Fukushima 137Cs eleases dispe sion modelling o e 1
he Pacific Ocean. Compa isons o models wi h wa e ,2
sedimen and bio a da a3
4
Decembe 14, 20185
Abs ac 6
A numbe o ma ine adionuclide dispe sion models (bo h Eule ian and La-7
g angian) we e applied o simula e 137Cs eleases om Fukushima Daiichi nuclea 8
powe plan acciden in 2011 o e he Pacific a oceanic scale. Simula ions ex ended9
o e wo yea s and bo h di ec eleases in o he ocean and deposi ion o a mosphe ic10
eleases on he ocean su ace we e conside ed. Dispe sion models included an embed-11
ded biological up ake model (BUM). Th ee ypes o BUMs we e used: equilib ium,12
dynamic and allome ic. Model esul s we e compa ed wi h 137Cs measu emen s13
in wa e (su ace, in e media e and deep laye s), sedimen and bio a (zooplank on,14
non-pisci o ous and pisci o ous fish). A easonable ag eemen in model/model and15
model/da a compa isons was ob ained.16
Keywo ds: Fukushima-Daiichi acciden ; dispe sion model; ocean; sedimen ; biological17
up ake model; caesium18
1 In oduc ion19
A e he 9.0 magni ude ea hquake and esul ing sunami occu ed on Ma ch 11 h, 2011,20
in Japan, significan amoun s o adioac i e ma e ial we e eleased o he en i onmen 21
om Fukushima Dai-ichi nuclea powe plan (FDNPP). Radionuclides eleased o he22
a mosphe e we e anspo ed eas wa d by a s ong je s eam and eached he coas o 23
No h Ame ica in ou days (Takemu a e al., 2011). A po ion o hese adionuclides was24
deposi ed on he Pacific Ocean su ace by we and d y deposi ion p ocesses. In addi ion,25
1
wa e used o cool a damaged nuclea eac o leaked in o he ocean (Kobayashi e al.,26
2013).27
Some exe cises compa ing nume ical model pe o mances when applied o simula e he28
137Cs eleases om FDNPP in he Pacific Ocean ha e been ca ied ou , as o ins ance in29
Masumo o e al. (2012). These au ho s ound disc epancies be ween he fi e pa icipa ing30
models and concluded ha hey we e due o he diffe en calcula ed cu en fields in he31
coas al wa e s o Japan, off Fukushima, which lead o diffe en adionuclide dis ibu ions.32
Diffe ences in ci cula ion fields we e caused by he diffe en ocean models and dispe sion33
model se ings used by he esea ch g oups. Howe e , a sys ema ic assessmen aimed a 34
in es iga ing he easons o diffe ences was no ca ied ou .35
The Science Council o Japan (SCJ, 2014) ca ied ou a simila in e compa ison s udy36
o 137Cs, wi h ele en models in ol ed. Again, significan diffe ences be ween models37
we e ound. Models we e diffe en in concep (Eule ian s. Lag angian), wi h diffe en 38
se ing and e en diffe en sou ce e ms. I was concluded ha a simple compa ison was39
no s aigh o wa d and consequen ly de ailed sys ema ic compa ison s udies, such as40
ones ha use he same adionuclide o cing wi h diffe en models and/o he same model41
wi h diffe en o cing scena ios, we e equi ed. This kind o in e compa ison exe cise was42
ca ied ou in he ame o IAEA (In e na ional A omic Ene gy Agency) MODARIA1p o-43
g am (Pe i´a˜nez e al., 2015a; 2016a). The MODARIA p ojec was unning om 2012 o44
2015 o make p og ess in he assessmen o adioac i e subs ances in he en i onmen and45
i s impac o man and bio a. Diffe en dispe sion models we e applied o simula e FDNPP46
eleases in he Pacific, using diffe en and also he same wa e ci cula ion fields. Simu-47
la ions wi h he same se o pa ame e s (like diffusion coefficien s o ins ance) we e also48
ca ied ou . I was ound ha he main sou ce o disc epancy be ween diffe en dispe sion49
models was due o he diffe en ci cula ion fields. Model/model and model/measu emen s50
1Modelling and Da a o Radiological Impac Assessmen s. Fu he in o ma ion can be ound he e:
h p://www-ns.iaea.o g/p ojec s/moda ia/de aul .asp?l=116
2
compa isons o bo h he dissol ed phase and bed sedimen s (no included in ea lie model51
compa ison exe cises) we e ca ied ou in such s udy.52
Al e na i ely, he same dispe sion model o ced wi h diffe en ci cula ion fields was53
es ed as well, al hough wa e /sedimen in e ac ions we e no included in his s udy54
(Kawamu a e al., 2017).55
An in e es ing exe cise was desc ibed in Made ich e al. (2018). In his case he56
same dispe sion model, unning wi h gene ic pa ame e s, was applied o desc ibe 137Cs57
dispe sion om Che nobyl NPP acciden in he Bal ic and Black seas, and FDNPP acci-58
den in he Pacific Ocean. The applied box model (POSEIDON-R; Lepica d e al., 2004;59
Made ich e al., 2014a; 2014b; Bezhena e al., 2016) con ained an embedded ood web60
model. Compa isons o model esul s wi h measu emen s in he h ee scena ios indica ed61
ha , wi h some es ic ions, he model could be used wi h gene ic pa ame e alues in62
adia ion eme gency si ua ions in a eas whe e limi ed in o ma ion is a ailable.63
MODARIA-II p og am2was launched by he IAEA in 2016 as a ollow-up o MODARIA.64
The wo k in compa ing nume ical model pe o mances when applied o simula e FDNPP65
eleases in he ocean was con inued in he ame o his p ojec . Ne e heless, spa ial ange66
and empo al ame o simula ions we e ex ended: wo yea long simula ions o e almos 67
he whole No h Pacific Ocean we e ca ied ou . In addi ion, ma ine dispe sion models68
con ain an in eg a ed biological up ake model (BUM) wi h ou componen s (phy oplank-69
on, zooplank on, non-pisci o ous and pisci o ous fish). Model/model and model/da a70
compa isons we e ca ied ou o wa e , bed sedimen s and biological componen s o he71
models, which has no been done be o e.72
Six ins i u es ha e pa icipa ed in he model compa isons. These a e he Ins i u e73
o Ma hema ical Machines and Sys em P oblem (IMMSP, Uk aine), Ko ea Ins i u e o 74
Ocean Science and Technology (KIOST, Rep. o Ko ea), ABme i (Slo akia), Uni e -75
si y o Se ille (USEV, Spain), Japan A omic Ene gy Agency (JAEA, Japan), and Ko ea76
