lable at ScienceDirect

Journal of Environmental Radioactivity 131 (2014) 4e18

Contents lists avai

Journal of Environmental Radioactivity

journal homepage: www.elsevier .com/locate / jenvrad

Regional long-term model of radioactivity dispersion and fate in theNorthwestern Pacific and adjacent seas: application to the FukushimaDai-ichi accident

V. Maderich a,*, R. Bezhenar b, R. Heling c, G. de With c, K.T. Jung d, J.G. Myoung d, Y.-K. Cho e,F. Qiao f, L. Robertson g

a Institute of Mathematical Machine and System Problems, Glushkov av., 42, Kiev 03187, UkrainebUkrainian Center of Water and Environmental Projects, Glushkov av., 42, Kiev 03187, UkrainecNRG, Utrechtseweg 310, 6800 ES Arnhem, The NetherlandsdKorea Institute of Ocean Science and Technology, 787, Haean-ro, Ansan 426-744, Republic of Koreae School of Earth and Environmental Sciences, Research Institute of Oceanography, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 151-741,Republic of Koreaf First Institute of Oceanography, 6 Xianxialing Road, Qingdao 266061, Chinag Swedish Meteorological and Hydrological Institute, SE-601 76, Norrköping, Sweden

a r t i c l e i n f o

Article history:Received 3 February 2013Received in revised form19 July 2013Accepted 23 September 2013Available online 11 October 2013

Keywords:Compartment modellingRadionuclide transfer in marine biotaHuman ingestion dosesFukushima Dai-ichi accident

* Corresponding author. Tel.: þ380 445266084; faxE-mail addresses: vladmad@gmail.com (V. M

gmail.com (R. Bezhenar), g.dewith@nrg.eu (G.(K.T. Jung), choyk@snu.ac.kr (Y.-K. Cho), qlennart.robertson@smhi.se (L. Robertson).

0265-931X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jenvrad.2013.09.009

a b s t r a c t

The compartment model POSEIDON-R was modified and applied to the Northwestern Pacific andadjacent seas to simulate the transport and fate of radioactivity in the period 1945e2010, and to performa radiological assessment on the releases of radioactivity due to the Fukushima Dai-ichi accident for theperiod 2011e2040. The model predicts the dispersion of radioactivity in the water column and in sed-iments, the transfer of radionuclides throughout the marine food web, and subsequent doses to humansdue to the consumption of marine products. A generic predictive dynamic food-chain model is usedinstead of the biological concentration factor (BCF) approach. The radionuclide uptake model for fish hasas a central feature the accumulation of radionuclides in the target tissue. The three layer structure of thewater column makes it possible to describe the vertical structure of radioactivity in deep waters. In total175 compartments cover the Northwestern Pacific, the East China and Yellow Seas and the East/JapanSea. The model was validated from 137Cs data for the period 1945e2010. Calculated concentrations of137Cs in water, bottom sediments and marine organisms in the coastal compartment, before and after theaccident, are in close agreement with measurements from the Japanese agencies. The agreement forwater is achieved when an additional continuous flux of 3.6 TBq y�1 is used for underground leakage ofcontaminated water from the Fukushima Dai-ichi NPP, during the three years following the accident. Thedynamic food web model predicts that due to the delay of the transfer throughout the food web, theconcentration of 137Cs for piscivorous fishes returns to background level only in 2016. For the year 2011,the calculated individual dose rate for Fukushima Prefecture due to consumption of fishery products is3.6 mSv y�1. Following the Fukushima Dai-ichi accident the collective dose due to ingestion of marineproducts for Japan increased in 2011 by a factor of 6 in comparison with 2010.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The assessment of radiological consequences after routine oraccidental release of radioactivity in the marine environment

: þ380 445263615.aderich), romanbezhenar@de With), ktjung@kiost.aciaofl@fio.org.cn (F. Qiao),

All rights reserved.

involves consideration of processes in the water, suspended andbottom sediment contamination, radionuclide uptake by biota andtransfer through the food chain, and dose assessments to humansand the biota. These processes are multi-scale in terms of time andspace. For long temporal and large spatial scales, analysis withcompartment (“box”) models has been a standard approach in thepast 40 y (e.g. CEC, 1990, 1995; Nielsen, 1995; Lepicard et al., 1998,2004; Iosjpe et al., 2002; Smith and Simmonds, 2009). In thesemodels, the instantaneousmixing in the compartments is assumed,

Fig. 1. Scheme of radionuclide transfer to marine organisms.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 5

and dispersion processes for radionuclides released in the marineenvironment are described by horizontal water exchange from onecompartment to another, by adsorption on suspended sediments inequilibriumwith thewater phase radioactivity, by sedimentation ofsuspended materials, by diffusion transfer of radioactivity betweenthe upper sediment layer and water, and by vertical redistributionof adsorbed radionuclides in the sediments through bioturbationphenomena. In most marine compartment models (e.g. CEC, 1990;Smith and Simmonds, 2009) the dose from ingestion of marineproducts is estimated using a bioconcentration factor (BCF)approach that calculates radionuclide concentrations in sea or-ganisms by implicitly assuming an equilibrium between theradionuclide concentration inwater and biota. However, accidentalreleases are characterised by short-term changes with strongspatial gradients of the environmental radioactivity concentration,and thus the use of conventional box models and the classical BCFapproach is less suitable. Three-dimensional transport models,coupled with circulation models, are best suited for short-termpredictions of radioactivity transport (e.g. Margvelashvily et al.,1997; Periáñez, 2005; Masumoto et al., 2012). However, thesemodels require large computational resources and detailed inputinformation. An intermediate approach, based on the numericalsolution of the Eulerian or Lagrangian transport equations forradioactivity (e.g. Goshawk et al., 2003; Nakano and Povinec, 2003,2012; Nakano, 2006), and using time-averaged current fields onstructured grids, is less cumbersome. Nevertheless, compartmentalmodels still have a strong potential to contribute to assessmentstudies because of their flexibility when constructing the com-partments in three dimensions. Such compartmental models canbe enhanced when hydrodynamic models are used for calculatingthe fluxes between the compartments, and when dynamic modelsare used for the transfer of radionuclides in the food web.

For the radiological assessment of radionuclide releases to theNorthwestern Pacific and adjacent seas (the East China and YellowSeas and the East/Japan Sea), the POSEIDON-R model (Lepicardet al., 2004) was modified and adapted. To the authors’ knowl-edge, this is the first application of a compartment model for thisregion, where 25 Nuclear Power Plants (NPP) are situated near thecoast. The fate and behaviour of radionuclides in the region bothprior to and after the Fukushima Dai-ichi NPP accident were eval-uated. First, the period 1945e2010 (fallout fromweapons tests) wassimulated, and then the period from 2011 to 2040 (Fukushima Dai-ichi accident) was considered. The paper is organized as follow: thedescriptions of the compartment model and of the dynamic foodweb model are given in Section 2. Section 3 presents the modelcustomization for the Northwestern Pacific and adjacent seas. Thevalidation results for period 1945e2010 are given in Section 4.Section 5 discusses the results of the radiological assessment for theradioactivity releases into the sea due to the Fukushima Dai-ichiaccident. Section 6 summarizes the findings.

2. Model description

2.1. Dispersion of activity in water and sediments

The marine environment is modelled as a compartmental sys-tem following the “Marina” methodology (CEC, 1990; Nielsen,1995; EC, 2002). The compartments describing the water columnare subdivided into a number of vertical layers also containingsuspended matter. The model assumes equilibrium between dis-solved and particulate radioactivity in the water column, describedby a distribution coefficient. The radionuclide concentration foreach compartment is governed by a set of differential equationsincluding the temporal variations of concentration, the exchangewith adjacent compartments and with the suspended and bottom

sediments, radioactive sources, and decay. The exchange betweenthe boxes is described by fluxes of radionuclides due to advectionand diffusion processes. The transfer of radionuclides in the bottomsediments is described by a three-layer model. The transfer ofradioactivity from the upper sediment layer (10 cm thickness) tothe water column is described by diffusion and bioturbation.Radioactivity in the upper sediment layer migrates downwards bydiffusion and by burial (caused by the ongoing settling of particlesfrom overlying water). The upwards transfer of radioactivity fromthe middle sediment layer to the top sediment layer occurs only bydiffusion. Burial causes an effective loss of radioactivity from themiddle to the deep sediment layers, from which no upwardmigration occurs. In the POSEIDON-R model, the progeny of long-lived radionuclides is taken into account; very short-liveddaughter products are assumed to be in equilibrium with thelong-lived parent radionuclide in the water phase and in the biotain terms of concentration, and are included in the dose modulewith a dose conversion coefficient including the short-lived prog-eny. In the case when equilibrium is not likely, the decay productsare modelled separately.

