Clima e-d i en a ia ions in p oduc i i y e eal adap i e s a egies in
Ibe ian co k oak ag o o es y sys ems
Jos
e Ca los P
e ez-Gi
on
a
,
*
, Emilio Ra ael Díaz-Va ela
b
, Ped o
Al a ez-
Al a ez
a
a
Depa men o O ganisms and Sys ems Biology, Poly echnic School o Mie es, Uni e si y o O iedo, E-33600, Mie es, As u ias, Spain
b
Resea ch G oup on Planning and Managemen in Complex Adap i e Socio-Ecological Sys ems (COMPASSES). School o Enginee ing, Uni e si y o San iagode
Compos ela, E-27002, Lugo, Spain
ARTICLE INFO
Keywo ds:
Ne p ima y p oduc ion
Ca bon use e ficiency
Clima e
Que cus sube
Ag o o es y sys em
ABSTRACT
Backg ound: Co k oak ag o o es y sys ems (AFS) ha e been managed o cen u ies by humans o p oduce co k
and o he goods and se ices and ha e ecen ly been ecognised as an impo an ese oi o biodi e si y
imp o emen and conse a ion. Howe e , despi e ha ing ecen ly been included as a na u al habi a o
communi y-wide in e es wi hin he EU Habi a s Di ec i e, hese sys ems a e in a c i ical si ua ion o decline.
Among o he ac o s, hey a e s ongly h ea ened by clima e change, he e ec s o which a e also expec ed o be
pa icula ly se e e in he Medi e anean egion. In his s udy, we aimed o e alua e he influence o clima e
a iabili y by examining p ima y p oduc ion indica o s and also o analyse whe he he geog aphical loca ion
may ha e a ole in he incidence o he ad e se e ec s o clima e.
Me hods: Co k oak AFS we e iden ified in he Fo es Map o Spain and he Land use map o Po ugal and ca e-
go ized on he basis o canopy co e . Seasonal clima e da a om 2001 o 2020 we e used o model ela ionships
wi h clima e p edic o s and p oximi y o he coas . Ho spo analysis was conduc ed o iden i y significan spa ial
clus e s o high- and low-e ficiency a eas.
Resul s: The esponses o he influence o clima ic condi ions di e ed among he a ious co k oak AFS ca ego ies,
pa icula ly in he o es ca ego y, which was less dependen on clima e a ia ions. Rela i e humidi y and wa e
a ailabili y we e he main d i e s o ne p ima y p oduc ion (NPP). Ca bon use e ficiency (CUE) was limi ed by
ela i e humidi y and sp ing empe a u e in open ecosys ems. P oximi y o he coas p o ed beneficial, especially
in yea s wi h ad e se wea he condi ions, bu was no a limi ing ac o o su i al o he ecosys em. Finally, he
esul s o he ho spo analysis suppo ed he o he findings, highligh ing high-e ficiency a eas close o he coas
and cold spo s g ouped in specific a eas o dispe sed inland.
Conclusions: Canopy plays a key ole in he influence o clima ic condi ions, pa icula ly in o es ca ego ies in
which a high densi y seems o gene a e mic oclima e condi ions. Wa e a ailabili y, bo h ia he soil and ai
mois u e, is he main d i e o p ima y p oduc ion, eflec ing di e en adap i e s a egies. The oceanic a mo-
sphe e may ac as a bu e in yea s o ex eme d ough .
1. In oduc ion
Clima e is conside ed one o he main d i e s o biodi e si y and
ecosys em change (Millennium Ecosys em Assessmen , 2005;IPBES,
2019) and is expec ed o become a majo s esso (Bella d e al., 2012;
U ban, 2015;Willei e al., 2019). The Medi e anean egion has been
conside ed one o he ho spo s in u u e clima e change (Gio gi, 2006),
and a p onounced inc ease in empe a u e (4 C–5C) and a conside able
dec ease in mean p ecipi a ion (o a ound 25%–30%), mainly in summe ,
a e expec ed (Gio gi and Lionello, 2008), hus se iously inc easing he
e ec s o summe d ough in he egion. D ough s will occu mo e
equen ly, and be o g ea e du a ion and in ensi y, wi h he added
agg a a ion in he Ibe ian Peninsula ha hey migh no be ollowed by
we win e s (B€
ohnisch e al., 2021). Fu he mo e, his end has been
confi med by he la es IPCC epo , which indica es ha he Medi e -
anean egion has been ge ing wa me and d ie in he las ew yea s
Abb e ia ions: AFS, ag o o es y sys ems; GPP, g oss p ima y p oduc ion; NPP, ne p ima y p oduc ion; CUE, ca bon use e ficiency; COS, Ca a de Uso e Ocupaç~
ao do
Solo de Po ugal Con inen al 2018; MFE, Fo es Map o Spain; TCD, ee co e densi y.
* Co esponding au ho .
E-mail add ess: jcpe ezgi [email protected] (J.C. P
e ez-Gi
on).
Con en s lis s a ailable a ScienceDi ec
Fo es Ecosys ems
jou nal homepage: www.keaipublishing.com/cn/jou nals/ o es -ecosys ems
h ps://doi.o g/10.1016/j. ecs.2022.100008
2197-5620/©2022 The Au ho s. Publishing se ices by Else ie B.V. on behal o KeAi Communica ions Co. L d. This is an open access a icle unde he CC BY license
(h p://c ea i ecommons.o g/licenses/by/4.0/).
Fo es Ecosys ems 9 (2022) 100008
(IPCC, 2021). This may cons i u e a h ea o he u u e o co k oak
ecosys ems, which a e al eady in a c i ical si ua ion o decline (Cos a
e al., 2009;Pin o-Co eia e al., 2011) in he Medi e anean egion,
whe e wa e a ailabili y is he main d i e o p ima y p oduc ion
(Reichs ein e al., 2007;Ga bulsky e al., 2010) and is a majo ecosys em
unc ion sensi i e o changes in clima e (Huang e al., 2019;Tang e al.,
2019;S ocke e al., 2019).
Co k oak (Que cus sube L.) is a ypically e e g een Medi e anean
ee species, wi h a ange o dis ibu ion expanding a ound he wes e n
Medi e anean basin, whe e he la ges popula ions o his species a e
ound, specifically in he sou hwes o he Ibe ian Peninsula (Día-
z-Fe n
andez e al., 1995;Pe ei a, 2007;Na a o Ce illo e al., 2013).