2h p://www-ns.iaea.o g/p ojec s/moda ia/moda ia2.asp?s=8&l=129
3
A omic Ene gy Resea ch Ins i u e (KAERI, Rep. o Ko ea).77
The me hodology is p esen ed in sec ion 2, whe e wa e ci cula ion used by models,78
sou ce e ms, and he o igin o expe imen al da a on 137Cs concen a ions a e desc ibed.79
Resul s a e p esen ed in sec ion 3. Some gene al discussion on model unce ain y and80
complexi y is finally included in sec ion 4.81
2 Me hods82
2.1 Hyd odynamics83
Wa e ci cula ion p o ided by FORA3model was used o calcula ions. This model,84
Fou -dimensional Va ia ional Ocean ReAnalysis o he Wes e n No h Pacific (FORA-85
WNP30), is he fi s -e e da ase co e ing he wes e n No h Pacific o e he las h ee86
decades (1982-2014) a eddy- esol ing esolu ion. I is a coope a i e wo k o he Japan87
Agency o Ma ine-Ea h Science and Technology (JAMSTEC) and he Me eo ological88
Resea ch Ins i ude, Japan Me eo ological Agency (JMA/MRI) using he Ea h Simula o 89
(Usui e al., 2017; Tsujino e al., 2010).90
The domain used in he p esen calcula ions ex ends 117◦E-160◦Wand15
◦N-65◦Nin91
longi ude and la i ude, espec i ely. Ho izon al esolu ion is 0.1oand he e a e 54 e ical92
le els (0-6300 m) wi h inc easing hickness om he su ace o he sea bo om. Mon hly93
clima ological da a om 2011 o 2014 we e used. Two yea long (Ma ch 11, 2011 o Ma ch94
11, 2013) simula ions we e made.95
The model domain showing wa e dep hs and an example o su ace wa e ci cula ion96
(a e aged alue o Ma ch 2011) can be seen in Fig. 1. The gene al la ge scale ci cula ion97
in he wes e n Pacific Ocean is domina ed by he in e ac ion be ween he Ku oshio and98
Oyashio cu en s. The Ku oshio Cu en is he wes e n bounda y cu en in he no h99
Pacific, which flows along he coas o Japan owa ds he no h and cu es o he cen al100
3h p://syn hesis.jams ec.go.jp/FORA/e/index.h ml
4
Pacific Ocean, hen o ming he so-called Ku oshio Ex ension. The Oyashio Cu en is101
a cold cu en which flows om he no h. These wo cu en sys ems con e ge in he102
coas al wa e s off Fukushima coas . Such con e gence leads o he gene a ion o uns eady103
eddies in he a ea. These ea u es may be seen in Fig. 1.104
2.2 Radionuclide sou ces105
Radionuclides we e di ec ly in oduced in o he Pacific Ocean om FDNPP. They we e106
also eleased o he a mosphe e; adionuclides which we e la e deposi ed on he sea107
su ace. Bo h sou ces we e conside ed in calcula ions.108
Di ec eleases o 137Cs a e gi en o he pe iod Ma ch 25 h, 2011, o Decembe 31 h,109
2011, and p esen ed in Fig. 2. They we e econs uc ed by JAEA as explained in de-110
ail in Kobayashi e al. (2013). Moni o ing da a om he web si e o Tokyo Elec ic111
Powe Company (TEPCO), ega ding he a ea nea he no he n and sou he n discha ge112
channels o he Fukushima Daiichi NPP (TEPCO, 2011), we e used o his pu pose.113
A mosphe ic deposi ion in he No h Pacific Ocean was ob ained om he a e aged114
alues om WSPEEDI-II (JAEA: Te ada e al., 2012) and LADAS (KAERI: Suh e al.,115
2006; Suh e al., 2009) a mosphe ic dispe sion models o he pe iod Ma ch 12 h, 2011,116
o June 1s , 2011. E en hough simula ions a e 2 yea long, mos deposi ion occu ed117
wi hin he fi s mon hs a e he acciden . Daily in eg a ed alues we e p o ided. As an118
example, he in eg a ed deposi ion o Ma ch 15 h, 2011, a e aged om bo h models, is119
p esen ed in Fig. 2.120
In addi ion, a p e-FDNPP acciden 137Cs uni o m backg ound o 1.5 Bq/m3was con-121
side ed o e he Pacific Ocean wa e s, in o de o ca y ou compa isons o model esul s122
wi h field measu emen s.123
5

2.3 Dispe sion models124
Some o he main cha ac e is ics o he dispe sion models which we e applied a e summa-125
ized in Table 1. Bo h Eule ian and Lag angian models we e used wi h diffe en pa am-126
e e iza ions o ho izon al and e ical diffusi i ies. The gene al cha ac e is ics and basic127
equa ions desc ibing he wo ypes o dispe sion models which we e applied a e p esen ed128
in appendix A.1 and A.2.129
A kine ic (dynamic) app oach was applied o desc ibe wa e /sedimen in e ac ions in130
bo h Eule ian and Lag angian models, which is based on a deso p ion coefficien and he131
dis ibu ion coefficien , kd, o he co esponding adionuclide (Pe i´a˜nez, 2005).132
All models used an equilib ium dis ibu ion coefficien o 2.0 m3/kg. This is he133
mean alue ecognized by IAEA (2004) o open ocean wa e s and is also in ag eemen 134
wi h measu emen s off Fukushima (Honda e al., 2012). The kine ic a e desc ibing135
elease om sedimen s, k2=1,16 ×10−5s−1, was de e mined o Cs om expe imen s136
(Nyffele e al., 1984). The kine ic a e desc ibing up ake (k1) is de i ed om k2and he137
dis ibu ion coefficien , as usually done (Pe i´a˜nez, 2005). A s ochas ic me hod is used o138
sol e up ake/ elease p ocesses in Lag angian models (Pe i´a˜nez and Ellio , 2002).139
Mos models include a biological up ake model (BUM). Fou species we e conside ed:140
phy oplank on, zooplank on, non-pisci o ous and pisci o ous fish. Th ee ypes o BUM141
we e used in he models: an equilib ium model based upon a concen a ion ac o CR142
(appendix B.1), a dynamic model (B.2) and an allome ic me hod (B.3). The BUM143
inco po a ed wi hin each physical dispe sion model is indica ed in Table 1 as a e e ence144
o he appendix whe e he co esponding BUM cha ac e is ics a e commen ed.145
2.4 Expe imen al da a146
Model esul s we e compa ed wi h a ailable 137Cs measu emen s in wa e a h ee diffe en 147
laye s, bed sedimen s and biological compa men s (zooplank on, non-pisci o ous and148
6
I/K THREETOX I/K Lag angian ESTE USEV SEA-GEARN LORAS
Model (IMMSP/KIOST) (IMMSP/KIOST) (ABme i ) (Uni . Se ille) (JAEA) (KAERI)
Model ype Eule ian Lag angian Lag angian Lag angian Lag angian Lag angian
Made ich B o chenko Pe i´a˜nez e Kobayashi Min e al.
