Mid-range atmospheric dispersion modelling. Intercomparison of simple models in EMRAS-2 project

Mid- ange a mosphe ic dispe sion modelling. In e compa ison o simple models in EMRAS-2 p ojec Raúl Pe ia ~ nez a , * , Ka hleen M. Thiessen b , Sohan L. Chouhan c , F ancesco Mancini d , Emilie Na a o e , Ge Sdouz , 1 , Dejan T i uno ic  g a Uni e si y o Se ille, ETSIA, C a U e a km 1, 41013, Se illa, Spain b Oak Ridge Cen e o Risk Analysis, 102 Donne D i e, 37830, Oak Ridge, Tennessee, USA c Canadian Nuclea Labo a o ies, K0J 1J0 Chalk Ri e , On a io, Canada d SOGIN S.p.A., Via To ino, 6, I-00184, Rome, I aly e Ins i u de Radiop o ec ion e de Sû e e  Nucleai e, 31 A enue de la di ision Lecle c BP17, F-922625, Fon enay-aux-Roses, F ance Aus ian Ins i u e o Technology, A-2444, Seibe sdo , Aus ia g Fede al Au ho i y o Nuclea Regula ion, Sheikh Zayed Fi s S ee , P.O. Box 112021, Abu Dhabi, Uni ed A ab Emi a es Keywo ds: A mosphe ic dispe sion Nuclea powe plan Model 137-Cs 131-I abs ac An in e compa ison o a mosphe ic dispe sion models has been ca ied ou o a hypo he ical acciden occu ing in a nuclea powe plan in he cen e o Spain. The acciden consis ed o a s eam gene a o ube up u e, and wo adionuclides ha e been conside ed o he exe cise: 137-Cs and 131-I. Me eo- ological condi ions and adionuclide elease a es we e supplied. Models p o ided deposi ion maps, imein eg a ed concen a ions in ai and a i al imes o he plumes o specific loca ions. The e ec o he me eo ological condi ions used in he modelling was clea , wi h di e en beha io o he plume wi h neu al s abili y s. s able condi ions. The p edic ed a i al imes o he plume a specific loca ions showed much less a iabili y han deposi ion and ai concen a ions. This a iabili y in pa eflec s he unce ain ies inhe en in a mosphe ic dispe sion modelling and in he selec ion o pa ame e alues, such as deposi ion eloci ies o di usi i ies. 1. In oduc ion En i onmen al assessmen models a e used o e alua ing he adiological impac o ac ual and po en ial eleases o adionuclides o he en i onmen . They a e essen ial ools o use in he egula- o y con ol o ou ine discha ges o he en i onmen and also in planning measu es o be aken in he e en o acciden al eleases. They a e also used o p edic ing he impac o eleases which may occu a in o he u u e, o example, om unde g ound adioac- i e was e eposi o ies. I is impo an o check, o he ex en possible, he eliabili y o he p edic ions o such models by com- pa ison wi h measu ed alues in he en i onmen o by compa ing wi h he p edic ions o o he models. The In e na ional A omic Ene gy Agency (IAEA) has been o ga- nizing p og ammes o in e na ional model es ing since he 1980s. The p og ammes ha e con ibu ed o a gene al imp o emen in models, in ans e da a and in he capabili ies o modelle s in Membe S a es. The possible benefi s o ca ying ou model ali- da ion and es ing a an in e na ional le el we e ecognized by he Swedish Radia ion P o ec ion Ins i u e, which sponso ed he Biosphe ic Model Valida ion S udy (BIOMOVS) and BIOMOVS II p og ammes s a ing in 1985 (BIOMOVS II,1996). BIOMOVS was he fi s in e na ional exe cise aimed a he es ing and alida ion o models o he p edic ion o adionuclide ans e h ough he en i onmen o humans. The Che nobyl acciden in 1986 c ea ed a enewed need o eliable assessmen s in many coun ies and p o ided an inc eased impe us o wo k in his a ea. I also o igi- na ed new da a se s ha could be pu o use o model es ing. As a consequence, he IAEA was p omp ed o s a a p og amme on he Valida ion o Model P edic ions (VAMP) in 1988, which concluded in 1996 (IAEA, 2000). Mo e ecen ly, En i onmen al Modelling o Radia ion Sa e y (EMRAS) p og am, unning om 2003 o 2007, was launched (IAEA, 2012). F om 2009 o 2011, he second s age o EMRAS was o ga- nized (EMRAS II). The U ban A eas Wo king G oup was o ganized *Co esponding au ho . E-mail add ess: [email p o ec ed] (R. Pe i a~ nez). 