Re iew
Sus ainable U ban D ainage Sys em (SUDS) modeling suppo ing
decision-making: A sys ema ic quan i a i e e iew
Pascual Fe ans
a,d,
⁎, Ma ía N. To es
b,c
, Ja ie Temp ano
a
, Juan Pablo Rod íguez Sánchez
e
a
Depa amen o de Ciencias y Técnicas del Agua y del Medio Ambien e, Uni e sidad de Can ab ia, Spain
b
Depa men o Ci il, S uc u al and En i onmen al Enginee ing, Uni e si y o Bu alo, USA
c
RENEW Ins i u e, Uni e si y o Bu alo, USA
d
Escuela de Ingenie ía de Bilbao, Uni e sidad del País Vasco UPV/EHU, Spain
e
Cen o de In es igaciones en Ingenie ía Ambien al (CIIA), Uni e sidad de Los Andes, Bogo á, Colombia
HIGHLIGHTS
•This pape aims o quan i a i ely ana-
lyze how SUDS-DSS a e being build
and applied.
•The ques ion was “Wha is he ole o
SUDS models on he decision-making
p ocess?”
•Da abase and snowballing sea ches
me hodswe eused oapp aise he
pape s.
•The esea ch ocus has shi ed om sim-
ple ep esen a ions o mo e sophis i-
ca ed ools.
•The e a e some aspec s ha e-
qui e special a en ion o SUDS-
DSS de elopmen .
GRAPHICAL ABSTRACT
abs ac a icle in o
A icle his o y:
Recei ed 6 July 2021
Recei ed in e ised o m 15 Sep embe 2021
Accep ed 15 Sep embe 2021
A ailable online 25 Sep embe 2021
Edi o : Fe nando A.L. Pacheco
Decision Suppo Sys ems (DSS) o Sus ainable U ban D ainage Sys ems (SUDS) a e a aluable aid o SUDS
widesp ead adop ion. These ools sys ema ize he decision-making c i e ia and elimina e he bias inhe en o ex-
pe judgmen , ab idging he echnical aspec o SUDS o non- echnical use s and decision-make s. Th ough he
collec ion and ca e ul assessmen o 120 pape s on SUDS models and SUDS-DSS, his e iew shows how hese
ools a e buil , selec ed, and used o assis decision-make s ques ions. The manusc ip classifies he DSS based
on he ques ion hey assis in answe ing, he spa ial scale used, he so wa e selec ed, among o he aspec s.
SUDS-DSS aspec s ha equi e mo e a en ion a e iden ified, including en i onmen al and social conside a ions,
SUDS ains pe o mance and c i e ia o selec ion, s ochas ici y o ain all, and u u e scena ios impac . Sugges-
ions o SUDS-DSS a e finally o e ed o be e equip decision-make s in acing eme ging s o mwa e challenges
in u ban cen e s.
© 2021 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Keywo ds:
U ban d ainage modeling
Wa e Sensi i e U ban Design (WSUD)
Op imized SUDS design
G een in as uc u e (GI)
Decision suppo ools
Sponge ci ies
Science o he To al En i onmen 806 (2022) 150447
⁎Co esponding au ho a : Escuela de Ingenie ía de Bilbao, Uni e sidad del País Vasco, Spain.
E-mail add esses: p e [email protected] (P. Fe ans), m o esc@bu alo.edu (M.N. To es), ja ie . emp [email protected] (J. Temp ano), pabl- [email protected]
(J.P. Rod íguez Sánchez).
h ps://doi.o g/10.1016/j.sci o en .2021.150447
0048-9697/© 2021 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Con en s lis s a ailable a ScienceDi ec
Science o he To al En i onmen
jou nal homepage: www.else ie .com/loca e/sci o en
Con en s
1. In oduc ion................................................................ 2
2. P e ious e iews.............................................................. 3
2.1. SUDSmodeling........................................................... 3
2.2. SUDS-DSS............................................................. 3
3. Me hodology............................................................... 4
4. Resul sanddiscussion........................................................... 4
4.1. Ca ego iza ionandhis o icalo e iew ................................................ 4
4.2. Geog aphicalsp ead......................................................... 5
4.3. Cases udyscale, ypologies,spa ialand empo almodeling esolu ion.................................. 6
4.4. Modelingme hodologies:so wa eande en sselec ion ........................................ 8
4.5. Ca ego ies,p ocessesmodeled,ands akeholde s............................................ 8
4.6. DSSinpu s,ou pu sandgene al amewo k.............................................. 8
4.6.1. Inpu sandou pu s..................................................... 8
4.6.2. SUDS-DSS amewo k................................................... 10
4.6.3. SUDS-DSSin ol ingop imiza ion.............................................. 10
5. Conclusionsandpe spec i es....................................................... 11
Disclosu es a emen ............................................................. 12
Funding................................................................... 12
Decla a iono compe ingin e es ........................................................ 12
AppendixA. Supplemen a yda a...................................................... 12
Re e ences.................................................................. 12
1. In oduc ion
Sus ainable U ban D ainage Sys ems (SUDS) a e an in eg a ed ne -
wo k o enginee ed ege a ed a eas and open spaces (i.e., g een oo s,
ain ga dens, po ous pa emen s, e c.) used o p o ec na u al ecosys em
p inciples and unc ions and o o e a wide a ie y o benefi s o people
and wildli e (Tang e al., 2021). SUDS a e a complemen o cen alized
con en ional sewe sys ems in as uc u e o minimize he hyd ological
u baniza ion impac s and inc ease esilience o ex eme ain all e en s
in u ban cen e s (Zhu e al., 2019). These s uc u es ha e he abili y o
a enua e ex eme ain all e en s (Tang e al., 2021) and a e known
o p o iding mul iple en i onmen al benefi s (Liao e al., 2013), includ-
ing clima e change impac s educ ion (Cou s and Hahn, 2015;Jones
and Sompe , 2014;Ghodsi e al., 2020;Rosebo o e al., 2021), along
wi h ecological and social benefi s and o he po en ial mone izable ben-
efi s in he long e m (Wol , 2003;Hamann e al., 2020).
SUDS a e usually e e ed o using se e al o he e ms, including
Bes Managemen P ac ices (BMPs), G een In as uc u e (GI) (Benedic
e al., 2006), blue-g een sys ems (Bozo ic e al., 2017), Low Impac
De elopmen (LID), sou ce con ol (Hamel e al., 2013), sponge ci y (Xia
e al., 2017), na u e-based solu ions (Kabisch e al., 2016;O al e al.,
2020), and Wa e Sensi i e U ban Design (WSUD) (Wong, 2006),
among o he s (Fle che e al., 2015;Cha zimen o e al., 2020). This se
o e ms is no s a ic since, as desc ibed in Fle che e al. (2015), hey
espond o e ol ing echnologies and he inco po a ion o o he fields
in o he u ban d ainage p ac ice, and ha e (some sub le, o he s d as ic)
di e ences in scope and p inciples. Fo he pu pose o his e iew,
e ms e e ing o ‘na u e-based s o mwa e managemen solu ions’
will be unified unde ‘SUDS’, and he di e ence in he scope hey encom-
pass will be o e looked.