Re e ence e al. (2016) e al. (2018) www.abme i .sk al. (2016b) e al. (2007) (2013)
Ho izon al Smago inskyaSmago insky Smago insky
diffusion o mula 10 m2/s o mula o mula 10 m2/s 10 m2/s
10−3m2/s o d<60 m
Ve ical 10−5m2/s o d>120 m
diffusion 10−4m2/s linea unc ion o 10−4m2/s 10−4m2/s 10−4m2/s 10−3m2/s
60 <d<120 m
Bed po osi y 0.6 0.6 0.6 0.6 0.6 0.7
Sedimen
hickness 0.05 m 0.05 m 0.05 m 0.05 m 0.05 m 0.1 m
Pa icle
densi y 2600 kg/m32600 kg/m32600 kg/m32600 kg/m32600 kg/m3
137Cs kd2m
3/kg 2 m3/kg 2 m3/kg 2 m3/kg 2 m3/kg 2 m3/kg
k2(s−1)3.17 ×10−8Made ich e al. (2017) 1.16 ×10−61.16 ×10−61.16 ×10−61.16 ×10−6
BUM B.2 B.2 B.1 B.2 no B.3
Table 1: Model main cha ac e is ics. dis wa e dep h and k2is he 137Cs deso p ion coefficien . aSee o ins ance Cushman-
Roisin and Becke s (2011). A selec ed e e ence is gi en o each model. The BUM ow indica es he appendix whe e some
de ails o he up ake model a e gi en: B.1 is an equilib ium model, B.2 is a dynamic model and B.3 is he allome ic me hod.
7
pisci o ous fish) in he su ace laye ( o 20 m dep h). The o he wo conside ed wa e 149
laye s a e 20-460 m and 460 m o he seabed.150
Measu emen s we e compiled om he ollowing e e ences: Honda e al. (2012),151
Cha e e e al. (2013), Kae iyama e al. (2013) o wa e ; he “Da abase o Radioac-152
i e Subs ance Moni o ing Da a”4 o sedimen s; Honda e al. (2012), Ki amu a e al.153
(2013) o zooplank on; Wada e al. (2016); Men e al. (2017); Johansen e al. (2014)154
o fish (pisci o ous and non-pisci o ous). Only da a o pelagic fish we e used. Sam-155
pled pelagic non-pisci o ous fish a e Eng aulis japonicus,E umeus e es,Clupea pallasii156
and Hypo hamphus sajo i. Sampled pelagic pisci o ous fish a e Hexag ammos sebas es,157
Toda odes pacificus,Snake macke el,Onco hynchus ke a,Ammody es japonicus,Se iola158
quinque adia a,Se iola quinque adia a,T achu us japonicus and Scombe japonicus.Wa-159
e samples collec ed in he di ec elease a ea ha e been fil e ed ou since he models a e160
gi ing a e age alue o adionuclide concen a ions o e boxes, as explained below.161
Loca ions whe e samples we e collec ed du ing he simula ion pe iod a e indica ed as162
do s in Fig. 3. The Pacific Ocean was di ided in o a numbe o boxes, p esen ed in Fig. 4,163
acco ding o gene al ci cula ion and he loca ion o he elease poin . Model esul s we e164
a e aged o each box and hen hese a e aged alues we e compa ed wi h measu emen s.165
Boxes in he elease a ea may be oo la ge o a de ailed s udy o adionuclide be-166
ha iou in such egion close o FDNPP. Howe e , i should be aken in o accoun ha 167
he dispe sion o FDNPP 137Cs eleases was s udied a a smalle spa io- empo al scale in168
a p e ious pape o he g oup (Pe i´a˜nez e al., 2015a); and model p edic ions and mea-169
su emen s we e compa ed in he a ea close o FDNPP (less han some 100 km away).170
The p esen wo k is complemen ing such p e ious pape , going o la ge spa ial and em-171
po al scales. Thus, la ge boxes a e used. In addi ion, i should be conside ed ha a172
model/da a compa ison o specific poin s in such a la ge domain is no easible wi h173
Lag angian models which elease indi idual pa icles, and i is be e o use a e ages o e 174
4h p://emdb.jaea.go.jp/emdb/en/
8
gi en a eas, which a e defined in iew o he physical oceanog aphy o he egion (Pe i´a˜nez175
e al., 2015a; 2015b; 2016a). Howe e , i should be no ed ha measu emen s we e no 176
dis ibu ed homogeneously in he ela i ely la ge conside ed boxes.177
3 Resul s178
As explained be o e, wo yea long simula ions we e ca ied ou ; om Ma ch 2011 o179
Ma ch 2014. Mon hly mean alues o 137Cs concen a ions in each box in Fig 4 we e180
p o ided by he models o he h ee wa e laye s, seabed sedimen s and he ou biological181
compa men s (su ace laye only).182
Model esul s and 137Cs measu emen s a e p esen ed in Fig. 5 o Fig. 12. Resul s a e183
p esen ed only o such boxes whe e measu emen s a e a ailable. Resul s o he abio ic184
and bio ic componen s o he models a e discussed sepa a ely in he ollowing subsec ions.185
3.1 Wa e and sedimen s186
Resul s o su ace wa e may be seen in Fig. 5 and Fig. 6, o boxes which a e a 187
om Japan and boxes loca ed close , a ound FDNPP, espec i ely. In boxes 1, 3, 5 and188
20 (Fig. 5) he e is a sligh inc ease in 137Cs concen a ions wi h espec o backg ound189
immedia ely a e he acciden , which mus be a ibu ed o a mosphe ic deposi ion. In190
