Fo es Ecology and Managemen 517 (2022) 120268
A ailable online 9 May 2022
0378-1127/© 2022 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/).
G ow h his o y o polla ded black popla s in a con inen al Medi e anean
egion: A pa adigm o anishing landscapes
J. Julio Cama e o
a
,
*
, Es e Gonz´
alez de And ´
es
a
, Michele Colangelo
a
,
b
, Chabie de Jaime
Lo en
c
a
Ins i u o Pi enaico de Ecología (CSIC), Apdo. 202, 50192 Za agoza, Spain
b
School o Ag icul u al, Fo es , Food and En i onmen al Sciences (SAFE), Uni e si y o Basilica a, 85100 Po enza, I aly
c
Pa que Cul u al del Chopo Cabece o del Al o Al amb a, 44159 Aguila del Al amb a, Spain
ARTICLE INFO
Keywo ds:
Dend oecology
Floodplain o es
Land-use changes
Polla ded ees
Populus nig a
Ri e low
Te uel
ABSTRACT
Polla ded woodlands a e iconic componen s o Eu opean u al landscapes. Polla ding is a adi ional manage-
men echnique used o ob ain imbe and i ewood. Howe e , hese woodlands a e subjec ed o di e en
s esso s in apidly depopula ing u al egions unde con inen al Medi e anean a eas whe e ipa ian black
popla (Populus nig a) polla ds a e widesp ead bu s ill unde s udied. Fi s , he apid u al depopula ion has lead
o he abandonmen o polla ding c ea ing ees wi h abundan , hick and hea y b anches. Second, clima e
wa ming and al e a ion o i e lows educe soil mois u e which could con ibu e o g ow h decline o ipa ian
polla ds. We in es iga ed how hese s ess ac o s ha e impac ed adial g ow h in se en si es loca ed in ou i e
basins o sou he n A ag´
on, no h eas e n Spain. We used dend och onology o econs uc he changes in adial
g ow h du ing he pe iod 1890–2020 and o assess how g ow h has esponded o changes in empe a u e,
p ecipi a ion and i e low a iabili y. Since 1900 g ow h supp essions, p obably co esponding o polla ding
e en s, occu ed in se e al decades (1910s, 1960s, 1980s and 1990s) and showed a high a iabili y among si es
and i e basins. Supp essions peaked du ing he 1940s, p obably associa ed o in ense polla ding ela ed o high
imbe and i ewood demand a e he Spanish Ci il Wa . Popla adial g ow h inc eased in esponse o high
p ecipi a ion and i e low om Ma ch o Ap il, pa icula ly in si es loca ed away o i e banks. Polla ding
abandonmen and declines in soil mois u e a ailabili y h ea en he pe sis ence o e e an black popla s. Ac i e
managemen h ough polla ding and p o ec ion should be implemen ed o p ese e hese iconic, an h opogenic
woodlands due o hei mul iple ecological and cul u al se ices.
1. In oduc ion
Socie y is demanding o es and woodland conse a ion and sus-
ainable use, bu we s ill lack his o ical and p ecise amewo ks o
di e en ia ing na u al s. dis u bed s ages (Fos e e al., 1996). Many
o es s ac oss he no he n hemisphe e co espond o cul u al land-
scapes which ha e been his o ically modi ied (Pe e ken, 1996; Rackam,
2006). Fo ins ance, o es s ac oss he Medi e anean Basin ha e been
subjec ed o a secula his o y o use and managemen (G o e and
Rackham, 2003). These o es s and woodlands ha e been clea ed,
d as ically al e ed and con e ed in o c oplands o g asslands, managed
o imbe o i ewood o ans o med h ough coppicing and polla ding
(Bobiec e al., 2018). Ne e heless, we s ill lack adequa e his o ical
assessmen s on hei pas g ow h dynamics and accoun s on hei cu -
en s a us. Re ospec i e s udies a e undamen o o ecas how hese
cul u al, an h opogenic o es s and woodlands will espond o he
decline in hei adi ional managemen , clima e wa ming and a i-
abili y (Fos e e al., 1996).
Polla ded s ands o b oadlea species can ha bo old ees which a e
able o s o e ca bon as woody issues o soil biomass o decades and
ep esen impo an ca bon ese oi s (Luyssae e al., 2008). The
egula cu ing o polla ded ees allows hem eaching much g ea e
ages han uncu ees (Read, 2008). These singula s ands also deli e
unique cul u al alues (Rozas, 2004), and ep esen biodi e si y ho -
spo s since hey p o ide habi a s o many plan o animal species. Fo
ins ance, hollow old oaks a e unique habi a s o a e, o en h ea ened
* Co esponding au ho .
E-mail add esses: [email p o ec ed] (J.J. Cama e o), [email p o ec ed] (E. Gonz´
alez de And ´
es), [email p o ec ed] (M. Colangelo), cdejaime@
educa.a agon.es (C. de Jaime Lo en).
Con en s lis s a ailable a ScienceDi ec
Fo es Ecology and Managemen
jou nal homepage: www.else ie .com/loca e/ o eco
h ps://doi.o g/10.1016/j. o eco.2022.120268
Recei ed 29 Ma ch 2022; Recei ed in e ised o m 29 Ap il 2022; Accep ed 1 May 2022
Fo es Ecology and Managemen 517 (2022) 120268
2
sap oxylic bee le species (Ranius and Jansson, 2000), and polla ded ash
ees a e he habi a s o a e epiphy e lichens and mosses (Moe and
Bo nen, 2000).