SUDS selec ion, design, and loca ion is a high-le el complexi y p ob-
lem ha elies on ools ha sys ema ically in oduce ele an in o ma-
ion, usually based on he bes a ailable ep esen a ion o he u ban
d ainage sys em. In such a complex endea o , modeling is necessa y
o p edic he beha iou o SUDS configu a ions ( ype, design, and loca-
ion) and app aise hei impac in he u ban sys em. Modeling, along
wi h o he ools like mul i-c i e ia ma ices and op imiza ion ools a e
pu oge he on amewo ks o aid SUDS decision-making equen ly e-
e ed o as Decision Suppo Sys ems (DSS).
The applicabili y o DSS in di e se fields esul s on nume ous defini-
ions, which al hough well-es ablished in a pa icula niche, a e con-
used when applied o an in e disciplina y field. P o ided ha SUDS
a e pa o bo h he u ban and en i onmen al sys ems, he e m DSS
ends o be used in e changeably wi h o he s such as En i onmen al
Decision Suppo Sys ems (EDSS) (Poch e al., 2004;Ma hies e al.,
2007;Reiche e al., 2015) and Planning Suppo Sys ems (PSS)
(Klos e man, 1997). Fo a discussion o he usage o hese e ms, e e
o Kapelan e al. (2005) and Te B ömmels oe (2013).
In he field o SUDS, he abo e-men ioned e ms a e ha dly sepa a-
ble since SUDS-DSS can be classified in o bo h highly complex sys ems
(EDSS - as defined in Poch e al. (2004)) and planning-ac ions- ela ed
(PSS). Building upon he DSS defini ion p o ided by Fox and Das
(2000), in his s udy he e m SUDS-DSS is used o e e o “an s uc u ed
se o ools (e.g., op imiza ion, a ificial in elligence, nume ical models,
s a is ical me hods, Geog aphical In o ma ion Sys ems (GIS)) o assis deci-
sion make s and p o ide ecommenda ions on SUDS design and spa ial
deploymen ”.
DSS is a aluable aid o SUDS widesp ead adop ion. They sys ema-
ize he decision-making c i e ia and elimina e he bias inhe en o
expe judgmen . By making SUDS decision-making less echnical,
SUDS-DSS encou age hei adop ion and inc ease hei impac a he
local, egional, and global scale (Bap is a e al., 2005). A ailable SUDS-
DSS gene ally aim o sol e p oblems o wo na u es 1) SUDS design
(p elimina y o de ailed) and 2) SUDS spa ial loca ion (selec ion and
placemen ), seeking he bes SUDS implemen a ion scena ios in e ms
o , a leas , wa e quan i y/quali y and ha ing ideal ma gins o cos -
benefi (Vei h e al., 2003).
Because he p ima y SUDS objec i e is he a enua ion o he hyd o-
logical cycle dis u bances, SUDS-DSS ely on he bes a ailable hyd o-
logical ep esen a ion o he s udy a ea and he SUDS s uc u e. U ban
D ainage Models (UDM) wi h SUDS modeling capabili ies a e com-
monly used o his pu pose, (K ebs e al., 2013;Kong e al., 2017).
While mechanis ic app oaches a e gene ally p e e ed o hyd ological
ep esen a ions, he e a e o he simplified app oaches used in SUDS-
DSS, chosen o p ac ical easons (e.g., as con e gence o s aigh o -
wa d coupling wi h o he DSS modules). Calib a ed UDMs a e ideal o
building a obus and eliable SUDS-DSS (Ha is e al., 2016;Beck e al.,
2017;Ellis, 2013;I fland e al., 2021) bu hey a e no always a ailable.
Fu he mo e, u u e scena io p ojec ions conside ing u baniza ion
ends and clima e change a e desi able bu no always included
(Wang e al., 2020).
Some SUDS-DSS emphasize he exploi a ion o SUDS en i onmen al
benefi s, aiming o find he bes configu a ion o maximize one o mo e
objec i es. Also, despi e s akeholde s' ele ancy, hese ac o s a e
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
2
seldom included in he ea ly s eps o he decision-making p ocess,
esul ing in SUDS designs and loca ions ha do no adjus o he expec-
a ions o hose who will benefi om he s uc u e (Raei e al., 2019).
Many o hese limi a ions esul om he eac i e app oach in which
SUDS-DSS a e concei ed and buil . When a DSS-SUDS esponds o he
pa icula i ies o he case s udy and i s empo al necessi ies, he
esul ing DSS 1) ails in cap u ing a holis ic and unbiased pe spec i e
and 2) i s applicabili y is cons ained o he case s udy ha mo i a ed
i s de elopmen (i.e., To es e al. (2016);Kulle e al. (2017)).
In a apidly ( egionally- ocused) e ol ing field like he SUDS-DSS, i
is necessa y o make pe iodical assessmen s o he s a e-o -a o in e -
na ionalize egional expe iences and b ing o wa d new pe spec i es
o he de elopmen and use o SUDS-DSS. This pape p esen s a quan-
i a i e and c i ical discussion on how modeling-based SUDS-DSS a e
being used o suppo decision-making. Th ough a wo-yea -long
(2019-[Ma ch]2021) app aisal o a icles in oducing o applying
u ban d ainage models o assis SUDS- ela ed planning ac ions, he
cu en s a e o he a was quan i a i ely e alua ed wi h he key objec-
i es o :
•Analyze and upda e he s a e o he a and la es - ends in modeling-
based SUDS-DSS esea ch.
•Quan i y and map he de elopmen and implemen a ion o modeling-
based SUDS-DSS.
•Unde s and he (modeling) p ac ices employed when using SUDS-
DSS.
The nex sec ions o his a icle a e s uc u ed as ollows. Sec ion 2
summa ises p e ious e iews. Sec ion 3 desc ibes each s ep o he sys-
ema ic quan i a i e e iew. Sec ion 4 is di ided in o six subsec ions
ha epo he quan i a i e esul s and discuss implica ions. Finally,
Sec ion 5 p o ides a c i ical pe spec i e and sugges s u u e esea ch di-
ec ions.