gene al, models p oduce his ini ial inc ease, which is abou one o de o magni ude abo e191
backg ound. In o he boxes (like 15 and 16), bo h models and measu emen s indica e p e-192
FDNPP acciden backg ound. Thus, eleases did no affec hese a eas in he conside ed193
empo al ame.194
In con as , high concen a ions a e ound close o FDNPP (Fig. 6). Fo some o he195
boxes (6, 7, 12) models and measu emen s show a end owa ds achie ing backg ound196
concen a ions a e app oxima ely one yea . The ini ial concen a ion inc ease abo e197
backg ound is abou wo o de s o magni ude. O he egions sou h om Japan (boxes 13198
9
concen a ions o e such boxes; as i was done o he Bal ic Sea model in e compa ison355
in Pe i´a˜nez e al. (2015b). Howe e , i should be aken in accoun ha in he icini y o 356
FDNPP measu emen s we e no dis ibu ed homogeneously in space. This can esul in357
o e es ima ion o expe imen al box-a e aged alues.358
Models ag ee in p edic ing a eas in he Pacific Ocean which we e affec ed by FDNPP359
eleases (di ec and/o a mosphe ic deposi ion) and egions which we e no . In addi ion,360
p edic ed concen a ions a e wi hin he same o de o magni ude in mos cases.361
Wi h espec o calcula ed 137Cs empo al ends in bio a, dynamic models end o362
unde es ima e concen a ions. Allome y and he equilib ium app oach esul s a e, in363
gene al, in be e ag eemen wi h obse a ions. This is explained by he highe 137Cs364
concen a ions in wa e p oduced by ESTE and LORAS models. Tempo al e olu ions o 365
137Cs concen a ions calcula ed h ough he diffe en app oaches a e diffe en , al hough366
he e is no enough expe imen al da a o assess which app oach leads o be e esul s.367
Howe e , i is clea ha dynamic models p o ide he known pa e n o delayed ise o 368
ac i i y concen a ion in bio a.369
Acknowledgemen 370
Wo k ca ied ou in he ame o IAEA MODARIA-II (Modelling and Da a o Radio-371
logical Impac Assessmen s) p og am. This wo k was pa ially suppo ed by he Na ional372
Resea ch Founda ion o Ko ea (NRF) and pa ially unded by he Ko ean Go e nmen 373
(MSIP) (MSIP: NRF-2017M2A8A4015253, NRF-2015M2A2B2034282). The au ho s a e374
indeb ed o all membe s o MODARIA-II wo king g oup 7 o use ul discussions held375
du ing g oup mee ings.376
16

6 Re e ences377
Bezhena , R., Jung, K.T., Made ich, V., Willemsen, S., de Wi h, G., Qiao, F.,378
2016. T ans e o adiocesium om con amina ed bo om sedimen s o ma ine o -379
ganisms h ough ben hic ood chain in pos -Fukushima and pos -Che nobyl pe iods.380
Biogeosciences 13, 3021-3034.381
B o chenko, I., Jung, K.T., Made ich, V., Kim, K.O., Ko ale s, S., 2018. Me hod382
o momen s o andom walk algo i hms. In p epa a ion.383
B own, J., Bo en zen, P., Dowdall, M., Sazykina, T., K yshe , I., 2004. The384
De i a ion o T ans e Pa ame e s in he Assessmen o Radiological Impac s on385
A c ic ma ine Bio a. ARCTIC 57(3), 279-289.386
Ca alho, F. P., 2018. Radionuclide concen a ion p ocesses in ma ine o ganisms:387
A comp ehensi e e iew. Jou nal o En i onmen al Radioac i i y 186, 124-130.388
Cha e e, A.M. B eie , C.F., Hende son, P.B., Pike, S.M., Rypina,I.I., Jayne, S.R.,389
Buessele , K.O., 2013. Radium-based es ima es o cesium iso ope anspo and390
o al di ec ocean discha ges om he Fukushima Nuclea Powe Plan acciden ,391
Biogeosciences 10, 2159-2167.392
Cushman-Roisin, B., Becke s, J.M., 2011. In oduc ion o Geophysical Fluid Dy-393
namics. Else ie .394
Heling, R., Koziy, L., Bulgako , V., 2002. On he dynamical up ake model de el-395
oped o he up ake o adionuclides in ma ine o ganisms o he POSEIDON-R396
model sys em. Radiop o ec ion 37(C1), 833-838.397
Higley, K.A., Domo o , S.L., An onio, E.J., Koche , D.C., 2003. De i a ion o 398
a sc eening me hodology o e alua ing adia ion dose o aqua ic and e es ial399
bio a. Jou nal o En i onmen al Radioac i i y, 66, 41-59.400
17
Honda M., Aono T., Aoyama M., Hamajima Y., Kawakami H., Ki amu a M., Ma-401
sumo o Y., Miyazawa Y., Takigawa M., Saino T., 2012. Dispe sion o a ificial402
caesium-134 and -137 in he wes e n No h Pacific one mon h a e he Fukushima403
acciden . Geochemical Jou nal 46, 1-9.404
IAEA, 2004. Sedimen dis ibu ion coefficien s and concen a ion ac o s o bio a405
in he ma ine en i onmen . Technical Repo s Se ies 422, Vienna.406
Johansen, M.P., Ruedig, E., Tagami, K., Uchida, S., Higley, K., Be es o d, N., 2014.407
Radiological dose a es o ma ine fish om he Fukushima Daiichi acciden : he fi s 408