Polla ding o b oadlea species ocused on ob aining wood and lea
odde (Ba gioni and Sulli, 1998). Howe e , he in ensi y and equency
o adi ional polla ding a e declining and his echnique has ceased in
mos places ac oss Eu ope (Read, 2003). The e o e, we need sui able
assessmen s o he si ua ion o neglec ed polla ded s ands o inc ease ou
ecological knowledge, p omo e hei conse a ion and de elop man-
agemen poli ics o p ese e hem since hey a e a cen al pa o he
Eu opean cul u al he i age. Cu en ly, e e an polla ds wi h abundan
hea y b anches o canopy dieback a e widesp ead. Mo eo e , in addi-
ion o he ending o adi ional polla ding hese ee popula ions ha e
o ace he ad e se e ec s o clima e ex emes as d ough s which a e key
d i e s o Medi e anean o es dynamics (Gazol e al., 2018).
In he his o ically g azed, seasonally d y and con inen al si es o
eas e n Spain (Te uel, A ag´
on), polla ded black popla (Populus nig a L.)
s ands ep esen ed a unique sou ce o wood and odde o u al, local
popula ions (De Jaime Lo en and He e o Loma, 2007). He e, he his-
o ical in ensi e use o sheep g azing and he ha sh en i onmen al
condi ions p oduced open landscapes domina ed by g asslands and
sh ublands (Fig. 1). Nea small i e s and s eams, black popla s we e
plan ed, managed and polla ded in win e e e y 12–15 yea s, whe eas
g asslands we e used o ca le illus a ing he sus ainable managemen
o ag o- o es y sys ems in clima ically ha sh en i onmen s (De Jaime
Lo en and He e o Loma, 2007). Nowadays, hese polla ded popla s a e
pa o unique cul u al landscapes, bu hey a e menaced by he lack o
managemen ( adi ional polla ding ceased in mos si es a e he 1960s;
bu see Fig. 1c). They also ace addi ional s esso s such as d ough s ess
due o clima e wa ming o lack o adequa e soil mois u e a ailabili y
caused by i e egula ion (e.g., dam building) o ag icul u al uses (De
Jaime Lo en, 2017). The e o e, his is a special case o polla ded
woodland because i is pa o ipa ian ecosys ems subjec ed o mul iple
s esso s (S ella and Bendix, 2019).
In his s udy we ollowed a dend oecological app oach o econs uc
pas dis u bance egimes (Lo ime and F elich, 1989). In pa icula , we
econs uc ed he g ow h his o y o polla ded black popla o es s and
woodlands in sou he n A ag´
on, no h eas e n Spain. We aimed: (i) o
cha ac e ize he size and g ow h pa e ns o polla ded black popla s, (ii)
o econs uc g ow h his o y as ela ed o changes in polla ding e-
quency and in ensi y, and (iii) o quan i y he black popla adial g ow h
esponses o clima e a iabili y and changes in i e low. We expec ed
o de ec peaks in g ow h supp essions du ing epochs wi h in ense pol-
la ding o imbe o i ewood ex ac ion. We also hypo hesized ha
polla ded ees loca ed in he d ies si es (e.g., a om i e banks) will
be he mos sensi i e o changes in g owing-season p ecipi a ion and
i e low.
2. Ma e ials and me hods
2.1. S udy a ea
We selec ed se en si es in ou i e basins loca ed in sou he n
A ag´
on (Za agoza and Te uel p o inces, eas e n Spain) whe e big,
heal hy polla ded popla ees we e ound (Table 1, Fig. 2). In his a ea
clima e condi ions a e Medi e anean con inen al wi h d y summe s and
cold win e s acco ding o da a om local clima e s a ions (Fig. S1).
Mean annual empe a u es anges 10.6–12.6 ◦C, wi h minimum alues
in Janua y (mean empe a u e is 1.6–4.4 ◦C) and wa mes condi ions in
July (mean empe a u e is 20.3–21.6 ◦C). Annual p ecipi a ion anges
380–420 mm wi h peaks in May and au umn and low alues in summe
(July, Augus ) and win e (Janua y, Feb ua y). In he s udy a ea, he
Fig. 1. Views o black popla s in old pho og aphs aken nea he Na a e e (a) and Fon ía (b) s udy si es and sampling si e loca ed in he (c) Allepuz s udy si e (no e
he ecen ly polla ded popla s whose b anches we e analysed). The images (a) and (b) we e aken in he 1920s and la e 1930s (soldie s co espond o he Spanish
Ci il Wa , which las ed om 1936 o 1939), espec i ely.
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
3
con inen ali y deg ee is ex eme o he Ibe ian Peninsula, and some o
he coldes empe a u es o Spain (ca. −30 ◦C) ha e been eco ded
he e (Aupí 2021).
