U ban Fo es y & U ban G eening 63 (2021) 127215
A ailable online 12 June 2021
1618-8667/© 2021 Published by Else ie GmbH. 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/).
P opo ion o non-na i e plan s in u ban pa ks co ela es wi h clima e,
socioeconomic ac o s and plan ai s
´
Al a o Bay´
on
a
, Osca Godoy
b
, No¨
elie Mau el
c
, Ma k an Kleunen
c
,
d
, Mon se a Vil`
a
a
,
e
,
*
a
Es aci´
on Biol´
ogica de Do˜
nana (EBD-CSIC), A . Am´
e ico Vespucio 26, 41092, Se illa, Spain
b
Depa men o Biology, Uni e si y o Cadiz, Cadiz, Spain
c
Ecology, Depa men o Biology, Uni e si y o Kons anz, Kons anz, Ge many
d
Zhejiang P o incial Key Labo a o y o Plan E olu iona y Ecology and Conse a ion, Taizhou Uni e si y, Taizhou, 318000, China
e
Depa men o Plan Biology and Ecology, Uni e si y o Se ille, Se ille, Spain
ARTICLE INFO
Handling Edi o : W Wendy McWilliam
Keywo ds:
Ha diness zones
In asi e species
O namen al plan s
Plan ai p o ile
ABSTRACT
U ban pa ks and ga dens p o ide cul u al and aes he ic se ices c i ical o human well-being. Ye , hey
ep esen one o he main easons o he in en ional in oduc ion o o namen al species, some o which can
escape and es ablish in na u al ecosys ems. Besides aes he ic easons, clima e and socioeconomic ac o s can also
modula e which species a e plan ed in u ban pa ks. He e, we e alua e he ela ionship be ween ai s o 486
o namen al woody species om 46 Spanish u ban pa ks and clima ic and socioeconomic a iables. We speci -
ically assessed how plan ai s, clima ic, and socioeconomic ac o s a e ela ed o he p opo ion o non-na i e
species and, among hem, o he p opo ion o es ablished non-na i e species in Spain. O e all, we ound clea
associa ions be ween species ai s and clima ic a iables. Mos no ably, pa ks wi h wa me win e s ha e mo e
plan species wi h conspicuous lowe s, whe eas pa ks wi h colde win e s and a mo e con inen al clima e ha e
mo e species wi h highe ole ances o cold and shade. Mos o he species eco ded in ou s udy a e non-na i e
(82 %). Highe p opo ions o non-na i e species in u ban pa ks we e posi i ely associa ed wi h owns wi h la ge
size homes and inhabi an s wi h highe median age bu nega i ely ela ed o pa ks wi h species wi h a highe
ha diness-zone ange. Mo eo e , a g ea e p opo ion o non-na i e species ha can es ablish in he na u al
ecosys ems was ound in pa ks wi h lowe con inen ali y condi ions. Ou esul s show ha Spanish u ban pa ks
ha e an o e whelming p opo ion o non-na i e woody species, some o which ha e he po en ial o es ablish,
and ha he a ia ion in hei p opo ions can be explained by clima ic, and socioeconomic ac o s.
1. In oduc ion
U ban pa ks and ga dens p o ide cul u al and aes he ic se ices
c i ical o human well-being (Bolund and Hunhamma , 1999; Hulme,
2007; Kendal e al., 2012), ye hey a e one o he main easons o he
in en ional in oduc ion o o namen al non-na i e plan species wo ld-
wide (Maye e al., 2017; an Kleunen e al., 2018). This sou ce o
non-na i e species in oduc ion is pa ly explained by he ac ha in ou
con inuous sea ch o no el y, we, as humans, a e ac i ely looking o
plan ea u es ha inc ease he aes he ic alue o u ban pa ks while
c ea ing a sense o place. In addi ion, p e ious wo k has shown ha such
sea ch is also modula ed by he socioeconomic con ex o he owns
wi hin a coun y. In gene al, he p esence o non-na i e plan species is
explained by clima e, human ac o s and a combina ion o bo h (Pyˇ
sek
e al., 2010). Fo ins ance, owns wi h high a e age incomes and high
de elopmen dedica e g ea e mone a y in es men o he in oduc ion
o no el species wi h highe aes he ic alue (Vaz e al., 2018) and o he
main enance o g een a eas, wi h mo e di e se species composi ion
(Pyˇ
sek e al., 2010). Likewise, densely popula ed egions a e cha ac-
e ized by highe p opagule p essu e, he e o e, mo e indi iduals a e
plan ed and mo e in asi e plan s p esen (Pino e al., 2005; Pyˇ
sek e al.,
2010).
Ne e heless, he deco a i e alue is only one eason o plan ing
non-na i e species in u ban pa k. O he impo an easons ela e o he
abili y o some non-na i e species o cope wi h en i onmen al s ess.
Commonly, pa k manage s selec species ha a e easy o es ablish and
main ain. This means inding species wi h pa icula ai s ha a e p e-
adap ed o he pa icula clima ic condi ions. Howe e , his
* Co esponding au ho .
E-mail add esses: [email p o ec ed] (´
A. Bay´
on), [email p o ec ed] (M. Vil`
a).
Con en s lis s a ailable a ScienceDi ec
U ban Fo es y & U ban G eening
jou nal homepage: www.else ie .com/loca e/u ug
h ps://doi.o g/10.1016/j.u ug.2021.127215
Recei ed 30 No embe 2020; Recei ed in e ised o m 2 June 2021; Accep ed 7 June 2021
U ban Fo es y & U ban G eening 63 (2021) 127215
2
managemen can b ing en i onmen al p oblems as i also acili a es
hei es ablishmen in na u al ecosys ems in he in oduced ange
(Gonz´
alez-Mo eno e al., 2014; Mau el e al., 2016; Pyˇ
sek e al., 2010).