h ee yea s ac oss he No h Pacific, En i onmen al Science and Technology 49(3),409
1277-1285.410
Kae iyama, H., Ambe, D., Shimizu, Y., Fujimo o, K., Ono, T., Yonezaki, S., Ka o,411
Y., Ma sunaga, H., Minami, H., Naka suka, S., Wa anabe, T., 2013. Di ec ob-412
se a ion o 134Cs and 137Cs in su ace seawa e in he wes e n and cen al No h413
Pacific a e he Fukushima Dai-ichi nuclea powe plan acciden , Biogeosciences414
10, 4287-4295.415
Kawamu a, H., Fu uno, A., Kobayashi, T., In, T., Nakayama, T., Ishikawa, Y.,416
Miyazawa, Y., Usui, N., 2017. Oceanic dispe sion o Fukushima-de i ed Cs-137417
simula ed by mul iple oceanic gene al ci cula ion models. Jou nal o En i onmen al418
Radioac i i y 180, 36-58.419
Ki amu a, M., Kumamo o, Y., Kawakami, H., C uz, E.C., Fujiku a, K., 2013. Ho -420
izon al dis ibu ion o Fukushima-de i ed adiocesium in zooplank on in he no h-421
wes e n Pacific Ocean. Biogeosciences 10, 5729-5738.422
Kobayashi, T., O osaka, S., Togawa, O., Hayashi, K., 2007. De elopmen o a423
nonconse a i e adionuclides dispe sion model in he ocean and i s applica ion424
18
o su ace cesium-137 dispe sion in he I ish Sea. Jou nal o Nuclea Science and425
Technology 44(2), 238-247.426
Kobayashi, T., Nagai, H., Chino, M., Kawamu a, H., 2013. Sou ce e m es ima ion427
o a mosphe ic elease due o he Fukushima Dai-ichi Nuclea Powe Plan acciden 428
by a mosphe ic and oceanic dispe sion simula ions. Jou nal o Nuclea Science and429
Technology 50, 255-264.430
Kusakabe, M., Oikawa, S., Taka a, H., Misonoo, J., 2013. Spa io empo al dis-431
ibu ions o Fukushima-de i ed adionuclides in nea by ma ine su ace sedimen s.432
Biogeosciences 10, 5019-5030.433
Lepica d S., Heling R., Made ich V. 2004. POSEIDON/RODOS model o adio-434
logical assessmen o ma ine en i onmen a e acciden al eleases: applica ion o435
coas al a eas o he Bal ic, Black and No h seas. Jou nal o En i onmen al Ra-436
dioac i i y 72, 1-2, 153-161437
Lynch D.R., G eenbe g D.A., Bilgili A., Mcgillicuddy, D.J., Manning, J.P., A e x-438
abale a A.L., 2015. Pa icles in he Coas al Ocean. Theo y and Applica ions.439
Camb idge Uni e si y P ess, NY.440
Made ich V., Heling R., Bezhena R., B o chenko I., Jenne H., Koshebu skyy V.,441
Kuschan A., Te le ska K.. 2008. De elopmen and applica ion o 3D nume ical442
model THREETOX o he p edic ion o cooling wa e anspo and mixing in he443
inland and coas al wa e s. Hyd ological P ocesses 22, 1000-1013.444
Made ich, V., Bezhena , R., Heling, R., de Wi h, G., Jung, K.T., Myoung, J.G.,445
Cho, Y.K., Qiao, F., Robe son, L., 2014a. Regional long- e m model o adioac-446
i i y dispe sion and a e in he no hwes e n Pacific and adjacen seas: applica ion447
o he Fukushima Dai-ichi acciden . Jou nal o En i onmen al Radioac i i y 131,448
4-18.449
19
Made ich V., Jung, K.T., Bezhena R., de Wi h, G., Qiao F., Casacube a, N.,450
Masqu´e, P., Kim, Y.H., 2014b. Dispe sion and a e o 90S in he No hwes e n Pa-451
cific and adjacen seas: Global allou and he Fukushima Dai-ichi acciden . Science452
o he o al En i onmen 494-495, 261-271.453
Made ich V., B o chenko I., D o zhak A., Ie din I., Koshebu sky, V., Pe ia˜nez, R.,454
2016. In eg a ion o 3D model THREETOX in JRODOS, implemen a ion s udies455
and modelling o Fukushima scena ios. Radiop o ec ion 51(HS2), S133-S135.456
Made ich, V., Jung, K.T., B o chenko, I., Kim, K.O., 2017. Mig a ion o adioac-457
i i y in mul i- ac ion sedimen s. En i onmen al Fluid Mechanics 17, 1207-1231.458
Made ich, V., Bezhena , R., Ta eda, Y., Aoyama, M., Tsumune, D., 2018. Sim-459
ila i ies and diffe ences o 137Cs dis ibu ions in he ma ine en i onmen s o he460
Bal ic and Black seas and off he Fukushima Dai-ichi nuclea powe plan in model461
assessmen s. Ma ine Pollu ion Bulle in 135, 895-906.462
Masumo o, Y., Miyazawa, Y., Tsumune, D., Kobayashi, T., Es ou nel, C., Ma saleix,463
P., Lane olle, L., Meh a, A., and Ga affo, Z.D., 2012. Oceanic dispe sion simula ion464
o Cesium-137 om Fukushima Dai-ichi nuclea powe plan , Elemen s, 8, 207-212.465
Men,W.,Deng,F.,He,J.,Yu,W.,Wang,F.,Li,Y.,Lin,F.,Lin,J.,Lin,L.,Zhang,466
Y., Yu, X., 2017. Radioac i e impac s on nek on species in he No hwes Pacific467
and humans mo e han one yea a e he Fukushima nuclea acciden , Eco oxicology468
and En i onmen al Sa e y 144, 601-610.469
Min, B.I., Pe i´a˜nez, R., In-Gyu Kim, Kyung-Suk Suh, 2013. Ma ine dispe sion470
assessmen o 137Cs eleased om he Fukushima nuclea acciden . Ma ine Pollu ion471