2. P e ious e iews
F om a dedica ed e ision o p e ious e iews, i was e idenced how
he SUDS esea ch conce ns and di ec ions ha e been shaped by he de-
elopmen o in e disciplina y esea ch, he b oadening o SUDS unde -
s anding as p o ide s o mul iple benefi s, and he la e inclusion o
u ban planning in he s o mwa e managemen field (Kulle e al.,
2017). Among he many e iews a ailable, hose ackling specifically
SUDS modeling o SUDS-DSS and published in he las wo decades
(1997–[Ma ch] 2021) we e conside ed.
2.1. SUDS modeling
Fi s u ban s o mwa e models lacked he abili y o model SUDS (see
Bu on e al. (2001) and Zoppou (2001) o a e iew). SUDS modeling
gained sha pe a en ion a e hei epo ed success in managing un-
o . In 2007, Ellio and T owsdale (2007) p esen ed he fi s SUDS
modeling e iew and p oposed a classifica ion based on hei pu pose:
planning, p elimina y o de ailed design. Thei e iew poin ed ou he
impo ance o empo al and spa ial esolu ion (pa icula ly he limi ed
abili y o s o mwa e models o p edic he flow a es om small ca ch-
men s), uno gene a ion, and pollu an s anspo modeling. Two o
hei main findings we e ha i) only hal o he SUDS models had a
g oundwa e /baseflow componen , and ii) he e was a deficiency o
ools ha ope a e e ec i ely a a la ge spa ial scales.
Ahiablame e al. (2012) p o ided a de ailed e iew o SUDS ep e-
sen a ion in compu a ional me hods. By ocusing on 4 SUDS ypes,
hey iden ified wo modeling app oaches: a p ocess ep esen a ion
(e.g., infil a ion, sedimen a ion, se ling) and a p ac ice ep esen a ion,
which uses an agg ega ion me hod o model he p ac ice as a uni . They
iden ified as a eas o u u e esea ch he scaling o SUDS p ac ices om
lo o wa e sheds and egional scales.
The u he efinemen o SUDS models' physical p ocesses
ep esen a ion (Kaykhos a i e al., 2018) was impulsed by he widening
o SUDS models' usage o u ban planning and decision-making.
Kaykhos a i e al. (2018) compa ed s o mwa e models' capabili ies
o ep esen ing he hyd ological and hyd aulic SUDS p ocesses and
poin ed ou he need o de elop mo e comp ehensi e SUDS models
allowing a ious applica ions (i.e., esea ch, concep ual, p elimina y
and de ailed design, and ope a ional suppo ).
Wi h Ahiablame e al. (2012), i was e idenced ha esea ch g ew
on SUDS models' applicabili y o u ban planning and policy-making.
The au ho s p omo e he de elopmen o easy- o-use SUDS-DSS ha e -
ec i ely suppo decision make s and in ol e s akeholde s, egula o s,
and policy-make s. As discussed p e iously, he a ac i eness o SUDS
o s o mwa e managemen is hei abili y o p o ide en i onmen al
benefi s beyond he hyd ological dimension (Capa ós-Ma ínez e al.,
2020). Consequen ly, many SUDS-DSS a e de eloped o help decision-
make s inco po a e addi ional c i e ia o placemen and design.
The e a e in o ma i e e iews ha ackled a b oade pe spec i e o
mode n s o mwa e managemen and UDMs. Fo example, Sal ado e
e al. (2015) s a ed ha many modeling app oaches a ge specificob-
jec i es and ha he le el o de ail in ep esen ing physical p ocesses
is no consis en . O he examples a e he wo ks by Bach e al. (2014)
and Ma uhah e al. (2018), who ocused on in eg a ed u ban wa e sys-
ems modeling. While Bach e al. (2014) classified in eg a ed UDMs a
one o ou deg ees o in eg a ion, Ma uhah e al. (2018) pe o med a
classifica ion conside ing social aspec s, ins i u ional dynamics, echni-
cal inno a ion, and local con ex s. These bigge -pic u e e iews ocused
on d ainage sys ems in eg a ion and in e ac ion wi h o he u ban sys-
ems a he han ocusing exclusi ely on SUDS.
2.2. SUDS-DSS
Le e e al. (2015) classified he SUDS-DSS acco ding o he ques ion
i assis s in answe ing: “How Much”,“Whe e”,and“Which”.To es e al.
(2016) ocused on he geog aphical dis ibu ion o SUDS-DSS and he
s o mwa e dimensions conside ed (e.g., quan i y, quali y, ecosys em
se ices). Bo h e iews ound case s udy specifici y and lack o flexibili y
we e d awbacks o mos SUDS-DSS. On he o he hand, Zhang and Chui
(2018) e iewed SUDS-DSS o spa ial decision-making and concluded
ha i s gene ic s uc u e couples a de ailed UDM and an op imiza ion
ool, which communica e i e a i ely un il a s op c i e ion is me .
O he e iews on SUDS-DSS include he wo ks by Zhou (2014) and
Jayasoo iya e al. (2020).Zhou (2014) made a compa ison o modeling
app oaches and decision-aid ools o assessing SUDS al e na i es. The
au ho classified DSS in o ypes o assessmen ools: i) Economic, ii) So-
cial, iii) En i onmen al, i ) Li e-Cycle Assessmen , and ) Heal h. Addi-
ionally, Zhou (2014) highligh ed he impo ance o clima e change
and u baniza ion impac s in SUDS design, and s a ed ha he u u e o
he field a e solu ions ha pu sue a balance be ween he cos o in es -
men and e ficien pe o mance (Zhou, 2014).
Mo e ecen ly, Jayasoo iya e al. (2020) e isi ed he impo ance o
balancing en i onmen al and economic goals and showed ha despi e
many s udies ha e ecognized s akeholde s' in ol emen impo ance,
none ha e ex ensi ely s udied he ele ancy o hei pa icipa ion.
Finally, he au ho s lis ed SUDS implemen a ion ba ie s, including
land owne ship and lack o in e es in nego ia ing land a eas o SUDS
placemen .
A seminal e iew ha showed he impo ance o SUDS as pa o he
u ban o m is he wo k by Kulle e al. (2017). The au ho s p oposed
ha SUDS loca ion should no be conside ed a one-way p ocess, bu
a he a wo-sided p oblem. By de ending ha “WSUD (SUDS) needs a
place as much as a place needs a WSUD”, hey p oposed he fi s -o -a-
kind sui abili y amewo k o SUDS planning. Kulle e al. (2017)
wen beyond in classi ying PDSS in o hei app oach owa ds SUDS, as
a pa o (a) he u ban wa e cycle, (b) he u ban o m, and (c) he
wa e go e nance.
