O iginal Resea ch A icle
Dis inc and combined impac s o clima e and land use scena ios on
wa e a ailabili y and sedimen loads o a wa e supply ese oi in
no he n Mo occo
Fa iha Chouk i
a
, Damien Raclo
a
,
b
,
*
, Mus apha Naimi
a
, Mohamed Chikhaoui
a
,
Jo~
ao Ped o Nunes
c
,F

ed
e ic Hua d
d
,C

ecile H
e i aux
e
, Mohamed Sabi
,
Yannick P
epin
a
,
b
a
IAV Hassan II, Depa men o Na u al Resou ces and En i onmen , Madina Al I ane, Raba , Mo occo
b
LISAH, Uni Mon pellie , INRAE, IRD, Ins i u Ag o, Mon pellie , F ance
c
CE3C - Cen e o Ecology, E olu ion and En i onmen al Changes, Faculdade de Ci^
encias, Uni e sidade de Lisboa, 1749-016, Lisboa, Po ugal
d
INRAE-Ag oClim, 228 Rou e de l’A
e od ome, CS 40 509, 84914, A ignon, Cedex 9, F ance
e
BRGM, Uni Mon pellie , 1039 Rue de Pin ille, 34000, Mon pellie , F ance
ENFI, bd Moulay Yousse , Tab ique , SALE, Mo occo
a icle in o
A icle his o y:
Recei ed 29 Augus 2019
Recei ed in e ised o m
17 Janua y 2020
Accep ed 24 Ma ch 2020
A ailable online 4 Ap il 2020
Keywo ds:
Clima e change
Land use change
SWAT
Runo
Rese oi managemen
Mo occo
abs ac
The objec i e o his s udy was o examine he impac s o clima e and land use changes on wa e
a ailabili y and sedimen loads o a wa e supply ese oi in no he n Mo occo using da a-in ensi e
simula ion models in a da a-sca ce egion. Impac s we e assessed by compa ing he simula ed wa e
and sedimen en e ing he ese oi be ween he u u e pe iod 2031e2050 and he 1983e2010 e e ence
pe iod. Th ee scena ios o land use change and wo scena ios o clima e change we e de eloped in he
Tle a wa e shed. Simula ions unde cu en and u u e condi ions we e pe o med using he Soil and
Wa e Assessmen Tool (SWAT) model. The simula ions showed ha clima e change will lead o a sig-
nifican dec ease in he annual wa e supply o he ese oi (16.9% and 27.5%) and in he annual
olume o sedimen en e ing he ese oi (7.4% and 12.6%), depending on he clima e change sce-
na ios es ed. The h ee scena ios o land use change will lead o a mode a e change in annual wa e
inflow in o he ese oi (be ween 6.7% and þ6.2%), while causing a significan dec ease in sedimen
en e ing he ese oi (37% o 24%). The combined impac s o clima e and land use changes will cause
a educ ion in annual wa e a ailabili y (9.9% o 33.3%) and sedimen supplies (28.7% o 45.8%). As
a esul , he li e ime o he ese oi will be ex ended, bu a he same ime, he isk o wa e sho ages
will inc ease, especially om July o Ma ch. The e o e, al e na i e wa e esou ces mus be conside ed.
©2020 In e na ional Resea ch and T aining Cen e on E osion and Sedimen a ion and China Wa e and
Powe P ess. P oduc ion and Hos ing 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/).
1. In oduc ion
Soil e osion is a se ious h ea o he en i onmen (Lal, 2003). I
causes on-si e damage, such as he de e io a ion o he physico-
chemical and biological p ope ies o soils, he loss o nu ien s,
he educ ion in ag icul u al p oduc i i y and he loss o a able
land. O -si e damage, such as i e sedimen deposi ion, ese oi
sedimen a ion, and channel sil ing, esul s in significan economic
loss (Pimen el e al., 1995). Soil wa e e osion is a widesp ead
phenomenon in Medi e anean coun ies (De F anchis &Ibanez,
2003;Kosmas e al., 1997) because o he combina ion o clima ic
and an h opogenic e osion-p one ac o s (Ga cía-Ruiz, Nadal-
Rome o, Lana-Renaul , &Begue ía, 2013;Raclo , Le Bissonnais,
Annabi, Sabi , &Sme ano a, 2018). As a esul , ese oi sil a ion
ep esen s a majo issue o many Medi e anean coun ies, espe-
cially in No h A ica (Ayadi, Abida, Djebba , &Mahjoub, 2010;
*Co esponding au ho . UMR LISAH, B^
a .24, 2 place Viala, 34060, Mon pellie
Cedex 2, F ance.
E-mail add esses: [email p o ec ed] (F. Chouk i), damien. aclo @i d.
(D. Raclo ), m.naimi@ia .ac.ma (M. Naimi), m.chikhaoui@ia .ac.ma (M. Chikhaoui),
[email p o ec ed] (J.P. Nunes), [email p o ec ed] (F. Hua d), c.he i aux@b gm.
(C. H
e i aux), [email p o ec ed] (M. Sabi ), [email p o ec ed]
(Y. P
epin).
Con en s lis s a ailable a ScienceDi ec
In e na ional Soil and Wa e Conse a ion Resea ch
jou nal homepage: www.else ie .com/loca e/iswc
h ps://doi.o g/10.1016/j.iswc .2020.03.003
2095-6339/©2020 In e na ional Resea ch and T aining Cen e on E osion and Sedimen a ion and China Wa e and Powe P ess. P oduc ion and Hos ing 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/).
In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153
Lahlou, 2000), whe e wa e a ailabili y is mainly based on he
su ace wa e mobiliza ion capaci ies. Because he Medi e anean
egion is conside ed a global “ho spo ”in e ms o clima e a i-
abili y and change, as well as he a e o land ans o ma ion p o-
cesses (Ga cía-Ruiz, L
opez-Mo eno, Vicen e-Se ano, Lasan a-
Ma ínez, &Begue ía, 2011;Gio gi &Lionello, 2008), i is c ucial o
an icipa e he u u es o ese oi in e ms o wa e a ailabili y and
li e ime.
