Expe iences om a case s udy o mul i-model applica ion
o assess he beha iou o pollu an s in he Dniepe –Bug
Es ua y
Luigi Mon ea,∗, La s H ˚akansonb, Raul Pe ia ˜nezc, Gennady Lap e d, Ma k Zheleznyake,
Vladimi Made ich , Giacomo Angelig, Vladimi Koshebu skyh
a ENEA, CR Casaccia, ia P. Anguilla ese, 301, 00100 Rome, I aly
b Ins i u e o Ea h Sciences, Uppsala Uni e si y, Sweden
c Depa amen o Fisica Aplicada, Uni e si y o Se illa, Spain
d Uk ainian Ins i u e o Hyd ome eo ology, Uk aine
e Ins i u e o Ma hema ical Machines and Sys em P oblems, Uk aine
Ins i u e o Ma hema ical Machines and Sys em P oblems, Uk aine
g ENEA, CR Casaccia, I aly
h Ins i u e o Ma hema ical Machines and Sys em P oblems, Uk aine
Keywo ds:
En i onmen al models
Coas al a eas
Con amina ion
Mul i-model app oach
abs ac
The p esen pape desc ibes he esul s o he applica ion o ou s a e-o - he-a models o
p edic he concen a ions o pollu an s in he abio ic componen s o he Dniepe –Bug Es u-
a y (Uk aine). The es ua y was con amina ed by he adioac i e subs ances in oduced in he
en i onmen ollowing he Che nobyl acciden . The scope, he me hodological app oaches
and he heo e ical ounda ions unde pinning he examined models a e p esen ed and
compa ed. The model pe o mances we e assessed by compa ison wi h a ailable empi -
ical da a o wa e con amina ion. The main ac o s influencing he inhe en unce ain y o
he models we e examined: incomple e knowledge, pauci y o ex ensi e da a se s ele an
o some en i onmen al quan i ies, he agueness and he ambigui y o ce ain in o ma ion
abou en i onmen al p ocesses ha can be ha dly pa ame e ised in quan i a i e way, e c.
Model pe o mances eflec he in insic unce ain y o knowledge conce ning he quan-
i a i e beha iou o he in ol ed en i onmen al p ocess, he ambigui y o in e p e a ion
and pa ame e isa ion o such p ocesses, he inhe en a iabili y o en i onmen al quan i-
ies, e c. The di ficul ies in selec ing he “bes pe o mance” model and he benefi s a ising
om a mul i-model app oach o a o d complex en i onmen al p oblems a e p esen ed and
discussed. Mul i-model app oach helps o ge an insigh in o complex p oblems o en i-
onmen al managemen , o p omo e co-ope a ion among modelle s and o p ofi by he
di e en pe spec i es o he models.
∗Co esponding au ho . Tel.: +39 06 30484645; ax: +39 06 30486716.
E-mail add ess: [email p o ec ed] (L. Mon e).
1. In oduc ion
The p esen pape , al hough dealing wi h he beha iou
o adioac i e con aminan s in coas al a eas, a o ds some
p oblems o gene al na u e conce ning he applica ion o
models o complex en i onmen al sys ems. I should be
ecognised ha he acciden a he Che nobyl nuclea
powe plan in 1986 occu ed when, owing o a g ea
deal o p e ious heo e ical and expe imen al s udies, some
disciplines, like adioecology, had eached ma u e s ages
o de elopmen . The e o e, he numbe o in e na ional
p ojec s launched o he assessmen o model applica-
ions o he en i onmen al con amina ion om he abo e-
men ioned acciden (BIOMOVS, 1991; IAEA, 2000) o e ed
he unique oppo uni y o lea ning mo e abou p ob-
lems o gene al ele ance conce ning he alidi y, he
usage and he use ulness o en i onmen al modelling
o sol ing complex p oblems o en i onmen al manage-
men . This wo k will y o discuss and analyse such
lessons.
Due o he biological alue o he coas al zone and he
di e en demands on he u ilisa ion o coas al wa e s, i
is easy o unde s and why so much in e es conce ns his
ecosys em and, mo e gene ally, he ma ine en i onmen
(Bax e e al., 1998; Aa k og, 1998; K yshe and Sazykina,
1995). In pa icula , es ua ies a e he objec o esea ch
and en i onmen al managemen in iew o hei high bio-
logical p oduc i i y, o hei economic exploi a ion and o
hei impo ance o he whole biological and ecological
cycles o many li ing species. Howe e , in he amewo k
o in e na ional p ojec s on model alida ions in adioe-
cology, much mo e e o s ha e been ocused on lakes
and i e s han on coas al a eas and seas. Recen ly, some
e iews o s a e-o - he-a models o he managemen o
lakes, ca chmen s and i e s pollu ed by adioac i e sub-
s ances ha e been p esen ed (Mon e e al., 2003, 2004,
2005a). These e iews we e done wi hin he hema ic ne wo k
EVANET-HYDRA (“E alua ion and Ne wo k o EC-Decision
Suppo Sys ems in he field o Hyd ological Dispe sion Mod-
els and o Aqua ic Radioecological Resea ch”; h p://in o.
casaccia.enea.i /e ane -hyd a) financed by he Eu opean
Commission.
In spi e o he a en ion ha coas al en i onmen
dese es, i seems ha no simila s udies a e a ail-
able o models aimed a p edic ing he beha iou o
adionuclides in es ua ine sys ems. One o he aims o
he p esen wo k is o b idge his gap. We will ou line
an assessmen o he me hodologies used by s a e-
o - he-a models ha can be applied o he coas al
en i onmen . A pa icula a en ion will be de o ed
o he assessmen o he esul s o an exe cise o he
applica ion o di e en coas al models o a specific
es ua ine scena io, he Dniepe –Bug Es ua y (DBE). The
benefi s om a mul i-modelling app oach will be anal-
ysed and assessed in iew o he di ficul ies, such as he
sho age o inpu da a and in o ma ion, encoun e ed by
modelle s when dealing wi h complex en i onmen al
p oblems.
2. O e iew o modelling app oaches
2.1. Gene al ema ks
In his sec ion, we will b iefly ou line he gene al, heo e i-
cal ounda ions unde pinning he models o assessing he
beha iou o pollu an s in es ua ine sys ems. Mo e de ailed
desc ip ions o he models objec o he p esen s udy a e
epo ed in Appendix A.
In p inciple, mos app oaches o p edic ing he mig a-
ion o adionuclides h ough es ua ies a e simila o he ones
implemen ed in models o lakes and i e s (Mon e e al., 2003,
2005a). I is, howe e , necessa y o accoun o he pa icu-
la en i onmen al condi ions and p ocesses ha significan ly
influence he beha iou o con aminan s in coas al en i on-
men s.