2.2. Uptake of activity by marine organisms and transfer throughthe food chain

The biota model used in POSEIDON-R is the simplified dynamicfood web model BURN (Heling et al., 2002) suitable for use inmanaging emergency situations after nuclear accidents, which arecharacterized by rapid temporal changes in environmental con-centrations. By grouping the marine organisms into a limitednumber of classes based on their trophic level and type of species,and by grouping the radionuclides into a limited number of classesassociated with the dominating tissue in which a radionuclidepreferentially accumulates, the number of input parameter is keptrather limited. Standard sets of input parameters are used to avoidthe necessity to collect site specific parameters for a large numberof different species, and for each possible radionuclide, as iscommonly required in complex ecological models (see e.g.Kumblad et al., 2006; Tateda et al., 2013).

In the BURN model, marine organisms are grouped intophytoplankton, zooplankton, fishes (two types: piscivorous andnon-piscivorous), crustaceans (e.g. detritus-feeders), and molluscs(filter-feeders). The corresponding scheme depicting the transfer ofactivity through the food chain is shown in Fig. 1. Additionally, allorganisms take radionuclides directly from water. The basic equa-tion connecting concentration of activity in predator Cpred withactivity concentration in food Cf is

dCpreddt

¼ aK1Cf þ bKwCw � K0:5Cpred; (1)

where t is time, K1 is food uptake rate, a is transfer coefficientthrough food, Kw is water uptake rate, b is transfer coefficient fromwater, Cw is activity concentration in water, K0.5 is the radionuclideelimination rate from the body of fish given by K0:5 ¼ ln 2T�1

0:5,where T0.5 is the biological half-life of the radionuclide. The activity

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e186

concentration in the food of a predator can be expressed by thefollowing equation, summing for a total of n prey types,

Cf ¼Xn

i¼1

Cprey;iPprey;idrwprey

drwprey;i; (2)

where Cprey,i is the activity concentration in prey of type i, Pprey,i ispreference for prey of type i, drwpred is the dry weight fraction ofpredator, and drwprey is the dry weight fraction of prey of type i. It isassumed that phytoplankton is the food for zooplankton withpreference factor 1.0; zooplankton is the food for non-piscivorousfish with preference factor 1.0; phytoplankton and zooplanktonare the food for molluscs with preference factors 0.8 and 0.2,respectively; phytoplankton and zooplankton are the food forcrustaceans with preference factors 0.2 and 0.8, respectively; non-piscivorous fish is the food for piscivorous fish with preferencefactor 1.0.

Values of the other parameters are given in Table 1. The dryweight fractions are derived from De Vries and De Vries (1988), K1the food uptake rate, is composed of the maintenance uptake andthe uptake of food for growth. Standard literature values have beenused for K1 and Kw (De Vries and De Vries, 1988). The biologicalhalf-life data are based on literature values compiled by Coughtreyand Thorne (1983). For radionuclides with longer physical half-livesthan the biological half-life, the biological half-life given in theliterature is used. For short-lived radionuclides the effective half-life should be calculated by correcting the biological retentiontime in the specific tissuewith the physical decay time. In this studythe transport and fate of 134Cs and 137Cs are considered and thebiological half-life is used.

The equations as described are generally applicable for fish if theradionuclide is homogeneously distributed throughout the body ofthe fish. In the case where a certain concentration of radioactivity isnot homogeneously distributed over all tissues of the organism theconcentration in the prey can be described by the followingequation

Cprey ¼Xm

k¼1

Ckfk; (3)

where Cprey is the concentration in a certain prey, fk is the weightfraction of k-th tissue and Ck is the concentration of the nuclide in k-th tissue, m is the number of tissues.

According to the review of radiological data (Coughtrey andThorne, 1983), every radionuclide accumulates in a specific tissue(target tissue). For example, radiocaesium accumulates in the flesh(muscles), while the actinides, plutonium and americium, accu-mulate in specific organs. It can be assumed that the target tissuecontrols the overall elimination rate of the nuclide (T0.5) in theorganism. The radioactivity in the food for the predator is then the

Table 1Parameters of dynamical food chain model.

Organism Parametersa

drw K1 (d�1) a Kw (m3 (kg d)�1) B T0.5 (d)

Phytoplankton 0.1 e e e e e

Zooplankton 0.1 1.0 0.2 1.5 0.001 5Molluscs 0.1 0.06 0.5 0.15 0.001 50Crustaceans 0.1 0.015 0.5 0.1 0.001 100Non-piscivorous fish 0.25 0.030 0.5 0.1 0.001 Table 2Piscivorous fish 0.3 0.0055 0.7 0.075 0.001 Table 2

a See text for definitions of parameters.

activity concentration in the target tissue diluted by the remainingbody mass of the prey fish, calculated by multiplying the predictedlevel in the target tissue by its weight fraction. To calculate theconcentration in the edible part of fish (flesh) from the calculatedlevels in the target tissues, a target tissue modifier (TTM) is intro-duced. This is based on tissue distribution information as reportedby Coughtrey and Thorne (1983), and confirmed by the conversionfactors fromYankovich et al. (2010). Values of described parametersfor the dynamic food-chain model are listed in Table 2. To calculatethe average levels in the body, in the case when whole fish isconsumed, the predicted level in the target tissue is multiplied bythe weight fraction of the target tissue.

Phytoplankton receives radionuclides only from the water viaadsorption and absorption processes. Due to the rapid uptake andshort retention time of radioactivity the concentration of radio-nuclides in phytoplankton is calculated using the BCF approach

Cphyt ¼ CFphytCw; (4)

where CFphyt ¼ 20 L kg�1 is the concentration factor (IAEA, 2004).The same approach is used for macro-algae, which is consumed as amarine product. The corresponding concentration factor CFma isCFma ¼ 50 L kg�1 (IAEA, 2004).

2.3. Description of sources of radionuclides

The model POSEIDON-R can deal with three types of radioactivereleases: (i) atmospheric fallout, (ii) point sources associated withroutine releases of nuclear facilities, located either directly at thecoast or inland at river systems, and (iii) point sources associatedwith accidental releases. For coastal discharges occurring into large(“regional”) compartments, it is useful to provide a more detaileddescription in the area close to the release point. For that purpose,“coastal” release compartments can be added to the regionalcompartmental system. These coastal compartments are nestedinto the regional compartments, and their physical characteristics(depth, sedimentation, etc.) can differ from those of the adjacentregional compartments. The location of the routine release point isfixed. The approach implies, however, some assumptions and re-strictions: (i) A “coastal” compartment has one vertical layer for thewater column, and a standard multi-layered bottom sedimentcompartment, (ii) A “coastal” compartment interacts with a surfacelayer of one regional compartment only (the depth of “coastalcompartment” is therefore less or equal to that of the surface layerof the regional compartment), (iii) Water exchange fluxes with theadjacent regional compartment are equal in both directions (i.e.only lateral diffusion is taken in account), (iv) Only one “coastal”release compartment can be added per regional compartment, and(v) A “coastal” compartment contains at most one source ofradioactivity.

When calculating the radionuclide concentrations in fish,random fish migration must be taken into account by including theproportion of the residence time of the fish between the “coastal”compartment, and the adjacent regional compartment. For thispurpose, the right hand side of equation (1) for radionuclide

Table 2Parameters for the fish in dynamical food chain model.

Target tissue Bone Flesh Organs Stomach

Weight fraction, f 0.12 0.80 0.05 0.03Target tissue modifier (TTM) 0.001 1 0.1 0.01Biological half-life of prey (d) 500 75 20 3Biological half-life of predator (d) 1000 200 40 5

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 7

concentration in fish, both in the coastal and outer regional com-partments, is extended by the term ðCout

fish � Ccoastfish Þ=Tmigr for coastal

compartment and by the term �ðCoutfish � Ccoast

fish Þ=ðdTmigrÞ, for theouter regional compartment. Tmigr is the characteristic time of fishmigration from a coastal compartment, depending on compart-ment scale and fish species, and d is the ratio of outer regionalcompartment volume to the coastal compartment volume.