The e is a g owing consensus in Ea h sys ems science ha
global empe a u es a e inc easing and will con inue o do so o e
he nex cen u y, leading o changes in global clima e egimes
(IPCC, 2014;Nunes &Nea ing, 2010). The p ojec ions o g eenhouse
gas inc eases a e expec ed no only o inc ease he global a e age
empe a u e bu also o influence p ecipi a ion pa e ns (IPCC,
2014). Al hough a iable om egion o egion, hese ends end
o inc ease he equency o ex eme e en s such as hea wa es and
high-in ensi y s o ms (Nunes &Nea ing, 2010). Li and Fang (2016)
ecen ly e iewed he di ec and indi ec impac s o clima e change
on wa e e osion. Di ec impac s a e mainly caused by changes in
p ecipi a ion (quan i y, in ensi y, spa ial and empo al dis ibu-
ion), and indi ec impac s a e ela ed o empe a u e inc ease: a
wa m clima e will a ec soil e osion mainly h ough changes in
ege a ion co e and soil mois u e. O he indi ec impac s a e
associa ed wi h an h opogenic o socioeconomic ac o s; he
combina ion o p ecipi a ion and empe a u e changes is likely o
be accompanied by changes in c op managemen , ag icul u al
planning, c op ypes and p ices. Va ia ions in ain all pa e ns (i.e.,
he amoun o ain all pe e en , in ensi y, equency and he sea-
sonali y o p ecipi a ion) could ha e a significan impac on he
hyd ological egime and soil e osion (Bangash e al., 2013;Li &
Fang, 2016;Nea ing e al., 2005;P uski &Nea ing, 2002;Zhang,
Nea ing, &Liu, 2005,2010). Lu e al. (2013) epo ed ha a 1%
change in p ecipi a ion esul ed in a 2% change in sedimen loads
and a 1.3% change in wa e inflows. O he s udies ha e shown ha a
simple change in ain all seasonali y can ha e significan e ec s on
soil losses (Chouk i e al., 2016).
Un o una ely, he complex in e ac ions be ween land use and
clima e make i di ficul o p edic he impac s o global change on
uno and e osion because o possible an agonis o syne gis ic
e ec s (Tome &Schilling, 2009). Fo ins ance, a decline in p e-
cipi a ion can cause less e osion, bu a he same ime, i can also
educe ege a ion co e ha will a ou e osion. This an agonis
phenomenon is discussed ex ensi ely in Nunes and Nea ing (2010).
The consequence o a decline in p ecipi a ion on uno can also be
e y complex, as highligh ed by he “Sahel pa adox”, i.e., he ac
ha he Sahel has wi nessed a pa adoxical inc ease in su ace wa e
despi e a gene al p ecipi a ion decline in ecen decades (Desc oix
e al., 2013, chap. 10; Gal, G ippa, Hie naux, Pons, &Ke goa , 2017).
The impac s a e e en mo e unexpec ed when conside ing land use
change induced by socioeconomic condi ions in addi ion o clima e
change (Nunes, Jacin o, &Keize , 2017).
A modelling app oach is a common and use ul way o p ojec
uno and soil e osion unde global change, as shown by he e-
iew o Li and Fang (2016). Clima e scena ios may be based on
ou pu s o gene al ci cula ion models (GCMs) o egional clima e
models (RCMs). GCMs a e unable o assess si e-specific clima e
impac s in a eliable way because o hei coa se spa ial esolu ion
(Hulme e al., 1993;Zhang, 2005). RCMs dynamically simula e
clima e cha ac e is ics a esolu ions o 10e50 km, aking in o ac-
coun ime- a ying a mosphe ic condi ions a he bounda y o a
specified domain (Wilby, 2007;Zhang e al., 2019a). Downscaling
me hods a e used o fill spa ial and empo al esolu ion gaps be-
ween clima e modelling and modelle needs (Wilby, Dawson, &
Ba ow, 2002;Zhang e al., 2019b). Land use scena ios a he
egional scale a e gene ally based on na a i e s o ylines such as
hose associa ed wi h he sha ed socioeconomic pa hways (SSPs)
es ablished a he global scale, as shown in O’Neill e al. (2017).As
wi h any p ospec i e wo k, he assump ions behind modelling
app oaches lead o unce ain ies (Ludwig &Roson, 2016), such as
hose ela ed o clima e p ojec ions, socioeconomic scena ios, da a
a ailabili y and quali y, and knowledge deficiency on bo h he
biophysical p ocesses in ol ed and he models used o he simu-
la ions (Gha a i, Keess a, Ghodousi, &Ahmadi, 2010;Nunes e al.,
2017).
In his s udy, he Soil and Wa e Assessmen Tool (SWAT model,
Nei sch, A nold, Kini y, Williams, &King, 2005) was applied o he
Tle a wa e shed o examine o he fi s ime he impac o global
change on he su ace wa e mobiliza ion capaci ies o a ese oi
in no he n Mo occo. The model was applied o cu en condi ions
using obse ed da a and o u u e condi ions using clima e inpu s
esul ing om a bias-co ec ed RCM and map inpu s esul ing om
a sepa a e land use scena io building exe cise. The impac s o
clima e o land use changes on uno and e osion ha e been
e alua ed sepa a ely o di e en ia e he dis inc impac s o clima e
change om he dis inc impac s o land use change and join ly o
iden i y he combined impac s o clima e and land use changes on
he s udied wa e shed. SWAT was selec ed as i has been p e iously
used o simila s udies in Medi e anean en i onmen s (Bucak
e al., 2017;Nunes e al., 2017;Se pa e al., 2015), and i has been
shown o be adap ed o he no he n Mo occo con ex wi hou
adjus men s o he han pa ame e iza ion o fi he local condi ions,
e.g., o simula ing he impac o clima e change on he ulne a-
bili y o wa e esou ces and c op pe o mance (B ouziyne e al.,
2018) o he impac o managemen p ac ices on uno and
e osion (B iak, Moussadek, Abouma ia, &M abe , 2016,2019).
The main objec i es o his s udy a e (1) he e alua ion o he
pe o mance o he SWAT model in e ms o uno and sedimen s
o a semi-a id Medi e anean wa e shed in he no he n egion o
Mo occo; (2) he quan ifica ion o he impac s o global change on
wa e a ailabili y and sedimen loads o he Ibn Ba ou a ese oi
by 2040; and 3) he assessmen o he implica ions o he wa e
mobiliza ion capaci ies o he ese oi and su ounding ag icul-
u al ac i i ies.
