169 (2021) 112515 Available online 21 May 2021 0025-326X/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Transport and dispersion of tritium from the radioactive water of the Fukushima Daiichi nuclear plant Chang Zhao a , b , c , d , Gang Wang a , b , c , d , Min Zhang a , b , c , d , Guansuo Wang a , b , c , d , Govert de With e , Roman Bezhenar f , Vladimir Maderich f , Changshui Xia a , b , c , d , Biao Zhao a , b , c , d , Kyung Tae Jung g , Raúl Peri´ a˜ nez h , Mohd Fadzil Akhir i , Chalermrat Sangmanee j , Fangli Qiao a , b , c , d , * a First Institute of Oceanography, Ministry of Natural Resources, Qingdao 266061, China b Laboratory for Regional Oceanography and Numerical Modeling, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266237, China c Key Laboratory of Marine Science and Numerical Modeling, Ministry of Natural Resources, Qingdao 266061, China d Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China e Nuclear Research and Consultancy Group (NRG), Arnhem, the Netherlands f Institute of Mathematical Machine and System Problems, Kyiv, Ukraine g Oceanic Consulting and Trading, Yangpyeong-ro, Seoul, Republic of Korea h Dpt. Física Aplicada I, ETSIA, Universidad de Sevilla, Sevilla, Spain i Institute of Oceanography and Environment, Universiti Malaysia Terengganu, Terengganu, Malaysia j Oceanography and Environment Division, Phuket Marine Biological Center, Phuket, Thailand ARTICLE INFO Keywords: Tritium concentration scenarios Fukushima Daiichi Nuclear Power Plant Three-dimensional numerical simulation Radioactive water ABSTRACT Japan recently announced plans to discharge over 1.2 million tons of radioactive water from the Fukushima Daiichi Nuclear Power Plant (FDNPP) into the Pacific Ocean. The contaminated water can poses a threat to marine ecosystems and human health. To estimate the impact of the plan, here, we developed a threedimensional global model to track the transport and dispersion of tritium released from the radioactive water of the FDNPP. The pollution scenarios for four release durations (1 month, 1 year, 5 years, and 10 years) were simulated. The simulation results showed that for the release in short-duration scenarios (1 month and 1 year), the peak plume with high tritium concentration shifted with the currents and finally reached the northeastern Pacific. For the long-duration scenarios (5 years and 10 years), the peak plume of the contaminated water was confined to coastal regions east of Japan. 1. Introduction Since the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident in March 2011, large amounts of artificial radionuclides have been released into the atmosphere and ocean (Aoyama et al., 2016a, 2016b, 2016c; Smith et al., 2015; Qiao et al., 2011; Buesseler et al., 2011; Zhao et al., 2015; Buesseler et al., 2017). The estimated total amount of the released radionuclides is approximately 520 PBq (Steinhauser, 2014), which includes more-volatile radionuclides and gases such as 131 I, 134 Cs, 137 Cs, and radioactive tritium. Most of the artificial radionuclides were released into the ocean by fallout and direct discharge before May 1, 2011. Subsequently, small, and sometimes unexpected, sources of the artificial radionuclides have continued entering the ocean (Sanial et al., 2017). The majority of the artificial radionuclides have been transported to the whole Pacific from energetic coastal waters off of Japan by ocean current (Buesseler et al., 2012; Behrens et al., 2012; Nakano and Povinec, 2012; Rossi et al., 2013; Miyazawa et al., 2013; Aoyama et al., 2013; Min et al., 2013; Povinec et al., 2013; Aoyama et al., 2016a, 2016b, 2016c). The arrival of the artificial radionuclides off the west coast of North America was documented in measurements taken in June 2013 (Smith et al., 2015). The artificial radionuclides have been transported to the marginal seas of Western Pacific Ocean (Suseno and Prihatiningsih, 2014; Suseno and Wahono, 2018; Zhou et al., 2018; Zhao et al., 2018). Moreover, these artificial radionuclides have been rapidly incorporated into marine organisms either by uptake from seawater or by food ingestion (Buesseler, 2012; Madigan et al., 2012; Alava and * Corresponding author at: First Institute of Oceanography, Ministry of Natural Resources, 6 Xian-Xia-Ling Road, Qingdao 266061, China. E-mail address: [email protected] (F. Qiao). Contents lists available at ScienceDirect Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul https://doi.org/10.1016/j.marpolbul.2021.112515 Received 3 May 2021; Received in revised form 14 May 2021; Accepted 15 May 2021 Marine Pollution Bulletin 169 (2021) 112515 2 Gobas, 2016; Bezhenar et al., 2016; De With et al., 2021). The operators of Tokyo Electric Power Company (TEPCO) have been pumping water into the damaged FDNPP to cool the nuclear reactors during the past 10 years, resulting in the storage of a huge volume of radioactive water in many large tanks. Underground water and rainwater seepage into the reactors have also contributed to the accumulation of radioactive water (Buesseler, 2020). Therefore, radioactive water of the FDNPP is still being produced continuously. According to the operator of TEPCO, about 1000 tanks used to store the contaminated water will run out of space by 2022, at which point they will contain nearly 1.3 million tons of radioactive water (TEPCO, 2020; TEPCO, n.d., Treated Water Portal Site). The Japanese government has decided to dump contaminated radioactive wastewater from the FDNPP into the ocean two years later. This decision is being fiercely opposed by international communities, local fishermen, and residents. Although most radionuclides in the water have been treated by an Advanced Liquid Processing System (ALPS), the water still contains radionuclides whose half-lives are comparatively long, such as tritium (half-life of 12.3 years), carbon-14 (half-life of 5730 years), ruthenium- 106 (half-life of 373.59 days), cobalt-60 (half-life of 5.26 years), strontium-90 (half-life of 28.79 years) etc. These radionuclides will contaminate marine ecosystems and threaten human health (Eyrolle et al., 2018; Buesseler, 2020; Shozugawa et al., 2020). To estimate the potential impact of Japan's plan to discharge the radioactive water into the ocean (mainly radioactive tritium), we established a threedimensional global model to track the large-scale oceanic transport of the tritium scenarios released from the radioactive water of the FDNPP. Four releasing numerical scenarios were simulated to produce contamination maps for different periods after discharge. A detailed discussion on the remaining uncertainties is also provided in this paper. 2. Model and numerical experiments design 2.1. Model description A three-dimensional global model based on the Princeton Ocean Model (POM) (Blumberg and Mellor, 1987) was developed to simulate tritium activity in global ocean. The model domain covers the quasiglobal ocean of 78◦S–65◦N in latitude and 0–360◦E in longitude, with a horizontal resolution of 0.5◦×0.5◦. The vertical resolution is 21 sigma levels with a fine resolution in the upper ocean and bottom layers. The model topography was constructed by using the global 5′×5′dataset, Etopo5, with the minimum depth set as 10 m. A cyclic boundary condition is used for the eastern and western boundaries, whereas a solid boundary is used in the north and south. A Mellor-Yamada 2.5 turbulence closure scheme is applied (Mellor and Yamada, 1982) for vertical viscosity and diffusion. The simulation is initialized with climatological temperature (T) and salinity (S) data in the January fields of the World Ocean Atlas dataset, and the initial velocity is set as 0. The forcing for spin-up is built on the atmospheric reanalysis product Comprehensive Ocean-Atmosphere Data Set (COADS), with a horizontal resolution of 1◦ ×1◦. After 20 years of spin-up, the model output provides the initial fields to simulate tritium transport in the oceans released from the FDNPP. Advection and diffusion are two major processes that determine the transport and dispersion of tritium in the ocean. Self-decay and adsorption by suspended matters are the processes responsible for removing the tritium dissolved in seawater. The self-decay process depends solely on half-life of tritium, while the scavenging process associated with the absorption onto suspended materials can