2.4. Numerical solution

The problem is described by a set of ordinary differentialequations, which may be written in a vector-matrix notation as:

dCdt

¼ ACþQ re; (5)

where C is the concentration vector, A is the coefficient matrix thatincludes water fluxes between compartments, parameters of thefood-chain model, etc, and Qre is the vector for the release term.Step-like variations of the release in time are assumed and theMatrix Exponential Method (Patten, 1971) is used to solve thissystem (Lepicard et al., 1998).

2.5. Dose to humans from ingestion of contaminated marineproducts

Individual effective dose rate (IDR), resulting from the ingestionof five categories (f) of marine products (piscivorous and non-piscivorous fish, crustaceans, molluscs and macro-algae), can becalculated for a given i-th compartment, from equation

IDRi ¼Xm

j¼1

DCFjX5

f ¼1

Intakeðf ÞCðf Þij ; (6)

where Cðf Þij is the activity concentration of j-th radionuclide in the

marine product of type f for i-th compartment, m is the number ofradionuclides under consideration, Intake(f) is the marine foodintake rate, and DCFj is the dose coefficient for j-th radionuclide,from ICRP (1995). The individual effective committed dose IDi iscalculated by the integration in time of Eq. (6) for a prescribed timeperiod.

Collective dose rate (CDR) resulting from ingestion of marineproducts is calculated as

CDR ¼Xm

j¼1

DCFjX5

f ¼1

4ðf Þ Xn

i¼1

Hðf Þi Cðf Þ

ij ; (7)

where n is number of compartments, Hðf Þi is the catch of marine

products of type f in i-th compartment, 4ðf Þ is the edible fraction forthemarine product of type f. Default values for the edible fraction of0.35, 0.15, 0.2 and 0.5 are used for crustaceans, molluscs, macro-algae and fish, respectively (Smith and Simmonds, 2009). Howev-er, they can differ by regions (e.g. edible fraction of macro-algae inFar-East countries is almost 1). Therefore, in the paper it is assumedto be equal 1. The edible fraction for fish is lower than the value forthe weight fraction 0.8 for flesh in Table 2 which relates to all tis-sues except the intestines, bones, and guts, and is therefore partlynot edible. Muscles tissues are about 65% of the total body weight(Yankovich, 2003), and it is assumed that about 20% of this musclefraction is not consumed, which means 50% of the entire fish. Thecollective effective dose is calculated by time integration of Eq. (7).

3. Model setup for Northwestern Pacific and adjacent seas

The model was customized for the Northwestern part of thePacific Ocean, for the East China and Yellow Seas, and the East/JapanSea (Fig. 2a). A total of 175 compartments cover this entire region(Fig. 2b). The horizontal size of the compartments along the coast toboth longitudinal and latitudinal directions is about 1�. Near theNPPs, these compartments become smaller. In the deep-sea re-gions, a three-layer compartmental system was built to describethe vertical and horizontal transport of radioactivity in the upperlayer (0e200 m), in the intermediary (ocean thermocline) layer(200e1000 m), and in the lower (abyssal) layer (>1000 m). Thevolumes of the compartments were calculated on the basis of thebathymetry of the region from ETOPO5.

The averaged water fluxes between compartments were calcu-lated for a ten year period (2000e2009), with moderate resolution(1/9� horizontally and 20 vertical levels), using the Regional OceanModelling System (ROMS). Details of model setup and forcing aregiven by Cho et al. (2009). Averaged over this period, currents in theupper and intermediate layers of the region are shown in Fig. 3. Thecharacteristic feature of this circulation is the Kuroshio Current,transporting water along the eastern coast of Japan to the openocean. The Kuroshio branch in the East/Japan Sea also transportswater to the north and further onto the Pacific Ocean. Thecomputed velocities have been interpolated to calculate the time-averaged advective and diffusive water fluxes through the verticalsurfaces of the connected compartments. Following Goshawk et al.(2003) the additional diffusion fluxes were included for boxes inthe Yellow Sea to take into account mixing by tides. The tidalvelocities were obtained from Choi (1986). Coefficients of bio-turbation (B) and diffusion (D) in the bottom sediments wereadopted from the “Marina” project (CEC, 1990). Concentrations ofsuspended sediments in the East/Japan Sea and Northwestern Pa-cific were 1 and 0.1 g m�3, respectively, and sedimentation rates inthe East/Japan Sea and Northwestern Pacific were 75 and10 g m�2 y�1, respectively. Concentration of suspended sedimentsand sedimentation rates in the Yellow and East China Seas weretaken from Choi et al. (2005).

4. Results of comparison of 137Cs dispersion modelling withobservation data from 1945 to 2010 for East China and YellowSeas and East/Japan Sea

4.1. Source of 137Cs

The main source of 137Cs in the Northwestern Pacific in theperiod 1945e2010 was from fallout due to atmospheric nuclearweapon tests. The fallout includes: (i) a global component, causedby the transport of radioactivity due to the general atmosphericcirculation and subsequent deposition on the surface of the sea and(ii) a regional component, caused by fallout from weapon testscarried out at Bikini Atoll and at the Enewetak Atoll, resulting incontamination of the surface layer of the ocean. Hamilton et al.(1996) estimated that ocean contamination of fallout due torunoff of radionuclides from terrestrial surfaces contributes nomore than 10% of the oceanic inventory from global fallout. For thisreason, and because few empirical data exist, the terrestrial sourceof 137Cs was not considered in our model. The annual deposition of137Cs on the ocean for period 1945e2005, compiled from UNSCEAR(2000), Nakano (2006) and Hirose et al. (2008), is shown in Fig. 4a.There are some differences between deposition rates for differentlatitudes but according to Nakano (2006) this difference is small forlatitude 30�e50� N. Therefore averaged deposition rates forNorthwestern Pacific were used.

Fig. 2. Bathymetry map of the Northwestern Pacific with adjacent seas (a) and the compartment system (b). The shaded boxes represent the deepwater boxes. The boxes with shadedlines present the areaswithdeposition from the FukushimaDai-ichi accident. The arrowswithnumbers showthe compartments representingestuaries of large rivers (174e theYangtzeRiver, 173 e the Huanghe River and 175 e the Han River). The NPPs are shown by filled circles. Letters “S” and “F” represent the Shimane NPP and the Fukushima Dai-ichi NPP.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e188

Fig. 3. Ten-years-averaged currents in the upper layer (a) and the intermediate layer (b) of the considered region.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 9

Fig. 4. Time variations of the annual deposition on the surface compiled from Nakano(2006) and Hirose et al. (2008) (a) and the boundary values for the 137Cs concentrationin the NW Pacific compiled from MARIS (2012) database and Kang et al. (1997) (b).

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e1810

The concentration of 137Cs in the upper layer of ocean at theeastern and southern boundaries of the computational domain(Fig. 4b) was estimated by using both observations from the MARIS(Marine Information System) database (MARIS, 2012), and obser-vations from Kang et al. (1997). These values represent both theeffect of global deposition of 137Cs on the Northeastern Pacific andregional effect of weapon tests carried out at Bikini Atoll and at theEnewetak Atoll. The concentration in the intermediate layer ofocean was set according to observed profiles of 137Cs in theNorthwestern Pacific (Nakano and Povinec, 2003) as 1/5, 1/3 and 1/2 of surface concentration values for the periods 1965e1974, 1975e1984 and 1985e2004. To predict the 137Cs concentration for theperiod 2005e2010, five-year-averaged deposition and the bound-ary concentrations over the period 2000e2004 were extrapolatedand corrected for radioactive decay. The time step for period 1945e2010 was 1 y.

4.2. Results

The calculated 137Cs concentrations in the compartments werecompared with observed data from various publications and fromthe MARIS (2012). These data were collected for the period 1960e2010 in the East China and Yellow Seas and the East/Japan Sea. Thecomparison of the predicted 137Cs concentration in the waters ofthe East China Sea with observations as published by Nagaya andNakamura (1992) and data from the MARIS database (for com-partments 17,19, 20 and 36) is given in Fig. 5. The plots demonstrate

a good agreement with the observed data. The fraction of theradiocaesium concentration caused by transport via the ocean fromthe location of the bomb tests in the Central North Pacific is largerthan the fraction from atmospheric deposition.

In Fig. 6 the model predictions for the Yellow Sea are comparedwith observations (Nagaya and Nakamura, 1992; Hong et al., 2006)and with data from the MARIS database compiled for compart-ments 44, 46, 63 and 71. Whereas, in the southern part of theYellow Sea, the ocean transport component dominates over theatmospheric deposition, in the northern part of Yellow Sea thecontribution of the atmospheric fallout equals the transport via theocean due to weak circulation of seawater. The concentration of137Cs in the bottom sediments in 2005 varied between 1 and9.5 Bq kg�1, showing good agreement with the observed range(Hong et al., 2006).