2. Ma e ial and me hods
2.1. Si e desc ip ion
Loca ed be ween Tangie and T
e ouan (No h Mo occo), he
Tle a wa e shed co e s an a ea o 180 km
2
(Fig. 1a) ups eam o he
Ibn Ba ou a ese oi buil in 1977 wi h an ini ial s o age capaci y o
43.5 Mm
3
o p o ide d inking wa e o he ci y o Tangie . The
ese oi is loca ed in he oo hills o he Ri Moun ains, which is an
in ensi e wa e e osion-p one a ea. Al hough i co e s only 6% o
Mo occo’s a ea, he Ri Moun ains and i s oo hills p o ide mo e
han 60% o he sedimen mobilized each yea h oughou he
coun y (Issa, Lech-Hab, Raissouni, &El A im, 2016). Land deg a-
da ion in he Tle a wa e shed has been desc ibed as ala ming in he
absence o soil conse a ion e o s (Me zouk, Fenji o, &Laouina,
1996). Thus, he s o age capaci y o he Ibn Ba ou a ese oi is
cons an ly dec easing (43.6 Mm
3
in 1978 o 29.1 Mm
3
in 2013),
which ep esen s an annual sil a ion o app oxima ely 0.4 Mm
3
/
yea and a loss o 33% o i s ini ial capaci y.
The wa e shed ele a ions ange om 23 o 672 m a he highes
poin wi h e ain slopes anging om 0 o 30% (Fig.1c). Wi h a sub-
humid o humid Medi e anean clima e and we win e s and d y
summe s (B iak e al., 2016), he wa e shed is loca ed in one o he
mos humid a eas in he coun y. A he Ibn Ba ou a dam, he
a e age annual p ecipi a ion was 692 mm o e he pe iod
1980e2010, and he po en ial e apo anspi a ion (PET) de i ed
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153142
om a Colo ado e apo a ion pan and a co ec ion ac o o 0.7 was
app oxima ely 1190 mm o e he pe iod 1983e2010. The a e age
minimum and maximum daily empe a u es o e he same pe iod
we e 14e23

C a he closes s a ion (Tangie ai po ). The wa e -
shed is cha ac e ized by a dominance o ma ly subs a e, indica ing
he low pe meabili y o he subsoil. The main li hos a ig aphic
ea u es o he Ri Moun ain idge a e well ep esen ed in he
wa e shed, namely, he o e lay o se e al bed ock laye s (flysch)
o ming he idge line abo e he na i e Tangie uni (Meloussa)
o ming he hills (Micha d, 1976). This li hological a ie y has
allowed he de elopmen o a a he impo an soil mosaic (Fig.1d)
composed o Cambisols, Ve isols, Flu isols, S agnosols, and Lep-
osols. The land use/co e (Fig. 1b) was qui e s able be ween 1980
and 2010. I was pa ially condi ioned by he mo pho-pedological
disposi ion. The sands one and flysch acies cons i u e he sub-
s a e o o es ed and ma o al lands ha a e some imes e y
deg aded, while he ma ly acies is mos ly cul i a ed. The o es and
he dense and clea ma o al occupy he highes pa s o he hills,
i.e., he ups eam pa o he wa e shed (Me zouk e al., 1996).
Ag icul u al lands a e dominan in he lowe pa s o he wa e -
shed, while bed ock ou c ops a e limi ed o an a ea nea he
ese oi . Ce eals (e.g., whea , ba ley and oa s) a e he dominan
c ops and a e some imes cul i a ed in o a ion wi h leguminous
c ops (e.g., chickpeas and aba beans) o in associa ion wi h oli e
ees o o he ui ees. Wa e and sedimen inpu s in o he
ese oi ha e been moni o ed a he wa e shed ou le since he
ins alla ion o he Ibn Ba ou a dam, and a ne wo k o ain gauges
was ins alled o cap u e he spa ial a iabili y o ain all. As in many
Medi e anean wa e sheds, di ec uno p ocesses a e p edomi-
nan , and he wa e shed ime esponse is ypically less han one
day.
2.2. Da ase used o SWAT implemen a ion
Relie and land use maps we e p oduced om emo e sensing
images and field su eys (see Table 1 o desc ip ion and Fig. 1 o
illus a ion). A de ailed soil desc ip ion o he s udy a ea made o
ag icul u al pu poses was composed o a soil map wi h a comple e
desc ip ion o he ep esen a i e p ofile o each soil ype,
including he maximum oo ing dep h o he soil p ofile and
s anda d soil pa ame e s o each soil laye (e.g., ex u e, o al and
o ganic ca bon con en , soil bulk densi y, and a ailable soil wa e
con en ). Addi ional soil desc ip ions and analyses we e pe o med
be ween 2010 and 2017 o c oss alida e he exis ing sou ce o da a
and quan i y hyd ological soil p ope ies such as wa e infil a ion
a es.
Conside ing clima ic da a (see Table 1 o desc ip ion and Fig. 3
o loca ion), ime se ies we e a ailable a eigh s a ions o daily
Fig. 1. (a) Geog aphic loca ion o he Tle a wa e shed (Lambe con o mal conical p ojec ion, Cla ke 1880), (b) land use/co e map (LU 2010), (c) al i ude map and (d) soil map (WRB,
2006).
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153 143
ain all and a ou gauge s a ions o daily empe a u es and
ela i e humidi y. Exis ing gaps in he daily ain all and empe a-
u e ime se ies o a gi en s a ion we e filled using mul iple linea
co ela ions be ween s a ions o ob ain con inuous se ies om
1980 o 2010. The daily empe a u es (minimum and maximum)
and p ecipi a ion amoun s a he di e en gauge loca ions we e
finally in e pola ed using an in e se dis ance weigh ing me hod
(IDW), which allows he p oduc ion o syn he ic daily ime se ies
on a egula g id wi h a esolu ion o 1 km. Da ase s om all he
clima ic s a ions we e conside ed in he in e pola ion p ocess
because hey can be cap u ed by he IDW in e pola ion unc ion
and imp o e in e pola ion a he edge o he wa e shed. Because
ela i e humidi y, sola adia ion and wind we e only a ailable a
he Tangie s a ion, hese h ee clima ic da ase s we e assumed o
be uni o m o e he whole wa e shed.