The species is well adap ed o mild Medi e anean clima es wi h A lan ic
influence, i.e. mild win e s and ho and d y summe s wi h high ela i e
humidi y (Pe ei a, 2007;Que o e al., 2008). The p esence o he species
in con inen al a eas depends on he e being some oceanic influence
(Na a o Ce illo e al., 2013). The op imum mean annual empe a u e
anges be ween 13 C and 19 C, and he species ole a es cold poo ly and
does no wi hs and pe iods o os , especially below 5C(Pe ei a,
2007;Gil and Va ela, 2008;Na a o Ce illo e al., 2013). The ain all
egime ole a ed is wi hin a wide ange, mainly g ea e han 500 mm and
eaching up o 2,400 mm (Gil and Va ela, 2008), and he species can
wi hs and up o 4 mon hs o summe d ough due o i s powe ul oo
sys em. Howe e , co k oak is e y sensi i e o wa e logging (Pe ei a,
2007). To deal wi h summe d ough , co k ees ha e de eloped deep
oo sys ems ha enable g oundwa e up ake (Da id e al., 2007;Piayda
e al., 2014) and physiological mechanisms ha p e en wa e loss, such
as e ficien s oma al con ol o anspi a ion (Na dini e al., 1999;
Media illa and Escude o, 2004;P
e ez e al., 2005;Da id e al., 2007;
Besson e al., 2014).
Cen u ies-long managemen o Q. sube , equen ly associa ed wi h
holm oak (Q. ilex L. and Q. o undi olia Lam.) and o a lesse ex en wi h
o he oaks (Q. aginea Lam. and Q. py enaica Willd.), has gi en ise o
mul i unc ional ag o o es y sys ems (AFS) (Jo e e al., 1999;Cos a
e al., 2009;Pin o-Co eia e al., 2011). In hese sys ems, known as
mon ados in Po ugal and dehesas in Spain, co k p oduc ion p e ails
oge he wi h he p oduc ion o o he goods and se ices, such as c ops
and o he non- imbe p oduc ions (mainly g azing and hun ing) (Pe ei a,
2007). Fo simplici y, we will e e o hese sys ems he ea e as dehesas.
The land co e pa e ns o dehesas a e simila o hose o sa anna,
cha ac e ized by he p esence o sca e ed ees in a ying densi ies
(A onson e al., 2009;Fonseca and Pin o-Co eia, 2009;Pin o-Co eia
e al., 2011). Howe e , he ee densi y is gene ally low, wi h he p es-
ence o he baceous o sh ub ege a ion in he unde s o y (Pe ei a e al.,
2007;Co eia e al., 2014,2016;Piayda e al., 2014). These sys ems ha e
been key elemen s o he landscape since ime immemo ial (Eichho n
e al., 2006;Fonseca and Pin o-Co eia, 2009), making aluable con i-
bu ions o he landscape and he en i onmen , ulfilling undamen al
unc ions and p ocesses such as p ima y p oduc ion, soil o ma ion and
egula ion o nu ien cycles o hyd ological flows, in addi ion o being
an impo an ese oi o biodi e si y imp o emen and conse a ion
(Plieninge and Wlb and, 2001;To alba e al., 2016). Dehesas ha e
ecen ly been included as a na u al habi a ype o communi y-wide in-
e es wi hin he EU Habi a s Di ec i e; in iew o hei conse a ion
s a us, hey ha e been ca ego ized as in se ious dange o disappea ance.
Al hough he p oblem is no new, al eady ha ing been epo ed in he
mid- wen ie h cen u y, co k oak AFS a e in a c i ical si ua ion o decline
(Cos a e al., 2009), and hei iabili y has been se iously h ea ened and
a ec ed by a ious ac o s (bo h na u al and human-induced) in addi ion
o clima e change (Aguile a e al., 2020), such as he p oli e a ion o pes s
and diseases (B asie e al., 1993;Gonz
alez e al., 2020), fi e ecu ence
(Sil a and Ca y, 2006;Guioma e al., 2015), lack o egene a ion,
change in land use, and land abandonmen (Pin o-Co eia and Mas-
ca enhas, 1999;Bugalho e al., 2011;Godinho e al., 2016). Thus,
ollowing a cascade eac ion (Haines-Young and Po schin, 2010), he loss
o ecological unc ions and p ocesses due o he impac o hese ac o s
(Sch €
o e e al., 2005,2019;Mooney e al., 2009), may igge a educ-
ion in he capaci y o p o ide ecosys em se ices, and consequen ly a
isk o human well-being.
The cu en s a e o co k oak AFS and he e ec s o clima e on he
de elopmen o hese ecosys ems can be assessed using indica o s. P i-
ma y p oduc ion indica o s, such as g oss p ima y p oduc ion (GPP) and
ne p ima y p oduc ion (NPP), ha e been widely modelled in ca bon
cycle- ela ed s udies a he global scale as hey a e clima e-sensi i e
(Huang e al., 2019;Tang e al., 2019;S ocke e al., 2019). GPP is he
o al amoun o ca bon s o ed by plan s, which akes in o accoun au o-
ophic espi a ion (Collal i and P en ice, 2019;Collal i e al., 2020).
Sub ac ion o au o ophic espi a ion gi es us he ne ca bon ans-
o med in o biomass (NPP). The NPP/GPP a io is a measu e o ca bon
use e ficiency (CUE), which ep esen s he e ficiency o plan s o
seques e ca bon om he a mosphe e h ough pho osyn hesis. In his
ype o assessmen , he use o open sou ce emo e sensing da a is e y
use ul as i p o ides con inuous, aluable in o ma ion on ecosys em
p oduc i i y o e la ge a eas.
The aims o he p esen s udy we e (i) o e alua e he influence o
clima e a iabili y on di e en co k oak AFS in he Ibe ian Peninsula
(ca ego ized by canopy co e ) using p oduc ion indica o s (NPP and
CUE) and (ii) o analyse whe he geog aphical loca ion may play a ole in
he incidence o he ad e se e ec s o clima e on hese ecosys ems.
2. Ma e ials and me hods
2.1. S udy a ea and clima e da a
The s udy ocused on mainland Po ugal and Spain (Ibe ian Penin-
sula) and did no include he islands (Fig. 1). The s udy egion is su -
ounded by wa e , mainly by he A lan ic Ocean and he Medi e anean
Sea, including 1,793 and 4,964 km o coas line o Po ugal and Spain,
espec i ely. The egion also comp ises 65% o he o al dis ibu ion
ange o Que cus sube (Caudullo e al., 2019).
The a ea is e y he e ogeneous in e ms o clima e and b oadly
speaking can be di ided in o h ee zones: d y clima e zones (widesp ead
in he sou h and sou heas ); empe a e zones wi h d y, ho summe s
(mos o he Ibe ian peninsula, i.e. app oxima ely 40% o i s su ace); and
empe a e zones wi h d y, empe a e summe clima es (mos o he
no heas o he Peninsula, as well as almos all o he wes coas o
mainland Po ugal) (AEMET, 2011).
Clima e da a we e ob ained by combining ERA5-Land mon hly
a e aged da a om 1981 o p esen (Mu~
noz Saba e , 2019)andERA5
mon hly a e aged da a on p essu e le els om 1979 o p esen
(He sbach e al., 2019). We download mon hly mean a iables om
2001 o 2020, o empe a u e, p ecipi a ion, ela i e humidi y, o al
e apo a ion, po en ial e apo a ion and olume o soil wa e a
di e en laye s (laye 1: 0–7cm;laye 2:7–28 cm; laye 3: 28–100 cm;
laye 4: 100–289 cm). All a iables we e p o ided a 0.10.1
spa ial esolu ion (ca. 9 km 9 km pixel size), excep ela i e hu-
midi y, which was dis ibu ed a 0.250.25spa ial esolu ion (ca.