The calculated inventories of 137Cs for the East China and YellowSeas for the period 1945e2005 are given in Table 3. The largestcontribution to the budget was from the Pacific Ocean inflow (93%),whereas the global fallout contribution was 7.2%, and 6.8% of 137Csdecayed. The outflow is divided between the Tsushima/Korea Strait(38.4 PBq) and an outflow (2.5 PBq) to the Pacific, south of Kyushuisland (Fig. 3). Note, that riverine input into the East China Sea wasestimated by Su and Huh (2002) as 0.83 PBq . In 2005 the amountsof 137Cs in the water and in the bottom sediment were 0.08 PBq(0.017% from total input) and 0.23 PBq (0.5% from total input),respectively. The modelling results show a low spatial variability of137Cs concentration in the East/Japan Sea. As seen in Fig. 7a, thepredicted time series of 137Cs in the surface layer of the sea arealmost identical for the different compartments (boxes 115, 133,156). They agree well with observations (Nagaya and Nakamura,1981, 1987; Miyake et al., 1988; MSA, 1995, 1998, 1999; Kang etal., 1997; Miyao et al., 1998; Aoyama and Hirose, 2004) carriedout in different locations in this sea except for two periods, duringwhich the global fallout was maximal. The first period, inwhich thelevels exceeded the model results, was due to atmosphericweapons tests (maximal fallout in 1963) and the second period wasafter the Chernobyl accident (peak fallout in 1986). In these periods,radionuclides are concentrated in a relatively thin surface layerwhereas the thickness of upper layer in the model is 200 m, leadingto under-predictions of the concentration. The comparison withlong term measurements by JCAC (2004) in the East/Japan Sea inthe area around the Shimane NPP (for location, see Fig. 2b) shows agood agreement (Fig. 7b). In agreement with themeasurement data(Nagaya and Nakamura, 1987; Hong et al., 1999; Otosaka et al.,2006) the predicted 137Cs concentration in the seafloor of theEast/Japan Sea in 2005 was in the range 0.03e7.3 Bq kg�1.

The dominant factor in the 137Cs budget of the East/Japan Sea, asseen in the Table 3, is inflow from the Tsushima/Korea Strait (90% oftotal influx), whereas the global fallout contributed 10%. Theoutflow of 137Cs via the Tsugaru Strait and Soya Strait was 36.6 PBq(85% of total influx), whereas 10% of 137Cs has decayed. In 2005, theamount of 137Cs in the water column and in the bottom sedimentswas 1.57 PBq (3.7% of total influx) and 0.12 PBq (0.2% of total influx),respectively. These estimates are comparable with inventory of137Cs as calculated by Nakano (2006), where inflow, fallout, decayand outflow were 44 PBq, 7 PBq, 3.9 PBq, 7.7 PBq and 38.3 PBq,respectively. The amount of 137Cs in the water column was 2.6 PBqin 2005 (Nakano, 2006).

The comparisons (Figs. 5e7) demonstrate that POSEIDON-R,customized for the Northwestern Pacific, the East China and Yel-low Seas and the East/Japan Sea, correctly describes the transportand fate of 137Cs in this region. The correlation between calculatedand observed values of 137Cs concentration in the water is shown inFig. 8. The corresponding correlation coefficient is 0.925, the root-mean-squared-error (RMSE) is 1.43 Bq m�3 and the bias error

Fig. 6. Comparison of the calculated 137Cs concentration in water of the Yellow Sea with measurements for compartments 44 (a), 46 (b), 63 (c) and 71 (d).

Fig. 5. Comparison of the predicted concentration of 137Cs in the waters of the East China Sea with observations from the MARIS (2012) database for compartments 17 (a), 19 (b), 20(c) and 36 (d).

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 11

Table 3Estimated 137Cs budget in PBq for the East-China Sea and Yellow Sea, and for theEast/Japan Sea for period 1945e2005.

Seas Influx Outflux Globalfallout

Bottomdeposition

Decayed Watercolumnin 2005

Bottomsedimentsin 2005

East Chinaand YellowSeas

41.4 40.9 3.2 0.57 3.05 0.08 0.23

East/JapanSea

38.4 36.6 4.3 0.23 4.3 1.57 0.12

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e1812

(mean difference between modelled and observed radioactivity) is0.06 Bq m�3.

Fig. 8. Correlation between the predicted and the measured 137Cs concentration in thesurface water from 1960s to 2005 (all compartments). The dashed lines correspond to

5. Application of POSEIDON-R to the Fukushima Dai-ichi NPPaccident

5.1. Source of 134Cs and 137Cs

The Fukushima Dai-ichi NPP accident was caused by an earth-quake of 9.0 on the Richter scale, which occurred onMarch 11, 2011,and was followed by a devastating tsunami. Large amounts of

Fig. 7. Comparison of the calculated 137Cs concentration in water of the East/Japan Seawith measurements for compartments 115, 133, 156 (a) and (b) for Shimane area of theEast/Japan Sea (see Fig. 2b for location).

ratio 2 and 1/2 of predicted to measured values.

radioactive materials were released into the atmosphere and intothe marine environment. High concentrations of 131I and 134,137Cswere detected in the water, in the bottom sediments and in themarine biota along the coastal waters of Japan. Therefore it isnecessary to assess the radiological impact on the population fromyears to decades. For this purpose, the POSEIDON-R model wascustomized with the same schematization as was used for theweapon test case study, with the addition of an extra coastalcompartment near the location of the Fukushima Dai-ichi NPP,within regional box 90 (see Fig. 2b). It was chosen to cover obser-vation data within a circular-shaped surface area of a radius 15 km,with a centre at the Fukushima Dai-ichi NPP. The area of the coastalcompartment (30 km along the coast and 15 km across it) was450 km2, which is 22 times smaller than the area of compartment90, while the volume of the coastal compartment is 150 timessmaller than the volume of compartment 90. The characteristictime of fish migration used from coastal compartment Tmigr was0.67 y.

The simulation for the period 1945e2010 described in the pre-vious section was continued for 2011e2040 with a source termestimated from the Fukushima accident. It was assumed thatrelease of activity directly to the ocean took place over the period1e10 April 2011. Amounts of 5 PBq of 134Cs, and 4 PBq of 137Cs weretransferred directly into the ocean. These quantities are in accor-dance with widely accepted source terms for the Fukushima acci-dent simulations (see Kawamura et al., 2011; Nakano and Povinec,2012; Masumoto et al., 2012; Tsumune et al., 2012). The atmo-spheric deposition data was obtained from simulations with theMATCH model where the dispersion of 137Cs for the period 12Marche5 April was computed. The ECMWF meteorological datawith a source term reported by Stohl et al. (2012) was used in thesimulation. MATCH is an Eulerian model for long-range transport ofradionuclides (Robertson et al., 1999). The amount of deposited137Cs in the computational domain was 8.5 PBq. The deposition of134Cs was estimated at 10.2 PBq using an activity ratio134Cs/137Cs ¼ 1.2 (NISA, 2011). Note that the atmospheric releasetook place two weeks prior to start of the time period of the model

Fig. 9. Comparison between calculated and observed 137Cs concentration in seawater(a) and in bottom sediments (b) around the Fukushima Dai-ichi NPP.

Fig. 10. Comparison between calculated and measured 137Cs concentration in non-piscivorous fish (a), piscivorous fish (b) and mollusks (c) in the coastal compartmentnear the Fukushima Dai-ichi NPP.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 13

scenario; however it did not affect the results since the integrationtime step for 2011e2012was onemonth. As in the previous section,the boundary concentrations over the period 2000e2004 wereextrapolated and corrected for radioactive decay to predict the137Cs concentration for the period 2005e2040. The time step forperiod 2013e2040 was 1 y.