1 P esen add ess: Hockegasse 24/23, A-1180, Vienna, Aus ia. wi hin his las p og amme, as pa o a heme en i led “Ap- p oaches o Assessing Eme gency Si ua ions”. The Wo king G oup has been buil on he wo k done by he U ban Remedia ion Wo king G oup o he fi s phase o he EMRAS P og amme (IAEA, 2012). The goal o he U ban A eas Wo king G oup is o es and imp o e he capabili ies o models used in assessmen o adioac- i e con amina ion in u ban se ings, including dispe sion and deposi ion e en s, sho - and long- e m con aminan edis ibu- ion ollowing deposi ion e en s, and po en ial coun e measu es o emedia ion e o s o educing human exposu es and doses. A mid- ange a mosphe ic dispe sion exe cise was ca ied ou by he g oup. I is based on a hypo he ical acciden a a nuclea powe plan and he esul ing p edic ed deposi ion in u ban en i onmen s up o 70 km downwind. The scena io assumed a 1-h elease om a up u e o a s eam gene a o ube, based on an acciden scena io de eloped by he Ins i u de Radiop o ec ion e de Sû e  e Nucl eai e (IRSN), and uses ac ual geog aphic and me eo ological in o ma ion o he T illo nuclea powe plan in cen al Spain. I should be commen ed ha complex a mosphe ic dispe sion models exis now. Fo ins ance, WSPEED I-II (Te ada e al., 2012) and LADAS (Suh e al., 2009) we e applied in he case o Fukushima acciden o e alua e deposi ion on he sea su ace o la e compu e ma ine dispe sion o hese eleased adionuclides (Pe i a~ nez e al., 2015). Also, a mosphe ic dispe sion models we e applied o e al- ua e he sou ce e m om Fukushima (Chino e al., 2011; Kobayashi e al., 2013). These models consis o a me eo ological p edic ion model coupled wi h an a mosphe ic dispe sion model. Ne e he- less, he objec i e o his wo k consis ed o es ing simple a mo- sphe ic dispe sion models, which do no equi e he me eo ological sub-model, and which can p o ide a e y as answe in case o an acciden . Since he final pu pose o he models gene ally is a dose es ima ion o he public, models a e based on conse a i e app oaches. The pu pose o his pape is o p esen he main esul s om his exe cise. The exe cise is desc ibed in Sec ion 2. Resul s a e p e- sen ed in Sec ion 3. 2. Me hods The nuclea powe plan chosen o he exe cise is T illo (TNPP), Fig. 1. Top: Gene al localiza ion o he s udy a ea. Bo om: Topog aphy o he domain (ele a ions in m abo e sea le el). in he cen al pa o Spain, abou 70 km no heas om Mad id me opoli an a ea and 46 km om Guadalaja a, which is a smalle own in cen al Spain loca ed be ween TNPP and Mad id (Fig. 1). TNPP s a ed ope a ion in 1987. The powe is 1043 MW, and he eac o is PW ype. Cooling is ca ied ou h ough wo owe s. Specific me eo ological condi ions, which would be ep esen- a i e o a wo s -case scena io, we e used o he simula ions. Simula ed wind fields 10 m abo e he g ound we e p o ided. Two si ua ions we e conside ed: one wi h a s able a mosphe e, and one wi h neu al s abili y. This allowed