31 km pixel size). These da a we e summa ized by season, i.e. o
win e , sp ing, summe and au umn mon hs.
Al hough p oximi y o he coas is no a clima ic a iable pe se,i isa
ac o ha a ec s clima e condi ions, wi h empe a u es being highe o
lowe and coas al a eas being we e han inland a eas. Thus, he sho es
dis ance be ween each Q. sube plo and he coas was compu ed.
2.2. Vege a ion p oduc i i y da a
The global MODIS da a collec ion was ob ained om he Land
P ocesses Dis ibu ed Ac i e A chi e Cen e (LP DAAC) da a pool. We
used he MOD17A2HGF.006 and MOD17A3HGF.006 p oduc s
(Running and Zhao, 2019a,2019b), which p o ide g oss p ima y p o-
duc ion (GPP) and ne p ima y p oduc ion (NPP) da a (in kg ca bon pe
m
2
) espec i ely, om 2001 o 2020, a 500-m esolu ion. GPP and NPP
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
2
alues o non- ege a ed o a ificial a eas we e excluded om he
analysis (Zhang e al., 2014), and he land pixel alues we e mul iplied
by a scale ac o o 0.0001 (Running and Zhao, 2015), as o de ed in he
me ada a file, o e u n he o iginal alue a hose pixels.
The GPP da a se (o iginally one o e e y 8 days) was used o
calcula e annual mean alues and ca bon use e ficiency (CUE)
(P
e ez-Gi
on e al., 2020). Annual CUE alues we e calcula ed as he
NPP/GPP a io, ep esen ing he e ficiency o plan s o seques e ca bon
om he a mosphe e h ough pho osyn hesis.
2.3. Plo selec ion
The digi al maps a ailable o Po ugal and Spain we e di e en , due
o he di e en pu poses o each. Fo Po ugal, we used he Ca a de Uso e
Ocupaç~
ao do Solo de Po ugal Con inen al 2018 (COS), de eloped wi h he
aim o cha ac e izing land co e in he coun y in 2018. In Spain, we used
he Fo es Map o Spain (MFE), a scale 1:25000 (MFE25) o 1:50000
(MFE50), depending on a ailabili y, as he map was c ea ed wi h he
in o ma ion cap u ed in na ional o es in en o ies ca ied ou a
di e en imes: he MFE50 p ojec was comple ed be ween 1997 and
2006, and he MF25 was de eloped a e 2007.
We selec ed a eas designa ed as “Flo es as de sob ei o”y“Supe ícies
ag oflo es ais (SAF) de sob ei o” om he COS maps as hose wi h
Q. sube p esence. To ha monize he selec ion c i e ia, we ep oduced he
classifica ion c i e ia used by COS in MFE, selec ing pa ches classified as
“Alco nocales”o “Dehesa”, p edomina ed by Q. sube and wi h a canopy
co e g ea e han 10%, he h eshold es ablished in he COS maps.
Finally, o ensu e ha GPP and NPP alues co esponded o Q. sube
AFS, we only selec ed he a eas i mo e han 80% o he MODIS pixels
we e ully occupied by a eas in which Q. sube AFS was p esen .
2.4. Ca ego iza ion
The social, economic and ecological impo ance o he co k oak has
led o i s managemen o e cen u ies, gi ing ise o di e en ee den-
si ies depending on he aim o he AFS (A onson e al., 2009;Fonseca and
Pin o-Co eia, 2009). T ee densi y seems o con ol edaphic (Galla do,
2003) and clima ic condi ions in his ype o ecosys em (Jo e e al.,
1999). I has ecen ly been demons a ed ha o es canopy has a bu -
e ing e ec on clima ic condi ions and hei a ia ions, e en gene a ing
mic oclima es (De F enne e al., 2021;Haesen e al., 2021).
Thus, in o de o add ess he di e en densi ies in he AFS, we used
he T ee Co e Densi y (TCD) 2018 om he high- esolu ion p oduc s
p o ided by he Cope nicus Land Moni o ing Se ice (Eu opean En i-
onmen Agency) o assign o each p e iously selec ed MODIS pixel he
a e age alue o he TCD pixels on which he MODIS pixel o e laps. The
selec ed pixels we e hen classified acco ding o TCD in o he ollowing
ca ego ies (Cos a e al., 2006): low-densi y AFS (<10%),
mode a e-densi y AFS (10%–25%), high-densi y AFS (25%–50%) and
o es s ands (>50%). Acco ding o Cos a e al. (2006) he a e age ex-
pec ed s and densi ies (in ees pe hec a e) a e 25–35 (18) o
low-densi y AFS, 36–42 (17) o mode a e-densi y AFS, 48–55 (18)
o high-densi y AFS, while he s and densi y o o es ca ego y will be
much highe . To check he ela ionship be ween TCD and s and densi y,
da a om he 3
d
Na ional Fo es In en o y (IFN3; om i s ac onym in
Spanish) we e used o assess whe he he s and densi y (numbe o ees
pe hec a e) inc eased wi h he TCD. Thus, based on TCD, he fi s h ee
ypes (low-, mode a e- and high-densi y AFS) can be conside ed open
o es sys ems, wi h low ee densi ies, while he o es ca ego y mo e
closely esembles a o es s and s uc u e, wi h canopy closu e expec ed
o c ea e a mic oclima e.
2.5. S a is ical analysis
We used mul iple linea eg ession o de e mine how co k oak AFS
p ima y p oduc ion was ela ed o clima e p edic o s and p oximi y o
he coas . Ho spo analysis was used o iden i y significan spa ial clus e s
o high- and low-e ficiency a eas.
Mul iple linea eg ession was used o model he ela ionships be-
ween p ima y p oduc ion indica o s (NPP and CUE) and he clima ic
a iables and p oximi y o he coas . An exhaus i e sea ch o he bes
p edic o subse s was pe o med using he b anch-and-bound algo i hm,
o es all possible combina ions o p edic o s o final selec ion o he
bes model (Na end a and Fukunaga, 1977). As his algo i hm e u ns he
bes model o each size, we limi ed he numbe o p edic o s o 5.
Candida e models we e compa ed using he adjus ed coe ficien o
de e mina ion (R
adj
2
) and he oo mean squa e e o (RMSE). To check
mul icollinea i y, we compu ed he a iance infla ion ac o (VIF) and
excluded p edic o s yielding VIF >10 (Mande ille, 2008).
Fo p oximi y o he coas , we used a uni a ia e linea eg ession
app oach and applied linea -log models. Spea man co ela ion coe fi-
cien was used o de e mine he s eng h o he linea ela ionship
Fig. 1. Range o dis ibu ion o Q. sube in he Ibe ian Peninsula and he di e en ypes o s udy plo s.