In he s udy a ea, ipa ian ege a ion in he we es si es is domi-
na ed by black popla , willows (Salix a ocine ea B o ., Salix pu pu ea L.,
Salix alba L., Salix eleagnos Scop.) and ush (Juncus acu us (L.) To . ex
Re z) nea he i e s and s eams. Plan ed ui ees a e also equen
such as walnu (Juglans egia L.). In he nea by, d ie loca ions he
landscape is domina ed by open g asslands and sh ublands (C a aegus
monogyna Jacq., Rosa canina L., Amelanchie o alis Medik., Genis a
sco pius (L.) DC., Thymus ulga is L., Cis us lau i olius L., Ephed a spp.)
wi h sca e ed junipe s (Junipe us communis L., Junipe us phoenicea L.). I
is equen ly obse ed ha junipe s ec ui below popla s due o ui
dispe sal and bi d pe ching. In we e , no h- o no h-wes o ien ed
slopes oaks a e ound (Que cus ilex L., Que cus aginea Lam.). Mos si es
ha e basic soils o med on limes ones and ma ls, some imes wi h
Table 1
S udy si es and s uc u al ea u es. Values a e means ±SE. Di e en le e s indica e signi ican (p <0.05) di e ences be ween si es based on Mann-Whi ney U es s.
Si e (code) Ri e basin Flow da a La i ude (N) Longi ude (W) Ele a ion (m a.s.l.) Dis ance o sampled ees o nea es i e bank (m)
Aguila de Al amb a (AG) Al amb a X 40◦35′22′’ 0◦48′01′’ 1320 1.3 ±0.6a
Allepuz (AL) Al amb a X 40◦30′28′’ 0◦45′12′’ 1328 30.0 ±8.7c
Fon ía (FO) Hue a 41◦00′42′’ 1◦05′43′’ 1192 1.5 ±0.8a
Gúda (GU) Al amb a X 40◦26′38′’ 0◦43′57′’ 1363 2.5 ±0.8a
Huesa del Común (HU) Aguas i as 41◦00′49′’ 0◦55′00′’ 839 2.0 ±0.7a
Ma ine a (MA) Aguas i as 41◦00′71′’ 0◦57′37′’ 868 6.9 ±0.5b
Na a e e (NA) Panc udo X 40◦55′55′’ 1◦13′37′’ 946 2.7 ±0.9a
Fig. 2. Loca ion o he s udy a ea and i e
basins in eas e n Spain ( he small map shows
he dis ibu ion o black popla in Eu ope
and he loca ion o he s udy a ea −g ey
box). In he map he se en s udy si es a e
indica ed and di e en colo s co espond o
he s udied i e basins ( om no h o
sou h): Hue a i e (b own a ea), Aguas i-
as i e (blue a ea), Panc udo i e (yellow
a ea) and Al amb a i e (pu ple a ea). Blue
lines and pa ches co espond o i e s and
ese oi s, espec i ely. The ed pa ches
show Za agoza and Te uel ci ies, A ag´
on e-
gion, no h-eas e n Spain. G een poin s
indica e polla ded black popla s. (Fo in e -
p e a ion o he e e ences o colou in his
igu e legend, he eade is e e ed o he
web e sion o his a icle.)
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
4
gypsum ou c ops (Na a e e s udy si e), and a ew si es ha e acid soils
(FO, Fon ía s udy si e).
In he s udy si es, ees ha e no been polla ded o he pas 20 yea s,
excep ing in he Na a e e si e (Table 1) whe e hal o sampled ees
we e polla ded 8 yea s ago. The lack o managemen o hese o es s is
ela ed o he apid loss o u al popula ion (Collan es and Pinilla, 2011)
due o he massi e mig a ion o ci ies, which peaked in he 1960s
(Fig. S2). Polla ding also declined because o he ad ances in building
and hea ing echnologies. In ac , he s udy a ea is loca ed in he Ibe ian
Range, which o ms pa o he so-called “Se anía Cel ib´
e ica”, one o
he mos spa sely popula ed a eas in Eu ope wi h mean densi y o 7
inhabi an s km
−2
(Bu illo-Cuad ado e al. 2019).
2.2. T ee sampling
Sampling was done om au umn 2020 o sp ing 2021. In each si e
we selec ed dominan , polla ded black popla ees and measu ed hei
diame e a 1.3 m, o al heigh and polla ding heigh using apes and a
lase ange inde (Nikon Fo es y P o 550 hypsome e ). We also
measu ed he dis ance be ween ees and he nea es i e bank. We ook
wo co es a 1.3 m om each ee using a P essle inc emen bo e . In
addi ion, main b anches we e sampled in Allepuz (AL) and Gúda (GU)
si es a polla ding heigh (Fig. 1c, Table 2). These b anches co e-
sponded o he pos -polla ding pe iod. In all si es, sampled ees we e
loca ed nea i e banks (Table 1). Howe e , ees sampled in Ma ine a
(MA) we e no so close o he i e bank and hose in si e AL we e
loca ed away om he i e and nea a o me ly used i iga ion di ch. In
con as , ees in si e GU we e loca ed close o he i e . The e o e, we
expec ed o ind highe b anch g ow h sensi i i y o p ecipi a ion and
i e low in he AL han in he GU si e.
2.3. Clima e and i e low da a
Mon hly mean empe a u e and p ecipi a ion da a we e ob ained
om he 0.1◦-g idded E-OBS clima e da ase (Co nes e al., 2018) since
long and homogeneous clima e eco ds a e sca ce in he s udy a ea
excep ing in pa icula loca ions such as Da oca, Calamocha and Te uel
clima e s a ions (see Fig. S1). Using hese clima e da a we calcula ed he
S anda dized P ecipi a ion-E apo anspi a ion Index (SPEI) (Vicen e-
Se ano e al., 2010). The SPEI quan i ies d ough leng h and se e i y a
mul iple empo al scales wi h alues below −1.5 usually indica ing se-
e e d ough . We calcula ed he SPEI o 1- o 48-mon h esolu ion.