Gene ally speaking, he o e all s uc u e o models o p e-
dic ing he beha iou o adionuclides in es ua ies comp ises
he ollowing sub-models:
(a) modules o p edic ing he physical, hyd ological, he
hyd aulics and he bio ic p ocesses ha occu in he
es ua ine sys em (such as wa e empe a u e and salin-
i y p ofiles, wa e fluxes and cu en eloci ies, e osion-
sedimen a ion p ocesses and dynamics o suspended
ma e in wa e , g ow h a es o o ganisms, idal cycles,
e c.) and a e supposed o influence he con aminan
mig a ion;
(b) modules o p edic ing he adionuclide ans e :
1. o he es ua y om i s ca chmen ;
2. h ough he abio ic componen s o he es ua y and
3. om he abio ic componen s o he bio a.
As o he o he wa e bodies, he main p ocesses ha con-
ol he mig a ion o con aminan s in es ua ies a e, basically,
he di usion o dissol ed subs ances due o wa e u bulen
mo ion (eddy di usion), he anspo due o he wa e cu -
en , he in e ac ion o dissol ed pollu an wi h suspended
ma e and bo om sedimen , he mixing p ocesses be ween
di e en laye s o wa e and, finally, he mig a ion o pollu-
an s om he wa e o he bo om sedimen s (sedimen a ion)
and om he bo om sedimen o he wa e ( e-suspension).
An example o sub-model o p edic ing hyd odynamic p o-
cesses is desc ibed in Appendix A (Uni e si y o Se ille model).
In gene al, se e al di e en configu a ions may be adop ed
o his kind o models. The mos gene al equa ions a e he
ull 3D hyd odynamic equa ions including ba oclinic e ms
(densi y di e ences). Howe e , “2D e ical dep h-a e aged”
models a e o en used when a e ical mixing o he wa e col-
umn can be hypo hesised (Pe ia ˜
nez e al., 1996). These models
p o ide p edic ions o he wa e cu en . Suspended sedimen
anspo is also desc ibed by anspo /di usion equa ion
accoun ing o wa e eloci y and e osion and deposi ion p o-
cesses (Eisma, 1993; Pe ia ˜
nez, 2002).
I is ob ious ha no all he models make use o hyd o-
logical/hyd aulic sub-models o assessing he quan i a i e
beha iou o he en i onmen al p ocesses influencing he
mig a ion o pollu an s h ough he componen s o an es u-
a y ecosys em. Some models make use o empi ical alues
o hose pa ame e s ha ela e he mig a ion o pollu an s o
he en i onmen al p ocesses and ha a e a e aged o e fini e
egions o space and in e als o ime ( o ins ance, mon hly
a e ages o ime-dependen quan i ies, a e age wa e fluxes
among di e en sec ions o a wa e body, e c.).
The sub-models o p edic ing he in e ac ion o adionu-
clide in he wa e column wi h suspended ma e and bo om
sedimen s show di e en deg ees o complexi y anging
om simple kd-based assump ions (i is supposed ha a
e e sible equilib ium be ween dissol ed and pa icula e
phases o adionuclide is quickly achie ed) o complex mul i-
s age in e ac ions (Pe ia ˜
nez, 2004; Ci oy e al., 2001). These
sub-models assess he complica ed p oblem o adionuclide
in e ac ion wi h sedimen pa icles om a ious pe spec i es
emphasising, o ins ance, he s ochas ic na u e o such a p o-
cess (Bø e zen and Salbu, 2002), he a ie y o physical and
chemical cha ac e is ics ha influence i s a iabili y (Ab il
and F aga, 1996; Ma sunaga e al., 2004; Smi h and Comans,
1996) o he di e en modelling app oaches (Mon e e al.,
2005b). As many o he men ioned aspec s we e desc ibed in
p e ious pape s (Mon e e al., 2003, 2005a), we do no deem
necessa y o epea he esul s o assessmen s ha ha e been
discussed in he scien ific li e a u e.
Some pa icula en i onmen al condi ions and p ocesses
a e ypical o he es ua ine ecosys em and significan ly influ-
ence he mig a ion o oxic subs ances. Fo ins ance, he di e -
en densi y be ween sea and esh wa e s gene a es a e ical
s a ifica ion ha a ec s he di usion o pollu an s h ough
he wa e column. Mo eo e , idal cycle is a u he ac o ha
should be conside ed and modelled o assessing bo h he
hyd odynamics o es ua ies and he dispe sion o con ami-
nan s h ough es ua ies (Dye , 1980).
The men ioned p inciples and app oaches unde pin he
g ea deal o models de eloped o p edic he dispe sion o
adionuclide h ough he ma ine and he coas al en i on-
men (P andle, 1984; B e on and Salomon, 1995; Salomon e
al., 1995; Schon eld, 1995; Ald idge e al., 2003; Ald idge, 1998;
Goshawk e al., 2003; Smi h e al., 2003; Ab il and Abdel-Aal,
2000; Ha ms, 1997; H˚
akanson, 1999, 2000; Ce ina e al., 2000;
Fishe e al., 1999).
Mos me hodologies o adioecological modelling a e sim-
ila o he ones used o p edic he dis ibu ion and he e ec
o o he kinds o oxican s, such as hea y me als, in wa e
Table1–Main ea u eso hemodels ha pa icipa ed in he exe cise
Model De elope Main model ea u es Ho izon
Coas Mab Uppsala Uni e si y,
Sweden
Gene ic, p ocess-based, dynamic, high emphasis
on ecological aspec s. TRP = mon hly a e ages,
SRP = he en i e coas al sys em
Compa men model o p edic mon hly a e age
alues o adionuclide concen a ions in wa e and
fish accoun ing o p e ailing ecological and
en i onmen al p ocesses when adionuclide
concen a ions in sea and di ec adionuclide
deposi ion o e he coas al a ea a e a ailable ( o
he specific applica ions also di ec adionuclide flux
om i e s a e supplied as inpu da a)
U. Se illa Uni e si y o Se illa,
Spain
2D (dep h-a e aged) based on
ad ec ion-di usion equa ion, high emphasis on
hyd odynamic p ocesses TRP and SRP: in
p inciple he model can supply e y de ailed
in o ma ion
2D model. The hyd ological cha ac e is ics o he
sys em ha influence he mig a ion o pollu an
(wa e cu en eloci y field) a e modelled om
undamen al equa ions accoun ing o
me eo ological condi ions
ENEA ENEA, I aly Gene ic, p ocess-based, dynamic, emphasis on
adionuclide flux balance and mig a ion o
sedimen TRP = mon hly a e ages SRP = o he
p esen applica ion he coas al sys em is
subdi ided in h ee sec o s
Compa men model o p edic mon hly a e age
alues o adionuclide concen a ions in deep and
su ace wa e accoun ing o p e ailing adionuclide
fluxes when adionuclide concen a ions in sea and
di ec adionuclide deposi ion o e he coas al a ea
a e a ailable. Balance o adionuclide in he sys em
is e alua ed using, as inpu da a (a) he adionuclide
fluxes om he i e s flowing in o he coas al
sys em; o (b) he deposi ion o adionuclide o e he
whole Dniepe ca chmen ( egional model o 90S )
THREETOX IMMSP, Uk aine 3D model based on ad ec ion-di usion
equa ion, high emphasis on hyd odynamic
p ocesses TRP and SRP: in p inciple he model
can supply e y de ailed in o ma ion
3D model o p edic ing concen a ion in wa e
when adionuclide inpu in o he sys em is known.