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
3
be ween p oximi y o he coas and clima ic a iables. A high R alue
indica es a s onge ela ionship, while a low R alue indica es he
opposi e. Posi i e R alues e eal he same end, while low R alues
e eal he opposi e end (Spea man, 1904).
Ho spo analysis was conduc ed by applying he local Ge is-O d Gi*
s a is ic (Ge is and O d, 1992) o iden i y high and low CUE a eas (ho spo s
and cold spo a eas) wi hin he selec ed pixels. The analysis only included
he CUE as his alue es ablishes he h eshold o he physiological ac i i y
o plan s (Am ho , 2000;Van Ie sel, 2003;Kei h e al., 2010). Local spa ial
au oco ela ion was measu ed o assess how each da a poin is su ounded
by o he da a poin s (neighbou hood) wi h simila ly high o low alues.
The me hod e u ns a Z-sco e and p- alue o each da a poin assuming a
no mal dis ibu ion. Z-sco es g ea e han 1.96 and less han 1.96 a e
s a is ically significan a p<0.05, while la ge posi i e and nega i e
alues indica e g ea e clus e ing.
The leaps lib a y (Lumley, 2013) implemen ed in he R so wa e
en i onmen (RCo eTeam,2020)wasused ofi mul iple linea eg es-
sion based on b anch-and-bound algo i hms. G aphical analyses we e
conduc ed wi h he ggplo 2 package (Wickham, 2009). Significan di e -
ences we e de e mined using he Wilcoxon-Mann-Whi ney es (a
α
¼
0.01).
3. Resul s
3.1. P ima y p oduc ion indica o ends
The 870420 AFS plo s we e dis ibu ed as ollows: 11.6% low-
densi y, 28.9% mode a e-densi y, 44.4% high-densi y and 15.1% o es
plo s. T ee densi y di e ed significan ly be ween he di e en g oups
based on o es in en o ies (Figu e S1 in supplemen a y ma e ial), excep
be ween he low-densi y and mode a e-densi y g oups. A la ge di e ence
in ee densi y be ween high-densi y plo s (TCD be ween 25% and 50%)
and o es plo s (TCD >50%) was no ed, and he e o e la ge di e ences
in he influence o clima ic condi ions we e expec ed.
The ends in GPP and NPP in ela ion o TCD showed ha he indi-
ca o alue and dispe sion inc eased wi h canopy co e (Fig. 2). The
mean NPP anged om 0.655 kg C∙m
2
∙y
1
o low-densi y plo s o
0.985 kg C∙m
2
∙y
1
o he o es plo s, wi h maximum alues o up o
2.21 kg C∙m
2
∙y
1
eached. In e sely, he CUE alue dec eased as he
canopy co e inc eased, wi h mean alues anging om 0.64 in low-
densi y plo s o 0.58 in o es plo s and maximum alues eaching up
o 0.72 we e eached in each o hese ypes o plo s; he minimum alue
was 0.30 in high-densi y and o es plo s.
Rega ding annual ends, he mean alues we e main ained (Fig. 3a).
In gene al, he dispe sion ange and a e age NPP inc eased wi h canopy
co e , while CUE dec eased as canopy co e inc eased. Howe e ,
gene alized sligh dec eases in NPP we e obse ed, pa icula ly in 2005,
bu we e no ela ed o CUE. The minimum mean NPP alues o he
his o ical se ies we e eco ded in 2005, eaching alues o 0.474 kg
C∙m
2
∙y
1
in he low-densi y plo s, 0.544 kg C∙m
2
∙y
1
in mode a e-
densi y plo s, 0.611 kg C∙m
2
∙y
1
in he high-densi y plo s and 0.898
kg C∙m
2
∙y
1
in he o es plo s. This ep esen s dec eases o mo e han
20% o low-, mode a e- and high-densi y plo s, while he dec ease in he
o es plo s was only 11%. The dec ease in NPP alues was immedia ely
eco e ed in he ollowing yea wi h inc eases o he same magni ude as
he dec eases. The mean CUE alues o his pe iod su e ed dec eases o
less han 2% ha las ed 2 yea s, i.e. eco e y o he alues be o e he
la ge dec ease in NPP occu ed in 2007 (no he la ges dec ease in he
his o ical ange). In 2012 and wi hou any p eceden , CUE alues below
0.3 we e obse ed o a o al o 97 samples in he o es ca ego y. Simila
findings we e obse ed in 2015, 2017 and 2020, in espec i ely 172, 87
and 201 samples in he o es ca ego y. Ra he han being an isola ed
e en , his seems o ha e become a mo e equen end in ecen yea s,
fi s a ec ing he high-densi y plo s in 2015.
3.2. Influence o clima e ac o s
The pa ame e es ima es and goodness o fi s a is ics o he mul iple
linea eg ession models based on canopy co e a e summa ized in
Table 1. NPP showed a good fi in all ca ego ies, wi h he a iance
explained inc easing wi h canopy co e (R
adj
2
anging om 0.64 o 0.72),
excep o o es ca ego y, which yielded he lowes a iance explained
(R
adj
2
¼0.59). On he o he hand, CUE also showed a good fi inlow-,
mode a e-, and high-densi y plo s, wi h he a iance explained inc easing
wi h canopy co e (R
adj
2
om 0.46 o 0.57) as obse ed o NPP. The
goodness o fi o CUE in o es ca ego y was e y low, which may indica e
ha he CUE is influenced by ac o s o he han clima e ac o s.
Rela i e humidi y in he summe mon hs ( hm_Summe ) and olume
o soil wa e in he fi s wo laye s o soil (swl1 and swl2) in he summe
mon hs we e he mos impo an p edic o s o NPP models o low-,
mode a e- and high-densi y plo s. The influence was simila in models:
NPP benefi ed om an inc ease in ela i e humidi y and olume o soil
Fig. 2. Box-and-whiske plo showing he mean ends in GPP, NPP and CUE. The do s ep esen ou lie s.
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
4
wa e o laye 2, bu was nega i ely influenced by an inc ease in olume
o soil wa e in laye 1. The mos impo an p edic o s in he o es plo s
changed sligh ly. Rela i e humidi y in he summe mon hs was main-
ained, bu he olume o soil wa e was eplaced by o al p ecipi a ion in
summe ( p_Summe ). In addi ion o changing he p edic o s, he influ-
ence also changed, wi h he lowe ain all du ing he summe mon hs
leading o highe NPP alues.
Rega ding CUE, bo h ela i e humidi y in he summe mon hs
( hm_Summe ) and mean empe a u e o sp ing mon hs ( 2m_Sp ing)
we e cons an h oughou he models, and ela i e humidi y had a posi i e
influence, as occu ed wi h NPP. The mean empe a u e o sp ing mon hs
nega i ely a ec ed he CUE. The a iables ha changed we e o al
e apo a ion in summe (e_Summe ) in he low-densi y model and olume
o soil wa e in laye 3 in sp ing (swl3_Sp ing) in he high-densi y model.
Fig. 3. Box-and-whiske plo showing he annual ends in GPP, NPP and CUE. The do s ep esen ou lie s. Dashed black lines indica e he h eshold o ecosys ems a
po en ial isk (CUE ¼0.3).