5.2. Results of the simulation and impact of additional releases

A comparison of the predicted 137Cs concentrations with mea-surements in the water and in the bottom sediments in the coastalregion around Fukushima Dai-ichi NPP is shown in Fig. 9. Twosources of 137Cs measurements have been used to compare withmodel predictions in Fig. 9a; time-series of 137Cs concentration inwater before the Fukushima accident were obtained from MEXT(the Japanese Ministry of Education, Culture, Sports, Science andTechnology) environmental radiation database (MEXT, 2010),whereas 137Cs measurements after the Fukushima accident wereobtained by TEPCO (Tokyo Electric Power Company) within an areawith a radius 15 km (TEPCO, 2013a). The data plotted in Fig. 9awereweekly averaged for 2011, whereas data were monthly averaged for2012e2013. The model predicts that after the accidental release137Cs concentrations in the seawater rapidly decrease to reachbackground levels in 2013. The modelling results and measure-ments before and after the accident agree rather well, except 2013,when the observed concentration of 137Cs decays slower thanpredicted. The phenomenon indicates the presence of an additional

continuous source of radioactivity. This source can be an under-ground leak of contaminated water from Fukushima Dai-ichi NPPand/or river run-off. Kanda (2013a,b) suggests a release from theNPP in 2012 of 3.6 TBq y�1 using the measurements of 137Cs con-centration in the harbour of NPP. According to Kanda (2013a) thetotal river flux of 137Cs in the prefectures Fukushima, Ibaraki andMiyagi is around 1.56 TBq y�1. We assume a continuous flux fromthe NPP of 3.6 TBq y�1 during the period 2012e2014. The river fluxwas distributed between the various compartments according toKanda (2013a). The results obtained with the additional sources aregiven in Fig. 9a. They agree with measurements and confirm thepresence of a continuous leak from the Fukushima Dai-ichi NPP.The influx from rivers does not affect the concentration in thecoastal box. The alleged ending of the flow of 137Cs in 2015 results ina return to background concentration within one year.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e1814

The model predicted 137Cs concentration in the bottom sedi-ments of the coastal compartment matched the measurementscarried out by MEXT before and TEPCO after the accident(Fig. 9b). Note that the results of measurements by TEPCO afteraccident show strong patchiness of 137Cs concentrations. Theavailable data (TEPCO, 2013b) suggest that the distributionfunction of 137Cs concentrations in the bottom sediments isapproximately lognormal. The expected geometric mean valueand standard deviation were obtained using maximal likelihoodestimator for lognormal distribution (Baker and Gibson, 1987).The 137Cs concentrations in the bottom sediments decrease, asexpected, significantly slower than in the seawater due to slowprocesses of sedimentewater exchange; elevated levels ofradioactivity in the sediments around Fukushima Dai-ichi NPPwill persist for a long time. Note, that the continuous sources of137Cs do not affect bottom contamination in the coastalcompartment.

In Fig. 10a and b, the predicted 137Cs concentration in fish in thecoastal compartment is compared with measurement data (MEXT,2010) before the accident and with data from the Japan FisheriesResearch Agency (JFRA, 2012) and from TEPCO (2013b) after theaccident. All data on different fish species for the area with a radius15 km around the Fukushima Dai-ichi NPP were grouped intonon-piscivorous and piscivorous fishes, and subsequently month-ly-averaged values are plotted in Fig. 10a and b. Due to strongvariations in the 137Cs concentration in fish the geometric meanvalues and standard deviations were also obtained using maximallikelihood estimator for lognormal distribution. Fig. 10a and b alsoshows the 137Cs concentration in fish calculated with a BCF value of100 L/kg. The BCF approach for piscivorous fish leads to a poordescription of the observed concentrations due to the assumptionof immediate equilibrium between fish and the seawater. The dy-namic food-web model, on the contrary, predicts a much later

Fig. 11. The predicted 137Cs concentration in water (a), bottom sediments (b) and fish (c) in bcompartments 90 (d) for the period 1950e2040.

return to the background level due to the delay caused by thetransfer of radioactivity throughout the food web. The BURNmodelpredicts that the concentration of 137Cs for non-piscivorous andpiscivorous fish returns to background concentrations only in 2014and 2016, respectively, instead of in the year 2013, as predicted bythe BCF approach. In contrast to the BCF approach, the modelpredicted concentrations of 137Cs in piscivorous fish agree withmeasurements that indicate 137Cs concentrations in non-piscivorous decreasing faster than those in piscivorous fish, dueto the relatively longer biological half-life of the top-predator andits position on the top of the food-web. However, the calculatedconcentration in non-piscivorous fish using the BCF approachshows good agreement with the measurement data for 2011e2012.As seen in Fig.10c the 137Cs concentrations in mollusks predicted bythe BURN model decay like the observed concentrations (JFRA,2012); whereas, the BCF approach, with the standard value forthe concentration factor Cmol

f ¼ 60 L kg�1 (IAEA, 2004), results in afaster decay than the observed concentrations. The additional fluxslows the decay of 137Cs concentration in molluscs from 2013, andin non-piscivorous and piscivorous fish from 2013 and 2014,respectively. The corresponding difference in concentration withand without additional sources reaches 1 Bq kg�1.

The predicted concentrations for 137Cs in water, in sedimentsand in fish in regional compartment 90 are shown in Fig.11aec. Thecalculated 137Cs concentration in water agrees well with the MEXTmeasurements around the Fukushima Dai-ichi NPP for the period1984e2010, and with results of the simulations by Nakano andPovinec (2012) for the Northwestern Pacific. Before 2000, theconcentration of 137Cs in the bottom sediments in compartment 90agrees well with measurements (MEXT, 2010), but in the lastdecade the calculated values exceed the observed values. The 137Csconcentration in piscivorous fish agrees well with data (MEXT,2010) for the period 1984e2010. The figure confirms that the

ox 90, and individual ingestion dose rate due to consumption of marine products from

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 15

change in the 137Cs concentration due to the Fukushima accident inthe relatively large compartment 90 is significantly less than thepulse in the smaller coastal compartment in Figs. 9 and 10. Theadditional flux does not affect 137Cs concentrations in water, sedi-ments and biota.

The individual ingestion dose rate due to consumption ofcontaminated marine products from compartment 90 was esti-mated on the basis of statistical data on the consumption of marineproducts in Japan as given by Nakano and Povinec (2012). Ac-cording to these data, the annual consumption of marine productsis 23.4 kg of fish, 2 kg of crustaceans, 1.3 kg of mollusks, and 3.7 kgof macro-algae per year. We assume that 20% of consumed fishspecies are piscivorous species. Fig. 11d shows the individualingestion dose rates due to consumption of marine products con-taining 134,137Cs from compartment 90. Note that marine productsfrom the highly contaminated coastal compartment (Fig. 10) werenot included in this dose estimation due to the ban of catchmentaround the Fukushima Dai-ichi NPP. The dose rate contribution ofthe two caesium-isotopes, 134Cs and 137Cs, for the year 2011 is2.26 mSv y�1 and 1.34 mSv y�1, respectively. The relative contribu-tion to the dose from the consumption of the different marine or-ganisms is 83% due to fish, 4% due to crustacean, 4% due to mollusk,and 9% due to macro-algae consumption. The calculated individualdose rates for 137Cs due to the Fukushima Dai-ichi accident are ofthe same order as the maximal dose rate from weapons tests in1957. The calculated maximal individual dose rate for 2011(3.6 mSv y�1) is higher than estimate as given by Nakano andPovinec (2012) (i.e. 1.7 mSv y�1). Note that the dose estimation(Nakano and Povinec, 2012) was performed only on the maximalvalues for the predicted radioactivity in the open NorthwesternPacific water, and the concentration in marine organism wascalculated using the BCF approach. However, when POSEIDON-R isused and the dose rate is calculated by means of the BCF approachthe maximal values for the dose rate in 2011 would be 3.0 and1.8 mSv y�1 for 134Cs and 137Cs, respectively. The values for the doserate estimates calculated with the dynamic food web model BURNare relatively low, due to delayed uptake to species caused by theradionuclide transfer throughout the food-chain instead of transferdirectly from the water as is implicitly the case in the BCF approach(see Fig. 10b). Finally, in the hypothetical case of a worse-casereference group in the Fukushima region that consumes only ma-rine products from the local compartment near the Fukushima NPP,the annual dose for 2011 would be much higher: 1.8 mSv y�1.

Fig. 12. The collective ingestion dose rate for the Japan, China and South Korea pop-ulation due to consumption of marine products in period 2000e2020.