assessmen o he e ec s o s abili y condi ions on adionuclide dispe sion. In bo h cases, he same geos ophic wind di ec ion is conside ed (no heas ). Geos ophic wind speeds o 3.0 m/s and 6.0 m/s we e used o he s able and neu al condi ions, espec i ely. The bounda y laye heigh was 1000 m o s able and 1500 m o neu al s abili y. Wind fields 10 m abo e he g ound we e ob ained om WIN- MOD model, de eloped a he Uni e si y o No h Wales (Jones, 1998). WINMOD calcula es such wind fields om he geos ophic wind and a mosphe e s abili y. Essen ially, he model diagnoses he local modifica ion o he wind field in egions o complex opog aphy. A geos ophic wind speed and di ec ion, as well as he a mosphe ic lapse a e and bounda y laye heigh , a e specified and he model i e a es he ho izon al momen um and empe a u e equa ions a he su ace owa ds a s eady s a e. Wind fields 10 m abo e he g ound a e p esen ed in Fig. 2 o bo h a mosphe ic condi ions. −3.8 −3.6 −3.4 −3.2 −3−2.8 −2.6 40.2 40.3 40.4 40.5 40.6 40.7 40.8 Longi ude La i ude • TNPP MADRID • Guadalaja a 500 600 700 800 900 1000 1100 1200 −3.8 −3.6 −3.4 −3.2 −3−2.8 −2.6 40.2 40.3 40.4 40.5 40.6 40.7 40.8 Longi ude La i ude • TNPP MADRID • Guadalaja a 500 600 700 800 900 1000 1100 1200 Fig. 2. Wind fields 10 m abo e he g ound p o ided by WINMOD model o s able ( op) and neu al (bo om) s abili y condi ions. Only one o each 16 p o ided ec o s is d awn o cla i y. Land ele a ions (colo scale) in m abo e sea le el. The same hypo he ical acciden was conside ed o bo h me e- o ological si ua ions. The adionuclides a e assumed o be eleased as gas, and only d y deposi ion is conside ed. Two adionuclides wi h di e en hal -li es a e conside ed, 137 Cs and 131 I, wi h all o he la e in molecula o m. Use o hese wo adionuclides wi h di e en hal -li es was conside ed enough o modelling in e - compa ison pu poses, al hough o cou se many mo e adionuclides and in di e en chemical o ms would be eleased du ing a eal acciden . The hypo he ical acciden conside ed he e consis s o a s eam gene a o ube up u e, a scena io which was de eloped by he Ins i u de Radiop o ec ion e de Sû e  e Nucl eai e (IRSN) o F ance. The du a ion o he elease is 1 h, and he elease a e is a iable o bo h adionuclides. Modelle s we e p o ided wi h elease da a o e he 1 h pe iod, shown in Fig. 3. Tempo al esolu ion o he elease da a is 60 s. To al (in eg a ed) eleases a e 6.42 10 11 Bq and 3.69 10 12 Bq o 137 Cs and 131 I, espec i ely. An e ec i e elease heigh o 50 m was conside ed. Modelle s we e asked o ca y ou a simula ion o e 10 h. Endpoin s o simula ions we e con ou maps o deposi ed ac i i y on he g ound and o ime-in eg a ed ac i i y concen a ions in ai a g ound le el, a he end o he simula ion. Modelle s we e also asked o p o ide a ime se ies o ac i i y concen a ions in ai a ou selec ed poin s ( wo in e media e poin be ween TNPP and Guadalaja a own, deno ed as IP1 and IP2, Guadalaja a and Mad id down own). These poin s a e shown in Fig. 1. Table 1p o ides a summa y o he fi e models used in he ex- e cise. Mo e in o ma ion abou indi idual models is p o ided in Appendix A-E and in he e e ences included in he Table. The models ep esen se e al di e en pu poses (e.g., eme gency assessmen and esea ch) and wo majo ypes o modelling ap- p oaches (Gaussian and Lag angian). One model (Ho spo ) p o- ided esul s in e ms o he dis ance down he plume cen e line, while he o he s we e