Daily s eamflow was de i ed om he hyd ological budge o
he ese oi , as explained in Raclo and Albe gel (2006). This
assessmen is based on he knowledge o he heigh -su ace and
heigh - olume cu es es ablished by ba hyme ic su eys. A di -
e en ia ion be ween di ec and delayed uno was de e mined
using he g aphic me hod on some indi idual e en s and au oma ic
fil e ing me hods (local minimum me hod, one pa ame e digi al
fil e and ecu si e digi al fil e included in he Web-based
Hyd og aph Analysis Tool a ailable a h ps://enginee ing.pu due.
edu/mapse e/WHAT/) on he whole daily uno da ase . Sedi-
men olumes en e ing he ese oi be ween wo ba hyme ies
we e es ima ed by di e ences be ween he ba hyme ic su eys
and a co ec ion using a apping e ficiency coe ficien o 93%
e alua ed om Heinemann’s o mula (Heinemann, 1981). The
olume o sedimen a i ing a he dam be ween wo measu es
was hen decon olu ed a a daily ime s ep, assuming ha he daily
solid flows (Sed) a e p opo ional o he maximum flow (Qmax)
and he daily o al uno olume (V o ) acco ding o Equa ion (1)
p oposed by Williams (1975):
Sed ðV o QmaxÞ0:56 KLS CP(1)
The main da ase s used as inpu o SWAT implemen a ion a e
summa ized in Table 1.
2.3. Implemen a ion o he SWAT model unde cu en condi ions
The SWAT model was implemen ed o e a 31-yea pe iod
(1980e2010) using he Ha g ea es me hod o PET es ima ion and
he da ase lis ed in Table 1, knowing ha empe a u e and ain all
ime se ies we e in e pola ed on he 1 km esolu ion g id be o e
being used as inpu da a. Thus, each subbasin in SWAT was assigned
wi h he closes bias-co ec ed clima ic ime se ies om he egula
g id. Fo each land use ype, he mos simila land use ype om he
SWAT land use da abase was selec ed, and only mino adjus men s
o some plan g ow h/managemen pa ame e s we e made.
In a e y adi ional spli -sample app oach (Kleme
s, 1986;
Re sgaa d, Hen iksen, Ha a , Schol en, &Kassahun, 2005), his
pe iod was di ided in o an ini ializa ion o wa m-up phase
(1980e1982), a calib a ion phase (1983e1996) and a model ali-
da ion phase (1997e2010). The calib a ion was ca ied ou manu-
ally in wo successi e s eps. The fi s s ep consis ed o adjus ing
pa ame e s ela ed o he su ace uno ou ing (SURLAG) and o
he in e ac ion among he unde g ound compa men , unsa u a ed
zone and flow in he s eam (GW_REVAP, GWQMN and ALPHA_BF)
o ep oduce as closely as possible he daily uno a he ou le o
he wa e shed o he calib a ion pe iod, ensu ing ha he
espec i e con ibu ions o di ec and delayed uno we e p ope ly
simula ed. Once uno was co ec ly eplica ed du ing he cali-
b a ion pe iod, he second s ep consis ed o adjus ing pa ame e s
ela ed o channel sedimen de achmen and ou ing (SPCON,
SPEXP, CH_COV1 and CH_COV2) o minimize he di e ences be-
ween measu ed and simula ed daily sedimen flows. The pe o -
mance o model p edic ion was assessed o he calib a ion and
alida ion pe iods using he ollowing s anda d s a is ical in-
dica o s (A nold e al., 2012): he de e mina ion coe ficien (R
2
), he
Nash-Su cli e coe ficien (NSE) and he bias indica o (PBIAS). Ac-
co ding o Mo iasi e al. (2007), a mon hly model simula ion can be
judged as “sa is ac o y”i NSE >0.50 and i PBIAS <±25% o
s eamflow and PBIAS <±55% o sedimen . In he p esen s udy,
he ecommended s a is ics p oposed by Mo iasi e al. (2007) ha e
been used o analyse SWAT simula ion pe o mance o bo h
mon hly and daily ime s eps. Fo he daily ime s eps, hese s a-
is ical indica o s we e also calcula ed a e smoo hing he
obse ed and simula ed daily alues using a mo ing a e age o e 3
days o explo e issues ela ed o daily disc e iza ion when unning
Table 1
Da ase used as inpu o SWAT implemen a ion.
Type Sou ce Pe iod Desc ip ion
Relie SPOT images (20 m esolu ion) 2014 20 m digi al ele a ion model (see Fig. 1c)
Land use/co e Classifica ion o a 2010 Landsa image
(30 m esolu ion), field su eys and
in e p e a ion o ae ial pho og aphs
(H
e i aux, Vina ie , Sabi , Guillo , &
Rinaudo, 2018)
2010 See Fig. 1b o land use/co e classes ( hese da a a e called
LU2010 in his pape )
Soil map and
desc ip ion
LCGS/Inypsa (1/50,000) 1989 See Fig. 1d o soil classes. Also includes a comple e desc ip ion o a ep esen a i e
p ofile by soil ype wi h s anda d soil pa ame e s o each soil laye .
Soil cha ac e is ics Soil sampling and analyses 2010e2017 Addi ional s anda d soil p ope ies ( ex u e, ca bon con en , wa e e en ion …)
and infil a ion a es
(de i ed om single- ing, double ings measu emen s o ain all simula ions)
Rain all ABHL - Loukkos Hyd aulic Basin
Agency
1980e2010 Daily ain all in 8 loca ions (see Fig. 3 o loca ions)
Tempe a u es Di ec ion o Na ional Me eo ology
and ABHL
1980e2010 Daily empe a u es (min., max., a e age) a ou gauge s a ions (including a comple e
clima ic s a ion a Tangie ai po )
Runo ABHL 1980e2010 Daily uno (Ibn Ba ou a s a ion a ca chmen ou le ) calcula ed by a ese oi hyd ological
balance
E osion ABHL þpos - ea men 1980e2010 Daily sedimen yield (Ibn Ba ou a s a ion a ca chmen ou le ) calcula ed by a ese oi
sedimen balance
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153144
SWAT a he daily ime s ep in he Medi e anean con ex .
2.4. Implemen a ion o he SWAT model unde u u e condi ions
2.4.1. Land use change scena ios
Th ee land use/co e scena ios ha e been elabo a ed by
H
e i aux e al. (2018) o 2040 and desc ibed by Fig. 2. These sce-
na ios a e he ou comes o a p ospec i e s udy conduc ed by a
mul idisciplina y eam on he s udied wa e shed. Only he main
lines o each scena io a e ep oduced he e:
Scena io S1, en i led “U baniza ion and indus ial de elop-
men ”, assumes s ong economic de elopmen in no he n
Mo occo, associa ed wi h a high en y o labou om o he e-
gions. U baniza ion is e y impo an nea he ecen ly c ea ed
ci y o Ch a a and along oads wi h he cons uc ion o houses,
second homes, ac o ies and logis ic pla o ms. This u baniza-
ion is mainly a he expense o ag icul u al lands, wi h he
excep ion o he e y di ficul - o-access wes e n zone.