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
3
Ellio and T owsdale (2007),Ahiablame e al. (2012) and
Kaykhos a i e al. (2018) ex ensi ely explo ed key aspec s o SUDS
modeling while Zhou (2014),Le e e al. (2015),To es e al. (2016),
Zhang and Chui (2018),Jayasoo iya e al. (2020) and Kulle e al.
(2017) ocused on SUDS-DSS axonomy and good p ac ices o SUDS-
DSS de elopmen . This e iew does no a emp o co e in de ail opics
al eady discussed in p e ious e iews, bu o build upon hese ecom-
menda ions o quan i a i ely analyze how modeling-based SUDS-DSS
a e being build and applied. Fo example, wha ques ions a e mo e e-
quen ly being answe ed wi h SUDS-DSS? How models a e being used in
p ac ice (scale o he cases o s udy, ime s eps, modeling windows, cal-
ib a ion p ocedu es, e c.). How SUDS-DSS de elopmen and usage a e
sp ead geog aphically?
3. Me hodology
A sys ema ic quan i a i e li e a u e e iew loca es, app aises, and
syn hesizes e idence o a specific issue limi ing bias by deciding specific
c i e ia o include and exclude s udies (Pe ic ew, 2001). The wo mos
widely used echniques o sys ema ically collec publica ions we e
used: da abase and snowballing sea ches (Badampudi e al., 2015). In
he fi s , a combina ion o keywo ds was used o sea ch in di e en da-
abases (Scopus, Web o Science -WOS, and Google Schola ); and in he
la e , new pe inen pape s we e iden ified h ough he e e ence lis
(backwa d) and ci a ions ( o wa d) o a seed-se o influen ial pape s
(Jalali and Wohlin, 2012;Fon echa e al., 2021).
The e iew ques ion add essed in his s udy was “Wha is he ole o
SUDS models on he decision-making p ocess?”The objec i es we e o i)
unde s and which SUDS models/so wa e a e mo e equen ly used and
how hey a e deployed o decision-making, ii) de e mine which ques-
ions he SUDS-DSS assis in answe ing (e.g., Which SUDS? Whe e?
How many?), iii) analyze he DSS capabili ies (e.g., op imiza ion, s ake-
holde s inclusion, unce ain y analysis).
Table 1 lis s he sea ch e ms used in he da abases. Pape s whose i le
ha e a leas h ee wo ds in di e en keywo d se s we e included in he
e iew. In his way, he inclusion o he key eligibili y c i e ia was gua an-
eed: “decision/ ool”,“SUDS”,“modeling”,and“s o mwa e ”. The bes
e o was ca ied ou o include a comp enhensi e se o key sea ch
e ms o he pape s app aisal, bu i is no gua an ee ha all e ms
ha e been included gi en he p oli e a ion and ola ili y o local e minol-
ogy. Simila ly, i is acknowledged ha much o he li e a u e on SUDS-DSS
applica ions is w i en in languages di e en o English, lea ing ou o he
e iew applica ions o non-English-speaking coun ies.
Once a fi s po en ial seed-se o significan pape s was ga he ed, he
inclusion c i e ia o he sea ch we e ha he pape mus be i) use ul o
decision-making (i.e., a ool o case s udy, no a amewo k, e iew, o
expe ience epo ), ii) specific o s o mwa e (al hough o he u ban
wa e cycle elemen s may be p esen ), and iii) include SUDS modeling.
All pape s ha answe he e iew ques ion and ulfill he inclusion
c i e ia we e collec ed. I he e iew ques ion was no answe ed a e
eading he whole documen o /and he inclusion c i e ia we e no ul-
filled, he pape was wi hd awn. The da a ex ac ed om each pape
was s o ed by filling fields in a e iew ool de eloped in Excel Visual
Basic (VB) o ease he in o ma ion wi hd awal.
App oxima ely 270 pape s we e collec ed using he keywo ds in he
sea ch engines and subsequen snowball o wa d and backwa d p oce-
du es. Only 120 a icles me he inclusion c i e ia and we e s udied in
dep h. The ollowing subsec ions summa ize he esul s ex ac ed
om hese 120 manusc ip s, bu only some will be e e enced as pa
o he bibliog aphy o his documen . Fo a comple e lis o he pape s,
please e e o Appendix 1.
4. Resul s and discussion
4.1. Ca ego iza ion and his o ical o e iew
Two b oad ca ego ies we e iden ified in he p elimina y assessmen
o he a icles, each c ea ing di e en ia ed esea ch ou comes: 1) a
SUDS model o 2) a DSS elying on a SUDS model. The e a e inpu s
and ou pu s in bo h ca ego ies, bu he fi s e e s o isola ed modeling
gene ally o assess SUDS pe o mance, while he la e couples he
SUDS model wi h o he modules o include cos s, s akeholde s, and sec-
onda y benefi s, o example. Ano he essen ial di e ence is he na u e
o he ou pu s. While he s and-alone SUDS model deli e s uno se ies,
pollu an s educ ions, o any o he pe o mance measu e, he DSS p o-
ides answe s o decision-make s ques ions (i.e., Which SUDS is ecom-
mended o he bes ? Wha loca ions a e sui able/op imal o SUDS?
Which SUDS mee s he pollu an educ ion a ge ?). Fig. 1 illus a es
he ela ion be ween he wo a icle ca ego ies and shows he coun
o SUDS models (ca ego y 1) and SUDS-DSS (ca ego y 2), he numbe
o a icles ha add ess he ques ions Which?,Whe e?,How many?,o as-
sis s he design o indi idual SUDSand ains. A single pape can assis in
answe ing mo e han one ques ion, so a manusc ip can be coun ed se -
e al imes in Fig. 1 (once pe ques ion assis ed).
The o al e iewed a icles spanned he pe iod comp ised be ween
1997 and 2021. The numbe o a icles had an inc easing end, s a ing
wi h jus a couple o publica ions pe yea , om 1997 o 2012, and hen
con inued inc easing om 2015 o 2020. The ype o ou pu was di e se
(i.e., he ques ion he ool assis s in answe ing). No ice in Fig. 2(a), ha
SUDS-DSS commonly answe ed a single ques ion, while since 2012,
he e is mo e ou pu di e si y; obse e ha since 2015, he ou pu s in-
clude 5 ca ego ies. These obse a ions eflec bo h he di e sifica ion o
SUDS-DSS use s and he b oadening o pe spec i e om SUDS “uni s” o
SUDS “sys ems”.