Bulle in 72, 22-33.472
Mon e, L., Hakason, L., Pe i´a˜nez,R., Lap e , G., Zheleznyak, M., Made ich, V.,473
20
Koshebu sky, V., 2006. Expe iences om a case s udy o mul i-model applica ion o474
assess he beha iou o pollu an s in he Dniepe -Bug es ua y. Ecological Modelling475
195, 247-263.476
Nyffele , U.P., Li, Y.H., San schi, P.H., 1984. A kine ic app oach o desc ibe477
ace elemen dis ibu ion be ween pa icles and solu ion in na u al aqua ic sys-478
ems. Geochimica Cosmochimica Ac a 48, 1513-1522.479
Pe i´a˜nez, R., Ellio , A. J., 2002. A pa icle- acking me hod o simula ing he480
dispe sion o non-conse a i e adionuclides in coas al wa e s. Jou nal o En i on-481
men al Radioac i i y 58(1), 13-33.482
Pe i´a˜nez, R., 2005. Modelling he dispe sion o adionuclides in he ma ine en i on-483
men : an in oduc ion. Sp inge , Be lin.484
Pe i´a˜nez, R., Bezhena , R., Byung-Il, Min, Duffa, C., Jung, K., Kobayashi, T.,485
Kyung-Suk, Suh, Lamego, F., Made ich,V., Nies,H., Os a h, I., Psal aki, M., 2015a.486
A new compa ison o ma ine dispe sion model pe o mances o Fukushima eleases487
in he ame o IAEA MODARIA p og am. Jou nal o En i onmen al Radioac i i y488
150, 247-269.489
Pe i´a˜nez, R. Bezhena , R., Iosjpe, M., Made ich, V., Nies, H., Os a h, I., Ou ola, I.,490
de Wi h, G., 2015b. A compa ison o ma ine adionuclide dispe sion models o he491
Bal ic Sea in he ame o IAEA MODARIA p og am. Jou nal o En i onmen al492
Radioac i i y 139, 66-77.493
Pe i´a˜nez, R., R. Bezhena , I. B o chenko, C. Duffa, K.T. Jung, T. Kobayashi, F.494
Lamego, V. Made ich, B.I. Min, H. Nies, I. Os a h, I. Ou ola, M. Psal aki, K.S.495
Suh, G. de Wi h, 2016a. Modelling o ma ine adionuclide dispe sion in IAEA496
MODARIA p og am: lessons lea n om he Bal ic Sea and Fukushima scena ios.497
Science o he To al En i onmen 569/570, 594-602.498
21

Pe i´a˜nez, R., Kyung-Suk Suh, Byung-Il Min, 2016b. The beha iou o 137Cs in he499
No h A lan ic Ocean assessed om nume ical modelling: eleases om nuclea uel500
ep ocessing ac o ies, edissolu ion om con amina ed sedimen s and leakage om501
dumped nuclea was es. Ma ine Pollu ion Bulle in 113, 343-361.502
P oehl, J.A., Lynch, D.R., McGuillicuddy, D.J., Ledwell, J.R., 2005. Modeling503
u bulen dispe sion on he No h Flank o Geo ges Bank using Lag angian pa icle504
me hods. Con inen al Shel Resea ch 25, 875-900.505
P o e , P.E., 2004. S ochas ic In eg a ion and Diffe en ial Equa ions (2nd ed.).506
Sp inge , ISBN 3-540-00313-4.507
Schmid -Nielsen, K., 1977. P oblems o scaling: locomo ion and physiological co -508
ela es. In: Scale Effec s in Animal Locomo ion, T.J. Pedley (edi o ). Academic509
P ess, New Yo k. pp. 1-21.510
SCJ, 2014. A e iew o he model compa ison o anspo a ion and deposi ion o 511
adioac i e ma e ials eleased o he en i onmen as a esul o he Tokyo Elec ic512
Powe Compny’s Fukushima Daiichi Nuclea Powe Plan acciden . Repo o he513
Sec ional Commi ee on Nuclea Acciden Commi ee on Comp ehensi e Syn he ic514
Enginee ing, Science Council o Japan.515
Suh, K.S., Jeong, H.J., Kim, E.H., Hwang, W.T., Han, M.H., 2006. Ve ifica ion516
o he Lag angian pa icle model using he ETEX expe imen . Annals o Nuclea 517
Ene gy 33, 1159-1163.518
Suh, K.S., Han, M.H., Jung, S.H., Lee, C.W., 2009. Nume ical simula ion o a long-519
ange dispe sion o a pollu an using Che nobyl da a. Ma hema ical and Compu e 520
Modeling 49, 337-343.521
Takemu a, T., Nakamu a, H., Takigawa, M., Kondo, H., Sa omu a. T., Miyasaka,522
22
T., Nakajima, T., 2011. A nume ical simula ion o global anspo o a mosphe ic523
pa icles emi ed om he Fukushima Daiichi Nuclea Powe Plan . SOLA 7, 101-524
104.525
TEPCO, 2011. h p://www. epco.co.jp/en/nu/ ukushima-np/ 1/index2-e.h ml.526
Te ada, H., Ka a a, G., Chino, M., Nagai, H., 2012. A mosphe ic discha ge and527
dispe sion o adionuclides du ing he Fukushima Dai-ichi Nuclea Powe Plan ac-528
ciden . Pa II: e ifica ion o he sou ce e m and analysis o egional-scale a mo-529
sphe ic dispe sion. Jou nal o En i onmen al Radioac i i y 112, 141-154.530
Tsujino, H., Mo oi, T., Ishikawa, I., Hi aba a, M., Nakano, H., Yamanaka, G.,531
Yasuda, T., Ishizaki, H., 2010. Re e ence Manual o he Me eo ological Resea ch532
Ins i u e Communi y Ocean Model (MRI.COM) Ve sion 3. Tsukuba (Japan): Me-533
eo ological Resea ch Ins i u e. Technical Repo s o he MRI, 59.534
Usui, N., Wakama su, T., Tanaka, Y., Hi ose, N., Toyoda, T., Nishikawa, S., Fujii,535