We ob ained h ee se ies o mon hly low da a co e ing he pe iod
1950–2017 and co esponding o he ollowing s a ions: Na a e e del
Río (Panc udo i e basin, 40◦55′53′’N, 1◦15′31′’ W, 904 m), Villalba
Al a (Al amb a i e basin, 40◦36′56′’N, 0◦58′28′’ W, 1060 m), and
Te uel (Al amb a i e basin, 40◦21′30′’N, 1◦07′13′’ W, 885 m). Ri e
low da a we e e ie ed om he Spanish Na ional Flow and Discha ge
Da abase (“Cen o de Es udios y Expe imen aci´
on de Ob as Públicas”,
da a a ailable a he webpage h ps://ceh.cedex.es). Acco ding o hese
da a, i e lows peak om Ma ch o May and each minimum alues
om July o Sep embe (Fig. S3a). This co esponds o a con inen al
Medi e anean plu ial low, which is cha ac e ized by maxima in sp ing
and au umn–win e and a minimum in summe . In his egime, low
depends on p ecipi a ion and geological subs a es o he s udy a ea
such as limes ones which can bu e he educ ion o low in summe
(Solans and Po , 2013). Annual lows a y be ween 15 and 40 hm
3
, bu
hey ha e declined in he s udy s a ions a e peaking in he we 1970s
(Fig. S3b).
2.4. T ee- ing wid h da a
Wood samples we e ai d ied and ca e ully sanded ollowing
dend och onological p ocedu es un il ee ing bounda ies we e isible
(F i s, 1976). Samples we e scanned a 2400 dpi esolu ion and isually
c oss-da ed. Then, ing wid hs we e measu ed wi h a 0.001 mm eso-
lu ion using scanned images and he CDend o so wa e (La sson and
La sson, 2018). The quali y o c oss-da ing was checked using he
COFECHA so wa e which calcula es mo ing co ela ions be ween in-
di idual se ies o ing-wid h alues and he mean si es se ies (Holmes,
1983).
2.5. T ee- ing wid h p ocessing
Since mos co es did no ha e pi h o inne mos , cu ed ings, we did
no use g aphical me hods o es ima e age a 1.3 m. We assumed ha
s em ch onological and geome ic cen e s coincided. The leng h o
missing adius was es ima ed as he di e ence be ween he geome ic
Table 2
S uc u al ea u es and adial g ow h da a o sampled ees. In si es AL and GU b anches (ALb , GUb ) we e also sampled and measu ed. Values a e means ±SE. EPS is
he Exp essed Popula ion Signal. Di e en le e s indica e signi ican (p <0.05) di e ences be ween si es based on Mann-Whi ney U es s.
Si e S uc u al da a T ee- ing da a
Dbh
(cm)
T ee
heigh
(m)
Polla ding
heigh (m)
No. ees
(No.
adii)
Age a 1.3
m (yea s)
Timespan Ring
wid h
(mm)
Fi s -o de
au oco ela ion
Mean co ela ion
wi h si e se ies
EPS
#
Nega i e
g ow h change
(%)
§
AG 108.6 ±
8.0c
15.6 ±
1.5a
3.3 ±0.2c 12 (24) 106 ±
12bc
1822–2020 2.79 ±
0.39b
0.63 0.34 0.59 19.6 ±1.6a
AL 111.6 ±
15.5c
27.4 ±
1.7c
2.8 ±0.2b 10 (20) 99 ±10b 1877–2020 3.05 ±
0.45b
0.65 0.49 0.82 25.2 ±2.4b
FO 72.3 ±
6.9b
19.0 ±
1.0b
2.8 ±0.1b 23 (46) 90 ±4a 1888–2020 3.34 ±
0.35b
0.72 0.57 0.84 42.5 ±4.1d
GU 118.3 ±
9.9c
19.1 ±
1.1b
2.4 ±0.1a 12 (24) 82 ±5a 1910–2020 2.93 ±
0.38b
0.69 0.47 0.81 33.4 ±3.6c
HU 60.3 ±
1.3a
15.6 ±
1.3a
2.1 ±0.2a 5 (10) 103 ±
15bc
1877–2020 1.64 ±
0.39a
0.62 0.36 0.61 39.8 ±4.2 cd
MA 81.5 ±
4.9b
14.0 ±
0.3a
3.0 ±0.1b 16 (32) 91 ±5a 1886–2020 2.32 ±
0.31a
0.53 0.40 0.76 34.5 ±3.3c
NA 76.1 ±
4.0b
15.6 ±
1.8a
3.4 ±0.1c 22 (44) 128 ±
14bc
1794–2020 1.71 ±
0.24a
0.65 0.38 0.79 15.4 ±1.4a
ALb – – – 7 (14) 54 ±1
¥
1961–2020 3.03 ±
0.59b
0.74 0.59 0.88 –
GUb – – – 10 (20) 57 ±1
¥
1948–2020 3.56 ±
0.47b
0.78 0.48 0.76 –
#
Calcula ed o he common pe iod 1970–2020.
§
Calcula ed o he common pe iod 1890–2020.
¥
Age o b anches sampled a polla ding heigh .
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
5
adius, a e emo ing ba k hickness, and he o al co e leng h
ollowing (No on e al., 1987). Then, he numbe o inne missing ings
was es ima ed by ex apola ing he mean adial g ow h a e om he
inne mos 10 ings (Rozas 2003, 2004).