SUDS ains we e only included in DSS om 2015, which can be ex-
plained by he e ol ed capabili y o models o simula e flow among
connec ed SUDS s uc u es. Simila ly, he “SUDS design”ques ion p e-
domina ed he ea ly de elopmen o he ools, bu wi h ime his
Table 1
Se s o keywo ds used o sea ch. Pape s whose i le ha e a leas h ee wo ds in di e en
keywo d se s we e conside ed o e iew.
Se 1 Se 2 Se 3 Se 4
Decision/ ool
keywo ds
SUDS keywo ds Modeling
keywo ds
S o mwa e
keywo ds
Assess* Bes managemen p ac ice* (BMP) Model* Runo
E ec i e* Sus ainable u ban d ainage sys ems
(SUDS)
–S o m*
Cos * G een in as uc u e (GI) –U ban
Heu is ic* Low impac de elopmen (LID) –Flood*
Managemen Wa e sensi i e u ban design (WSUD) –Plu ial
Op im* Na u e based solu ions –Rain all
Objec i e* Blue g een sys ems ––
Planning Sponge ci ies ––
Suppo * Bio e en ion ––
Tool* Infil a ion ––
Decision* Re en ion ––
–De en ion ––
*Any wo d con aining he oo -wo d signaled by * also makes pa o he se . Fo example,
“assess*”includes he wo ds “assesing”,“assessed”,and“*objec i e”includes “mul i-ob-
jec i e”o “mul iobjec i e”.Fig. 1. SUDS models embedded in a SUDS Decision Suppo Sys em (DSS).
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
4
in e es declined, gi ing space o o he aspec s, such as “which”,
“whe e”o “how many”SUDS we e sui able o op imal. This
obse a ion is ied o he change in he s udy scale in e es and
dimensions unde s udy. Fig. 2(b) shows ha la ge s udy a eas
(i.e., subca chmen and ca chmen ) gained a en ion o e he las
yea s in compa ison wi h smalle scales, while he numbe o s udies
ocusing on household and neighbo hood scales has dec eased o e
he las 6 yea s. The ci y-scale decision-making appea ed o he fi s
ime in 2005 wi h he s udy de eloped by Mak opoulos and Bu le
(2005), which used non-s uc u al SUDS o po able wa e consump ion
educ ion. This shi in he spa ial scale can be explained by he expo-
nen ial g ow h o compu a ional capabili ies (Bu ge e al., 2014),
which allowed he modeling so wa es o include bigge spa ial scales
o e ime wi hou incu ing in longe p ocessing imes. When app ais-
ing he pape s, i was consis en ly ound as a ecommenda ion o u u e
s udies he de elopmen o SUDS-DSS capable o assis ing decision-
making a he ci y-scale (Mak opoulos and Bu le , 2005;Chen e al.,
2017;Zubelzu e al., 2020). Fu he mo e, hese a icles pay special a -
en ion o he impo ance o including op imiza ion and s akeholde s
ba gaining models o decision-making a wa e shed and ci y-scales,
and also make a special highligh on he impo ance o including
economic, social and en i onmen al dimensions. Simila ly, Fig. 2
(c) shows ha he dimensions conside ed in decision-making di e si-
fied wi h ime. In 2006, SUDS-DSS we e al eady conside ing he
economic aspec along wi h he uno quan i y and quali y, while he
en i onmen al and social dimension appea ed mo e ecen ly and con-
inue gaining impo ance (Al es e al., 2020).
4.2. Geog aphical sp ead
Fig. 3 shows ha he majo i y o he s udies we e de eloped in Asia,
wi h 58% o he a icles, ollowed by No h Ame ica (27%), Eu ope wi h
(14%), Oceania wi h (6%) and Sou h Ame ica (6%). The coun ies wi h
he la ges con ibu ions we e China, Uni ed S a es o Ame ica (USA),
I an and Aus alia, wi h 28%, 20%, 11%, and 5% espec i ely. The es o
he coun ies had a lowe coun (less han 3% om he o al) and 36%
when agg ega ed. Based on hese numbe s, i was possible o iden i y
he u gen need o less-de eloped coun ies o inc ease he SUDS-
DSS scien ificp oduc i i y(Fe ans e al., 2018), conside ing he e is a
high po en ial o exis ing models o be implemen ed in hese coun ies
(McClymon e al., 2020). De eloped coun ies can con ibu e o closing
his gap h ough collabo a i e in e na ional p ojec s (e.g. Resou ce
B andia, euPOLIS, euPOLIS, 2020, e c.), whe e o he coun ies' expe ise
can accele a e hei lea ning cu e.
As expec ed, he ques ion he ool assis s in answe ing and he SUDS
aspec s conside ed in each coun y a e di e se. While mo e-de eloped
coun ies, like he USA, Canada, and Aus alia included mo e aspec s be-
sides uno quan i y and quali y, less-de eloped coun ies ocused al-
mos exclusi ely on hese wo. An excep ion is China, wi h se e al
included aspec s, and B azil, which pu special a en ion o he social
Coun
45
40
35
30
25
20
15
10
5
0
Yea
(a)
Queson add essed o e me (coun ).
1997
1999
2004
2005
50
2006
2007
2010
40
2011
2012
2013
30
2014
2015
2016
2017
20
2018
2019
2020
10
2021
0.0 0.2 0.4 0.6 0.8 1.0
F ac ion
0
Yea
(b)
Spaal scale modeled o e me (p opo -
on)
(c)
Dimensions o e me (coun )
Design
How Many
T ains
Whe e
Which
Ca chmen
Ci y
Household
Neighbo hood
Subca chmen
Economic Analysis
En i onmen al bene i s
Quali y Modelling
Quan i y Modelling
Social bene i s
Yea
Coun
1997
1999
2004
2005
2006
2007
2009
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
1999
2004
2005
2006
2007
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
Fig. 2. Resea ch in e es e olu ion o e ime. Ca ego ies in panel (a) defined as ollows: Design (p elimina y o de ailed design), How Many (numbe o sui able/op imal s uc u es), T ains
(o de o in e connec ed ypologies), Whe e (spa ial alloca ion o SUDS), Which (op imal/sui able SUDS ypologies). Panel (c) shows he aspec s (dimensions) conside ed o SUDS
decision-making using a imeline. Obse e ha ea ly yea s ocused exclusi ely in quan i y and quali y modeling.
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
5
aspec . This finding is explained by he u gen challenges (e.g., u ban
flooding and ecei ing wa e bodies quali y impai men ) ha a e s ill
o o e come in less-de eloped coun ies. Once again, hese esul s ein-
o ce he need o de eloped coun ies o gene a e collabo a i e en i-
onmen s in which he e is oom o sha e expe iences and ga he
in o ma ion.