able o p o ide esul s wi h e e ence o he local geog aphy. Fo one model (RASCAL), he ange o he model p edic ions did no ex end o he dis ance o Mad id in his exe cise. Fou models we e used o p o ide esul s o bo h s able a mo- sphe ic condi ions and neu al s abili y; he fi h (JRODOS) was 0 10 20 30 40 50 60 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2x 109 Time (minu es) Release a e (Bq/s) 137Cs 131I Fig. 3. Radionuclide elease a es o he s eam gene a o ube up u e acciden . The o igin o ime is a bi a y. Table 1 Summa y o main model cha ac e is ics. ADDAM RASCAL USe JRODOS HOTSPOT Type o model Gaussian Gaussian þLag angian pu Lag angian Gaussian þsimplified pu Gaussian Calcula ion ange use specified adius 80 km <100 km <100 km >10 m, <100 km Release heigh 50 m 50 m 50 m 10 m 50 m S abili y class E (s able), D (neu al) E, D E, D E E, D Wind ec o s summed ou side he code Limi ed numbe used As p o ided As p o ided E: 3 m/2; D: 6 m/s Di usion Ho izon al: 60 m 2 /s High oughness: Ka ls uhe-Jülich c coe ficien s B iggs a Pasquill-Gi o d cu es b Ve ical: 30 m 2 /s Mode a e oughness: Mol d Pasquill-Gi o d D y deposi ion Cs: 0.01 m/s Cs: 0.04 m/s eloci y I: 0.008 m/s 0.003 m/s No applicable In e nally de e mined I: 0.22 m/s Release ime-s ep 1 h 1 h 1 min 0.5 h 1 h Calcula ion ime-s ep One single s ep 15 min 10 s 1 h One single s ep Simula ion ime No applicable 10 h 10 h 7 h No applicable Topog aphy Fla No used As p o ided No applicable Rugosi y 0.40 m 0.20 m 0.40 m Ci y e ain Re e ence Scheie and Chouhan, 2009 Ramsdell, 2012 NP e Ie din e al., 2012 Homann and Aluzzi, 2014 a B iggs (1971). b Gi o d (1968). c Geiss e al. (1981). d Pani z e al. (1989). e No published. used only o s able condi ions. One model (USe ) used ime- dependen sou ce e m in o ma ion as p o ided, in 1-min in- c emen s. Th ee models conside ed he en i e elease in a single 1- h ime s ep, while he o he model (JRODOS) conside ed i in wo hal -hou ime s eps. The models also di e ed in hei handling o he in o ma ion abou wind speed and di ec ion and in he d y deposi ion eloci ies ha we e used. 3. Resul s and discussion Con ou maps o p edic ed deposi ion o 137 Cs a e shown in Figs. 4e8. Two se s o p edic ions (USe and ADDAM, Figs. 4 and 5) clea ly show he e ec o me eo ological condi ions: unde s able condi ions, he p edic ed plumes in e sec ed Mad id, bu wi h neu al s abili y, he plume bypassed Mad id. The JRODOS plume o s able condi ions also in e sec ed Mad id (Fig. 6). P edic ions wi h RASCAL did no ex end as a as Mad id, bu he plo s sugges ha he plume would ha e in e sec ed Mad id wi h s able condi- ions and bypassed Mad id wi h neu al s abili y (Fig. 7). Ho spo p o ided esul s in e ms o he plume cen e line a he han he local geog aphy, bu he esul s do show highe deposi ion a he downwind o he s able condi ions (Fig. 8). P edic ed alues o deposi ion o 137 Cs and 131 I a specific lo- ca ions a e p o ided in Tables 2 and 3. As expec ed solely om he dis ances, he gene al endency was o he highes deposi ion a In e media e Poin -1 (IP1), hen IP2, Guadalaja a, and Mad id. P edic ed di e ences be ween bo h adionuclides eflec bo h he di e en sou ce e ms and ( o some models) di e en deposi ion eloci ies o he wo adionuclides. Fo 137 Cs deposi ion in Mad id, p edic ed alues o s able Fig. 4. Calcula ed 137 Cs deposi ion by USe model o s able ( op) and neu al (bo om) s abili y condi ions. condi ions (4 