The hyd odynamics is simula ed on he base o 3D,
ime-dependen , ee su ace, p imi i e equa ion
model The hyd ological cha ac e is ics o he
sys em ha influence he mig a ion o pollu an
(wa e cu en eloci y field) a e modelled om
undamen al equa ions accoun ing o
me eo ological condi ions. The model accoun o
he p ocesses go e ning he exchange o
adionuclide be ween wa e and suspended/bo om
sedimen
TRP: ime esol ing powe , SRP: spa ial esol ing powe .
sys ems (Jø gensen, 1979; Ci oy e al., 2000). On he o he
hand, he en i onmen al p ocesses, such as sedimen a ion,
sedimen e-suspension, idal dynamics, wa e s a ifica ion,
e c., ha con ol he beha iou o a con aminan in wa e , a e
common o bo h adioac i e and non- adioac i e subs ances
(H˚
akanson e al., 2004). Ne e heless, pa icula chemical p o-
cesses a ec ing he beha iou o specific con aminan s need
o be conside ed and app op ia ely modelled (me cu y is an
ob ious example; Raja e al., 2004). The si ua ion is e en mo e
complica ed o pollu an s, like o ganic compounds, whose
a e is in ima ely linked o he cycling o o ganic ma e in
he aqua ic ecosys em (DeB uyn and Gobas, 2004; Gobas, 1993;
Jimenez-Mon ealeg e e al., 2002) o is con olled by p ocesses
o deg ada ion like pho olysis and hyd olysis (Mossman and
AlMulki, 1996) ha do no a ec he beha iou o adionclides
o hea y me als.
Today, many use - iendly so wa e ools a e a ailable o
implemen models in compu e codes and many models ha e
been de eloped and es ed. Modelle s can exchange da a and
in o ma ion assessing and compa ing hei esul s in a e y
dynamic and in e ac i e en i onmen domina ed by he in o -
ma ion echnologies. How can we ake ad an age o all ha ?
2.2. Ou line o model ea u es
The main ea u es o he models used o he p esen exe cise
a e desc ibed in Appendix A and summa ised in Table 1.I can
be use ul o classi y models acco ding o hei ho izon and o
hei ime and spa ial esol ing powe s (TRP and SRP) in iew
o hei scopes and aims.
Ho izon, he ange o in o ma ion (knowledge ob ained
om in es iga ion and s udy, ins uc ions and eques ed
inpu da a) ha he model is mean o p ocess and o he ou -
comes ha is mean o p edic . The ho izon defines he s a -
ing and he end poin s o a model. Fo ins ance, a model can
use as inpu da a he deposi ion o adionuclide on o he wa e
su ace and he ca chmen a ea o he es ua y. The model
ho izon is di e en when he inpu da a a e he deposi ion
on o he wa e su ace and he empi ical alues o adionu-
clide flowing om he ibu a y i e s. Ob iously in he fi s
example, he model aim is mo e ambi ious and he model is
mo e in o ma i e as i does no equi e in o ma ion (inpu o
con aminan om i e s) ha can be di ficul o impossible
o ob ain o some scena ios ( o ins ance, in case o an acci-
den ). Ne e heless, a la ge ho izon gene ally implies a highe
unce ain y.
The esol ing powe o a model is a measu e o he le el
o de ail o i s p edic ions. The “ ime esol ing powe ” (TRP)
is he abili y o a model o p edic di e ences in he sys em
beha iou o e a gi en in e al o ime. In o he wo ds, he
model is supposed o desc ibe he a e age beha iou o he
sys em o e a defined ime in e al. Example o ac o s a ec -
ing he TRP a e he in e als o ime necessa y o assu e he
damping o some ansien p ocesses which a e no in ended
o be modelled in a su ficien de ailed ime scale ( o ins ance,
he ime necessa y o app oach a homogeneous dis ibu ion o
he con aminan in he wa e column, a condi ion ha , ob i-
ously, is no immedia ely achie ed, when box models a e used;
he so p ion and de-so p ion p ocesses o adionuclide on sus-
pended ma e when he model is based on he hypo hesis o
an “ins an aneous” equilib ium be ween he dissol ed and he
pa icula e con aminan phases, e c.).
Simila ly he spa ial esol ing powe (SRP) is he abili y o
a model o p edic di e ences in he sys em beha iou o e
a gi en spa ial g id. Fo ins ance, he SRP o box models o
p edic ing he spa ial dis ibu ion o a subs ance in wa e is
he size o he boxes. The ho izon, he TRP and he SRP a e “a
p io i” cha ac e is ics o a model. They a e planned and se up
by model de elope s acco ding o hei expe judgemen . I is
qui e ob ious ha he aim is o p ope ly iden i y such model
a ibu es o maximise he p edic i e powe o he model.
3. Model applica ions
3.1. In oduc o y ema ks: he ideal case
The main aim o he p esen wo k is o analyse he pe o -
mance o di e en models o applica ion o eal coas al sys-
ems, accoun ing o he a ailable empi ical in o ma ion and
inpu da a and in iew o he benefi om a mul i-modelling
app oach. The e o e, ou scope is somewha mo e han a sim-
ple alida ion s udy.
I can be use ul o summa ise he possible op ions o he
p oblem-sol ing p ocess in iew o en i onmen al modelling
applica ions.
The simples kind o exe cise ha can be pe o med is wha
we can call a “school-boy p oblem”. Candida es a e eques ed
o answe a specific ques ion on he basis o a comple e se o
p elimina y da a and in o ma ion. The answe is uni ocal and
he exe cise has only wo possible ou comes: he answe can
be igh o w ong. A mo e elabo a e exe cise consis s in o -
mula ing a p oblem ha supplies no only he necessa y inpu
da a and in o ma ion bu also some mo e da a ha a e unnec-
essa y o answe he p oblem i sel . I is qui e ob ious ha he
p e ious examples o p oblems a e ypical o academic exe -
cises. The e a e only wo kinds o ac o s playing di e en oles
in hese exe cises: an “omniscien ” e e ee and one o mo e
candida es. I is qui e ob ious ha , in eal ci cums ances, he
si ua ion is mo e complica ed.
In ela ion o complex, na u al sys ems, we can summa ise
a adi ional pe spec i e as ollows: a heo e ical “ideal model”
inhe en o he na u e does exis : he “MODEL”. The aim o sci-
ence is o e eal such a “MODEL”. Due o objec i e di ficul ies,
only app oxima e ealisa ions o such a “MODEL” a e a ailable
al hough hese can be incessan ly imp o ed. The di e en
ealisa ions o he “MODEL” can be empo a ily accep ed o
defini ely ejec ed by “ e ifica ion– alsifica ion” p ocedu es.