For the calculation of the collective ingested dose rate due to theconsumption of marine products available statistical records forJapan, Korea and China on fishery production (NFRDI, 2000; BFMA,2010; MAFF, 2011; MFAFF, 2012) have been used to distribute thefishery production over the POSEIDON-R compartments (seeTables A1eA3 in Appendix). The import/export of marine products,market dilution and the losses due to food preparation and cookinghave been excluded, which leads to a conservatively high value forthe dose. The calculations for the period 2000e2020 show (Fig. 12)that collective dose rate for Japan due to consumption of marineproducts after the Fukushima accident increased in 2011 by a factorof 6 (31.7 humanSv y�1) in comparison with 2010(5.1 humanSv y�1). After 2011 it takes 4 y before the collective doserate returns to the background level of 4.8 manSv y�1. Note that thecontribution of atmospheric fallout from the Fukushima accident tothe collective ingestion dose in 2011 is 83%, exceeding the effect ofthe direct release to the ocean (17%). This can be explained by thefact that although the activity concentration in marine products inthe compartments close to the NPP is higher than in the com-partments in the open ocean, the fish catch in compartments wherethe direct release dominates the 137Cs concentration is relativelylow in comparison with the catch in larger open ocean compart-ments (shaded boxes in Fig. 2b) where deposition governs the 137Csconcentration.

6. Conclusions

The POSEIDON-R compartmentmodel predicts the dispersion ofradioactivity in the water column, the transfer of radioactivity tosediments and throughout the marine food web, and subsequentlycalculates both the individual and collective dose to the populationresulting from the consumption of marine products. A predictivegeneric dynamic food-chain model, BURN, was used instead of thecommonly used BCF approach. The radionuclide uptake model forfish was based on radionuclide concentrations in the tissue withthe highest accumulation and longest retention time. The modifiedmodel was customized for the Northwestern Pacific and the adja-cent seas where about 300 nuclear units are expected to be locatedby the year 2030. In total 175 compartments were used to representthe Northwestern Pacific, the East China and Yellow Seas, and theEast/Japan Sea. Water fluxes between the compartments werecalculated by averaging the three-dimensional currents computedby the circulation model ROMS over a 10-y period.

The model was validated with observed 137Cs data in water forthe period 1945e2010. The source terms were both regional andglobal; the regional source term was from the weapon tests onEnewetak and Bikini Atolls, and the global source term was fromglobal 137Cs deposition following atmospheric nuclear weapontests. Dominating the 137Cs budget was the transport by oceancurrents from weapons test areas in the Pacific Ocean. The modelpredictions and observations agree well; the correlation coefficientbetween predicted and observed concentrations of 137Cs in thesurface water was 0.925, with a value for the RMSE of 1.43 Bq m�3

and for the BE of 0.06 Bq m�3.The model was used to perform a radiological assessment for

the period 2011e2040, based on radioactive releases to the oceanfrom the Fukushima Dai-ichi accident. To describe the local effectsof the accidental release of the Fukushima Dai-ichi NPP, a local-scale coastal compartment was defined in accordance with PO-SEIDON’s methodology to describe activity transport, depositionand food web transfer around the release location. The predictedbackground concentrations of 137Cs inwater, bottom sediments andfish in the area around the Fukushima Dai-ichi NPP appear to agreewell with observations before the accident (1984e2010). Calculatedconcentrations of 137Cs in the water, bottom sediments and marine

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e1816

organisms in the coastal box after the accident are close to themeasurements in 2011e2013, as performed by the Japaneseagencies. The agreement for water is achieved when an additionalcontinuous flux of 3.6 TBq y�1 is used that represent undergroundleakage of contaminated water from the Fukushima Dai-ichi NPP.As expected, the dynamic food web model reproduces the mea-surements better than BCF approach. Due to delay in transferthroughout the food web from lower to higher trophic levels, theconcentration of 137Cs for piscivorous fish returns to backgroundconcentrations in 2016.

For 2011, the calculated individual dose rate for Japanese in-habitants due to consumption of marine products from coastalwater of Fukushima Prefecture is higher than the maximal indi-vidual dose rate as observed in the year 1959 (weapon tests).Following the Fukushima Dai-ichi accident, the collective dose dueto ingestion of marine products for Japan increased in 2011 by afactor 6 in comparison with 2010.

The validation of the POSEIDON-R model using historical data,and the comparison of the model results with the observation datafrom the Fukushima Dai-ichi accident, demonstrate the potentialfor compartmental models to contribute to dose assessmentsof radioactivity in the marine environment. With further develop-ment and improvement of the model the attention should be givento refinement of the model parameters and input data on the basis

Appendix

Table A1Korean marine products catchment (tons y�1).

Box number Fish Crustacean Molluscs Algae

6 610,000 e e e

19 162 e e e

26 1182 e e e

27 30,188 e e e

32 2218 640 e e

33 92,974 4484 e e

34 4121 427 e e

43 468 e e e

44 110,235 1010 1230 e

45 21,864 1259 351 e

46 65,510 369 43 e

47 109 e e e

49 25,091 427 e e

50 8874 369 109 e

51 1725 e e e

52 530 e e e

57 9252 e 175 e

58 11,460 427 43 e

59 47,833 213 e 24060 1930 213 27,608 249,11661 123,252 78 27,608 249,11662 131,900 3 62,255 208663 1597 e 21 e

Table A2Japanese marine products catchment (tons y�1).

Box number Fish Crustacean Molluscs Algae

3 102,472 7154 50,242 11,8844 102,472 7154 50,242 11,88420 5105 622 365 544221 42,357 231 47 535 5105 622 65 4236 5105 622 365 544237 24,336 3397 44 16438 24,336 3397 44 16439 76,836 3397 544 3564

of detailed observations following the Fukushima Dai-ichi accident(especially the amount of underground leakage from the NPP) andmore detailed description of the benthic food web (Tateda et al.,2013). The lognormal distribution of concentrations of caesium inthe bottom sediments and in the fish is of special interest. Our in-terest for the next study is to include the effects of seasonal vari-ation of circulation in the region, since the marginal seas in theregion are Monsoon-dominated.

Acknowledgements

Once this work has been submitted, our co-author Rudie Helingpassed away. His contribution to the development of a dynamicmodel was invaluable. We dedicate this article to his memory. Thiswork was supported by KIOST project “Development of the marineenvironmental impact prediction system following the disastrousenvironmental event”, CKJORC (ChinaeKorea Joint Ocean ResearchCenter) Project for Nuclear Safety, FP7-Fission-2012 project PRE-PARE “Innovative integrative tools and platforms to be prepared forradiological emergencies and post-accident response in Europe”and it is in frame of IAEA MODARIA Programme “Modelling anddata for radiological impact assessments”. This article benefitedfrom the comments and suggestions of three anonymousreviewers.

Box number Fish Crustacean Molluscs Algae

64 5929 213 e e

65 e 86 4155 12,07667 447 592 e e

68 2757 e 21 e

69 5980 e e e

70 e 7136 19,257 11,17571 4571 4270 21 e

72 71 548 e 31876 122 e e e

96 655 e e e

97 591 e e e

99 38 52,422 62,286 2085100 6317 52,422 98,925 2085101 412 e e e

102 27 e e e

105 1453 901 99,436 12,064113 20,441 429 48,114 12,064114 158 e 64 e

121 4610 e 3267 e

122 2081 e 17,706 e

123 e e 3207 e

125 e e 22,453 e

126 e e 32 e

Box number Fish Crustacean Molluscs Algae

104 18,798 3190 187 33107 14,551 1784 551 156108 64,489 4952 580 45109 18,798 3190 187 33110 15,531 572 996 249111 4888 1324 135 205112 32,316 10,470 418 1072116 18,798 3190 187 33117 4770 1324 135 25

Table A3Chinese marine products catchment (tons y�1).