be approximated by taking the approach used by Tsumune (Tsumune et al., 2003a; Tsumune et al., 2003b). Comprehensively, the governing equation can be presented as ∂ C ∂ t=1 D ∂ ∂σ [KH D ∂ C ∂σ ]+FC−( ∂ Cu ∂ x+ ∂ Cv ∂ y+1 D ∂ Cw ∂σ )−λC −Kd ρ s( σ D)ws 1 D ∂ C ∂σ (1) with, FC= ∂ qx ∂ x+ ∂ qy ∂ y(2) qx=AH ∂ C ∂ x,qy=AH ∂ C ∂ y(3) where x,y, and z are the conventional Cartesian coordinates; u, v, and w are the zonal, meridional, and vertical velocities, respectively; C is the concentration of tritium (in Bq/m 3 ); t is time (s); σ =z− η D, where D =H + η , H =H(x,y) is the bottom topography, and η = η (x,y,t) is the surface elevation. Thus, σ ranges from σ =0 at z = η to σ = − 1 at z = − H. A H and K H are the horizontal diffusion and vertical diffusion coefficients of tritium, respectively. The Smagorinsky's formula is used to evaluate A H (Smagorinsky, 1963). K H is calculated by a Mellor-Yamada 2.5 turbulence closure schemes. Further, λ is the radioactive decay constant, which is about 1.8 ×10 −9 s −1 (half-life of 12.43 years) for tritium; K d is the distribution coefficient of the nuclides from water to suspended materials, 2.0 ×10 −3 m 3 /g for tritium; w s is the settling velocity of suspended materials which is set as 1.0 ×10 −3 m/s; and ρ s (z) is the concentration of the suspended materials, which is about 0.25 g/m 3 in the surface layer (Tsumune et al., 2003b). Its vertical distribution is expressed by ρ s(z) = ρ s(0) × 10−0.0005z.(4) The contaminated water mixed with seawater instantly after release, and thus the evaporation process at the sea surface could be ignored. 2.2. Numerical experiment design After 20-year's spin-up of the oceanic circulation, the tritium released from the FDNPP was added into the model to simulate its transport and dispersion. The model was forced by the climatology of surface wind stress and net heat fluxes from COADS. The total amount of tritium contained in the tanks, as reported, is around 1.0 PBq (PBq = 10 15 Bq) (TEPCO, 2020; Buesseler, 2020) and will be approximately 1.2 PBq in 2022. The Japanese government plans to dump the radioactive water into the ocean in two years. Thus, we set the initial value as 1.2 PBq in our numerical simulations. As the discharge procedure has not been announced yet, we ran the model through four scenarios with different release durations of the radioactive wastewater (Table 1). As the tritium itself will also decay during the release duration, total amount of the tritium will decrease with an increasing release duration. This was considered by setting different initial values for the total amount of tritium in the four simulation experiments. As the model Table 1 Design of numerical experiments. Experiment name Release duration of the radioactive water Total amount of tritium (PBq) 1m One month 1.20 1a One year 1.17 5a Five years 1.04 10a 10 years 0.91 C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 3 horizontal resolution is 0.5◦×0.5◦, one grid of the model could cover the whole discharge area. Therefore, the tritium discharge site was chosen as the nearest grid point to the Fukushima Station at 141.25◦E, 37.75◦N,. As Japan plans to dump the contaminated water directly into the ocean, we set the discharge depth as the upper 5 m below the sea surface, which is the first layer of our model grid. 3. Results and discussion 3.1. Surface distribution of tritium in the North Pacific The horizontal potential polluted areas and patterns at the sea surface from 1 year to 10 years after tritium emitted from the FDNPP with four emission scenarios are presented in Figs. 1 to 2. The main tritium plume emitted from coastal Japan was transported rapidly along the Kuroshio extension during the first year (Fig. 1, a1 and b1). Subsequently, the contaminated water spread gradually and reached the northeastern Pacific Ocean within 4–5 years with the North Pacific Current (Fig. 1, a3 and b3). Almost the entire North Pacific Ocean was contaminated after 10 years owing to the advection and diffusion processes (Fig. 1, a4 and b4). It should be noted that the shift of the contaminated seawater core was different in the four emission scenarios. For the release in