Table 1
Resul s o mul iple linea eg ession showing he bes models ob ained o NPP and CUE.
AFS Ca ego y Dependen a iable Independen a iable Pa ame e es ima e S d. E o RMSE R
fi
2
Low-densi y NPP (Kg C∙∙m
–2
∙∙y
–1
)(In e cep ) 0.6889 0.0004 0.1289 0.6438
hm_Summe 0.1845 0.0005
sw l1_Summe 0.2162 0.0012
sw l2_Summe 0.2119 0.0013
CUE (In e cep ) 0.6426 0.0001 0.0239 0.4606
e_Summe 0.0058 0.0001
hm_Summe 0.0210 0.0001
2m_Sp ing 0.0096 0.0001
Mode a e-densi y NPP (Kg C∙∙m
–2
∙∙y
–1
)(In e cep ) 0.7529 0.0002 0.1123 0.6832
hm_Summe 0.1702 0.0003
sw l1_Summe 0.1795 0.0007
sw l2_Summe 0.1817 0.0007
CUE (In e cep ) 0.3065 0.0000 0.4719 0.5039
hm_Summe 0.4485 0.0011
2m_Sp ing 0.1717 0.0010
High-densi y NPP (Kg C∙∙m
–2
∙∙y
–1
)(In e cep ) 0.8058 0.0002 0.1152 0.7243
hm_Summe 0.1828 0.0002
sw l1_Summe 0.1696 0.0005
sw l2_Summe 0.1740 0.0005
CUE (In e cep ) 0.6336 0.0000 0.0226 0.5713
hm_Summe 0.0206 0.0000
sw l3_Sp ing 0.0044 0.0000
2m_Sp ing 0.0098 0.0000
Fo es NPP (Kg C∙∙m
–2
∙∙y
–1
)(In e cep ) 0.9337 0.0005 0.1827 0.5747
hm_Summe 0.2047 0.0005
p_Summe 0.0342 0.0003
CUE (In e cep ) 0.5828 0.0002 0.0701 0.1146
sw l1_Summe 0.0232 0.0003
sw l3_Summe 0.0223 0.0003
2m_Summe 0.0295 0.0002
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
5
3.3. P oximi y o he coas
The ela ionship be ween he p ima y p oduc ion indica o (NPP and
CUE) and he p oximi y o he coas dec eased sha ply o sho dis-
ances, slowing down o becoming asymp o ic as he dis ance inc eased.
The e o e, a linea -log model was cons uc ed: he goodness o fi s a-
is ics o NPP and CUE a e shown in Fig. 4. In gene al, he R
adj
2
alues o
each indica o and g oup a e simila o he R
adj
2
alues ob ained o cli-
ma ic ac o s. In he case o NPP, he ou comes we e e y clea , wi h he
indica o alues inc easing as he dis ance o he coas dec eased. A
weake ela ionship was ob ained o CUE, al hough he end con inued
o be obse ed. Finally, o es CUE was no ela ed o he p oximi y o
coas , appa en ly due o a la ge numbe o poin s nea he coas wi h low
CUE alues, which may also explain he lack o any ela ionship be ween
CUE and clima ic ac o s. Annual models (Figu es S2, S3, S4 and S5 in
Supplemen a y ma e ial) ollowed he same ends shown he e. How-
e e , dependency was highe in some yea s, i.e. 2005 and 2017.
P oximi y o he coas was highly and nega i ely co ela edwi h sp ing
and summe ela i e humidi y ( ¼0.67 and 0.81, espec i ely),
nega i ely co ela ed wi h a e age empe a u e o au umn and win e (
¼0.52 and 0.54, espec i ely) and posi i ely co ela ed wi h summe
a e age empe a u e ( ¼0.48). The a iable was no co ela ed o only
sligh ly co ela ed wi h p ecipi a ion and soil wa e con en .
3.4. Ho spo analysis
The Gi* s a is ics e ealed clus e s o Q. sube AFS wi h high and low
CUE wi hin he s udy a ea, espec i ely co esponding o ho spo and
cold spo a eas (Fig. 5). In gene al, he la ges ho spo a eas occu ed in
he sou hwes o Po ugal and sou h and no heas o Spain, always close
o he coas , wi h la ge clus e s in low-, mode a e- and high-densi y plo s,
mainly in he Alen ejo and Alga e egions (Po ugal). The cold spo
a eas we e gene ally loca ed in inland a eas, usually spa sely dis ibu ed,
bu o ming la ge clus e s in high-densi y and o es plo s. He e we can
dis inguish h ee la ge clus e s mainly loca ed in Los Alco nocales na -
u al pa k (sou h o Spain), Sie a de Ho nachuelos and Sie a No e de
Se illa na u al pa ks (no h o he Guadalqui i alley) and sou h o
Cace es (cen al Spain). Fo es cold spo s we e almos absen om AFS
plo s in Po ugal.
Compa a i e box-and-whiske plo s indica e significan di e ences
be ween ho spo and cold spo CUE a eas ega ding he main
independen a iables selec ed in CUE mul iple linea eg ession models
(Fig. 6). The end shown is he same o he di e en ypes o AFS, and
he di e ences be ween ho spo s and cold spo s we e sligh ly smalle
only in he o es ca ego y. Ho spo s we e iden ified in coas al zones
while cold spo s we e ound in inland a eas (Fig. 6a). The a e age dis-
ance o he coas was abou 10 km o ho spo s and mo e han 120 km
o cold spo s, excep o he o es ca ego y, o which he dis ances we e
sho e , a ound 30 km. The dispe sion was also ema kable as ho spo s
a ied e y li le and he alues we e g ouped e y close o he mean,
wi h he in e se obse ed o in cold spo s. Rela ed o he abo e, ela i e
humidi y du ing he summe mon hs was highe in ho spo s (Fig. 6b),
wi h a e age alues a ound 68%–70%, while o cold spo s i was a ound
47%–48%, excep o he o es ca ego y, o which he alues we e
sligh ly highe (61%). Finally, he ends in a e age sp ing empe a u e
(Fig. 6c) we e e y simila , ega dless o he canopy co e , wi h alues
be ween 16.5 C and 18 C in he ho spo s and be ween 1 C and 2 Cin
he cold spo s.
4. Discussion
4.1. Changes wi h s and densi y
The mul i unc ional cha ac e o adi ionally managed co k oak
ag o o es y sys ems has condi ioned hei s uc u e and composi ion
(Bugalho e al., 2011), mainly cha ac e ized by a low ee densi y wi h
he p esence o he baceous o sh ub ege a ion in he unde s o y.