Y., Taka suki, Y., Iga ashi, H., Ishikawa, Y., Ku agano, T., Kamachi, M., 2017.536
Fou -dimensional Va ia ional Ocean Reanalysis: A 30-yea high- esolu ion da ase 537
in he wes e n No h Pacific (FORA-WNP30). Jou nal o Oceanog aphy 73(2),538
205-233.539
Vi es i Ba lle, J., Wilson, R.C., Wa s, S.J., Jones, S.R., McDonald, P., Vi es-540
Lynch, S., 2008. Dynamic model o he assessmen o adiological exposu e o541
ma ine bio a, Jou nal o En i onmen al Radioac i i y 99, 1711-1730.542
Vi es i Ba lle, J., Be es o d, N., Beaugelin-Seille , K., Bezhena , R., B own, J., Jing-543
Jy Cheng, Cujic, M., D ago ic, S.S., Duffa, C., Fie e , B., Hosseini, A., Jung, K.T.,544
Kamboj, S., Dong-Kwon Keum, Kobayashi, T., K yshe , A., Le Poi e, D., Made ich,545
V., Byung Il Min, Pe i´a˜nez, R., Sazykina, T., Suh, K.S., Yu, C., Wang, C., Heling,546
23
R., 2016. In e -compa ison o dynamic models o adionuclide ans e o ma ine547
bio a in a Fukushima acciden scena io. Jou nal o En i onmen al Radioac i i y548
153, 31-50.549
Wada, T., Fuji a, T., Nemo o, Y., Shimamu a, S., Mizuno, T., Soh ome, T., Kami-550
yama, K., Na i a, K., Wa anabe, M., Ha a, N., Oga a, Y., Mo i a, T., Iga ashi, S.,551
2016. Effec s o he nuclea disas e on ma ine p oduc s in Fukushima: An upda e552
a e fi e yea s. Jou nal o En i onmen al Radioac i i y 164, 312-324.553
Wes G.B,, B own, J.H., Enquis , B.J., 1997. A gene al model o he o igin o 554
allome ic scaling laws in biology. Science 276, 122-126.555
A Physical dispe sion models556
Physical dispe sion models (bo h Eule ian and Lag angian) a e based on he same gene al557
p inciples and equa ions; and hen pa icula ized as p esen ed in Table 1. Thus, hose558
common gene al desc ip ions a e gi en below.559
A.1 Eule ian models560
In Eule ian models he diffe en ial equa ions gi ing empo al and spa ial e olu ion o he561
adionuclide concen a ions in diffe en s a es (e.g. dissol ed in wa e column and po e562
wa e in sedimen s, fixed on he suspended and bo om sedimen e c) a e sol ed. The563
gene al compac o m o hese equa ions o concen a ion o adioac i i y Cαin s a e564
αpe uni o olume (Bq m−3)o pe uni o mass(Bqkg
−1) a e w i en in Ca esian565
coo dina es as:566
∂Cα
∂ +∂(uαCα)
∂x +∂( αCα)
∂y +∂(wαCα)
∂z =∂
∂x Kh
∂Cα
∂x +∂
∂y Kh
∂Cα
∂y +
24
+∂
∂z K
∂Cα
∂z +
n

β=1
kβαCβ+Sα−λCα(1)
whe e (x, y, z) a e Ca esian coo dina es, uα, αand wαa e componen s o flow field o he567
adionuclide in he s a e α. In gene al, eloci y can diffe o diffe en s a es (e.g. due he568
p esence o se ling eloci y o suspended sedimen o o be ze o in he bo om deposi ).569
Khand K a e u bulen o molecula diffusi i ies in he ho izon al and e ical di ec ions570
espec i ely, and/o biodiffusi i y in he bo om deposi , which a e a iable in ime and571
space. The e m n
β=1 kβαCβdesc ibes fi s o de eac ions be ween he adionuclides in572
diffe en s a es, kβα a e kine ic ans e coefficien s and kαα =−n
β=1 kαβ o α=β;Sα
573
is he adionuclide sou ce e m and λis he adionuclide decay cons an . Equa ions o 574
he wa e column and bo om sedimen laye a e linked by fluxes o ac i i y.575
A.2 Lag angian models576
In Lag angian models he eleased ac i i y is ep esen ed by a numbe o pa icles, each577
one equi alen o a gi en amoun o ac i i y (Bq). The pa h ollowed by each pa icle is578
calcula ed and adionuclide concen a ions a e ob ained om he numbe o pa icles pe 579
olume o mass uni . The equa ions desc ibing a ia ions o pa icle (in s a e α) posi ion580
o e each ime inc emen d a egi enby heI ˆo (P o e , 2004) s ochas ic diffe en ial581
equa ions:582
dx =uαd +∂Kh
∂x d +2KhdWx,(2)
dy = αd +∂Kh
∂y d +2KhdWy,(3)
dz =wαd +∂K
∂z d +2K dWz,(4)
whe e uα, αand wαa e eloci y componen s on coo dina e axis (x, y, z) o s a e α;583
Wx,W
y,W
za e independen componen s o he s ochas ic mo ion ( he Wiene p ocess).584
25
120 140 160 180 200
20
30
40
50
60
Longi ude
La i ude
Su ace wa e
135 140 14
5
35
36
37
38
39
40
41
42
Longi ude
La i ude
In e media e wa e
135 140 145
35
36
37
38
39
40
41
42
Longi ude
La i ude
Deep wa e
135 140 14
5
35
36
37
38
39
40
41
42
Longi ude
La i ude
Sedimen
120 140 160 180 200
20
30
40
50
60
Longi ude
La i ude
Zooplank on
135 140 145
35
36
37
38
39
40
41
42
Longi ude
La i ude
Non−pisci o ous ish
120 140 160 180 200
20
30
40
50
60
Longi ude
La i ude
Pisci o ous ish
Figu e 3: Loca ions o sampling poin s o all conside ed en i onmen al compa men s.