4.3. Case s udy scale, ypologies, spa ial and empo al modeling esolu ion
In ecen yea s, he e has been a apid de elopmen o compu a-
ional capabili ies, which allowed he ep esen a ion o de ailed p o-
cesses a la ge scales. Fo SUDS models, hese ad ances pe mi ed
efinemen s in p ocess ep esen a ion and fine spa ial and empo al
esolu ion. P e iously, i was shown ha he spa ial scale ocus has
shi ed in ime, pa ially because he modeling capabili ies ha e allowed
he inclusion o mo e de ail and mo e complexi y, bu also because
decision-make s needs ha e e ol ed. E e y day, decision-make s ely
mo e upon so wa e o make decisions, while expe judgmen has
been g adually eplaced by sys emic p ocedu es ha educe he amoun
o bias and manual wo k, gene ally speeding up he p ocess (Ha ab
e al., 2020).
F om he app aised pape s, 55% used a ca chmen -scale, 22% a
subca chmen -scale, ollowed by neighbo hood-, household-, and ci y-
scale, wi h 12%, 6%, and 5%, espec i ely. Fig. 4(a) shows ha he
s udy a ea has a ange o 6 o de s o magni ude, wi h a minimum
alue o 0.01 hec a es (ha), a maximum alue o 650,000 ha, and a s an-
da d de ia ion o 80,000 ha. The ca chmen -scale has he la ges ange
and numbe o ou lie s (comp ising 6 o de s o magni ude). Fig. 4
(a) shows ha he a ea is no de e minan o he spa ial uni o analysis,
since he same s udy a ea (e.g., 100 ha) can be classified in o ci y, ca ch-
men , sub-ca chmen , o neighbo hood. The smalles scale ound was
he household, wi h a mean a ea o 0.1 ha and he la ges was he
ci y-scale, wi h a mean o 1000 ha.
Fig. 4(b) shows ha mos DSS answe ed he ques ion “Which SUDS”,
dis ega ding he scale. In gene al, all ques ions can be answe ed a a
neighbo hood-scale o bigge , while o he household-scale, he only
ques ions add essed we e “Design”(de ailed dimensioning and loca-
ion) and “Which SUDS”. The household-scale has a la ge p opo ion
o “Design”, which was expec ed conside ing ha smalle a eas makes
i mo e a ainable o each a highe le el o de ail.
Fig. 4(c) shows he equency o he mos common SUDS ypologies
ound in he e iew. Cu iously, he mos used ypology is one ha does
no use g een a eas di ec ly (pe meable pa emen s), ollowed by
g assed swales, bio- e en ion cells, and ain ba els. In o al, nea ly 30
a icles included SUDS in hei models, bu do no speci y which ype
(e.g. Jia e al., 2012;Raei e al., 2019;Rod íguez-Sinobas e al., 2018). I
was ound ha all ypologies had simila p opo ions wi h espec o
he ques ion hey assis in answe ing (Fig. 4(c)), showing ha he
ools do no di e en ially add ess he ques ions depending on he ypol-
ogies hey include. Fig. 4(e) shows ha (as expec ed) some o he ypol-
ogies a e mo e equen ly used in la ge scales: cons uc ed we lands,
de en ion and infil a ion basins, d y de en ion, and bio- e en ion
ponds; while some o he ypologies ha e mo e flexibili y o be used in
la ge, medium, and small scales: s o age anks, ain ba els, infil a ion
enches, and g assed swales. Specifically o he household-scale, he
mos popula ypologies we e g een oo s, ain ga dens, pe meable
pa emen s, and bio- e en ion s a egies. Con a y o he scale, he
land use (e.g., comme cial, indus ial, esiden ial, ec ea ional open
space) showed no ends ega ding ypologies; all ypologies we e p es-
en in simila p opo ion. Only 4% o he a icles (5 pape s) allowed he
inclusion o SUDS ains ( ypologies sequen ially in e connec ed) de-
spi e he li e a u e ecommends ains o inc ease he s uc u es' e fi-
ciency in managing uno (e.g., Bas ien e al., 2010). Those a icles
ha did conside SUDS ains, selec ed he ain componen s based on
expe s knowledge and he epo ed e ficiency o indi idual SUDS o
con ol a ge pollu an s (e.g., Xu e al., 2017;Jayasoo iya e al., 2016;
Za a e al., 2017), ins ead o ollowing an s anda d p ocedu e.
The ime s ep used o he calcula ions showed a high a iabili y
among he di e en s udies, anging om 1 min o 2 h. Fig. 4
(d) shows a sca e plo o he size o he s udy a ea and he modeling
ime s ep, using colo codes o he so wa es. The ou lie in Fig. 4
(d) (30 days ime s ep) co esponds o Chang e al. (2011), who pe -
o med an own-de eloped model based on wa e balance equa ions
wi h a simula ion pe iod o 50 yea s.
Dis ega ding he so wa e, he e is no e iden a ea- ime end in
Fig. 4(d). This can be a ibu ed o di e ences in he models complexi y
e en when he same so wa e is used. The figu e shows ha he SWMM
so wa e is sca e ed along he wo axis, as opposed o o he so wa es
ha a e clus e ed in a eas o he plo (e.g. L-THIA-LID is ound in la ge
a eas and la ge ime s eps only).
A p e ious s udy, Sal ado e e al. (2015) e iewed 100 UDMs o
compa e he space- empo al esolu ions. The au ho s iden ified wo
clus e s: ca chmen -scale applica ions (la ge empo al esolu ions)
and small-size cases o s udy. The au ho s ound he fines empo al
Fig. 3. Li e a u e geog aphical dis ibu ion.