To es ima e pe iods o g ow h supp ession, which could be a ibu ed
o pas polla ding (c . Rozas, 2004), we calcula ed nega i e g ow h
changes (NGC) ollowing Nowacki and Ab ams (1997):
NGC = [(M1 − − M2)/M2 ] × 100 (1)
whe e M1 and M2 a e p e ious and subsequen 10-yea ing-wid h
medians calcula ed on indi idual se ies and shi ed e e y yea
(Fig. S4). We conside ed a 50% h eshold in NGC obse ed in a leas
50% o ees in each si e o de ining majo g ow h supp essions,
whe eas a 25%- h eshold in a leas 50% o ees was ega ded as
mode a e supp essions. The NGC se ies we e ob ained o he common
pe iod 1890–2020, i.e. hey we e plo ed o he 1900–2010 pe iod.
Analogously, we also calcula ed posi i e g ow h changes (PGC) and
es ima ed he numbe o majo and mode a e eleases ollowing:
PGC = [(M2 − − M1)/M1 ] × 100 (1)
Bo h NGC and PGC we e calcula ed on co es’ ing-wid h se ies bu
no on b anches’ se ies since hese c oss-sec ions co esponded o he
pe iod a e he las polla ding was done abou 55 yea s ago (see
Table 2). Finally, we calcula ed Spea man co ela ions be ween si es’
mean se ies o NGCs o PGCs o assess among-si e simila i y in g ow h
supp essions o eleases, espec i ely.
To s udy g ow h esponses o clima e a e emo ing size-, on ogeny-
o dis u bance- ela ed ends, we de ended ee- ing wid h se ies. We
selec ed cubic smoo h spline wi h a leng h o 0.67 o he se ies and a 0.5
esponse cu -o . Then, obse ed alues we e di ided by i ed alues o
ob ain ing-wid h indices. Indexed se ies we e subjec ed o p e-
whi ening o emo e empo al au oco ela ion and hen a e aged
using bi-weigh obus a e ages o ob ain p e-whi ened, esidual ch o-
nologies. To assess wi hin-si e g ow h cohe ence we calcula ed he mean
co ela ion o he indi idual se ies wi h he mean si e se ies. We also
calcula ed he Exp essed Popula ion Signal (EPS) o he common, bes -
eplica ed pe iod 1970–2020. The EPS was used o assess how eplica ed
and cohe en was each si e se ies o ch onology (Wigley e al., 1984). To
cha ac e ize and compa e he si e se ies o ing-wid h da a, we calcu-
la ed he ing wid h and he i s -o de au oco ela ion which measu es
yea - o-yea g ow h pe sis ence (F i s, 1976).
2.6. S a is ical analyses
To compa e indi idual linea g ow h ( ing-wid h) ends among si es
conside ing he pe iod 1990–2020 we used he K uskall-Wallis es
ollowed by pai ed Mann-Whi ney es s. To e alua e he ela ionships
among si es’ se ies o ing-wid h indices we calcula ed a P incipal
Componen Analysis (PCA) on he co a iance ma ix conside ing co es’
and b anches’ indices o he common and bes - eplica ed pe iod
1970–2020. We kep he i s (PC1) and second (PC2) p incipal com-
ponen s because hey indi idually accoun ed o a leas 15% o
a iance.
The ela ionships be ween he si es’ se ies o ing-wid h indices s.
mon hly clima e (mean maximum and minimum empe a u es, o al
p ecipi a ion), SPEI and low da a we e based on Pea son co ela ions.
Co ela ions we e calcula ed o he pe iod 1950–2020 in he case o
s em co es and o he pe iod 1970–2020 in he case o b anches da a.
Co ela ions we e ob ained om p io Oc obe (yea p e ious o he
yea o ee- ing o ma ion) o cu en Sep embe when he la e g owing
season occu s. To assess he empo al s abili y be ween g ow h indices
and selec ed p ecipi a ion o low da a we calcula ed 20-yea mo ing
co ela ions be ween hese se ies.
All analyses we e pe o med in he R en i onmen o s a is ical
compu ing (R De elopmen Co e Team, 2021). The package dplR (Bunn,
2010; Bunn e al., 2021) was used o de end and analyse ee- ing wid h
se ies, he package TRADER was used o calcula e NGCs and PGCs
(Al man e al., 2014), and he K uskall-Wallis es , he SPEI was calcu-
la ed using he R package SPEI (Begue ía e al., 2014), he PCA and
clima e-g ow h ela ionships (co ela ions) we e calcula ed using he
PAST e . 4.0 so wa e (Hamme e al., 2001).
3. Resul s
3.1. Size and age o polla ded ees
The mean diame e o all sampled popla s was 87.5 cm ( ange
39.3–196.0 cm), hei mean heigh was 16.8 m ( ange 6.7–35.0 m) and
he mean polla ding heigh was 2.9 m ( ange 1.5–4.8 m; Table 2). The
ees wi h hickes s ems we e ound in Aguila de Al amb a (AG),
Allepuz (AL) and Gúda (GU) si es and he highes indi iduals we e
ound in he AL si e. The highes mean polla ding heigh (3.3–3.4 m)
was ound in AG and Na a e e (NA) si es, and he lowes (2.2–2.4 m) in
GU and Huesa del Común (HU) si es. T ee diame e and heigh we e
posi i ely ela ed (Fig. 3a), whe eas ee heigh and polla ding heigh
we e nega i ely ela ed (Fig. 3b).