models) di e ed by a ac o o 28. Fo neu al s abili y, all p edic ed alues o Mad id we e 0 (USe ) o a ac o o abou 10e100 lowe han o s able condi ions. Fo 131 I, p edic ed alues in Mad id o s able condi ions di e ed by a ac o o abou 60. As wi h 137 Cs, p edic ed alues o 131 I o neu al s abili y we e 0 o a ac o o abou 10e100 lowe han o s able condi ions. These alues a e consis en wi h he con ou maps, which showed he plume bypassing Mad id in he case o neu al s abili y. Fo Guadalaja a (5 models), p edic ed alues o deposi ion wi h s able condi ions we e wi hin a ac o o 12 o 137 Cs and 24 o 131 I. Fo neu al s abili y, p edic ed alues we e wi hin a ac o o 165 o 137 Cs and 26 o 131 I, indica ing mo e a iabili y among he models o neu al s abili y. ADDAM and USe p edic ed highe deposi ion in Guadalaja a o neu al s abili y han o s able condi ions, while Ho spo and RASCAL p edic ed highe deposi ion wi h s able condi ions. USe model plume bypassed IP2 wi h neu al s abili y. ADDAM p edic ed deposi ion a bo h IP1 and IP2 o bo h s able condi ions and neu al s abili y, wi h a dec ease om IP1 o IP2 o abou a ac o o 9 o s able condi ions and a ac o o 3 o neu al s abili y. Su p isingly, RASCAL model plume bypassed IP1 (0 deposi ion) unde s able condi ions, and he e is no deposi ion immedia ely downwind TNPP (Fig. 7). This was a ibu ed o he me eo ological p ep ocesso included in his model (Appendix B): only a pa o he p o ided wind ec o s we e used. This p obably does no allow o desc ibe app op ia ely he plume beha io in he ini ial s ages o dispe sion. Time-in eg a ed concen a ions in ai a e no shown. Howe e , he gene al endency was a dec ease wi h dis ance, as expec ed. Thus he highes ai concen a ions we e ound a In e media e Poin -1 (IP1), hen IP2, Guadalaja a, and Mad id. One excep ion was JRODOS, which p edic ed sligh ly highe concen a ions o pa s o Mad id. Also, RASCAL’s p edic ed plume bypassed IP1 wi h s able condi ions and p edic ed highe concen a ions o 131 I in Guada- laja a han o IP2 o bo h se s o me eo ological condi ions. Fo Mad id, he esul s again show a la ge dec ease (a ac o o ~10 o Fig. 5. Calcula ed 137 Cs deposi ion by ADDAM model o s able ( op) and neu al (bo om) s abili y condi ions. mo e) om he alues p edic ed o s able condi ions o he alues p edic ed o neu al s abili y, consis en wi h a p edic ed plume bypassing Mad id in he case o neu al s abili y. An impo an ou pu om a mosphe ic dispe sion models, ela ed o eme gency managemen , is he a i al ime o he adioac i e plume o a gi en poin . P edic ed alues o he app oxima e ime o a i al o he p edic ed plume a specific lo- ca ions a e p o ided in Table 4. Fo any gi en loca ion and s abili y class, p edic ed imes o a i al a e wi hin a ac o o abou 2 o each o he , o hose plumes p edic ed o each he loca ion. As desc ibed p e iously, RASCAL p edic ed ha he plume would bypass IP1 unde s able condi ions, and USe p edic ed ha he plume would bypass bo h IP2 and Mad id o he case o neu al s abili y. Fo plumes p edic ed o each a gi en loca ion, imes o a i al we e sho e o neu al s abili y han o s able condi ions. In gene al, ADDAM p edic ed he longes imes o a i al o a gi en Fig. 6. Calcula ed 137 Cs deposi ion (Bq/m 2 ) by JRODOS model o s able condi ions. Fig. 7. 