I we accep he abo e poin o iew, we should conclude
ha , when pe o ming blind es exe cises o model alida-
ion and compa ison, he modelle s ha e o p o e ha he
pe o mances o he models hey de eloped a e close o he
ideal one by showing ha model esul s fi empi ical da a. The
modelle ha go he esul s mo e close o he empi ical ou -
come wins he game. Mo eo e , in gene al, a “ ule” exis s o
decide whe he he model is “ igh ” o “w ong”. Such a ule
is based on s a is ical es s o significance o asce ain i he
model ou pu a e in ag eemen wi h he empi ical da a when
he unce ain y o he model ou come and o he measu e-
men s a e accoun ed o .
I is ou o he scope o his pape o deba e abou
he exis ence o such an “ideal” model, ne e heless i is
qui e ob ious ha he assessmen o he co ec ness o he
ealisa ion o he “ideal” model is based on se e al assump-
ions: expe imen s can be pe o med and epea ed “ad libi-
um” ( ep oducibili y); unce ain ies o empi ical da a can be
educed a will by epea ed measu emen s (accu acy); i is
possible o pe o m “c ucial expe imen s” ha ing wo pos-
sible ou comes, alse o ue, o assessing he alidi y o
a model ( alsifica ion). The abo e assump ions s aigh o -
wa dly imply ha e e mo e accu a e empi ical da a can
be compa ed wi h e e mo e eliable model ou pu (i is
su ficien ha modelle s and expe imen alis s a e cle e
enough o achie e high quali y esul s in hei espec i e
fields).
F om a heo e ical poin o iew he assessmen o he
unce ain y o he model ou pu has a sound ma hema ical
ounda ion. Acco ding o he mos gene al o mula ion, any
model can be defined as ollows:
L(m1,...,m
n)X=Y(1)
whe e Lis an ope a o ac ing on ec o X( he unknown quan-
i ies ha should be p edic ed and ha depends on ime and
space posi ion x), ec o Yis he se o inpu unc ions and
m1,...,m
na e a se o pa ame e s. Toge he wi h Eq. (1) ini-
ial/bounda y condi ions should be conside ed. Model pa am-
e e alues, ini ial/bounda y condi ions and inpu unc ions
a e ob ained om expe imen al e alua ions co esponding,
espec i ely, o he empi ical cha ac e is ics o he examined
en i onmen al sys em, o he ini ial s a us o en i onmen al
sys em and o he con amina ion and o he “ o cing unc-
ions” co esponding o he pollu ion o he sys em om
ex e nal sou ces. Simila ly, epea ed expe imen s, model es ,
hypo hesis e ifica ions gi e he oppo uni y o deepening ou
knowledge o he complex se o quan i a i e ules and laws
ha unde lie he many p ocesses occu ing in he en i on-
men . This co esponds o he imp o emen o in o ma ion
abou he s uc u e and he pa ame e s o ope a o L.In he
ideal case, ini ial/bounda y condi ions, model pa ame e s and
inpu unc ions a e cha ac e ised by expec ed alues and he
ele an p obabili y dis ibu ions. When hese dis ibu ions
a e known, i is possible o de e mine he a e age alues and
he dis ibu ion o he model ou pu by analy ical ma hema i-
cal echniques o by nume ic p ocedu es such as Mon e Ca lo
ou ines. Mo e op imis ically, epea ed expe imen al ac i i ies
and he imp o emen o measu emen me hods can allow one
o educe he “dispe sion” o he empi ical alues associa ed
wi h he model pa ame e and he ini ial/bounda y condi-
ions.
Thus, o a gi en inpu , he “dispe sion” (unce ain y)
o he ou pu alues can be lowe ed a will. This is he
mos adi ional and eassu ing ecipe o managing model
unce ain y.
3.2. The p ac ical case: applica ions o s a e-o - he-a
models
Un o una ely, we ha e o cope wi h less op imis ic ci cum-
s ances:
(a) he a ailable alues o model pa ame e s (chiefly he
ans e pa ame e s) o many en i onmen al sys ems a e
scan y, consequen ly i is almos impossible o ob ain hei
s a is ical dis ibu ions (i is p e e able o say ha , in gen-
e al, only ew empi ical e alua ions o many pa ame e s
a e a ailable and hese should be conside ed as “ e e ence”
alues);
(b) inpu unc ions can be a ec ed by significan unce ain y
ha canno be igo ously quan ified chiefly in connec ion
wi h he eme gency phase o an acciden ;
(c) he s uc u e i sel o he model ( he ope a o L) is unce -
ain.
T adi ional app oaches o assessing model alidi y a e
no applicable o en i onmen al modelling. Mos knowledge
is o en ob ained by acciden al e en s (like he Che nobyl
inciden ), whe eas labo a o y expe imen s can ne e ep o-
duce he complex en i onmen al p ocesses and si ua ions
occu ing in eal ci cums ances (Pe e s, 1986). I is, in gene al,
impossible o pe o m ad hoc expe imen s o educe unce -
ain ies and o alsi y a model. The e o e, ep oducibili y,
accu acy and alsifica ion p inciples can be di ficul o apply.
Gene ally, di e en modelle s make use o di e en
expe imen al da a and in o ma ion o de elop, o es and
o calib a e hei models. All ha can make i di ficul o
pa ame e ise in a quan i a i e and uni ocal way he key
p ocesses and mechanisms egula ing he alue o a a ge
a iable. In p inciple, i is, he e o e, ha d o selec a unique,
op imal model ( he model closes o he “ideal” one).
Se e al au ho s ha e discussed hese p oblems
(Ho nbe ge and Spea , 1981). Acco ding o he “equifi-
nali y” p inciple, i was claimed ha , gi en a ce ain le el
o p ocess-unde s anding, di e en model s uc u es and
pa ame e alues can be equally accep able in assessing he
beha iou o complex en i onmen al sys ems (Be en and
F ee , 2001). In o he wo ds, we a e usually dealing wi h many
plausible models whose esul s a e expe es ima es en ailing
ce ain commi men based on scien ific knowledge (Giles,
1981).
The e a e se e al o he di ficul ies ha modelle s ace o
he applica ion o en i onmen al models such as agueness,
ambigui y and incomple eness o in o ma ion and inpu da a
ele an o he examined en i onmen al scena io. Inpu da a
demands om modelle s a ely co espond o he inpu da a
o e ed by expe imen alis s. In spi e o all ha , we ha e he
u ge o ge ing he bes om he a ailable in o ma ion o he
mos e ec i e exploi a ion o exis ing knowledge in iew o
p ac ical applica ions.