Scena io S2, en i led “Tle a ui baske ”, assumes mode a e
de elopmen o Ch a a and i s su oundings and e y low in-
dus ial de elopmen . In his con ex , he s a e conside s he
de elopmen o ag icul u e as a p io i y and suppo s he
implemen a ion o ag icul u al de elopmen p ojec s o ien ed
owa ds ain ed a bo icul u e and ag o o es y (oli e, fig and
walnu ), beekeeping, goa a ming and ma ke ga dening (i i-
ga ion in he icini y o Ibn Ba ou a dam), as well as he c ea ion
o ag icul u al coope a i es.
Scena io S3, en i led “One oo in he ci y, one oo in he
coun yside”, is based on he de elopmen o he ci y o Ch a a
combined wi h an ambi ious u al de elopmen p og amme o
e y small neighbou ing illages called doua s, enabling he
popula ion o emain in he doua s and de i e li ings om bo h
ag icul u e and ac i i ies in u ban a eas (Ch a a , Tangie and
he indus ial a ea). Coope a i es ha e been se up o de elop
he local p oduc s sec o , such as honey, goa cheese, p ickly
pea s and a oma ic plan s.
Fig. 2. (a) Map o he h ee land use scena ios (S1 o S3) by 2040 ( om H
e i aux e al., 2018) and (b) land use dis ibu ion o he LU2010 and S1 o S3 scena ios.
Fig. 3. Geog aphic loca ion o he clima ic da ase .
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153 145

2.4.2. Clima e change scena ios
Two clima e scena ios based on wea he o ecas p ojec ions
om F ance’s ALADIN-Clima (Ai e Limi 
ee Adap a ion dynamique
D
e eloppemen In e Na ional) egional model ha e been es ed.
They co espond o p ojec ions based on RCPs (Rep esen a i e
Concen a ion Pa hways) 8.5 and 4.5, i.e., o adia i e o cings o 8.5
and 4.5 W m
2
by 2100. The ime se ies p o ided by ALADIN-Clima
a e a daily ime s eps and e e o minimum and maximum em-
pe a u es (T), ela i e humidi y (H ), ain all (R), global adia ion
(Rg) and wind speed (W) 2 m abo e he g ound. Conside ing he
loca ion o he wa e shed, he di e en clima ic s a ions used and
he na i e esolu ion o 12 km o ALADIN-Clima , he da a p o ided
by ALADIN-Clima on 7 g id cells we e selec ed (Fig. 3).
Fo each g id cell, simula ed his o ical ime se ies om 1950 o
2005 and ime se ies associa ed wi h he wo RCP scena ios o e
he pe iod 2006 o 2100 we e conside ed. A common s a is ical
downscaling echnique, desc ibed in de ail in Hua d e al. (2019),
was used o co ec biases in clima ic ime se ies be ween he
ou pu s o he ALADIN egional clima e model and he local in si u
clima ic obse a ions a ailable wi hin o nea he Tle a wa e shed.
The bias co ec ion was based on a quan ile mapping wi h b eak-
poin s, i.e., a me hod e y simila o he ARRM (Asynch onous
Regional Reg ession Model) app oach p oposed by S one , Hayhoe,
Yang, and Wuebbles (2013).Table 2 p o ides an example o s an-
da d s a is ics on he daily bias-co ec ed ain all ime se ies a he
Ibn Ba ou a clima ic s a ion (see Fig. 3 o loca ion), which enables a
global compa ison be ween cu en and u u e daily ain all con-
di ions ha ha e been used as inpu da a in SWAT.
The bias-co ec ed daily empe a u es (minimum and
maximum) and p ecipi a ion in he u u e clima e a he di e en
si es we e finally in e pola ed o p oduce, o each clima e scena io,
u u e daily ime se ies on he same egula g id wi h a esolu ion
o 1 km and wi h he same in e pola ion p ocess as ha o cu en
clima ic condi ions.
2.4.3. Implemen a ion o SWAT unde global change scena ios
The impac o global change scena ios was e alua ed by unning
SWAT o e he pe iod 2021 o 2050 wi h pa ame e s calib a ed
o e he cu en pe iod. Compa ed o he e e ence scena io, he
only changes in SWAT implemen a ion consis ed o conside ing i)
clima e and/o land use change scena ios and ii) a wa m-up pe iod
o 10-yea s (2021e2030). The land use scena io consis ed o a
change in land use ype only, and no change in ag icul u al p ac ice
o a gi en land use ype has been implemen ed be ween cu en
and u u e condi ions (i.e., he SWAT managemen file was un-
changed o a gi en land use be ween he cu en and u u e sim-
ula ions). Finally, wa e and sedimen en e ing he ese oi by
2040 we e quan ified by a e aging he simula ion ou pu s be ween
2031 and 2050.
2.5. E alua ion o he global change impac s
Se e al combina ions be ween he cu en and u u e land use
and clima e condi ions (Table 3) we e simula ed using SWAT. The
simula ed uno and sedimen yield en e ing he ese oi o each
combina ion was compa ed o he e e ence combina ion o
quan i y he dis inc and combined impac s o changes in land use
and clima e. Fo example, he dis inc impac o land use change
was analysed by compa ing SWAT ou pu s based on he combina-
ion be ween he h ee land use scena ios and he 1983e2010 cli-
ma ic condi ions (combina ion name ‘S1’ o ‘S3’in Table 3) wi h
SWAT ou pu unde cu en condi ions (combina ion name ‘Base-
line’in Table 3).
3. Resul s and discussion
3.1. Model pe o mance unde cu en condi ions
Visually, a good co espondence be ween mon hly simula ed
and obse ed uno and e osion fluxes can be obse ed o bo h
he calib a ion and he alida ion pe iods (Fig. 4).
