Fukushima 137Cs releases dispersion modelling over 1 the Pacific Ocean. Comparisons of models with water, 2 sediment and biota data

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