The oldes sampled indi iduals we e ound in he NA (227 yea s) and
AG (198 yea s) si es. The NA si e can be conside ed an old-g ow h s and
wi h nine ou o wen y wo sampled ees (41%) ha ing ages o e 200
Fig. 3. Main s uc u al cha ac e is ics o he sampled polla ded black popla
ees: (a) posi i e ela ionship be ween diame e a b eas heigh (Dbh) and ee
heigh , and (b) nega i e ela ionship be ween ee heigh and polla ding heigh .
S a is ics co esponding o linea eg essions (con inuous lines) a e shown in
he uppe igh co ne o each plo . The dashed lines show he 95% p edic-
ion in e als.
J.J. Cama e o e al.

Fo es Ecology and Managemen 517 (2022) 120268
6
yea s. The age o b anches sampled in si es AL and GU was e y uni o m
(54–57 yea s; see Table 2 and Fig. 4).
3.2. G ow h pa e ns and ends
The highes mean g ow h a es we e ound in AL and FO si es
(3.05–3.34 mm), whils he lowes a es co esponded o HU and NA
si es (1.64–1.71 mm). The se ies om MA and HU si es showed he
lowes i s -o de au oco ela ion alues, whe eas b anches’ se ies
showed he highes alues ollowed by FO and GU ees’ se ies (Table 2).
The highes mean NGCs we e ound in FO and HU si es, whe eas he
lowes alues we e ound in NA and AG si es.
Acco ding o he K uskall-Wallis es s, he e we e signi ican di e -
ences among si es ega ding indi idual g ow h ends o he pe iod
1990–2020 (H =46.76, p <0.001). All si es p esen ed mean nega i e
g ow h ends wi h lowes alues (mean ±SE) o si es FO (-0.006 ±
0.001 mm y
−1
), AG (-0.005 ±0.001 mm y
−1
), GU and HU (bo h
(-0.004 ±0.001 mm y
−1
). In con as , NA (0.002 ±0.001 mm y
−1
), AL
and MA (bo h 0.001 ±0.001 mm y
−1
) si es p esen ed posi i e ends.
As expec ed, b anches’ se ies also p esen ed nega i e ends (bo h si es,
−0.005 ±0.001 mm y
−1
).
O e all, he 1890–2020 pe iod was well eplica ed conside ing all
si es (Fig. 4), albei he EPS alues we e usually below he 0.85-
h eshold o en conside ed in ee- ing s udies du ing he common
pe iod 1970–2020 (Table 2). The highes EPS alues we e ound o he
FO (0.84), AL (0.82) and GU (0.81) si es. These si es also showed he
highes mean co ela ions o indi idual se ies wi h si e se ies. The
lowes EPS alues co esponded o AG (0.59) and HU (0.61) si es. The
only mean se ies o ch onology wi h EPS abo e he a o emen ioned
h eshold co esponded o he b anches om si e AL wi h an EPS o
0.88, con i ming a high cohe ence o ha se ies. B anches showed high
mean co ela ions o indi idual se ies wi h si e se ies.
3.3. G ow h changes as p oxies o polla ding
The NGCs peaked in he 1910s in si es AG and NA, and om he la e
Fig. 4. G ow h da a o black popla s based on he mean se ies o ee- ing wid h calcula ed o each si e. G ow h da a we e ob ained o s em co es (black ba s) and
b anches c oss-sec ions (g ey ba s) in si es AL and GU. Values a e means ±SE. The ba s show he annual numbe o measu ed adii ( igh y axes) wi h black and g ey
ills co esponding o s em and b anch adii, espec i ely.
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
7
1930s o ea ly 1940s in si es AL, FO, GU, HU and MA (Fig. 5). In si es AL
and NA, o he NGCs peaks we e obse ed in he mid o la e 1960s and
om he la e 1980s o ea ly 1990s. In si es FO and HU, NGCs also
inc eased in he 2000s. The e was a high numbe o mode a e and majo
supp essions in he ea ly 1920s, ea ly 1950s, mid 1970s and la e 1990s.
Rega ding he PGCs, hey peaked om he la e 1970s o ea ly 1980s in
si es AG, AL, MA and NA (Fig. S5). P e ious peaks we e also de ec ed in
he 1920s (si es AL, GU, HU and NA), 1940s (si es AL, FO and NA),
1950s (si es AG, HU, MA and NA) and om he 1960s o he 1970s in
si es GU and HU. The e was a high numbe o mode a e and majo e-
leases in he mid-1920s, ea ly 1950s, mid-1970s and la e 1990s.
The NGCs se ies o si es FO, GU, HU and MA we e posi i ely ela ed,
showing all hem he NGC peaks in he la e 1930s o ea ly 1940s
(Table 3). The si es FO, HU and MA we e loca ed nea by (Fig. 2). The
PGCs se ies o si es AG, FO, GU, HU, AL and MA we e posi i ely ela ed
(Table 3).
Fig. 5. Se ies o nega i e g ow h changes showing po en ial polla ding e en s co esponding o se e e g ow h educ ions (high alues o nega i e g ow h change)
and numbe o ees showing mode a e (g ey ba s) and majo (black ba s) g ow h supp essions ( igh y axes).