The s a is ical indica o s (Table 4) confi m he good pe o -
mance o uno and sedimen yield simula ions a he mon hly
ime s ep. The esul s ob ained he e o e show ha SWAT is able o
co ec ly ep oduce he mon hly inpu s o wa e and sedimen in o
he ou le ese oi o he s udied wa e shed. When conside ing
simula ion esul s a daily ime s eps, he pe o mance o SWAT o
ep oduce daily uno and sedimen yield was no as good as ha
a mon hly ime s eps. This highligh s he di ficul ies in ge ing he
iming igh in SWAT, especially in a poo da a con ex . The
disc epancy be ween daily obse ed and simula ed alues may
esul om p oblems in he wa e ans e module bu i may also
be induced by he daily spli ing in he obse ed ime se ies.
Indeed, pe o mance indica o s a e highly impac ed by daily da a
spli ing when wa e shed ime esponse and hyd og aph di usi i y
a e bo h e y low. Because he mean esponse ime is only
app oxima ely 6 h in he s udied wa e shed, he simula ed uno o
a ain all pulse ha occu s a he beginning o he day will be
egis e ed in he same day, whe eas he simula ed uno o he
same ain all slo ha occu s a he end o he day will be egis e ed
in he nex day. The e is he e o e a non-sys ema ic delay o 1 day
be ween ain all and uno ha canno be aken in o accoun in
daily ime s ep modelling. The in e p e a ion o he indica o s o
pe o mance on he 3-day smoo hed obse ed and simula ed
imese ies p o ided in e es ing in o ma ion on he likely eason o
he disc epancy. Indeed he pe o mance indica o s ob ained a e
smoo hing o e 3 days (Table 4) we e significan ly highe han
hose ob ained wi hou smoo hing, which indica es ha daily
spli ing pa ially explained he di ficul ies in ge ing he iming
igh in SWAT in he Medi e anean con ex . Fo sedimen yield,
howe e , a poo pe o mance a he daily ime s ep was ob ained
o he alida ion pe iod, e en a e smoo hing o he da a, which
indica es ha p oblems in he sedimen anspo module
emained. A simila si ua ion has been obse ed in o he s udies
(Nunes e al., 2018), which p oposed as a likely explana ion ha he
model pe o med well in he e osion simula ion, bu p oblems in
he sedimen anspo module led o imp ecisions in he daily
disc e iza ion o sedimen yield, which we e a e aged ou a he
mon hly ime s ep.
SWAT simula ion be ween 1983 and 2010 showed ha , o an
annual a e age p ecipi a ion o 791 mm o e he en i e Tle a
wa e shed, 59% (466 mm) was e u ned o he a mosphe e by
Table 2
S anda d s a is ics on he bias-co ec ed daily ain all ime se ies on he pe iod 1983e2010 (REF) and he pe iod 2031e2050 (RCP4.5 and RCP8.5) a he Ibn Ba ou a clima ic
s a ion. No e ha only daily ain all alues o e 2 mm we e conside ed.
Minimum Fi s qua ile Median Mean Thi d qua ile Maximum
REF (1983e2010) 2.00 4.52 9.10 13.38 17.80 103.00
RCP4.5 (2041e2060) 2.09 4.42 8.99 12.91 16.99 81.41
RCP8.5 (2041e2060) 2.10 4.40 8.48 12.04 15.96 72.93
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153146
e apo anspi a ion, and 8% (62.14 mm) pe cola ed in o he deep
aqui e (i.e., he pa o he aqui e no con ibu ing o s eam flow)
and 33% (262.86 mm) con ibu ed o s eam flow. The annual
s eam flow a e was di ided in o di ec uno (192.93 mm), la e al
flow con ibu ion (22.46 mm) and shallow aqui e con ibu ion
(47.47 mm). This means ha SWAT es ima ed a 74% con ibu ion o
o al annual uno o di ec flow and a 26% con ibu ion o in-
di ec flow. These alues we e consis en wi h he dis ibu ion o
s eam flow be ween di ec uno and delayed uno es ima ed
using he g aphic me hod (be ween 20 and 25% o indi ec flow) o
di e en au oma ic fil e ing me hods (be ween 15 and 25% o in-
di ec flow). SWAT was he e o e able o ep oduce he hyd ological
p ocesses in a semi-a id wa e shed such as he Tle a wa e shed, as
al eady shown by B iak e al. (2016) in he e y close Kalaya
wa e shed.
3.2. Dis inc and combined impac s o global change on uno and
sedimen yield
Table 5 summa izes he main annual a e age e ms o he wa e
and sedimen budge a he ese oi ou le unde he baseline
si ua ion and he di e en combina ions o land use and clima e
change scena ios by 2040.
3.2.1. Impac o land use change
Simula ion esul s showed ha he change in land use in
Table 3
The es ed combina ion o clima e and land use changes.
Combina ion name Land use Clima e
REFERENCE Baseline LU2010 1983e2010
Land use change only S1 S1 1983e2010
S2 S2
S3 S3
Clima e change only RCP4.5 LU2010 RCP4.5
RCP8.5 RCP8.5
Combined change in land use and clima e S1_RCP4.5 S1 RCP4.5
S1_RCP8.5 RCP8.5
S2_RCP4.5 S2 RCP4.5
S2_RCP8.5 RCP8.5
S3_RCP4.5 S3 RCP4.5
S3_RCP8.5 RCP8.5
Fig. 4. Compa ison be ween obse ed and SWAT simula ed mon hly uno (m
3
/s) and sedimen yield (10
5
ons) o (a and c) he calib a ion pe iod and (b and d) he alida ion
pe iod.
Table 4
Pe o mance o SWAT in ep oducing measu ed uno and e osion fluxes o daily and mon hly ime s eps and o he calib a ion and alida ion pe iods ( alues in i alics and in
b acke s o he daily ime s ep co espond o he pe o mance indica o s ob ained a e smoo hing o e 3 days, as explained in he Me hods sec ion).