The e o e, we should conside a complex scena io ha
includes se e al ac o s: fi s ly he “cus ome s” ( hose who
a e o mula ing ques ions o which modelle s should answe );
secondly, he da a/in o ma ion supplie s (“scena io de elop-
e s”); and hi dly, he modelle s. In his espec , he p esen
exe cise is no simply aimed a e alua ing he assessed models
in o de o ank he quali y o hei pe o mances. In his wo k,
we a e conce ned wi h he e alua ion o po en ial ad an ages
o a mul i-modelling app oach o he managemen o a com-
plex en i onmen al p oblem when inpu da a a e insu ficien
and non-uni ocal. In such condi ions, he so-called “expe

judgmen ” is an essen ial ac o . The main ques ion is: can a
mul i-model app oach con ibu e o achie e an expe -based
consensus conce ning he p edic ion o he e olu ion o he
conside ed en i onmen al sys em?
3.3. Desc ip ion o he Dniepe –Bug Es ua y case s udy
The Dniepe –Bug Es ua y (DBE) (Fig. 1) is he la ges o
all he Black Sea Es ua ies (su ace a ea = 1006.3 km2, ol-
ume = 4.24 km3). The DBE wa e sys em consis s o he Dniepe
Es ua y and he Bug Es ua y. The leng h o he DBE is 63 km
wi h a wid h o up o 15 km. The DBE is connec ed wi h he
Black Sea h ough he S ai o Ki nbu n. The a e age dep h
o he DBE is 4.4 m. The e is a na ow, 10–12m deep chan-
nel sui able o shipping along he es ua y o he Black Sea.
The bo om is co e ed mainly by clay (50%) and sand. The
main ac o a ec ing he egime o he sys em is he p o-
cess o mixing esh i e wa e s wi h saline ma ine wa e s.
This o ms he saline wedge in he es ua y, which in he
summe mon hs can each Khe son ci y. S a ifica ion in he
es ua y anges om almos none in he eas e n pa a he
Dniepe mou h o a defined wo-laye sys em in he wes e n
ma ine pa o he DBE. These p ocesses a e highly depen-
den on season. The egime o his d owned- i e es ua y
a ies om s a ified o pa ially mixed. The a e age discha ge
o he Dniepe anges om abou 400 o abou 6000 m3s−1
in sp ing, whe eas a e age discha ge o he Sou he n Bug
anges om 80 o 1000m3s−1. The a e age mean wa e e en-
ion ime in he es ua y can be es ima ed o he o de o 1
mon h. The Dniepe discha ge, unlike he Sou he n Bug, is
no simple seasonal because i is egula ed om Kakho ka
ese oi dam placed a 70km om Dniepe mou h. The e-
o e, in he summe he saline wedge pene a es much u he
in o he es ua y han in he sp ing. Also, he salini y o he
uppe s a um in he summe is much highe han in he
sp ing. In addi ion o he esh wa e inpu , o he key ac-
o s go e ning he anspo o con aminan s a e wind and
Fig. 1 – Ou line o he Dniepe –Bug Es ua y. The boxes
co espond o he sec o s o which empi ical adiological
da a we e a ailable: da a ele an o Wes DBE (Table 5);
da a ele an o Sou he n Bug (Table 3); da a ele an o
Eas DBE (Table 6); da a ele an o Dniepe mou h (Table 2);
da a ele an o Black Sea (Table 4).
sea le el a iabili y. The es ua y is ice co e ed in Janua y o
Feb ua y and wind su ges cause sho - e m excu sions o sal
wedge in o he i e mou hs. All hese ac o s o ce compli-
ca e 3D ime-dependen s a ified flows in DBE (Kos yani syn,
1964).
The ollowing inpu da a we e a ailable o modelle s:
•Ba hyme y o DBE and coas al a ea o he Black Sea we e
p o ided (2 km g id).
•Daily wa e discha ges o Dniepe Ri e and S. Bug Ri e in
1984–1987.
•Sea le el in Kinbou n S ai (Ochaki ) 1986–1987.
•Daily sea le el in Kinbou n S ai (Ochaki ) 1986–1987.
•Daily empe a u e and salini y (pp ) a wa e su ace.
•Th ee-hou alues o wind, ai empe a u e, ela i e humid-
i y and cloudiness (0–10).
•Su ey da a in 1986 o empe a u e and salini y measu e-
men s.
•Su ey da a in 1987 o empe a u e, salini y and eloci y.
O he pa ame e s o DBE hyd ology we e no a ailable:
empe a u e in he Dniepe Ri e and S. Bug Ri e , empe -
a u e and salini y p ofiles in he Kinbou n S ai . In o ma ion
was supplied on hei seasonal changes.
3.4. Radionuclide da a
The Dniepe –Bug Es ua y was con amina ed by adionu-
clides in oduced in he en i onmen ollowing he acciden
occu ed a he Che nobyl nuclea powe plan . The con ami-
na ion was caused by di ec deposi ion on o he es ua y and by
he adioac i e subs ances anspo ed by Ri e Dniepe and,
o a lesse ex en , by Ri e Bug.
Whe eas e y much da a ele an o he mo phome y,
he hyd ology and he me eo ological condi ions o DBE we e
a ailable, adiological inpu da a we e mo e di ficul o find.
Exis ing da a we e ga he ed by wo di e en Uk ainian Ins i-
u es (Ins i u e o Ma hema ical Machines and Sys em P ob-
lems and Uk ainian Ins i u e o Hyd ome eo ology) on he
basis o expe imen al campaigns ca ied ou by se e al lab-
o a o ies (Kani e s e al., 1997; Ka ich e al., 1993). I should
be no iced ha a significan e o o selec ion and e alua ion
o he empi ical da a we e pe o med o assu e he quali y o
hese da a se s.
The inpu da a we e he ime dependen adionuclide con-
cen a ions in Black Sea, in Ri e Dniepe and in sou he n
Bug ou le . Table 2 shows adionuclide concen a ions in Ri e
Dniepe , he main ibu a y o he coas al a ea. Tables 3 and 4
show he concen a ions o adionuclides in Sou he n Bug ou -
le and No h wes Black Sea (inpu da a). Tables 5 and 6 show
da a o measu ed adionuclide concen a ions in wa e o he
Wes and Eas pa o DBE (da a o he assessmen o he
model pe o mances). Abb e ia ions in he ables indica e he
Ins i u es ha collec ed samples and pe o med he adiolog-
ical measu emen s:
•IBSS, Ins i u e o Biology o he Sou he n Seas, Se as opol.
•IEM, Ins i u e o Expe imen al Me eo ology (Typhoon),
Obninsk.
•MHI, Ma ine Hyd ophysical Ins i u e, Se as opol.
Table 2 – Fluxes o 137Cs and 90S om Ri e Dniepe o he es ua y
Da e Da a ga he ed by ins i u e o ma hema ical
machines and sys em p oblems
Da a ga he ed by uk ainian
ins i u e o hyd ome eo ology
137Cs (dissol ed) (Bqm−3)137Cs (suspended) (Bqm−3)90S (Bq m−3)137Cs ( o al) (Bq m−3)90S (Bq m−3)
P e-acciden 3.0 23–33
May 1986 27.8 20.4 18.5 76.0
June 1986 12.2–14.1 14.8 37 4–11/16/7 26–92
July 1986 7.8 12.2 100 7.0 58–61
Augus 1986 7.4 12.2 78
Sep embe 1986 6.7 4.4 56
Oc obe 1986 7.0 4.1 63–89
No embe 1986 7.4 3.7 59–104 10.0 37–55
Decembe 1986 7.8 3.7 89
Janua y 1987
Feb ua y 1987 14.1 48.