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
6
Design
How Many
T ains
Whe e
Which
Pe meable
Pa emen s
G assed Swales
Bio e en ion Cells
Rain Ba els
G een Roo s
Bio e en ion Ponds
In il a ion T enches
Rainga dens
No Di e en ia ed
Cons uc ed We land
S o age Tanks
In il a ion Basins
D y Ponds
G een Space
De en ion Basins
T ee Pi s
Sand Fil e s
A ea (ha)
10
5
10
3
10
1
10
−1
Subca chmen
Neighbo hood
Household
Ci y
Ca chmen
0.0 0.2 0.4 0.6 0.8 1.0
F ac ion
Spa ial Scale
(a)
A ea (ha) s. spaal scale (b) Spaal scale s. queson add essed
100
80
10
5
60
10
4
40
10
3
20
10
2
0
10
1
10
0
So wa e
L
-THIA-LID
SWMM
SUSTAIN
Own-de eloped
MIKE URBAN
We Spa-U ban
Sewe GEMS
10
−1
Typology
10
0
10
1
10
2
10
3
10
4
Time S ep (Minu es)
(c)
SUDS ypologies equency
(d)
A ea (ha) s. modeling me s ep (min)
colo -coded by Sowa e
T ee Pi s
S o age Tanks
Sand Fil e s
Rainga dens
Rain Ba els
Pe meable Pa emen s
No Speci ied
In il a ion T enches
In il a ion Basins
G een Space
G een Roo s
G assed Swales
D y Ponds
De en ion Basins
Cons uc ed We land
Bio e en ion Ponds
Bio e en ion Cells
0.0 0.2 0.4 0.6 0.8 1.0
F ac ion
(e)
Spaal scale pe SUDS ypology
Design
How
Many
T ains
Whe e
Which
Ca chmen
Ci y
Household
Neighbo hood
Subca chmen
Coun
A ea (Ha)
Tipology
Spa ial Scale
Fig. 4. Case s udy (land use and ypologies) and empo al esolu ions pe spa ial scale. Ca ego ies in panels (b, c) defined as ollows: Design (p elimina y o de ailed design), How Many
(numbe o sui able/op imal s uc u es), T ains (o de o in e connec ed ypologies), Whe e (spa ial alloca ion o SUDS), Which (op imal/sui able SUDS ypologies).
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
7
and spa ial esolu ions o be 1 s and 10 m
2
, bu hese UDMs did no ha e
he capabili ies o modeling SUDS. When compa ing ou esul s o
Sal ado e e al. (2015), i was e idenced ha ca ego y 2 a icles (s udies
de eloping/applying a DSS) we e only being implemen ed in wha
hese au ho s call “ca chmen -scale applica ions,”meaning ha SUDS-
DSS a e s ill in he la ge scale o u ban d ainage modeling in e ms o
empo al and spa ial modeling g anula i y. As will be discussed in
Sec ion 4.4, he decision on he empo al and spa ial esolu ion is also
ela ed o SUDS model capabili ies and he selec ion o e en -based o
con inuous simula ion.
4.4. Modeling me hodologies: so wa e and e en s selec ion
The S o m Wa e Managemen Model (SWMM) (Rossman, 2010)
de eloped by he US En i onmen al P o ec ion Agency (EPA) was he
mos equen ly used in he pape s app aised (46% o he s udies). The
second mos common so wa e we e own-de eloped non-comme cial
models, wi h 16% o he s udies, ollowed by L-THIA-LID (Pu due-
Uni e si y, 2016) (4%) and SUSTAIN (Shoemake e al., 2009)(2%).
The es o he s udies, which ep esen he 20% o he o al a icles,
used o he so wa e (e.g., MIKE URBAN (DHI, 2008), MUSIC (eWa e ,
2020), SUDSLOC (Via a ene e al., 2011), GISP (Mee ow and Newell,
2017), ReVISIONS (Ha g ea es e al., 2019), SSANTO (Kulle e al.,
2019), U banBEATS (Bach e al., 2013), WSCT (Zhang e al., 2020),
e c.) each ep esen ing less han 2% o he o al numbe o publica ions.
SWMM, MIKE URBAN, and L-THIA-LID we e he only so wa e capable
o modeling SUDS ains (see Fig. 5(a)). SWMM and he own-
de eloped models we e used in simila p opo ions o add ess all ques-
ions, while he o he so wa es we e used o ackle mo e a ge ed
ques ions.
The mos basic app oaches used in own-de eloped models consis ed
o wa e balance calcula ions, s a is ical analyses, o GIS-based ools
(Wang and Wang, 2018;Zhen e al., 2006;Lee e al., 2010;Yang and
Bes , 2015). O he s op ed o nume ic algo i hms o sol e hyd ological
and hyd aulic di e en ial equa ions (e.g., Beck e al., 2017;Pe ez-Pedini
e al., 2005;Gülbaz and Kazezyılmaz-Alhan, 2017;W igh e al., 2018). Fi-
nally, some s udies elied on linea p og amming o embed he hyd olog-
ical equa ions in he op imiza ion model. This las app oach equi es a
simplified ep esen a ion o he p ocesses occu ing in he SUDS s uc-
u e. Fo example, Seb i e al. (2016) es ima ed SUDS impac using a a -
ian o he Imp o ed Ra ional Hyd og aph me hod (IRH), and Chang e al.
(2011) and To es e al. (2020) used wa e balance equa ions. Rega dless
o he complexi y in hese s udies, i was iden ified ha he necessi y o
de eloping own models and ools is d i en by he lack o flexibili y o
he a ailable so wa e, in pa icula when da a is limi ed and i s o ma in-
compa ible wi h he equi ed inpu s.
Rega ding he ype o empo al simula ion, 58% o he s udies used
e en -based simula ions, 28% used con inuous simula ion, and 14% pe -
o med a compa ison analysis using bo h app oxima ions. Fig. 5
(b) shows he p opo ion o con inuous/e en -based app oaches o
each spa ial scale; as he spa ial scale dec eases, he p opo ion o s ud-
ies pe o ming con inuous simula ion g ows. Dis ega ding whe he
e en -based o con inuous, 88% o he s udies app aised used a de e -
minis ic app oach, 9% a s ochas ic app oach, 3% did a compa a i e anal-
ysis o bo h app oaches. These esul s e idence ha he compu a ional
esou ces needed o pe o m ime- and esou ce-consuming simula-
ions (con inuous and s ochas ic app oaches) a e a ailable in seldom
cases.
Fig. 5(c) and (d) p esen s box-plo s wi h he numbe o e en s
(e en -based) o he numbe o yea s (con inuous-simula ion) ana-
lyzed o each DSS ques ion add essed. Fo e en -based, he numbe
o e en s anged be ween 1 and 10, wi h ou lie s up o 20 and a maxi-
mum alue o 53. I was ound ha 56% o he s udies used design ain-
all e en s wi h e u n pe iods ( anging om 5 o 50 yea s), 37% used
ep esen a i e his o ical ain all e en s, 6% employed syn he ically gen-
e a ed e en s, and only 3% based hei analyses on o ecas ed e en s. On
he o he hand, Fig. 5(d) shows ha he majo i y o he s udies using
con inuous-simula ion analyzed om 1 o 25 yea s. Fig. 5(e) p esen s
a s acked ba plo di e en ia ing he ype o simula ion pe o med pe
ques ion add essed. The “Design ools”used con inuous simula ion in
a highe p opo ion han he es o he ca ego ies, wi h 46% o he s ud-
ies using a con inuous simula ion. The my iad o modeling se ings and
p ecipi a ion e en s selec ion e idence ha he e is s ill no consensus
on good p ac ices o modeling-based SUDS decision-making.