Consequen ly, he con ibu ion o each pa o he s uc u e o p ima y
p oduc ion is di e en . In he dehesa ecosys em, unde s o y species
con ibu e be ween one hi d and one hal o he o al GPP, and conse-
quen ly, abou hal o wo- hi ds o he GPP is con ibu ed by ees
(Dubbe e al., 2014;Co eia e al., 2016). The e o e, in low-, mode a e-
and high-densi y ca ego ies he con ibu ion o biomass should be
conside ed in ela ion o he ecosys em as a whole, a he han in ela ion
o isola ed species as e.g. in monospecificQ. sube s ands. Howe e , in he
o es ca ego y, ees con ibu e mo e han hal he alue o he p ima y
p oduc ion indica o and hus es ablish a ela ionship wi h he species.
The esul s ob ained a e sensi i e o changes in canopy co e , high-
ligh ing di e en ends be ween he di e en ca ego ies as he canopy
co e inc eases and especially wi h he o es ca ego y, whe e he e ec s
may be a ibu ed o he species. Bo h GPP and NPP alues inc eased wi h
canopy co e . The highes GPP and NPP alues we e associa ed wi h he
Fig. 4. Linea -log models o NPP and CUE g ouped by canopy co e ca ego ies. No e ha “x”in he model o mula indica es log(x).
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
6
o es ca ego y and ela ed o a la ge numbe o ees and a g ea e
con ibu ion o he o al amoun o plan biomass o ees o he in-
dica o s han he baceous o sh ub ege a ion. By con as , he opposi e
e ec was obse ed o influence o he canopy co e on he CUE, wi h
a e age CUE alue dec easing and he o al numbe o low- alue ou lie s
inc easing as canopy co e inc eased. These sligh ly low CUE alues,
ela i e o he ca ego ies wi h lowe ee canopy densi y, a e consis en
wi h he s and de elopmen , as hal s in g ow h a e no ollowed by hal ed
espi a ion, hus leading o a dec ease in CUE (Collal i and P en ice,
2019).
Howe e , he inc easingly common e y low ex eme CUE alues in
2012, 2015 and 2020 in high-densi y AFS and o es ca ego ies a e o
Fig. 5. Loca ion o ho spo ( ed poin s) and cold spo (blue poin s) a eas o Q. sube AFS based on CUE and g ouped by canopy co e ca ego ies: a) low-densi y, b)
mode a e-densi y, c) high-densi y and d) o es . (Fo in e p e a ion o he e e ences o colou in his figu e legend, he eade is e e ed o he Web e sion o
his a icle.)
Fig. 6. Box-and-whiske plo compa ing a) p oximi y o he coas , b) summe ela i e humidi y and c) sp ing empe a u e in ho spo s and cold spo CUE a eas.
S a is ical significance: ns: p>0.05; *: p0.05; **: p0.01; ***: p0.001; ****: p0.0001. The black do s ep esen ou lie s. The ed do s ep esen mean alues.
(Fo in e p e a ion o he e e ences o colou in his figu e legend, he eade is e e ed o he Web e sion o his a icle.)
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
7
pa icula conce n. These alues a e e y close o he es ablished
h eshold o 0.2, which may limi he physiological ac i i y o plan s
(Am ho , 2000;Van Ie sel, 2003;Kei h e al., 2010) by inc easing he
addi ional espi a o y cos , educing g ow h and new issue o ma ion, o
e en leading o plan collapse, hus indica ing ha ecosys ems will
po en ially be a isk (P
e ez-Gi
on e al., 2020). Howe e , such low
alues we e no obse ed in he ca ego ies wi h lowe canopy co e (low-
and mode a e-densi y ca ego ies), which may be ela ed o he sca ci y o
ees. The unde s o ey ege a ion in co k oak AFS may mainly comp ise
annual species o c ops, he g ow h pe iod o which depends on he
amoun o ain all and i s seasonal dis ibu ion, o sh ub ege a ion well
adap ed o he Medi e anean summe d ough s ess (Co eia e al.,
2016). Howe e , his is no possible wi h ees, as he wea he condi ions
do no de e mine whe he he ees a e ali e o dead, as ees can ole a e
a ange o condi ions, wi h modified main enance cos s a ec ing ca bon
assimila ion. This di e ence would he e o e explain he absence o low
CUE alues in he low- and mode a e-densi y ca ego ies.
The s udy findings also highligh he highe CUE o managed eco-
sys ems han o unmanaged ecosys ems (Fe n
andez-Ma ínez e al.,
2014;Campioli e al., 2015), sugges ing ha he conse a ion and u u e
o hese sys ems - as well as he biodi e si y hey main ain and he
ecosys em se ices hey p o ide - a e s ongly dependen on human
managemen (Pe ei a and da Fonseca, 2003;Bugalho e al., 2011).
Howe e , some changes appea o be aking place in hese ecosys ems as
he same pa e n was no obse ed in o es plo s and high-densi y AFS
wi hin he his o ical pe iod conside ed, indica ing ha a change in some
ac o has been accen ua ed in ecen yea s. He e, di e en d i e s such
as economic (e.g. declining p ofi abili y o adi ional dehesa p oduc s o
ag icul u al in ensi y), socio-cul u al (e.g. u al exodus), poli ical (e.g.
a ailabili y o access o di ec paymen s o EU Common Ag icul u al
Policy (CAP)), echnological (e.g. new echnics) and na u al (e.g. clima e
change) ac o s may be a ec ing hese sys ems in a complex and simul-
aneous way (Pin o-Co eia e al., 2011;Plieninge e al., 2021).
4.2. Responses o NPP and CUE o clima e influence
P e ious s udies ha e claimed ha wa e is he main d i e o p ima y
p oduc ion in co k oak AFS, mainly linking p oduc i i y o he limi a-
ions caused by summe d ough s (Pe ei a e al., 2007;Piayda e al.,
2014;Ramos e al., 2015;Co eia e al., 2016). Ou findings also indica e
wa e limi a ions in summe as he main d i e o NPP models in all
dehesa ca ego ies, bu o a g ea e ex en in open ecosys ems, i.e. ca e-
go ies wi h low canopy co e (<50%), han in sys ems wi h a o es s and
ype s uc u e. This sugges s ha summe is a c i ical season o co k oak
AFS, ega dless o he canopy co e . Fu he mo e, al hough wa e
a ailabili y was also a limi ing ac o in he CUE models in ou s udy,
ewe limi a ions we e obse ed han in he NPP models, as o he ac o s
we e in ol ed.
The p oduc i i y o any plan depends on i s abili y o main ain
pho osyn he ic issues wi h an adequa e wa e le el. When wa e is
limi ing, wa e loss in plan s is minimized ia anspi a ion, wi h closu e
o s oma a es ic ing he en y o CO
2
and hus also limi ing pho osyn-
hesis (Media illa and Escude o, 2004;Rzigui e al., 2018;Da id-Sch-
wa z e al., 2019;G ossio d e al., 2020). Al hough his is a ypical
adap i e mechanism in Medi e anean species, such as Q. sube and
Q. ilex, p o ec ing agains summe d ough , ees gene ally
display s oma al con ol o e anspi a ion (Na dini e al., 1999;P
e ez
e al., 2005;Da id e al., 2007;Besson e al., 2014); d ough also
induces lea senescence in he unde s o ey ege a ion, and all o hese
ac o s lead o a dec ease in p ima y p oduc ion le els in he ecosys em
(Pe ei a e al., 2007). Rela i e humidi y plays a key ole in his p ocess
(closely ela ed o apo p essu e defici : G ossio d e al., 2020), because
he anspi a ion a e alls as ela i e humidi y inc eases. This explains
why ela i e humidi y in he summe mon hs is he mos impo an
a iable in all o ou models and why i is posi i ely ela ed o NPP. The
ela ionship wi h CUE is clea , as he main enance cos inc eases wi h he
anspi a ion a e (Am ho , 2000;Van Ie sel, 2003), and he main enance
cos will he e o e be lowe a high ela i e humidi y, and he CUE will
inc ease.