Ma ch 1987
Ap il 1987 3–11/5.2 660–470
May 1987
June 1987 5.6–19.2 3.7–6.7 40–278
July 1987
Augus 1987
Sep embe 1987 8.1–14.8 5.2
Oc obe 1987
No embe 1987
Decembe 1987 300–330
Table 3 – Concen a ion o adionuclides in he Sou he n
Bug ou le (da a ga he ed by Uk ainian Ins i u e o
Hyd ome eo ology)
Da e 137Cs (Bq m−3)90S (Bq m−3)
P e-acciden 3.5–7 (NOSS) 7–10.5 (NOSS)
June 1986 16 (UCME)
Oc obe 1986 10 (UCME)
Decembe 1987 3.5 (UCME)
In pa en heses he ac onyms o he Ins i u es ha ca ied ou sam-
pling and measu emen s.
•UCME, Uk ainian Cen e o Ma ine Ecology, Odessa.
•NOSS, Nikolae skaya Oblas Sani a y S a ion, Nikolae .
•CGO, Cen al Geophysical Obse a o y, Kie .
3.5. The exe cise
The main goal o he exe cise is o simula e he esponse
o a mul i-ac o sys em comp ised o modelle s and da a
supplie s in he occasion o an en i onmen al eme gency.
The e o e, he ac o s a e u ged o ca y ou a coope a-
i e e o a he han hampe ed by a conflic ing compe-
i ion.
Modelle s we e asked o supply p edic ions o bo h 137Cs
and 90S in wa e on he basis o he supplied da a and/o
using u he in o ma ion om o he sou ces hey deemed
use ul and us wo hy. Among he inpu da a no supplied
by he “scena io de elope s”, modelle s deemed impo an he
deposi ions o adionuclides o e he es ua y. These da a we e
ob ained om he li e a u e (De Co e al., 1998) o 137Cs.
Da a o deposi ion o 90S o Che nobyl o igin we e es ima ed
on he basis o he ela i e low mobili y in a mosphe e o his
adionuclide compa ed wi h 137Cs (a negligible o e y low 90S
deposi ion was he common hypo hesis o he modelle s).
4. Discussion
As p e iously s a ed many models o p edic ing he mig a-
ion o adionuclides h ough la ge es ua ies equi e he p e-
limina y assessmen o hyd ological cha ac e is ics such as
he wa e cu en field, he wa e empe a u e and salin-
i y p ofiles, e c. Examples o esul s o such kind o models
a e epo ed in Figs. 2 and 3. Compa isons o model ou pu
Table 4 – Concen a ion o adionuclide in he No h-wes Black Sea (da a ga he ed by Uk ainian Ins i u e o
Hyd ome eo ology)
Da e 137Cs (Bq m−3)90S (Bq m−3)
P e-acciden 18 (IEM), 15 (IBSS) 22 (IEM), 18-20 (MHI)
June 1986 55–80 (IEM), 125–185 (IBSS) 26–33 (IEM), 30–130 (IBSS)
Oc obe 1986 110 (UCME)
Ap il 1987 33–140 (IBSS) 19–63 (IBSS)
June 1987 52 (UCME)
In pa en heses he ac onyms o he Ins i u es ha ca ied ou sampling and measu emen s.
Table 5 – Concen a ion o adionuclides in Wes DBE
( alida ion da a; da a ga he ed by Uk ainian Ins i u e o
Hyd ome eo ology)
Da e 137Cs (Bq m−3)90S (Bq m−3)
P e-acciden 3.5–7 (NOSS) 7–10.5 (NOSS), 28±5 (MHI)
June 1986 12–34 (UCME) 26–41 (IBSS)
Oc obe 1986 14–18 (UCME)
Ap il 1987 14–30 (UCME) 220 (IBSS)
Decembe 1987 220 (IBSS)
In pa en heses he ac onyms o he Ins i u es ha ca ied ou sam-
pling and measu emen s.
Table 6 – Concen a ion o adionuclides in he Eas DBE
( alida ion da a; he da a we e ga he ed by Uk ainian
Ins i u e o Hyd ome eo ology)
Da e 137Cs (Bq m−3)90S (Bq m−3)
P e-acciden 3.5–7 (NOSS) 7–10.5 (NOSS), 28±5 (MHI)
June 1986 12–19 (UCME) 26–41 (IBSS)
Oc obe 1986 14–18 (UCME)
Ap il 1987 3–10 (UCME) 400 (IBSS)
Decembe 1987 260 (IBSS)
In pa en heses he ac onyms o he Ins i u es ha ca ied ou sam-
pling and measu emen s.
and empi ical da a o adionuclide concen a ion in wa e a e
epo ed in Figs. 4 and 5.
F om he assessmen o many models (Mon e e al., 2005a)
i seems qui e ob ious ha a main ac o o unce ain y is ep-
esen ed by he di ficul ies o p edic ing quan i a i ely he
complex in e ac ion wi h bo om sedimen o he pollu an
in wa e (con aminan di usion om he wa e column o
sedimen , sedimen a ion and e-mobilisa ion). Ne e heless
in he p esen exe cise, due o he as dynamic o he wa e
wi hin he es ua y and he ela i ely low in e ac ion o con-
side ed adionuclide wi h sedimen in sea wa e , his di ficul y
did no significan ly a ec he model pe o mances. Indeed, in
Fig. 2 – A g aphical ou pu showing he esul s o he
hyd ological model de eloped by he Uni e si y o Se illa.
spi e o he di e en app oaches, pa ame e alues and algo-
i hms used by he models (one o hese has qui e neglec ed
he in e ac ion o adionuclides wi h sedimen s), he e a e
no significan di e ences in model pe o mances ha can be
a ibu ed o he me hodologies employed o p edic ing he
con aminan emo al om he wa e column due o he men-
ioned p ocesses.
The modelle s made use o di e en alues o adionu-
clide deposi ion on o he es ua y. An es ima ed alue o
2000–5000 Bq m−2o 137Cs ini ial deposi ion on o he DBE was
ob ained om g aphical da a epo ed by De Co e al. (1998).
The di e ences among he ou pu o he models a ini ial ime
eflec he unce ain y o he assumed 137Cs ini ial deposi-
ion. Un o una ely, empi ical e alua ions o 90S deposi ion
a e no a ailable. Al hough se e al e idences indica e negligi-
ble con amina ion le els o he en i onmen due o he ini ial
all-ou o 90S in a eas a om he Che nobyl powe plan ,
one o he modelle s used a cau iona y, conse a i e alue o
1000 Bq m−2. This gi es eason o he p edic ed ini ial peak o
90S con amina ion in wa e (Fig. 5).