Time s ep R
2
NSE PBIAS (%)
Calib a ion Valida ion Calib a ion Valida ion Calib a ion Valida ion
Runo Daily 0.55 (0.76) 0.47 (0.69) 0.52 (0.71) 0.43 (0.62) 2(-2) 2(-2)
Mon hly 0.92 0.84 0.89 0.81 33
E osion Daily 0.56 (0.67) 0.40 (0.55) 0.40 (0.59) 0.01 (0.25) 10 (-10) 37 (-37)
Mon hly 0.84 0.70 0.74 0.52 10 37
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153 147
scena io S1 will p o ide a þ6.2% inc ease o annual wa e inflow
in o he ese oi . This annual inc ease in wa e supply is consis en
wi h he significan de elopmen o u banized a eas ( om 9%
cu en ly o 18% o he wa e shed su ace a ea in S1) and a mod-
e a e inc ease in e apo anspi a ion (þ4 mm), which esul s om a
sligh inc ease in he ui ee a ea (2%). No e ha he inc ease in
bo h uno and e apo anspi a ion was compensa ed by a signifi-
can dec ease in he amoun o wa e ha pe cola ed in o he deep
aqui e (i.e., g oundwa e losses in Table 5), pa ially due o he
de elopmen o u baniza ion ha p e en s pe cola ion. In he wo
o he land use scena ios, he annual wa e supply o he ese oi
will dec ease by 6.7% o S2 and by 3.2% o S3. This dec ease in
uno is mainly explained by he weak de elopmen o u banized
a eas ( om 9% o 13% and 14% o he wa e shed su ace a ea in S2
and S3), which would no compensa e o he inc ease in e apo-
anspi a ion due o ege a ion changes. Fo example, bo h o es ed
(þ7%) and ui ee (þ7%) a eas inc eased a he expense o ag i-
cul u e (20%) in he S2 scena io, and i is known ha an inc ease
in canopy (mo e plan biomass and high lea a ea index) a ou s
in e cep ion and e apo anspi a ion (Wang &Kalin, 2011). The
changes in ege a ion in scena ios S2 and S3 will also gene a e a
dec ease in wa e pe cola ion o he deep aqui e .
Rega ding e osion, SWAT p edic ed a significan dec ease in
sedimen yield ( om 25 o 37%) and sedimen concen a ion
( om 25 o 33%) o he 3 es ed land use scena ios. The g ea e
dec ease eco ded o scena io S2 can be explained by a dec ease in
e osion-p one land use (c ops dec ease om 46% o 26% o he o al
a ea o he wa e shed), combined wi h an inc ease in mo e soil-
p o ec i e land use ( ui ee ac i i y inc eases om 6% o 15%
and o es /ma o al om 36% o 43%). The dec ease in e osion in S3
is due o a dec ease in c oplands ( om 46 o 40%) and an inc ease in
ma o al a ea, whe eas he dec ease in e osion o S1 is linked o
he dec ease in ag icul u al a ea and an inc ease in non-e odible
a eas ela ed o an inc ease in u ban and indus ial ac i i ies.
In summa y, he h ee land use change scena ios will impac he
wa e supply o he ese oi in a mino way, whe eas hey will
significan ly dec ease he a e o sedimen inpu in o he ese oi .
These esul s a e in line wi h he esul s o Ca alho-San os e al.
(2016) and Nunes e al. (2017), who epo ed a low impac o
land use change scena io on uno a es in se e al wa e sheds o
Po ugal. The esul s a e also in line wi h s udies epo ing a highe
impac o land use o c op managemen changes on e osion a he
han on uno (Se pa e al., 2015).
3.2.2. Impac o clima e change
Fo he s udied wa e shed, he clima e change condi ions will
esul in an inc ease in a e age empe a u es (þ1.3

C oþ2

C) and
a dec ease in mean in e annual p ecipi a ion (6.2% o 11.6%) o
RCP4.5 and RCP8.5, espec i ely (Table 5, combina ion name
‘RCP4.5’and ‘RCP8.5’). SWAT simula ions showed ha e apo ans-
pi a ion will be impac ed by less han 1%, whe eas he annual wa e
supply o he ese oi will significan ly dec ease by 16.9% and
27.5% o RCP4.5 and RCP8.5, espec i ely. The dec ease in wa e
supplies will be logically mo e significan unde RCP8.5 han unde
RCP4.5. SWAT also simula ed a dec ease in he amoun o wa e
pe cola ion o he deep aqui e because o less soil sa u a ion
induced by less ain all. The esul s on e osion flows a e less ex-
pec ed. Indeed, he annual olume o sedimen en e ing he
ese oi will also dec ease, bu o a much lesse ex en han ha o
uno , wi h educ ions o 7.4% and 12.6%, espec i ely. This esul is
su p ising because g ea e changes in sedimen yields han in
wa e flows a e gene ally epo ed in he li e a u e (Lu e al., 2013).
The inc ease in he a e age annual sedimen concen a ion in he
s eam flow o 11.4% and 20.6% o RCP4.5 and RCP8.5, espec i ely
(Table 5), pa ially o se s he annual dec ease in uno . The eason
o he inc ease in sedimen concen a ion is discussed below.
The mon hly e olu ion o SWAT ou pu s (Fig. 5) p o ides mo e
p ecise insigh in o he clima e change impac on wa e and sedi-
men en e ing he Ibn Ba ou a ese oi . Fi s , he seasonal ain all
pa e n is expec ed o be g ea ly modified, wi h mo e p ecipi a ion
in sp ing and summe and significan ly less p ecipi a ion in
au umn and win e .
Fig. 5 also shows ha u u e p ecipi a ion amoun s can be e y
di e en o some mon hs be ween RCP4.5 and RCP8.5. Fo
example, he p ojec ed ain all by RCP8.5 is 50 mm highe han he
p ojec ed ain all by RCP4.5 in Decembe . Howe e , a seasonal
compa ison o he occu ence o ain all g ea e han 20 mm be-
ween he e e ence pe iod and he wo RCPs (Fig. 6) does no show
a significan change in he occu ence o majo ain alls.
The u u e mon hly uno pa e ns will also be g ea ly modified,
wi h an inc ease in uno amoun s in sp ing and a significan
educ ion du ing au umn and win e ega dless o he RCP
conside ed. The change in mon hly uno pa e ns is globally
simila o he change in ain all pa e ns. Howe e , he uno
esponse o ain all change is mo e complex since i appea ed o be
nonlinea . Fo example, he sligh dec ease in p ecipi a ion in
Janua y o he wo RCPs (i.e., 7% o RCP4.5 and 14% o RCP8.5)
will lead o a la ge dec ease in uno (i.e., 21% o RCP4.5
and 32% o RCP8.5). In Ma ch, a significan inc ease in ain all will
only gene a e a sligh uno inc ease.