When soil wa e is a ailable o plan s, anspi a ion a es a e mainly
con olled by clima ic ac o s. Howe e , when he soil wa e becomes
limi ing, he anspi a ion a e alls (Ga dne and Ehlig, 1963). Thus, in
he NPP models, in addi ion o being a ec ed by ela i e humidi y, he
low-, mode a e- and high-densi y ca ego ies we e also posi i ely a ec ed
by wa e olume in he second laye o soil (swl2; be ween 7 and 28 cm)
and nega i ely by soil wa e olume in he fi s 7 cm (swl1). In open
ecosys ems such as dehesas, he fi s laye o soil is pa icula ly sensi i e
o wa e loss h ough e apo a ion, which will o m pa o he ela i e
humidi y o he ai ; hus, a ou ing an inc ease in e apo a ion would
benefi he ecosys em o a g ea e ex en . F om he poin o iew o wa e
abso p ion by plan s, wa e is mo e beneficial in deepe laye s, whe e he
oo s o mos ees and o he unde s o y ege a ion occu (Da id e al.,
2007;Baldocchi e al., 2010;Co eia e al., 2016). This explains he
posi i e ela ionship wi h he p esence o wa e in he second laye o
soil. Al hough g ea e limi a ion o he amoun o wa e in he hi d laye
o soil was expec ed due o he loca ion o mos oo s ( om 28 cm o 1 m),
we belie e ha hese findings may be explained by daily fluc ua ions in
soil wa e con en due o hyd aulic li p ocesses (Da id e al., 2007). The
wa e ises o he uppe soil laye s whe e i becomes a ailable bo h o he
oak ee oo s and unde s o ey ege a ion (Ma a~
n
on e al., 2009).
Tempe a u e only appea s o influence CUE, especially he a e age
empe a u e in sp ing. Tempe a u e plays a undamen al ole because i
a ec s bo h pho osyn hesis and au o ophic espi a ion (R
a
), he a es o
which inc ease exponen ially wi h empe a u e, hus inc easing he
main enance cos (Ryan, 1991;Ryan e al., 1994). As his only occu s in
low-, mode a e- and high-densi y plo s, and as Q. sube is adap ed o
Medi e anean clima es cha ac e ized by high empe a u es, his pa e n
is associa ed wi h he ege a i e pe iod o he unde s o ey ege a ion
(Dubbe e al., 2014;Co eia e al., 2016). Highe empe a u es in ea ly
sp ing may accele a e he ac i a ion and ge mina ion p ocesses in annual
plan s, a ou ing g ea e pho osyn he ic ac i i y and he e o e g ea e
CUE. Howe e , highe empe a u es in la e sp ing a e also likely o ha e
he opposi e e ec , sho ening he du a ion o pho osyn he ic ac i i y by
ad ancing senescence. Fu he mo e, empe a u e a ec s bo h he plan
pa o he ecosys em and also o he clima ic ac o s such as ela i e
humidi y and soil wa e con en .
Finally, he findings sugges s uc u al and composi ional di e ences
be ween low-, mode a e- and high-densi y AFS and he o es plo s.
Unde o es canopy co e , he amoun o pho osyn he ically ac i e a-
dia ion and wind speed a e educed, di ec ly educing he a ia ions in
humidi y and empe a u e and ex eme e en s (De F enne e al., 2021;
Haesen e al., 2021). The e ec s o canopy h ough all on he soil cha -
ac e is ics a e also modified by he con ibu ion o o ganic ma e
(Ma a~
n
on e al., 2009). All o his is ansla ed in o di e ences in p ima y
p oduc ion and he associa ed ac o s. Thus, i was only possible o model
he NPP esponse, and he goodness o fi was lowe han ha o he
p e ious model; i was no possible o model he CUE, possibly because
his a iable does no depend on he p edic o s used. Simila ly, he un-
expec ed and nega i e e ec o summe ain all on NPP may be ela ed o
he ac ha he basic needs o humidi y and wa e a e co e ed in hese
s ands. Hea y s o ms occu a he end o he summe , which may cause
wa e logging ha is damaging o he ecosys em (Pe ei a, 2007). The
s o ms can also p o oke fluc ua ions anging om flooding o wa e
deficiency, he la e o which is a ou ed in pe iods o high empe a u es
and can po en ia e he sp ead and in ec i i y o Phy oph ho a species
(Gonz
alez e al., 2020).
4.3. Oceanic influence as a bu e o ex eme d ough s
The s udy findings highligh ed he influence o oceanic clima ic
condi ions on he NPP o co k oak AFS, wi h p oximi y o he coas
p o iding a clea benefi . Simila ly, he dec ease in NPP wi h inc easing
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
8
dis ance om he coas does no seem o be limi ing o hese ecosys ems
bu a he is asymp o ic, ei he because he ee species is wi hin i s
dis ibu ion ange (Gil and Va ela, 2008;Caudullo e al., 2019)o
because i is adap ed o he d ies clima ic condi ions (Da id e al., 2007;
Besson e al., 2014).
The s udy findings eflec a g adien o change along he dehesa
ca ego ies. In he low-densi y ca ego y, he highes NPP alues, mainly
con ibu ed by he unde s o y (Dubbe e al., 2014;Co eia e al., 2016),
a e lowe han o he ca ego y immedia ely abo e he same dis ance
om he coas , up o he highes alues ound in he o es ca ego y. The
o es ca ego y, coinciding wi h he ecological esponse o Q. sube ,is
benefi ed by a mild oceanic clima e wi h high ela i e humidi y (be ween
65% and 80%) (Que o e al., 2008), which is also consis en wi h he
obse ed di e ences be ween ho spo s and cold spo s (Fig. 6). The e o e,
unde op imal clima e condi ions a g ea e numbe o ees will be mo e
p oduc i e. In addi ion, he high and nega i e co ela ion be ween he
dis ance o he coas and he summe ela i e humidi y again sugges s
ha ela i e humidi y in he summe mon hs is a limi ing ac o o all
co k oak AFS as i is closely linked o s oma al closu e (Da id-Schwa z
e al., 2019). A simila bu sligh ly weake pa e n was obse ed o CUE,
he le els o which a ied wi h he p oximi y o he coas wi hin anges o
dis ance ha gene ally do no a ec he physiological unc ioning o he
ecosys ems. In his las case, he mic oclima e gene a ed by he o es (De
F enne e al., 2021;Haesen e al., 2021) seems o be su ficien o main-
ain he condi ions necessa y o adequa e ca bon assimila ion, wi hou
depending on he coas al influence.