137 Cs deposi ion con ou s calcula ed by RASCAL model o s able ( op) and neu al (bo om) s abili y condi ions. Fig. 8. Calcula ed 137 Cs deposi ion con ou s by Ho spo model o s able ( op) and neu al (bo om) s abili y condi ions. Table 2 Compa ison o p edic ions o deposi ion o 137 Cs (Bq/m 2 ). NR: no epo ed. Model IP1 IP2 Guadalaja a Mad id S able condi ions (E) ADDAM 9030 964 44.8 1.03 Ho spo NR NR 20 13 JRODOS NR NR 14 28 RASCAL 0 174 166 NR USe 4348 233.5 56.86 10.14 Neu al condi ions (D) ADDAM 4970 1490 381 0.0165 Ho spo NR NR 2.3 1.4 RASCAL 267 69.8 122 NR USe 7414 0 144.3 0 Table 3 Compa ison o p edic ions o deposi ion o 131 I (Bq/m 2 ). NR: no epo ed. Model IP1 IP2 Guadalaja a Mad id S able condi ions (E) ADDAM 43,800 4920 256 6.66 Ho spo NR NR 610 390 JRODOS NR NR 110 72 RASCAL 0 964 922 NR USe 4312 149.7 38.98 11.45 Neu al condi ions (D) ADDAM 23, 400 7120 1880 0.0837 Ho spo NR NR 71 45 RASCAL 1480 389 677 NR USe 7395 0 146 0 loca ion and USe and RASCAL he sho es imes. 4. Conclusions Pa icipan s in his es exe cise s a ed wi h he same in o - ma ion, and while he e is gene al ag eemen in he esul s, he e a e also some ob ious di e ences. I is impo an o iden i y and explain he easons o hese di e ences. As desc ibed abo e, pa - icipan s a ied in hei handling o he sou ce e m ( ime-depen- den o all a once) and wind field da a, and in hei selec ion o pa ame e alues such as deposi ion eloci y ( o hose models ha used a gi en pa ame e ). The e ec o he me eo ological condi ions used in he model- ling is ob ious, wi h di e en beha io o he plume wi h neu al s abili y s. s able condi ions. In his exe cise, he mos impo an di e ence was ha he p edic ed plumes ended o bypass Mad id ( he la ges ci y in he es egion) wi h neu al s abili y bu in e sec ed i wi h s able condi ions. P edic ions a ied as o whe he he plume eached o bypassed some o he in e media e poin s. I is impo an o no e ha e en when all modelle s used he same s abili y class, he pa h o he p edic ed plume could a y among pa icipan s. Va iabili y be ween model p edic ions was highe o deposi- ion and ai concen a ions (no shown) han o o he endpoin s. This a iabili y in pa eflec s he unce ain ies inhe en in a mo- sphe ic dispe sion modelling and in selec ion o pa ame e alues. Fo example, o he h ee models equi ing he use o selec alues o deposi ion eloci y, he selec ed alues anged om 0.003 o 0.04 m/s o 137 Cs and om 0.008 o 0.22 m/s o 131 I(Table 1). Models may also ha e a ied in he heigh used o calcula ion o he ai concen a ions. Al hough unce ain ies in indi idual model p edic ions we e no es ima ed, he esul s gi e an idea o he le el o unce ain y ha should be acknowledged in dealing wi h modelling esul s in impo an assessmen s. In his sense, a mul i- model app oach, as desc ibed by Mon e e al. (2008), may be o in e es when en i onmen al p ocesses a e complex. Th ough his app oach, he conclusions ha ob ain he g ea es deg ee o consensus among modelle s a e made e iden and he aspec s ha a e subjec o dispu e and which should he e o e be handled ca e ully also become clea . The p edic ed a i al imes o he plume a specific loca ions showed much less a iabili y, usually less han a ac o o 2 among h ee o ou se s o model p edic ions. In p ac ice, his is an impo an endpoin , in ha i p o ides an es ima e o he ime a ailable o e acua ing an a ea o ge ing people o shel e . The esul s also show he dependence o his endpoin on me eo o- logical condi ions. Acknowledgemen Wo k ca ied ou in he ame o IAEA EMRAS-2 p ojec (En i onmen al Modelling o Radia ion Sa e y-2). Appendix A ADDAM