Box number Fish Crustacean Molluscs Algae Box number Fish Crustacean Molluscs Algae

12 52,958 29,487 708,616 184,771 40 10,784 16,616 147,823 815613 52,958 29,487 708,616 184,771 41 10,784 16,616 147,823 815614 165,684 91,936 49,495 374 42 56,728 28,532 15,969 21315 52,958 29,487 708,616 184,771 43 56,728 28,532 15,969 21316 165,684 91,936 49,495 374 53 10,784 16,616 147,823 815617 165,684 91,936 49,495 374 54 10,784 16,616 147,823 815622 8086 19,882 152,900 10,068 55 159,332 109,005 845,594 126,87223 8086 19,882 152,900 10,068 56 56,728 28,532 15,969 21324 267,606 168,208 31,908 474 57 56,728 28,532 15,969 21325 267,606 168,208 31,908 474 66 159,332 109,005 845,594 126,87226 267,606 168,208 31,908 474 67 159,332 109,005 845,594 126,87228 8086 19,882 152,900 10,068 73 49,289 87,507 300,799 e

29 8086 19,882 152,900 10,068 74 159,332 109,005 845,594 126,87230 1282 9908 247 e 75 103,855 109,918 916,690 124,19232 267,606 168,208 31,908 474 76 103,855 109,918 916,690 124,192

Table A2 (continued )

Box number Fish Crustacean Molluscs Algae Box number Fish Crustacean Molluscs Algae

47 10,489 751 1573 511 118 32,316 10,470 418 107248 31,989 751 1853 40,911 119 11,966 1522 352 85351 133,925 10,508 1524 833 120 11,966 1522 352 85352 148,225 10,508 3274 7833 128 4770 1324 135 2578 52,457 231 47 905 129 48,070 6183 382 7579 59,944 2433 883 2684 130 11,966 1522 352 10380 249,963 7940 127,601 4407 131 3597 1553 231 1981 67,402 10,947 1210 457 132 3597 1553 231 1982 152,402 10,947 33,210 38,457 134 2287 263 99 1083 210,865 3480 30,227 5435 135 2287 263 99 1084 217,865 3480 37,227 41,435 136 2287 263 99 1085 171,270 7918 1803 671 137 2287 263 99 1086 171,270 7918 1803 7871 138 2287 263 99 1087 29,643 1474 112 2 149 50,969 27,563 1420 163488 29,643 1474 112 2 150 50,969 27,563 54,485 185989 29,643 1474 112 2 151 102,472 7154 72,387 16,38490 42,015 3010 688 1 156 102,472 7154 50,242 11,88491 185,939 21,486 89,646 105,443 157 102,472 7154 50,242 11,88492 185,939 21,486 1246 1443 158 102,472 7154 72,387 16,38493 50,969 80,628 54,485 1859 160 102,472 7154 50,242 11,88494 102,472 7154 72,387 16,384 161 102,472 7154 50,242 11,88495 102,472 7154 72,387 16,384 162 102,472 7154 72,387 16,38496 15,254 1096 2993 67,592 164 102,472 7154 50,242 11,88497 24,193 2489 2196 45,353 165 102,472 7154 72,387 16,38498 14,811 1784 551 3356 170 29,643 1474 112 2101 21,943 2489 2196 353 171 29,643 1474 112 2102 14,551 1784 551 156 172 42,015 3010 688 1103 64,529 4952 820 405

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e18 17

References

Aoyama, M., Hirose, K., 2004. Artificial radionuclides database in the Pacific Ocean:HAM database. Sci. World J. 4, 200e215.

Baker, M.A., Gibson, C.H., 1987. Sampling turbulence in the stratifiedocean: statistical consequences of strong intermittency. J. Phys. Oceanogr. 17,1817e1836.

BFMA e Bureau of Fisheries, Ministry of Agriculture, China, 2010. China FisheryStatistical Yearbook. Chinese Agriculture Publishing Company, ISBN 978-7-109-14862-8. http://www.docin.com/, 129pp. (in Chinese).

CEC, 1990. The Radiological Exposure of the Population of the European Communityfrom Radioactivity in North European Marine Waters. Project ‘Marina’. Com-mission of the European Communities, Bruxelles. EUR 12483.

CEC, 1995. Methodology for Assessing the Radiological Consequences of RoutineRelease of Radionuclides to the Environment. RP 72, EUR 15760.

Cho, Y.-K., Seo, G.-H., Choi, B.-J., Kim, S., Kim, Y.-G., Youn, Y.-H., Dever, E.P., 2009.Connectivity among straits of the Northwestern Pacific marginal seas.J. Geophys. Res. 114, C06018. http://dx.doi.org/10.1029/2008JC005218.

Choi, B.H., 1986. Tidal computations for the Yellow Sea. Coast. Eng. 6, 67e81.

Choi, B.H., Mun, J.Y., Ko, J.S., Yuk, J.H., 2005. Simulation of suspended sediment inthe Yellow and East China Seas. China Ocean Eng. 19, 235e250.

Coughtrey, P.J., Thorne, M.C., 1983. Radionuclide Distribution and Transport inTerrestrial and Aquatic Ecosystems: a Critical Review of Data, vol. 2. A. A. Bal-kema, Rotterdam.

De Vries, D.J., De Vries, M.B., 1988. Modelling the Fate of HCB and PCB153 in theLakes Ketelmeer and IJsselmeer. Report T250 Delft Hydraulics.

EC e European Commission, 2002. MARINA II. Update of the MARINA Project on theRadiological Exposure of the European Community from Radioactivity in NorthEuropean Marine Waters. Radiation Protection 132. European Commission,Luxembourg.

Goshawk, J.A., Clarke, S., Smith, C.N., McDonald, P., 2003. MEAD (part I) e amathematical model of the long-term dispersion of radioactivity in shelf seaenvironments. J. Environ. Radioact. 68, 115e135.

Hamilton, T.F., Milliés-Lacroix, J.-C., Hong, G.H., 1996. 137Cs(90Sr) and Pu isotopes inthe Pacific Ocean: sources and trends. In: Guéguéniat, P., Germain, P.,Métivier, H. (Eds.), Radionuclides in the Oceans. Inputs and Inventories. LesEditions de Physique, Les Ulis, pp. 29e58.

Heling, R., Koziy, L., Bulgakov, V., 2002. On the dynamical uptake model developedfor the uptake of radionuclides in marine organisms for the POSEIDON-R modelsystem. Radioprotection 37 (C1), 833e838.

V. Maderich et al. / Journal of Environmental Radioactivity 131 (2014) 4e1818

Hirose, K., Igarashi, Y., Aoyama, M., 2008. Analysis of the 50-year records of theatmospheric deposition of long-lived radionuclides in Japan. Appl. Radiat. Isot.66, 1675e1678.

Hong, G.-H., Lee, S.-H., Kim, S.-H., Chung, C.-S., Baskaran, M., 1999. Sedimentaryfluxes of 90Sr, 137Cs, 239,240Pu and 210Pb in the East Sea (Sea of Japan). Sci. TotalEnviron. 237/238, 225e240.

Hong, G.H., Chung, C.S., Lee, S.-H., Kim, S.H., Baskaran, M., Lee, H.M., Kim, Y.I.,Yang, D.B., Kim, C.K., 2006. Artificial radionuclides in the Yellow Sea: inputs andredistribution. Radioact. Environ. 8, 96e133.

IAEA, 2004. Sediment Distribution Coefficients and Concentration Factors for Biotain the Marine Environment. Technical Report Series No. 422.

ICRP e International Commission on Radiological Protection, 1995. Age-depen-dent Doses to the Members of the Public from Intake of Radionuclides. Part5: Compilation of Ingestion and Inhalation Coefficients. Publication 72. ICRP,Paris.

Iosjpe, M., Brown, J., Strand, P., 2002. Modified approach to modeling radiologicalconsequences from releases into the marine environment. J. Environ. Radioact.60, 91e103.

JCAC e Japan Chemical Analysis Center, 2004. Environmental Radioactivity Data-base. Available at: http://search.kankyo-hoshano.go.jp/servlet/search.top (inJapanese).

JFRA e Japan Fisheries Research Agency, 2012. Results of the Inspection onRadioactivity Materials in Fisheries Products. Available at: http://www.jfa.maff.go.jp/e/inspection/index.html.

Kanda, J., 2013a. Cesium-137 dispersion and biological transfer in coastal waters offFukushima. In: Seminar on Radioecological Studies after the Fukushima Acci-dent, Stockholm, 10 January 2013. Available at: http://www.sfrek.se/presentation2.pdf.

Kanda, J., 2013b. Continuing 137Cs release to the sea from the Fukushima Dai-ichinuclear power plant through 2012. Biogeosci. Discuss. 10, 3577e3595.

Kang, D.-J., Chung, C.S., Kim, S.H., Kim, K.-R., Hong, G.H., 1997. Distribution of 137Csand 239,240Pu in the surface waters of the East Sea (sea of Japan). Mar. Pollut.Bull. 35, 305e312.

Kawamura, H., Kobayashi, T., Furuno, A., In, T., Ishikawa, Y., Nakayama, T., Shima, S.,Awaji, T., 2011. Preliminary numerical experiments on oceanic dispersion of 131Iand 137Cs discharged into the ocean because of the Fukushima Dai-ichi nuclearpower plant disaster. J. Nucl. Sci. Technol. 48, 1349e1356.

Kumblad, L., Kautsky, U., Naeslund, B., 2006. Transport and fate of radionuclides inaquatic environments e the use of ecosystem modelling for exposure assess-ments of nuclear facilities. J. Environ. Radioact. 87, 107e129.