The findings o he ho spo analysis we e consis en wi h hese p e-
ious esul s, highligh ing high-e ficiency a eas nea he coas line and
cold spo s g ouped in specific a eas o dispe sed inland. Howe e ,
conside ing he CUE alue, some clus e s o poin s sugges some deg ee
o isk o he ecosys em (P
e ez-Gi
on e al., 2020). These CUE alues a e
e y low (gene ally below 0.3) bo h in high-densi y and o es ca ego ies
(Fig. 4) and a e pa icula ly no able in he la e a sho dis ances o he
coas line, in he sou h o Spain (Fig. 5), and conside ing he annual dis-
ibu ion was mo e p onounced in 2012, 2015, 2016, 2017 and 2020
(Fig. 3). Acco ding o he his o ic d ough da abase o Spain (Vice-
n e-Se ano e al., 2017), d ough s we e eco ded in 2005, 2012, 2015,
2017 and 2019, wi h hose occu ing in 2005 and 2017 being he mos
ex eme. Fo hese wo yea s in pa icula , he goodness-o -fi o ou
linea -log models inc eased significan ly (Figu es S2, S3, S4 and S5 in
Supplemen a y ma e ial), which sugges s ha in ex eme d ough yea s,
p oximi y o he coas may bu e co k oak AFS om ex eme clima e
condi ions. This u he confi ms he sugges ions o Piayda e al. (2014),
who a gued ha he high apo p essu e defici ( ela ed o empe a u e
and ela i e humidi y (G ossio d e al., 2020) ound in co k oak eco-
sys ems in Po ugal may be a consequence o he lack o en y o oceanic
ai masses.
5. Conclusions
The esponses o he di e en TCD-based ca ego ies o co k oak AFS
o p ima y p oduc ion indica o s sugges ha canopy co e and hence
ee densi y play a key ole in he influence o clima ic condi ions. Fo es
plo s can main ain mic oclima ic condi ions ha make hem less
dependen on en i onmen al condi ions, while AFS plo s wi h an open
ecosys em (lowe densi ies) depend on mac oclima e condi ions.
The e o e, wi hin he same ecosys em, he esponse o clima e change
may a y depending on ee densi y.
Rega ding he influence o clima e a iabili y, ou findings showed ha
he esponses o he ecosys ems eflec he ecological ai s and he
di e en adap i e s a egies used by he componen ees and unde s o y
plan s o su i e d ough seasons, whe e wa e (soil o ai mois u e) is he
main d i e o p ima y p oduc ion. Rela i e humidi y is associa ed wi h
anspi a ion and wa e loss h ough closu e o s oma a, which will a y
depending on he se e i y o wa e deficiency, limi ing pho osyn hesis o a
g ea e o lesse ex en . A he same ime, bo h he hyd aulic li ing p o-
cesses and he deep oo s allow he ees o ake ad an age o he
g oundwa e om he deepe laye s and make i a ailable o unde s o y
ege a ion wi h shallow, bu no supe ficial, oo s. Tempe a u e only
seems o influence CUE in open ecosys ems (low-, mode a e-, and high-
densi y) in which he unde s o y laye makes a g ea e con ibu ion o
p ima y p oduc ion. In pa icula , an inc ease in sp ing empe a u e could
ad ance he g owing season bu could also sho en he g ow h pe iod o
annual plan s and inc ease he main enance cos .
Se e al ac o s a ec hese ecosys ems and do so in a complex way,
and i is he e o e di ficul o isola e he indi idual e ec s. Fo example,
ela i e humidi y and he p oximi y o he coas a e closely ela ed, hus
influencing he ca bon balance in co k oak AFS. Ou findings show ha
p oximi y o he coas imp o es p oduc i i y le els and may also bu e
clima e condi ions in ex eme d ough yea s, educing he associa ed
ad e se e ec s and he isk o he ecosys em. The e o e, in u u e clima e
change scena ios, in which he Medi e anean egion is expec ed o be
one o he mos se e ely a ec ed and d ough s a e expec ed o be mo e
equen , p olonged, and in ense, an impo an isk o loss o ecosys ems
and hei associa ed unc ions will appea . This will a ec all ecosys ems,
al hough inland ecosys ems - whe e he fi s dis u bances ha e al eady
been de ec ed - may no be bu e ed agains he oceanic influence and
would he e o e be pa icula ly a ec ed, while coas al a eas, such as
sou hwes e n Po ugal, may cope be e .
E hics app o al and consen o pa icipa e
We ha e no e hical conce ns o decla e. All au ho s consen ed o
pa icipa e on his manusc ip .
Consen o publica ion
All au ho s ag ee wi h he con en o his manusc ip and i s publi-
ca ion in Fo es Ecosys ems.
A ailabili y o da a and ma e ial
The da ase s analysed du ing he cu en s udy a e eely a ailable
om he ollowing hos s. ERA5-Land mon hly a e aged da a om 1981
o p esen and ERA5 mon hly a e aged da a on p essu e le els om 1979
o p esen can be accessed om he Cope nicus Clima e Change Se ice
(C3S) Clima e Da a S o e (CDS) (h ps://doi.o g/10.24381/cds
.68d2bb30 and h ps://doi.o g/10.24381/cds.6860a573). MODIS
MOD17A2HGF.006 and MOD17A3HGF.006 p oduc s can be accessed
om he Land P ocesses Dis ibu ed Ac i e A chi e Cen e (LP DAAC)
da a pool (h ps://lpdaac.usgs.go /p oduc s/mod17a2hg 006/and
h ps://lpdaac.usgs.go /p oduc s/mod17a3hg 006/). Fo es Map o
Spain can be accessed om he Minis e io pa a la T ansici
on Ecol
ogica y
el Re o Demog
afico webpage (h ps://www.mi eco.gob.es/es/ca o
g afia-y-sig/ide/desca gas/biodi e sidad/m e.aspx). Ca a de Uso e
Ocupaç~
ao do Solo de Po ugal Con inen al 2018 can be accessed om he
Sis ema Nacional de In o maç~
ao Geog
afica O Di eç~
ao-Ge al do Te i-
o io (h ps://snig.dg e i o io.go .p / ndg/s /po /ca alog.sea ch).
Funding
JCPG is in eceip o a “Se e o Ochoa”PhD G an p o ided by he
Go e nmen o P incipado de As u ias (PA-18-PF-BP17-026).
Au ho s’con ibu ions
JCPG, PAA, and EDV oge he concei ed he esea ch idea; JCPG
collec ed and analysed da a and w o e he ini ial d a o he manusc ip ;
all au ho s con ibu ed o he manusc ip e iew and edi ing. All au ho s
ead and app o ed he final manusc ip .
J.C. P
e ez-Gi
on e al. Fo es Ecosys ems 9 (2022) 100008
9