ADDAM (Ballyk e al., 2003) is a Gaussian plume model which conside s he ollowing a mosphe ic dispe sion phenomena (all ha e been alida ed): plume ise, downwash, and en ainmen (e ec i e elease heigh ); umiga ion; eflec ion om an ele a ed in e sion; anspo and dispe sion (plume b oadening and plume di usion); we and d y deposi ion and plume deple ion; adioac- i e decay and build-up; ex e nal exposu e due o cloudshine (including a fini e cloud model) and g oundshine; and in e nal exposu e due o inhala ion. The model uns s ochas ically wi h espec o me eo ological da a, buoyan eleases can be conside ed and he la e al dispe sion coe ficien can be calcula ed om he s anda d de ia ion in he wind di ec ion. Me eo ological condi ions a e cons an and he e ain is fla . The domain o in e es is b oken down in o 16 sec o s and a use -specified numbe o adial dis ances. Then, o a gi en elease, dilu ion ac o s (concen a ion di ided by elease a e) and doses a e calcula ed a each dis ance in he a ec ed sec o s. Some limi a ions a e ha he model is no sui able o pu e- leases ( elease du a ion should be g ea e han he a el ime) and ha calm winds canno be ea ed. B RASCAL RASCAL (Ramsdell, 2012) is a adiological assessmen ool o use in eme gency esponse applica ions. I consis s o modules ha es ima e acciden sou ce e ms o nuclea powe plan s and o he nuclea uel cycle acili ies; anspo , dispe sion and deposi ion o adionuclides; and doses. I also includes a me eo ological p e- p ocesso ha p epa es me eo ological da a o use by he a mo- sphe ic anspo modules. RASCAL is a Lag angian ajec o y Gaussian-pu dispe sion model. The e a e wo a mosphe ic s abili y fields. One consis s o Pasquill-Gi o d s abili y classes (Gi o d, 1968), and he o he consis s o he in e se Monin-Obukho leng h (Monin and Obukho , 1954). The Monin-Obukho leng h is es ima ed om he Pasquill-Gi o d s abili y class using a g aphical ela ionship be ween Monin-Obukho leng h, s abili y class, and su ace oughness. RASCAL accep s h ee p ecipi a ion condi ions: no p ecipi a ion, ain, and snow. E e y hou , he p ecipi a ion g id is upda ed using hou ly obse a ions. The plume is ep esen ed by a se ies o pu s eleased a 5-min in e als. Each pu con ains he ac i i y eleased du ing a 5-min pe iod. The heigh o elease is he sum o he ac ual elease heigh and final plume ise. The dispe sion pa ame e s a e a unc ion o dis ance a elled and a mosphe ic s abili y using nume ical ap- p oxima ions o he Pasquill-Gi o d dispe sion cu es. Deposi ion is calcula ed using a sou ce deple ion model wi h a cons an d y deposi ion eloci y. We deposi ion is calcula ed using a simple washou model wi h cons an washou coe ficien s. C USe The model, no published, is based on he app oach by Ellio (1999).Essen ially, i is a Lag angian h ee-dimensional pa icle- acking dispe sion model. Radioac i e decay and u bulen di u- sion a e simula ed using s ochas ic me hods. A ho izon al and a e ical di usion coe ficien mus be specified. Pa icles a e assumed o be deposi ed when hei heigh is smalle han 10 cm o e he g ound le el, hus a deposi ion eloci y is no equi ed. Table 4 Compa ison o p edic ions o plume a i al imes (min). NR: no epo ed. Model IP1 IP2 Guadalaja a Mad id S able condi ions (E) ADDAM 67 150 417 850 JRODOS NR 120 240 420 RASCAL bypassed 90 210 NR USe 40 100 210 460 Neu al condi ions (D) ADDAM 33 83 233 483 RASCAL 15 60 150 NR USe 20 bypassed 130 bypassed