short-duration scenarios (1 month and 1 year, Fig. 1), the contaminated water core shifted with the current; it started in coastal Japan, was transported to the middle North Pacific Ocean after 2 years, and finally reached the west coast of America after 4–5 years. However, for the release in long-duration scenarios (5 years and 10 years, Fig. 2), the contaminated water core was constrained near the eastern coastal region of Japan. For the release in short-duration scenarios, either 1 month or 1 year, the potential polluted area and its concentration were almost identical. The core polluted area (with the concentration greater than 0.1 Bq/m 3 ) extended from eastern Japan offshore to 175◦E in 1 year (Fig. 1, a1 and b1), and then extended eastward at a speed of about 15◦per year (Fig. 1, a2 and b2). Five years later, the core of tritium polluted area arrived at the coast of the North America (Fig. 1, a3 and b3). Subsequently, the Fig. 1. Simulated tritium concentrations in sea surface water by Exp. 1m (a1–a4) and Exp. 1a (b1–b4). The symbol ● in a1 shows the stations for analyzing the chronological change of tritium (Fig. 5). Dashed lines show the sections for analyzing the vertical distribution of tritium (Figs. 3–4 and Figs. 6–9). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 4 contaminated water spread to nearly the whole North Pacific, including the East China Sea, the northern South China Sea, and Philippines offshore etc. After 10 years, the contaminated water with a tritium concentration greater than 0.001 Bq/m 3 could pollute the whole South China Sea and the Indonesian seas, and could even expand to the southern hemisphere (Fig. 1, a4 and b4). The tritium polluted core area in the contaminated seawater covered the whole east coast of the North America. Moreover, the polluted water can be transported to the Indian Ocean through Indonesia Throughflow. When the release duration was extended to 5 years (Exp. 5a) or 10 years (Exp. 10a), the contaminated seawater still extended eastward with nearly the same speed but of a lower concentration (Fig. 2). The polluted areas were almost identical to those in the short releaseduration scenarios (Exp. 1m and Exp. 1a) but of different patterns. The core polluted area with tritium concentrations greater than 0.1 Bq/ m 3 was confined to the east coast of Japan, but still spread to the eastern boundary of the North Pacific when the release was completed for Exp. 5a (Fig. 2, a4). For Exp. 10a, the North Pacific is almost equally polluted at the end of the whole release of 10 years (Fig. 2, b4). 3.2. Vertical distribution of tritium along section of 37.5◦N In addition to the horizontal spreading being mainly determined by advection and diffusion processes, the vertical redistribution of tritium due to the vertical turbulent mixing and convection cannot be ignored because of its irreversible accumulation in subsurface biological processes (Rafter et al., 2012). According to the FDNPP location and the main transport route of the tritium, 37.5◦N section was chosen to examine the vertical distribution of tritium. Figs. 3 and 4 show the simulated vertical distribution of radioactive tritium in the four scenarios for 1 year, 2 years, 5 years and 10 years after starting emissions. In the four scenarios, the polluted water core with the center of tritium concentration greater than 1 Bq/m 3 could be rapidly mixed in the upper 600 m because of the vertical mixing processes during the first year after release (Fig. 3, a1 and b1, Fig. 4, a1 and b1). In the next 2–10 years, the core of polluted water gradually penetrated into the deeper layer in the upper 1000 m, with the core pollution area deepening with release time (Fig. 3, a2–a4 and b2–b4, Fig. 4, a2–a4 and b2–b4). Moreover, the vertical mixing body of the polluted water extended from Fig. 2. Simulated tritium concentrations in sea surface water by Exp. 5a (a1–a4) and Exp. 10a (b1–4). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 5 the Japanese coast to the east Pacific at a speed of about 15◦per year. Similar to the horizontal distribution, for the release in shortduration scenarios either for 1 month or 1 year (Fig. 3), the core of polluted area shifted eastward, and vertically meanwhile, with the advection and convection processes. It started at the coast of Japan, and then transported to the mid North Pacific Ocean after 2 years, and finally reached the west coast of America after 4 or 5 years in the upper 1000 m. However, for the long release scenarios (5 years and 10 years, Fig. 4), the core of polluted area and the depth were constrained near the east coast of Japan. Thus, we concluded that the whole North Pacific Ocean in the upper 1000 m would contribute to the dilution of the contaminated water at least 10 years after the start of release. Moreover, a long discharge duration would have a greater impact near the discharge site area. 3.3. Chronological change of tritium concentration in sea surface water Four stations were selected to analyze the chronological change in the tritium concentration based on its transport route and affected areas: east of the East China Sea (a: 130.25◦E, 25.25◦N), the east coast of Fukushima (b: 145.25◦W, 37.75◦N), the mid North Pacific (c: 180.25◦W, 37.75◦N), and west coast of North America (d: 230.25◦E, 37.75◦N) (Fig. 1, a1). The chronological change in the tritium concentration in surface waters at these four stations is shown in Fig. 5. The tritium concentration at station b reached a peak value quickly and then dropped sharply in Exp. 1m (Fig. 5, b), while it maintained a relatively high level until the end of the release for the other three numerical experiments. The tritium dispersed into the Pacific Ocean would be partly transported back to the station along the south branch back to Fig. 3. Simulated tritium concentrations along the 37.5 ◦N section by Exp. 1m (a1–a4) and Exp. 1a (b1–b4). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 6 the Kuroshio Current. Therefore, the tritium concentration showed an increase 3 years after the start of release (black and blue lines in Fig. 5b). Similar patterns could be observed in the mid North Pacific (Fig. 5, c). The tritium concentration east of the East China Sea maintained a low level (<0.04 Bq/m 3 ) because of a weak southward current (Fig. 5, a). The contaminated water was dispersed into the west coast of North America after 4–5 years. There was a slight increase in the tritium concentration in the first 1 to 2 years, which was then maintained at a certain concentration for a prolonged period (Fig. 5, d). 3.4. The radioactive tritium water in China Seas In this section we discuss the influence of contaminated seawater on the China Seas and adjacent ocean regions. The southward branch of the tritium plume would intrude into the China Seas mainly through the Luzon strait and the strait of Ryukyu Islands. Thus, two sections along 20.5◦N and 27.5◦N (Fig. 1, a1) were selected to examine the vertical transport of contaminated water, which is presented in Figs. 6 and 7. According to the previous analysis (Figs. 1 to 5), 0.001 Bq/m 3 was taken as the threshold for the contaminated seawater. Because the southward plume will be blocked by the Kuroshio Current at sea surface, the radioactive tritium water would mainly enter the China Seas through the middle layer (400–600 m). For the section along the 20.5◦N (through the Luzon strait), unlike the transport to the Pacific Ocean, the radioactive tritium water was transported mainly at a depth of 400–600 m, which should be correlated with the fact that the Pacific mode water is controlled by tidal mixing, winter cooling, and buoyance loss (Oka et al., 2012). For the release in short-duration scenarios (Exp. 1m and Exp. 1a), the tritium concentration in the core polluted water (>0.1 Bq/m 3 ) was much higher than that of the long duration experiments (<0.1 Bq/m 3 ) (Exp. 5a and Exp. 10a). The polluted water expanded to the coast of China in less than 3 years for Fig. 4. Simulated tritium concentrations along the 37.5 ◦N section by Exp. 5a (a1–a4) and Exp. 10a (b1–b4). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 7 Exp. 1m and Exp. 1a and more than 5 years for Exp. 5a and Exp. 10a. Owing to the shallow structure of the coastal waters, it is more prone to biological accumulation processes, and eventually, leading to an increased biological impact. Comparatively, the tritium water along the section 27.5◦N (strait of Ryukyu Islands) showed a higher concentration than that through the Luzon strait, with the peak value being more than 0.3 Bq/m 3 . Moreover, the vertical transport of contaminated water was more evenly distributed in depth during the first 3 years (Figs. 8 and 9). All four scenarios showed agreement that the tritium contaminated water would reach the coast of China seas within 3 years (See Fig. 10.) 4. Conclusions In this scenario study, we simulated the spatiotemporal transport of tritium contaminated seawater released from the FDNPP by using a three-dimensional global model. Assuming a total release of 1.2 PBq of tritium into the ocean with four release duration scenarios, we concluded that the majority of tritium with radioactive water of the FDNPP would be mixed and diluted rapidly in the coastal waters off Japan, and then transported eastward along Kuroshio extension. It would arrive at the Pacific coast of North America within 4–5 years. Moreover, the southward plume of the contaminated water would be blocked by the strong Kuroshio Current at the sea surface. The radioactive tritium water would enter the Pacific mode water through vertical mixing such as winter cooling and buoyancy loss. This part of tritium would be transported to the China Seas through the Luzon strait and the strait of Ryukyu Islands. The polluted water can be transported to the Indian Ocean through the Indonesia Throughflow (Fig. 3 of Zhao et al., 2015). Different discharge durations of radioactive water would lead to different dispersion patterns in the oceans. If the discharge duration is short, as in 1 month or 1 year in our simulations, the core polluted water would shift with the currents and finally arrive the coastal area of the US and Canada in the northeastern Pacific, with larger polluted areas. For long release durations, the core polluted area would be more prone to be constrained around the emission site. Although some data (TEPCO, 2020) about tritium in the radioactive water of the FDNPP have been issued, there are still considerable uncertainties regarding the amount of tritium in the water. As shown in our results, the release durations have impact on the tritium concentraion and distribution. Moreover, the release rates and locations of discharge will also affect the accuracy of simulated results. However, the amount of the leakage would have no effect on the transport of tritium and its patterns. If there is a release, the radioactive plume would transport to entire North Pacific Ocean in any case. The monitoring site should be set near the emission site for short release-duration scenarios or along the Kuroshio Current for long release-duration scenarios. Further studies should be performed on the biological impact from the release of the radioactive water. One thing should be noted is that the tritium background concentrations in the sea are not considered in the simulations. According to Kaizer et al. (2018) and Aoyama et al. (2021), the background concentration of tritium in the surface water of the North Pacific Ocean is around 50 Bq/m 3 . However, there are other relevant aspects need to be considered. The peak tritium concentration may be underestimated due to the relatively coarse model resolution (0.5◦) and the reduced mesoscale eddies in the present model. And other radionuclides in the radioactive water behave differently with tritium in the ocean, and they are more readily incorporated into marine biota or seafloor sediments. A higher resolution regional model is needed for further studies. Fig. 5. Chronological change in the tritium concentrations in the sea surface water of the North Pacific at (130.25 ◦E, 25.25◦N), (145.25◦W, 37.75◦N), (180.25◦W, 37.75◦N) and (230.25◦E, 37.75◦N). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 8 Fig. 6. Simulated tritium concentrations along the 20.5 ◦N section by Exp. 1m (a1–a4) and Exp. 1a (b1–b4). C. Zhao et al. Marine Pollution Bulletin 169 (2021) 112515 9 Fig. 7. Simulated tritium concentrations along the 20.5 ◦N section by Exp. 5a (a1–a4) and Exp. 10a (b1–b4). C. Zhao et al.