32

120 130 140 150 160 170 180 190 200
15
20
25
30
35
40
45
50
55
60
65
Longi ude
La i ude
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figu e 4: Box di ision o he Pacific Ocean o model/da a compa isons.
33
200 400 600
10−1
100
101
102
C (Bq/m3)
BOX 1
USEV
ESTE
SEA−GEARN
I/K Lag .
THREETOX
LORAS
200 400 600
10−1
100
101
102
BOX 2
200 400 600
10−1
100
101
102
C (Bq/m3)
BOX 3
200 400 600
10−1
100
101
102
BOX 4
200 400 600
10−1
100
101
102
C (Bq/m3)
BOX 5
200 400 600
10−1
100
101
BOX 15
200 400 600
10−1
100
101
C (Bq/m3)
Day a e Jan 1, 2011
BOX 16
200 400 600
10−1
100
101
102
103
Day a e Jan 1, 2011
BOX 20
Figu e 5: Model p edic ions and measu emen s o 137Cs concen a ions in su ace wa e
o some boxes in he Pacific.
34
200 400 600
10−1
100
101
102
103
C (Bq/m3)
BOX 6
USEV
ESTE
SEA−GEARN
I/K Lag .
THREETOX
LORAS
200 400 600
10−1
100
101
102
103
BOX 7
200 400 600
10−1
100
101
102
103
C (Bq/m3)
BOX 9
200 400 600
100
101
102
103
104
105
BOX 10
200 400 600
100
102
104
106
108
C (Bq/m3)
BOX 11
200 400 600
10−1
100
101
102
103
BOX 12
200 400 600
10−1
100
101
C (Bq/m3)
Day a e Jan 1, 2011
BOX 13
200 400 600
10−1
100
101
Day a e Jan 1, 2011
BOX 14
Figu e 6: Model p edic ions and measu emen s o 137Cs concen a ions in su ace wa e
o some boxes in he Pacific.
35
200 400 600
100
101
102
103
104
105
C (Bq/m3)
BOX 10; Su ace wa e
USEV
ESTE
SEA−GEARN
I/K Lag .
THREETOX
LORAS
200 400 600
100
102
104
106
BOX 11; Su ace wa e
200 400 600
10−1
100
101
102
103
104
C (Bq/kg)
Day a e Jan 1, 2011
BOX 10; Sedimen
200 400 600
10−1
100
101
102
103
104
Day a e Jan 1, 2011
BOX 11; Sedimen
Figu e 7: Model p edic ions and geome ic means o 137Cs concen a ions measu ed o
each mon h in boxes 10 and 11, o su ace wa e and sedimen s. Geome ic s anda d
de ia ions a e no d awn because hey a e oo small compa ed wi h he e ical scales.
36
200 400 600
100
101
102
103
C (Bq/m3)
BOX 10, In e media e
USEV
ESTE
SEA−GEARN
I/K Lag .
THREETOX
LORAS
200 400 600
100
101
102
103BOX 11, In e media e
200 400 600
10−1
100
101
C (Bq/m3)
Day a e Jan 1, 2011
BOX 10, Deep
200 400 600
10−1
100
101
Day a e Jan 1, 2011
BOX 11, Deep
Figu e 8: Model p edic ions and measu emen s o 137Cs concen a ions in in e media e
and deep wa e s o some boxes in he Pacific.
37

200 400 600
10−2
10−1
100
101
C (Bq/kg)
BOX 9
USEV
ESTE
SEA−GEARN
I/K Lag .
THREETOX
LORAS
200 400 600
10−1
100
101
102
103
104
Day a e Jan 1, 2011
BOX 10
200 400 600
10−1
100
101
102
103
104
105
C (Bq/kg)
Day a e Jan 1, 2011
BOX 11
Figu e 9: Model p edic ions and measu emen s o 137Cs concen a ions in bed sedimen s
o some boxes in he Pacific.
38
200 400 600
10−2
10−1
100
101
102
C (Bq/kg)
BOX 1
USEV
ESTE
THREETOX
LORAS
200 400 600
10−2
10−1
100
101
Day a e Jan 1, 2011
BOX 7
200 400 600
10−2
10−1
100
101
102
C (Bq/kg)
Day a e Jan 1, 2011
BOX 12
Figu e 10: Model p edic ions and measu emen s o 137Cs concen a ions in zooplank on
o some boxes in he Pacific (Bq/kg we weigh ).
39
100 200 300 400 500 600 700
10−2
10−1
100
101
102
C (Bq/kg)
BOX 10
USEV
ESTE
THREETOX
LORAS
100 200 300 400 500 600 700
10−1
100
101
102
103
C (Bq/kg)
Day a e Jan 1, 2011
BOX 11
Figu e 11: Model p edic ions and measu emen s o 137Cs concen a ions in non-pisci o ous
fish (pelagic) o some boxes in he Pacific (Bq/kg we weigh ).
40
200 400 600
10−2
10−1
100
101
C (Bq/kg)
BOX 3
USEV
ESTE
THREETOX
LORAS
200 400 600
10−2
10−1
100
101
102
BOX 6
200 400 600
10−2
10−1
100
101
102
C (Bq/kg)
BOX 7
200 400 600
10−2
10−1
100
BOX 8
200 400 600
10−1
101
103
C (Bq/kg)
BOX 10
200 400 600
10−1
101
103
Day a e Jan 1, 2011
BOX 11
200 400 600
10−2
10−1
100
C (Bq/kg)
Day a e Jan 1, 2011
BOX 15
Figu e 12: Model p edic ions and measu emen s o 137Cs concen a ions in pisci o ous
fish (pelagic) o some boxes in he Pacific (Bq/kg we weigh ).
41