Fig. 5 shows he esul s om an applica ion o he ENEA
model a a egional scale. The inpu da a o his ” egional
model” was he deposi ion o 90S on o he ca chmen o he
Dniepe sys em a ound he Che nobyl powe plan . The con-
ibu ion o con aminan o he es ua y om Ri e Dniepe
was e alua ed by an applica ion o model MARTE (see he
ENEA model desc ip ion in Appendix A) o he whole basin o
he i e . In spi e o he la ge ho izon o his pa icula model
applica ion, he p edic ions o he ime beha iou o adionu-
clide concen a ion in he es ua y we e wi hin he ange o
he esul s ob ained by he o he models. This is a u he e i-
dence ha he anspo o pollu an s cha ac e ised by weak
adso p ion on bo om sedimen can be p edic ed wi h a su fi-
cien accu acy al hough mig a ion occu s o e la ge dis ances.
An impo an ques ion is whe he o no all he app oaches
used can be in eg a ed in a single model. Al hough his is
conside ed ca ego ical in many adi ional fields, in p inci-
ple, i is no so ob ious in case o en i onmen al models. The
pe spec i es, he ho izons, he ea u es, he s uc u es o he
models a e o en no ully consis en . The in eg a ion o hese
Fig. 3 – Tempe a u e and salini y along DBE sec ions in
summe simula ed by model THREETOX.
Fig. 4 – Compa ison o he esul s o he model ou pu wi h empi ical da a o 137Cs concen a ion in wa e .
di e en models in a comp ehensi e “supe -model” can be,
he e o e, ha d o achie e.
The examined models make use o significan ly di e en
alues o simila pa ame e s (Table 7). As su ficien in o ma-
ion on he s a is ical dis ibu ions o he empi ical alues o
hese pa ame e s a e no a ailable, he esul s o he models
canno be amed in a p obabilis ic pe spec i e o ins ance
in e ms o mean alues and confidence le els. As a ma e
o ac , i seems mo e p ope o s a e ha he ou pu o each
model is one o he possible ou comes ha he communi y o
expe s deems wo hy o commi men .
No hing is mo e ins uc i e han a compa a i e g aph, like
Figs. 4 and 5, o show he in o ma ion a ailable o he “cus-
ome s”.
F om he figu es i is qui e clea ha , despi e he abo e-
men ioned di ficul ies, he e is an appa en consensus among
modelle s conce ning he ime beha iou o he adionuclide
in wa e . Figu es make in ui i e wha in o ma ion ob ained
om he models can be pe cei ed wo hy o consensus and in
which measu e model ou pu should be conside ed illus a i e
o he empi ical ou comes. Fo ins ance, a delay o adionu-
clide concen a ion peak in wa e is p edic ed o 90S , whe eas
o 137Cs he concen a ion in wa e shows a clea decline on
ime.
The ange o a iabili y o model ou pu is compa able wi h
he ange o a iabili y o he empi ical concen a ions. This
clea ly confi m ha model pe o mances eflec he in in-
sic unce ain y o knowledge conce ning he quan i a i e
beha iou o he in ol ed en i onmen al p ocess, he ambigu-
i y o in e p e a ion and pa ame e isa ion o such p ocesses,
he inhe en a iabili y o en i onmen al quan i ies, e c.
When he exe cise was ca ied ou , da a o adionuclide
concen a ion in he es ua y compa men s we e a ailable
om he scien ific li e a u e. This occu ence made impos-
sible a blind alida ion o he assessed models. Consequen ly,
modelle s we e asked o un hei models a oiding p elimi-
na y calib a ion. I is e y di ficul o app ecia e how much he
a ailabili y o con amina ion da a influenced he model appli-
Table 7 – Selec ion o pa ame e alues used by he models
Pa ame e Model Value (m2s−1)
Ve ical mixing coe ficien (su ace-deep wa e s) Coas Mab 10−6(o de o magni ude)
ENEA 1.9 ×10−7–1.9 ×10−6(depend-
ing on he dis ance om he
ibu a y i e )
Ho izon al di usion coe ficien U. Se illa 0.58
Mixing coe ficien sea wa e -coas al wa e s Coas Mab 23
ENEA 116
Ha ms, I.H., 1997. Modelling he dispe sion o 137Cs and 239Pu
eleased om dumped was e in he Ka a Sea. J. Ma ine
Sys . 13, 1–19.
Ho nbe ge , G.M., Spea , R.C., 1981. An app oach o he
p elimina y analysis o en i onmen al sys ems. J. En i on.
Manage. 12, 7–18.
IAEA, 2000. Modelling o he ans e o adiocaesium om
deposi ion o lake ecosys ems. Repo o he VAMP Aqua ic
Wo king G ou IAEA-TECDOC-1143.
Vienna.
Jimenez-Mon ealeg e, R., Ve degem, M.C.J., an Dam, A.,
Ve e h, J.A.J., 2002. Concep ualisa ion and alida ion o a
dynamic model o he simula ion o ni ogen
ans o ma ions and fluxes in fish ponds. Ecol. Model. 147,
123–152.
Jø gensen, S.E., 1979. Modelling he dis ibu ion and e ec o
oxic subs ances. Ecol. Model. 6, 181.
Kani e s, V.V., Voi sekho ich, O.V., Simo , 1997. Radioac i e
con amina ion o he Black and Azo Seas. In:
Voi sekho ich, O.V. (Ed.), Radioecology o Wa e Objec s o
he Che nobyl NPP Acciden Impac A ea.
Che nobylin e in o m, Kie , pp. 127–151 (in Russian).
Ka ich, I.Yu., Niki in, A.I., Medine s, V.I., Lepeshkin, V.I.,
Kabano , A.I., Semko, N.N., Bazhano , V.N., 1993. Dynamics
o he adioac i e con amina ion caused by CNPP acciden
on obse ed da a 1986–1990. In: Bo zilo , V.A., K yshe , I.I.
(Eds.), Ecological and Hyd ophysical Consequences o he
Nuclea Acciden s. Hyd ome e ological Publ, Moscow, pp.
57–61 (in Russian).
Koshebu sky, V., Made ich, V., Nes e o , O., Heling, R., 2004.
Modelling o hea dispe sion in inland wa e s and coas al
a eas o sea. Appl. Hyd omech. 6 (78), No. 4 (in Russian).
Kos yani syn, M.N., 1964. Hyd ology o he Dniepe and
Sou he n Bug Es ua y. M. Gid ome eoizda , 336 p. (in
Russian).
Koziy, L., Made ich, V., Ma g elash ili, N., Zheleznyak, M., 1998.
Th ee-dimensional model o adionuclide dispe sion in he
es ua ies and shel seas. J. En i on. Model. So wa e 13
(5–6), 413–420.
Koziy, L., Made ich, V., Ma g elash ili, N., Zheleznyak, M., 2000.
Nume ical modelling o seasonal dynamics and adionuclide
anspo in he Ka a Sea. In: Oceanic on s and Rela ed
Phenomena. Kons an in Fedo o In . Memo ial Symp., IOC
Wo kshop Rep. Se ies, N 159, UNESCO’2000, pp. 296–
301.