Lepicard, S., Raffestin, D., Rancillac, F., 1998. POSEIDON: a dispersion computer codefor assessing radiological impacts in European seawater environment. Radiat.Prot. Dosim. 75, 79e83.

Lepicard, S., Heling, R., Maderich, V., 2004. POSEIDON-R/RODOS models forradiological assessment of marine environment after accidental releases:application to coastal areas of the Baltic, Black and North Seas. J. Environ.Radioact. 72, 153e161.

MAFF e Ministry of Agriculture, Forestry and Fisheries, Japan, 2011. The 85th Sta-tistical Year Book of Ministry of Agriculture, Forestry and Fisheries (2009e2010), XI Fishery, Catch Volumes and Values per Sea Region/Prefecture forMajor Types of Fish. http://www.maff.go.jp/e/tokei/kikaku/nenji_e/85nenji/index.html.

Margvelashvily, N., Maderich, V., Zheleznyak, M., 1997. Threetox e a computer codeto simulate three dimensional dispersion of radionuclides in stratified waterbodies. Radiat. Prot. Dosim. 73, 177e180.

MARIS eMarine Information System, 2012. Data Available at: http://maris.iaea.org/.Masumoto, Y., Miyazawa, Y., Tsumune, D., Tsubono, T., Kobayashi, T., Kawamura, H.,

Estournel, C., Marsaleix, P., Lanerolle, L., Mehra, A., Garraffo, Z.D., 2012. Oceanicdispersion simulations of 137Cs released from the Fukushima Dai-ichi nuclearpower plant. Elements 8, 207e212.

MEXT e Japanese Ministry of Education, Culture, Sports, Science and Technology,2010. Environmental Radiation Database. Available at: http://search.kankyo-hoshano.go.jp/servlet/search.top (in Japanese).

MFAFF eMinistry of Food, Agriculture, Forestry and Fisheries, Korea, 2012. FisheriesInformation Service. Available at: http://www.fips.go.kr/?gclid¼CLXy3Kusj6ICFUELHAodlygpbg (in Korean).

Miyake, Y., Saruhashi, K., Sugimura, Y., Kanazawa, T., Hirose, K., 1988. Contents of137Cs, plutonium and americium isotopes in the southern ocean waters. Pap.Meteorol. Geophys. 39, 95e113.

Miyao, T., Hirose, K., Aoyama, M., Igarashi, Y., 1998. Temporal variation of 137Cs and239,240Pu in the Sea of Japan. J. Environ. Radioact. 40, 239e250.

MSA e Hydrographic Department, Maritime Safety Agency, Ministry of Transport,Japan, 1995. Annual Reports of Radioactivity Survey for 1993.

MSA e Hydrographic Department, Maritime Safety Agency, Ministry of Transport,Japan, 1998. Annual Reports of Radioactivity Survey for 1996.

MSA, Hydrographic Department, Maritime Safety Agency, Ministry of Transport,Japan, 1999. Annual Reports of Radioactivity Survey for 1997.

Nagaya, Y., Nakamura, K., 1981. Artificial radionuclides in the western NorthwesternPacific (I): 90Sr and 137Cs in the deep waters. J. Oceanograph. Soc. Jpn. 37, 135e144.

Nagaya, Y., Nakamura, K., 1987. Artificial radionuclides in the western NorthwesternPacific (II): 137Cs and 239,240Pu inventories in water and sediment columnsobserved from 1980 to 1986. J. Oceanograph. Soc. Jpn. 43, 345e355.

Nagaya, Y., Nakamura, K., 1992. 239þ240Pu and 137Cs in the East China and the YellowSeas. J. Oceanography 48, 23e35.

Nakano, M., Povinec, P.P., 2003. Oceanic general circulation model for the assess-ment of the distribution of 137Cs in the world ocean. Deep-sea Res. II 50, 2803e2816.

Nakano, M., 2006. Simulation of the advectionediffusion-scavenging processes for137Cs and 239,240Pu in the Japan Sea. Radioact. Environ. 8, 433e448.

Nakano, M., Povinec, P.P., 2012. Long-term simulations of the 137Cs dispersion fromthe Fukushima accident in the world ocean. J. Environ. Radioact. 111, 109e115.

NFRDI e National Fisheries Research and Development Institute, Korea, 2000.Ecology and Fishing Ground of Fishery Resources in EEZ. Yaemunsa, Busan,317pp. (in Korean).

Nielsen, S.P., 1995. A box model for North-East Atlantic coastal waters comparedwith radioactive tracers. J. Mar. Syst. 6, 545e560.

NISA e Nuclear and Industrial Safety Agency, 2011. Regarding the Evaluation of theConditions on Reactor Cores of Unit 1, 2 and 3 Related to the Accident atFukushima Dai-ichi Nuclear Power Station. Tokyo Electric Power Co. Inc..Available at: http://www.nsr.go.jp/archive/nisa/english/press/2011/06/en20110615-5.pdf.

Otosaka, S., Amano, H., Ito, T., Kawamura, H., Kobayashi, T., Suzuki, T., Togawa, O.,Chaykovskaya, E.L., Lishavskaya, T.S., Novichkov, V.P., Karasev, E.V., Tkalin, A.V.,Volkov, Y.N., 2006. Anthropogenic radionuclides in sediment in the Japan Sea:distribution and transport processes of particulate radionuclides. J. Environ.Radioact. 91, 128e145.

Patten, B.C. (Ed.), 1971. Systems Analysis and Simulation in Ecology, vol. 1. AcademicPress., pp. 81e82.

Periáñez, R., 2005. Modelling the Dispersion of Radionuclides in the Marine Envi-ronment. Springer-Verlag, Heidelberg.

Robertson, L., Langner, J., Engardt, M., 1999. An Eulerian limited-area atmospherictransport model. J. Appl. Meteorol. 38, 190e210.

Smith, J.G., Simmonds, J.R. (Eds.), 2009. The Methodology for Assessing theRadiological Consequences of Routine Releases of Radionuclides to the Envi-ronment Used in PC-CREAM 08 HPA-RPD-058 Report. Health ProtectionAgency, Didcot.

Stohl, A., Seibert, P., Wotawa, G., Arnold, D., Burkhart, J.F., Eckhardt, S., Tapia, C.,Vargas, A., Yasunari, T.J., 2012. Xenon-133 and caesium-137 releases into theatmosphere from the Fukushima Dai-ichi nuclear power plant: determinationof the source term, atmospheric dispersion, and deposition. Atmos. Chem. Phys.12, 2313e2343.

Su, C.-C., Huh, C.-A., 2002. 210Pb, 137Cs and 239,240Pu in East China Sea sediments:sources, pathways and budgets of sediments and radionuclides. Mar. Geol. 183,63e178.

Tateda, Y., Tsumune, D., Tsubono, T., 2013. Simulation of radioactive cesium transferin the southern Fukushima coastal biota using a dynamic food chain transfermodel. J. Environ. Radioact. 124, 1e12.

TEPCO (Tokyo Electric Power Company), 2013a. Detection of Radioactive Materialsfrom Seawater near Fukushima Dai-ichi Nuclear Power Station, with Attach-ments. Press Releases http://www.tepco.co.jp/en/press/corp-com/release/.

TEPCO, 2013b. Handouts at Press Conference. http://www.tepco.co.jp/en/nu/fukushima-np/handouts/indexold-e.html.

Tsumune, D., Tsubono, T., Aoyama, T., Hirose, K., 2012. Distribution of oceanic 137Csfrom the Fukushima Dai-ichi nuclear power plant simulated numerically by aregional ocean model. J. Environ. Radioact. 111, 100e108.

UNSCEAR e United Nations Scientific Committee on the Effect of Atomic Radiation,2000. Exposures of the public from man-made sources of radiation. In: Sourcesand Effects of Ionizing Radiation. United Nations, New York.

Yankovich, T., 2003. Towards an improved ability to estimate internal dose to non-human biota. Development of conceptual models for reference non-humanbiota. Protection of the environment from ionising radiation. In: The Develop-ment and Application of a System of Radiation Protection for the Environment.IAEA, Vienna, pp. 365e373.

Yankovich, T., Beresford, N., Wood, M., Aono, T., Anderson, P., Barnett, C.L.,Bennett, P., Brown, J.E., Fesenko, S., Fesenko, J., Hosseini, A., Howard, B.J.,Johansen, P., Phaneuf, M.M., Tagami, K., Takata, H., Twining, J.R., Uchida, S., 2010.Whole-body to tissue-specific concentration ratios for use in biota dose as-sessments for animals. Radiat. Environ. Biophys. 49, 549e565.