The mon hly sedimen yield pa e ns unde he u u e clima e
will also be g ea ly modified, wi h a significan inc ease in
Table 5
In e annual a e age p ecipi a ion, e apo anspi a ion, uno , sedimen yield and sedimen concen a ion o he e e ence si ua ion (baseline) and o he di e en combi-
na ions o land use change (LUC) and clima e change (CC).
Combina ion name P ecipi a ion E apo-
anspi a ion
G oundwa e
losses
Runo Sedimen yield Sedimen
concen a ion
mm
D
(%) mm
D
(%) mm
D
(%) mm
D
(%) /ha
D
(%) g/l
D
(%)
Baseline 791.0 466.0 62.1
D
(%) 26.9 1,02
S1 791.0 0 470.1 þ0.9 41.6 33.0 279.3 þ6.2 20.4 24.4 0,73 28.6
S2 791.0 0 491.1 þ5.4 54.7 11.9 245.2 - 6.7 16.9 36.9 0,69 32.6
S3 791.0 0 483.6 þ3.8 52.8 15.0 254.6 - 3.2 19.6 27.3 0,77 24.8
RCP4.5 741.6 6.2 470.2 þ0.9 53 14.7 218.4 16.9 24.9 7.4 1,14 þ11.4
RCP8.5 699.2 11.6 467.1 þ0.2 41.6 33.0 190.5 27.5 23.5 12.6 1,23 þ20.6
S1_RCP4.5 741.6 6.2 470.5 þ1.0 34.3 44.8 236.8 9.9 19.2 28.7 0,81 20.8
S1_RCP8.5 699.2 11.6 469.5 þ0.8 21.3 65.7 208.4 20.7 18.2 32.4 0,87 14.6
S2_RCP4.5 741.6 6.2 492.8 þ5.8 45.9 26.1 202.9 22.8 15.7 41.5 0,77 24.4
S2_RCP8.5 699.2 11.6 490.2 þ5.2 33.7 45.7 175.3 33.3 14.6 45.8 0,83 18.6
S3_RCP4.5 741.6 6.2 485.6 þ4.2 43.9 29.3 212.1 19.3 18.3 32.3 0,86 15.7
S3_RCP8.5 699.2 11.6 483.0 þ3.6 31.6 49.1 184.6 29.8 17.2 36.3 0,93 8.9
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153148
sedimen yield amoun s in Ma ch o bo h RCPs and a significan
educ ion du ing au umn, while a educ ion in sedimen yield will
be obse ed om May o Decembe . Du ing Janua y and Feb ua y,
sedimen yield will inc ease o RCP8.5 and dec ease o RCP4.5.
The inc ease in sedimen yield o RCP8.5 in Janua y combined wi h
a dec ease in ain all highligh s he complex and nonlinea
esponse o sedimen yield o ain all change. This esul is in line
wi h p e ious s udies ha ha e al eady shown ha e osion in he
Medi e anean con ex was highly nonlinea (Gonz
alez-Hidalgo,
Pẽna-Monn
e, &de Luis, 2007).
Analysing he mon hly a e age sedimen concen a ion unde
cu en and u u e clima es (Fig. 5) highligh s a significan inc ease
in sedimen concen a ion in he s eamflow, especially om
Janua y o Ma ch. Since no significan change in he seasonal
occu ence o majo e en s has been de ec ed, i is mo e likely ha
he la ge inc ease in sedimen concen a ion om Janua y o Ma ch
is he esul o a educ ion in soil p o ec ion by ege a ion. Indeed,
he sha p d op in ain all om Oc obe o Janua y will esul in bo h
a dec ease and delay in ege a ion de elopmen , which will a ou
soil de achmen by ain all du ing he same mon hs and he
ollowing mon hs (Feb ua y and Ma ch). The dec ease in ege a-
ion de elopmen is confi med by he dec ease in biomass simu-
la ed by SWAT, especially o c oplands, o cha ds, o es s and
ma o als.
Mos exis ing s udies ha ha e analysed he impac o clima e
change on uno and e osion a e gene ally based on he 2000
Special Repo on Emission Scena ios (SRES) named by amily (A1,
A2, B1 and B2). To ela e he esul s o hese exis ing s udies o
hose o ou s udies based on he RCPs defined in 2010, we can
conside ha A2 is qui e simila o RCP8.5, B1 is qui e simila o
RCP4.5, and A1B and B2 a e in e media e scena ios be ween RCP4.5
and RCP8.5. Mo e de ails on he compa ison be ween SRES and RCP
can be ound a h ps://www.globalchange.go /b owse/
mul imedia/emissions-concen a ions-and- empe a u e-
p ojec ions. Mos o hese s udies ha e indica ed ha sligh a ia-
ions in he amoun o p ecipi a ion can ha e significan e ec s on
annual a e age uno and ha clima e change is he e o e a main
ac o de e mining o al uno in a wa e shed (Yang, 2013). Fo
example, Bussi, F anc
es, Ho el, L
opez-Ta az
on, and Ba alla (2014)
assessed he impac o clima e change on he hyd ological and
sedimen ological cycles o he 
Ese a Ri e ca chmen (Spain) unde
A2 and B2 scena ios (2070e2100). They ound ha he o al wa e
yield is expec ed o dec ease by 40 and 35% unde he A2 and B2
scena ios, espec i ely, while he o al p ecipi a ion is expec ed o
dec ease by only 13 and 12%. Nunes e al. (2017) simula ed a
dec ease in uno o app oxima ely 25% in a Medi e anean
wa e shed o Vale do Gaio, sou h o Po ugal, unde he wo clima e
scena ios A1B and B1 o e he pe iod 2071e2100, while he
amoun o p ecipi a ion dec eased by only 9%. The impac o
clima e change on sedimen yield desc ibed in o he Medi e a-
nean s udies o en showed an inc ease in e osion a es in esponse
o a dec ease in ain all. Fo example, Pa oissien e al. (2015)
Fig. 5. Mon hly SWAT simula ed alues o ain all, uno , sedimen yield and sedimen concen a ion a he Tle a wa e shed ou le du ing he pe iod 1983e2010 (REF) and he
pe iod 2031e2050 (RCP4.5 and RCP8.5).
Fig. 6. Seasonal occu ence o majo ain all e en s using boxplo s based on ain all
e en s g ea e han 20 mm du ing he pe iod 1983e2010 (REF) and he pe iod
2031e2050 (RCP4.5 and RCP8.5).
F. Chouk i e al. / In e na ional Soil and Wa e Conse a ion Resea ch 8 (2020) 141e153 149