K one R.B., 1962. Flume s udies o he anspo o sedimen in
es ua ial p ocesses. Final Repo . Hyd aulic Enginee ing
Labo a o y and Sani a y Enginee ing Resea ch Labo a o y,
Uni e si y o Cali o nia, Be keley, 120 p.
K yshe , I.I., Sazykina, T.G., 1995. Radiological consequences o
adioac i e con amina ion o he Ka a and Ba en s Seas. J.
En i on. Radioac . 29, 213–223.
Ma g elash ili, N., Made ich, V., Zheleznyak, M., 1997.
THREETOX: a compu e code o simula e h ee-dimensional
dispe sion o adionuclides in s a ified wa e bodies.
Radia . P o ec . Dosime . 73, 177–180.
Ma g elash ili, N., Made ich, V., Zheleznyak, M., 1999.
Simula ion o adionuclide fluxes om he Dniepe –Bug
Es ua y in o he Black Sea. J. En i on. Radioac . 43, 157–
171.
Ma g elash ili, N., Made ich, V., Yuschenko, S., Zheleznyak, M.,
2002. 3D modelling o he mud and adionuclide anspo
in Che nobyl cooling pond and Dniepe –Bug Es ua y. In:
Win e we p, J.C., K anenbu g, C. (Eds.), Fine Sedimen s
Dynamics in he Ma ine En i onmen P oceedings o
INTERCOH-2000. Else ie , pp. 595–
610.
Ma sunaga, T., Nagao, S., Ueno, T., Takeda, S., Amano, Hika u,
Tkachenko, Y., 2004. Associa ion o dissol ed adionuclides
eleased by he Che nobyl acciden wi h colloidal ma e ials
in su ace wa e . Appl. Geochem. 19, 1581–1599.
Mon e, L., Sal a o i, L., Spezzano, P., 1995. Modelling
adionuclide beha iou in deep lakes o he I alian Alpine
egion: seasonali y e ec s and compa ison wi h deep
olcanic lakes o Cen al I aly. Technical epo ENEA,
RT/AMB/95/12, Rome.
Mon e, L., 2001. A gene ic model o assessing he e ec s o
coun e measu es o educe he adionuclide con amina ion
le els in abio ic componen s o esh wa e sys ems and
complex ca chmen s. En i on. Model. So wa e 16, 669–
690.
Mon e, L., B i ain, J.E., H˚
akanson, L., Heling, R., Smi h, J.T.,
Zheleznyak, M., 2003. Re iew and assessmen o models
used o p edic he a e o adionuclides in lakes. J. En i on.
Radioac . 69, 177–205.
Mon e, L., B i ain, J.E., H˚
akanson, L., Smi h, T.J., an de Pe k,
M., 2004. Re iew and assessmen o models o p edic ing
he mig a ion o adionuclides om ca chmen s. J. En i on.
Radioac . 75, 83–103.
Mon e, L., Boye , P., B i ain, J.E., H˚
akanson, L., Lepica d, S.,
Smi h, J.T., 2005a. Re iew and assessmen o models o
p edic ing he mig a ion o adionuclides h ough i e s. J.
En i on. Radioac . 79, 273–296.
Mon e, L., Deside i, D., G imani, C., Angeli, G., 2005b. Modelling
he long- e m beha iou o adiocaesium and
adios on ium in wo I alian lakes. J. En i on. Radioac . 80,
105–123.
Mossman, D.J., AlMulki, N., 1996. One-dimensional uns eady
flow and uns eady pes icide anspo in a ese oi . Ecol.
Model. 89, 259–267.
Pa heniades, E., 1965. E osion and deposi ion o cohesi e soil.
J. Hyd . Di . ASCE. 91, 105–139.
Pe ia ˜
nez, R., 2002. Modelling he suspended ma e dynamics
in a ma ine en i onmen using a h ee-dimensional s
coo dina e models: applica ion o he eas e n I ish Sea.
Appl. Ma h. Model. 26, 583–601.
Pe ia ˜
nez, R., 2004. Tes ing he beha iou o di e en kine ic
models o up- ake/ elease o adionuclides be ween wa e
and sedimen s when implemen ed in a ma ine dispe sion
model. J. En i on. Radioac . 71, 243–
259.
Pe ia ˜
nez, R., Ab il, J.M., Ga c´
ıa-Le´
on, M., 1996. Modelling he
dispe sion o non-conse a i e adionuclides in idal wa e s.
Pa 1: concep ual and ma hema ical model. J. En i on.
Radioac . 31, 127–141.
Pe e s, R.H., 1986. The ole o p edic ion in limnology. Limnol.
Oceanog . 31, 1143–1159.
Polika po , G.G., Timoschuk, V.I., Kulebakina, L.G., 1988.
Concen a ion o 90S in he aqua ic en i onmen o lowe
Dniepe owa d he Black Sea. In: Dopo idi (P oceedings) o
Na ional Academy o Sciences o Uk aine, se . B, N3, pp.
75–76.
Polika po , G.G., Li ings on, H.D., Kulebakina, L.G., Buessele ,
K.O., S okozo , N.A., Casso, S.A., 1992. Inflow o Che nobyl
90S o he Black Sea om he Dniep Ri e . J. Es ua ine,
Coas al Shel Sci., V 34, 315–320.
P andle, D., 1984. A Modellino S udy o he Mixing o 137Cs in
he seas o he Eu opean con inen al shel , ol. A310.
Philosophical T ansac ions o he Royal Socie y, London, pp.
407–436.
Raja , R., Zaga , D., Ce ina, M., Akagi, H., Yano, S., Tomiyasu, T.,
Ho a , M., 2004. Applica ion o h ee-dimensional me cu y
cycling model o coas al seas. Ecol. Model. 171, 139–
155.
Salomon, J.C., B e on, M., Gueguenia , P., 1995. A 2D long e m
ad ec ion dispe sion model o he Channel and sou he n
No h Sea. Pa B. T ansi ime and ans e unc ion om
Cap de La Hague. J. Ma ine Sys . 6, 515–527.

Schon eld, W., 1995. Nume ical simula ion o he dispe sion o
a ificial adionuclides in he English Channel and he
No h Sea. J. Ma ine Sys . 6, 529–544.
Smi h, C.N., Goshawk, J.A., Cha les, K., McDonald, P., Leona d,
K., McCubbin, D., 2003. Mead (Pa 2), p edic ions o
adioac i i y concen a ions in he I ish Sea. J. En i on.
Radioac . 68, 193–241.
Smi h, J.T., Comans, R.N.J., 1996. Modelling he di usi e
anspo and emobilisa ion o 137Cs in sedimen s: he
e ec s o so p ion kine ics and e e sibili y. Geochimica e
Cosmochimica Ac a 60, 995–1004.
Van Rijn, L.C., 1984. Sedimen anspo . Pa II. Suspended
load anspo . J. Hyd. Eng . 110, 1613–1641.