Table 3
Spea man co ela ions calcula ed be ween he si e’s se ies o nega i e ( alues
shown abo e he diagonal) and posi i e ( alues shown below he diagonal)
g ow h changes conside ing he common pe iod 1890–2020. Signi ican (p <
0.05) alues a e shown in bold. See si es’ codes in Table 1.
AG AL FO GU HU MA NA
AG 0.27 −0.12 −0.13 0.12 0.18 ¡0.21
AL 0.12 −0.07 ¡0.36 0.15 ¡0.24 0.32
FO 0.28 −0.11 0.50 0.55 0.23 0.13
GU 0.33 ¡0.28 0.61 0.45 0.61 −0.18
HU 0.46 0.31 0.56 0.43 0.25 0.09
MA 0.62 ¡0.34 0.33 0.60 0.25 −0.12
NA 0.03 0.46 0.09 −0.06 0.13 −0.01
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
8
3.4. G ow h esponses o clima e, d ough and i e low a iabili y
We ob ained se ies o ing-wid h indices o s em co es and b anches
c oss-sec ions co e ing he 1950–2020 and 1970–2020 pe iods,
espec i ely (Fig. S6). The PCA showed ha se ies o he nea by MA,
HU, FO, NA and AG si es (see Fig. 2) we e aligned along he PC1, which
accoun ed o 36.1% o he o al a iance, whe eas co e and b anches
se ies o he sou he n AL and GU si es we e aligned along he PC2, which
accoun ed o 19.8% o he a iance (Fig. S7). Some yea s we e high-
ligh ed by he PCA showing e y low PC1 sco es and co esponding o
low g ow h a es such as 1986, 1995–1996, 2005 and 2012 (Fig. 6),
whils high a es we e ound in 2020 and 1979 (Fig. S5). In si es AL and
GU, g ow h a es we e high in 1988–1989 and e y low in 1982, 2001
and 1990. In hese wo si es, co ela ions be ween s em and b anch se-
ies o ing-wid h indices we e highly signi ican (p <0.001) in he AL (
=0.78) and GU ( =0.83) si es conside ing he common pe iod
1970–2020.
Conside ing he indexed se ies om nea by si es MA, HU, FO, NA
and AG we ound ha popla g ow h a es we e no e y esponsi e o
clima e a iabili y (Fig. S8). In addi ion, esponses changed among
si es. We ound posi i e co ela ions be ween MA ing-wid h indices and
win e maximum empe a u es and Ma ch-Ap il p ecipi a ion. In AG
he e was a posi i e co ela ion be ween indices and p io Oc obe
minimum empe a u es, bu a nega i e co ela ion wi h Sep embe
minimum empe a u es in NA. In FO, indices esponded nega i ely o
high Janua y minimum empe a u es and ele a ed p io No embe
p ecipi a ion.
We ound mo e and s onge co ela ions be ween ing-wid h indices
om b anches and i e low as compa ed o p ecipi a ion in si e AL
(Fig. S9). B anch adial g ow h was enhanced by high p ecipi a ion and
i e low om Ma ch o Ap il (Figs. 6 and 7). High low om May o
June also imp o ed g ow h. In si e AL, s em adial g ow h was also
enhanced by high Ma ch ( =0.24) and Ap il ( =0.28) p ecipi a ion
and ele a ed Ap il ( =0.25) and June i e low ( =0.29), whe eas no
signi ican co ela ion was ound be ween g ow h indices om co es
and empe a u e, p ecipi a ion o low da a in si e GU. Howe e , b anch
g ow h indices and Sep embe low we e posi i ely ela ed in his si e
(Figs. 7 and S9).
We only ound signi ican (p <0.05) co ela ions be ween ing-wid h
indices and SPEI alues in he FO and AL si es. In si e FO he highes
co ela ion ( =0.41) was ound o 45-mon h May SPEI, whe eas in si e
AL he highes co ela ion ( =0.31) was ound o 2-mon h Ap il SPEI.
In e es ingly, his co ela ion was simila o co es and b anches om
si e AL ( =0.48) when conside ing he same common pe iod o bo h
co es and b anches (1961–2020).
The associa ions be ween b anch g ow h indices and Ap il p ecipi-
a ion o low in he esponsi e si e AL we e s ong bu no s able
h ough ime (Fig. S10). They peaked om he mid o la e 1980s in he
case o p ecipi a ion and in he 1990s in he case o i e low. Howe e ,
co ela ions declined, pa icula ly in he case o low, in he 2000s. A e
emo ing he in luence o p ecipi a ion on low, we s ill ound he same
weakening in g ow h- low co ela ions con i ming his was due o low
educ ions a he han o a decline in p ecipi a ion (Fig. S3b).
4. Discussion
We ound a peak in g ow h educ ions (nega i e g ow h change)
du ing he 1940s in i e ou o se en si es. Such ma ked and las ing
g ow h declines could be a ibu ed o polla ding (c . Rozas 2004). We
also de ec ed ha g ow h o polla ded ees loca ed in d ough -p one
si es such as MA, whe e ees we e ound a om i e banks, and
AL, whe e ees we e ound nea a o me ly used i iga ion di ch, was
sensi i e o changes in Ma ch o Ap il p ecipi a ion and i e low. The
di e en g ow h ends obse ed among si es a e 1990, when clima e
became pa icula ly wa me in he s udy a ea, sugges di e en u u e
dynamics. Fo ins ance, i could be expec ed ha si es wi h he mos
nega i e g ow h ends and e y esponsi e o d ough (e.g., si e FO)
could become mo e ulne able o loss i ali y as compa ed wi h si es
wi h posi i e g ow h ends (e.g., si e NA).