4.5. Ca ego ies, p ocesses modeled, and s akeholde s
Table 2 shows ha 82% o he s udies included wa e quan i y, 53%
wa e quali y, 28% economic analysis, and only 8% and 3% included en-
i onmen al and social benefi s. Wi hin he en i onmen al aspec s, he
mos common a e ai quali y and ene gy sa ings (e.g., Chang e al.,
2011) and in he social aspec he ec ea ion, accep abili y, and ameni y
(e.g., Jia e al., 2012). Table 3 shows he p ocesses mo e equen ly
modeled in he quan i y and quali y dimensions besides he ain all-
uno p ocess (which was included in all he a icles). No su p isingly,
he p ocesses mo e equen ly modeled a e infil a ion and e apo ans-
pi a ion since hese wo a e esponsible o he uno olume educ ion
and he peak flow fla ening. O seconda y impo ance we e he
g oundwa e flow and sedimen a ion p ocesses. The fi s is equen ly
neglec ed, despi e i s p o en impo ance in SUDS e ficiency (Zhen
e al., 2004;Xu e al., 2020b), because o he lack o local da a, while
he la e is less included in p opo ion o o he p ocesses gi en ha i
is i ele an in uno quan i y assessmen s.
The majo i y o he SUDS-DSS app aised had scena io modeling
(65%) (i.e., compa ing SUDS spa ial configu a ions using pe o mance
me ics). Howe e , mos o he s udies did no conside clima e change
no u baniza ion ends p ojec ions (only 12% had his capabili y). I is
highly ecommended o e-di ec e o s o include u u e p ojec ions
in SUDS-DSS since i is expec ed ha clima e change and u baniza ion
a es will play a majo ole in u u e u ban hyd ology, pa icula ly in
la ge u ban cen e s (Xu e al., 2020a;Salda iaga e al., 2020).
P e ious wo ks iden ified ha a key aspec o gua an ee a success ul
decision-making p ocess is he ea ly inclusion o s akeholde s pe spec-
i es and p e e ences (e.g., Jayasoo iya e al., 2020;Ahiablame e al.,
2012;To es e al., 2020). Howe e , i was ound ha 87% o he s udies
do no include s akeholde s. Table 4 shows ha om he 13% (16 a i-
cles ou o he 120 e iewed) ha did conside one o se e al s ake-
holde s, he mos common a e local au ho i ies (31%), u ili ies (13%),
neighbo s (13%), poli icians (6%), and En i onmen al Agencies (EA)
(6%); 31% o he a icles include a leas one s akeholde , bu do no
s a e which one. F om he pape s e iewed, none conside ed he
opinion/p e e ences o he communi y membe s, who ul ima ely a e
impac ed by he decisions. I is highly ecommendable ha he s ake-
holde s' posi ions a e included o decision-making, pa icula ly he
communi ies.
4.6. DSS inpu s, ou pu s and gene al amewo k
4.6.1. Inpu s and ou pu s
This subsec ion is dedica ed o he a icles in ca ego y 2, DSS elying
on a SUDS model and including addi ional dimensions besides he hy-
d ological. A o al o 83 a icles lie in his ca ego y, compiling 4 ypes
o inpu a iables: hyd o-me eo ological (e.g., p ecipi a ion, uno ,
empe a u e o e apo anspi a ion), s udy-si e (e.g., land uses, impe -
meabili y, slope, infil a ion a e, p esence o con en ional d ainage sys-
ems), wa e quali y (e.g., loads o pollu an s like nu ien s, o ganic
ma e , solids, and hea y me als), and economic (e.g., SUDS and land
cos s and mone a y quan ifica ion o en i onmen al se ices).
The mos equen inpu a iable o SUDS-DSS a e he hyd o-
me eo ological a iables, wi h a pe cen age o inclusion be ween 40
and 60%. The s udy si e ea u es also p esen ed high pe cen ages o
inclusion (30-50%). The uno quali y (build-up and wash-o
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
8
pa ame e s) a iables had a lowe equency: 25%, 13%, 12% and 2% o
o al suspended solids, o al phospho us, o al ni ogen, and hea y
me als, espec i ely. Finally, among he economic inpu a iables,
SUDS cos s we e included in 40% o he s udies, bu o he economic
indica o s (e.g., economic e u n, ne p esen alue, e c.) exhibi
pe cen ages o inclusion lowe han 2%.
Rega ding he e olu ion o e ime o he inpu a iables, i was
no iced ha om 2000 o 2010 he mos equen ly included
ca ego y was he s udy si e aspec s, wi h some educed usage o
economic indica o s. F om 2010 o wa d, he wa e quali y inpu s
s a ed o be included, showing an inc easing end o e he yea s.
The economic aspec s also showed an inc easing end o e ime.
The hyd o-me eo ological a iables showed a s eady end o inclu-
sion o e ime. These ends eflec bo h he inc eased da a a ail-
abili y and he d i ing in e es on SUDS om a mo e holis ic
pe spec i e.
Design
How Many
T ains
Whe e
Which
Yea s o he Da a Base
90
75
60
45
30
15
0
Ci y
Ca chmen
Subca chmen
Neighbo hood
Household
So wa e
0.0 0.2 0.4 0.6 0.8 1.0
F ac ion
(a)
Queson add essed among he diffe en
sowa e
(b)
Tempo al simulaon among he
diffe en spaal scales
50
40
10
1
30
20
10
10
0
Design T ains Which
Whe e How Many
Ques ion Add essed
0
Whe e How Many
T ains Which
Design
Ques ion Add essed
(c) Numbe o e en s simula ed o (d) Yea s simula ed o each quesons
each quesons add essed add essed.
Design
How Many
T ains
Whe e
Which
0.0 0.2 0.4 0.6 0.8 1.0
F ac ion
(e) Type o simulaon o he diffe en
quesons add essed.
Con inuos E en Based
Coun
Numbe o E en s
AnnAGNPS
CANOE So wa e
Compa ison
GIFMod
GIS-SWMM
GSSHA
L-THIA-LID
MIKE URBAN
MODFLOW
MUSIC
Own-de eloped
PCSWMM
SEWSYS
SUDSLOC
SUSTAIN
SWMM
Sewe GEMS
U banBEATS
WABILA
Wa e Sensi i e Ci ies Toolki
We Spa-U ban
WinSLAMM
Ques ion Add essed
Spa ial Scale
Con inuos E en Based
Fig. 5. So wa e, simula ion, and empo al esolu ion o each ques ion add essed.
P. Fe ans, M.N. To es, J. Temp ano e al. Science o he To al En i onmen 806 (2022) 150447
9