4.1. G ow h his o y o polla ded popla s
In he s udy egion he e was a demog aphic su ge in he 1940s. The
inc ease o coal p ice a e Wo ld Wa I, he economic legacies o he
Spanish Ci il Wa (1936–1939) and he au a ky poli ics o he subse-
quen dic a o ship lead o an inc ease in he use and demand o i ewood
du ing he 1940s (In an e-Ama e and I ia e-Go˜
ni, 2017). Acco ding o
hese au ho s, he i ewood consump ion a e in he s udy a ea (abou 5
kg hab
-1
day
−1
) was among he highes in Spain om he 1930s un il
1955 and apidly declined a e wa ds when ossil uels became widely
used, u al elec i ica ion ex ended and popula ion massi ely mig a ed
o ci ies. The local imbe demand also inc eased a e he Spanish Ci il
Wa o ebuilding o es o ing damaged o des oyed u al houses,
ba ns and co als using popla beams since one o he Spanish Ci il Wa
on lines was es ablished in he s udy a ea se e ely a ec ing illages’
in as uc u es du ing one yea (De Jaime Lo en and He e o Loma,
Fig. 6. G ow h a es in popla b anches inc eased as Ma ch o Ap il p ecipi a ion did as shown by he posi i e co ela ion ound be ween bo h a iables in AL si e.
J.J. Cama e o e al.
Fo es Ecology and Managemen 517 (2022) 120268
9
2007). Such his o ical changes could explain he peak in polla ding in-
ensi y o he s udy popla si es d i en by he inc eased demand o local
imbe and i ewood. G ow h supp essions we e also obse ed in he
1960s, 1980s and 1990s, bu he nega i e g ow h change ne e eached
he in ensi y obse ed in he 1940s. Such conside able u al mig a ion o
ci ies du ing he 20 h cen u y a ec ed mos Eu opean coun ies, bu he
depopula ion a e was excep ionally high in some Spanish egions
(Collan es and Pinilla, 2021). Some o hese a eas con ain mos black
popla polla ds (e.g., Te uel p o ince) which explains hei cu en
ulne abili y since polla ding was igh ly linked o u al popula ions and
imbe o wood ex ac ion (De Jaime Lo en and He e o Loma, 2007).
4.2. G ow h esponses o clima e and i e low
Popla s a e ph ea ophy e species. In he case o black popla , i s
g ow h mainly depends on ph ea ic wa e and declines when wa e a-
bles a e supp essed (S´
anchez-P´
e ez e al., 2008; Singe e al. 2013;
2014). Wa e sho age leads o b anch abscission, canopy dieback,
g ow h decline and e en mo ali y in ipa ian popla species (Sco
e al., 1999; Rood e al., 2000; 2007). These ea u es suppo ou co -
ela ion analyses showing a igh connec ion be ween adial g ow h o
polla ded popla s and Ma ch-Ap il p ecipi a ion and i e low. These
esponses we e ound in si es subjec ed o luc ua ing wa e ables ei he
a om i e banks (e.g., si e MA) o nea o me ly used i iga ion
di ches (e.g., si e AL). In he s udy a ea, d y sp ing o summe condi ions
co espond o wa me empe a u es and lowe - han-no mal p ecipi a-
ion (Gazol e al., 2018), which educe he quan i y o uno con ib-
u ing o i e discha ge (Solans and Po , 2013). Since b anch g ow h
was no so a ec ed by polla ding as s em g ow h, b anches showed
s onge esponses o p ecipi a ion and low a iabili y han s ems. The
b anch esponsi eness o clima e and low changed h ough ime
because b anch g ow h decoupled om i e low om he 2000s on-
wa ds. Such weakening was no p obably due o compe i ion among
b anches, bu o se e e and pe sis en low educ ion (Fig. S3b), o en
associa ed o unsus ainable i e managemen due o ag icul u al wa e
demand (De Jaime Lo en and He e o Loma, 2007).
We ound ha annual p ecipi a ion and i e low became decoupled
om he la e 1990s o ea ly 2000s as i e low dec eased (Figs. S3b,
S11). Since annual p ecipi a ion showed no signi ican end om 1950
o 2020 and Ap il p ecipi a ion showed a posi i e end (Kendall
τ
=
0.18, p =0.02), he nega i e endency o i e low may be explained by
clima e wa ming. In ac , annual empe a u e has signi ican ly inc eased
in he s udy a ea (
τ
=0.44, p <0.001). Pa icula ly, we ound ha
maximum Ap il empe a u e has inc eased (
τ
=0.29, p <0.001). This
could lead o a highe e apo a i e demand and explain he nega i e
associa ion wi h g ow h obse ed in some si es such as FO. In ac , black
popla is cha ac e ized by high sap low a e and canopy wa e e apo-
a i e demand (Lambs and Mulle , 2002; S´
anchez-P´
e ez e al., 2008).
Fig. 7. Popla b anch adial g ow h a es and i e low we e posi i ely associa ed as shown he posi i e co ela ions ound be ween bo h a iables in GU (a) and AL
(b) si es. No e he loga i hmic scales o mon hly low da a co esponding o he Te uel s a ion in he Al amb a i e .
J.J. Cama e o e al.