In gene al e ms, i has been well-documen ed ha he dis ibu ion o
woody species along clima ic g adien s is modula ed by he pa icula
o gans and whole-plan ai s ha allow species o ole a e en i on-
men al s esso s such as os , shade, and d ough (Rueda e al., 2017;
Zanne e al., 2014). The e o e, when a non-na i e species is in oduced
in a no el a ea, he in e ac ion o clima ic condi ions wi h he species
ai s de e mine which species will pe sis (Haeuse e al., 2018, 2017;
Dullinge e al., 2017; an de Veken e al., 2008). This in e ac i e
p ocess is no di e en in he case o u ban lo as, al hough ga dening
p ac ices (e.g. wa e ing, d ainage sys ems o excess wa e , p uning) can
o e come o compensa e o some o hese clima ic limi a ions.
The in oduc ion o non-na i e woody species in u ban pa ks can
pose a conse a ion p oblem because some o he species can escape,
es ablish popula ions (i.e. na u alize) in na u al ecosys ems and become
in asi e. Se e al in asi e woody species ha e he po en ial o dis up
he unc ioning o na u al ecosys ems causing en i onmen al impac s, as
well as socio-economic and human heal h p oblems (Hulme, 2007;
Niineme s and Pe˜
nuelas, 2008; Pa ke e al., 1999). Al hough he
numbe o na u alized species has inc eased exponen ially du ing he
las decades (Seebens e al., 2017), we s ill ha e a poo unde s anding o
he d i e s in luencing he in oduc ion o non-na i e plan species in o
he u ban lo a compa ed o he d i e s in luencing he la e s ages o
he in asion p ocess. Along he in asion p ocess om anspo o
sp ead and in asion (Blackbu n e al., 2011; Colau i and MacIsaac,
2004), i has been well documen ed ha species ai s, clima ic condi-
ions and socioeconomic ac o s ul ima ely in luence in asion success.
Ne e heless, hese ac o s a e also likely o in luence he p obabili y
ha a species is in oduced somewhe e in he i s place, and i is wo h
no ing ha he species ai s de e mining in asion success migh no
necessa ily be he same as hose de e mining hei delibe a e in oduc-
ion, since he la e depends di ec ly on human pe cep ion, p e e ences
and uses.
Wi h his knowledge a hand, wha emains o be explo ed is how
hese h ee main componen s (i.e. species ai s, clima e and socioeco-
nomic con ex ) explain a ia ion in he p opo ion o non-na i e o na-
men al species (non-na i e species numbe / o al species numbe ) in
u ban pa ks. Indeed, he ela i e impo ance o each ac o is likely o
a y ac oss a la ge e i o y i he e is a ia ion in clima ic and socio-
economic a iables. Fo ins ance, se e e clima ic condi ions migh limi
he impo ance o socioeconomic a iables in de e mining he p opo -
ion o non-na i e o namen al species due o di ec en i onmen al
il e ing p ocesses (K a e al., 2015). Howe e , wi h a elaxa ion o
s ess ul condi ions due o mild clima ic condi ions o indi ec e ec s
media ed by human ga dening ac i i ies, human popula ion size o hei
income is expec ed o in luence he p opo ion o non-na i e o namen al
species plan ed in u ban pa ks. C i ically, hese ex insic ac o s a e
going o be media ed by plan ai s (Vaz e al., 2018). The e o e,
knowledge o he ai p o iles o o namen al plan communi ies in
conce wi h clima ic and socioeconomic co ela es will inc ease ou
unde s anding o he p ocesses d i ing he in oduc ion and composi-
ion o non-na i e lo as in u ban pa ks, and he es ablishmen o hese
non-na i e species. In hese lines, p e ious wo k has shown ha socio-
economic ac o s explain as much o he a ia ion o he dis ibu ion o
non-na i e species es ablished in na u al ecosys ems as clima ic ac o s
(Essl e al., 2011; Pyˇ
sek e al., 2010). Howe e , we do no know o wha
ex en he p opo ion o woody species na u alized in na u al ecosys-
ems is associa ed wi h median species ai alues o he pa ks, and hei
clima ic and socioeconomic a iables.
In his s udy, we analyze he species plan ed in u ban pa ks ha a y
in size and ounding yea . We speci ically ask (1) how clima ic and so-
cioeconomic cha ac e is ics o he own in which he pa ks a e loca ed,
de e mine he pa ks’ plan - ai p o iles and (2) how plan o igin (na i e
s. non-na i e) and he in asion s a us (es ablished s. no es ablished in
na u al ecosys ems) o he species plan ed in pa ks a e associa ed wi h
plan ai s, clima ic and socioeconomic a iables? To answe hese
ques ions, we ocused ou s udy on he peninsula e i o y o Spain
(he ea e Spain) o se e al easons. Fi s , he o namen al use o non-
na i e plan species in Spain is one o he main pa hways o in oduc-
ion o po en ial in asi e species (Sanz Elo za e al., 2004). Wi h his
de ailed in o ma ion, we can assess he ai p o iles o he communi ies
o non-na i e u ban lo as ha a e mo e likely o escape om pa ks and
become in asi e in na u al landscapes. Recen ly, 83 in asi e o po en-
ially in asi e species we e iden i ied in nu se y ca alogues, a ailable
o comme ce, including eigh egula ed in asi e species (Bay´
on and
Vil`
a, 2019). The exis ence o hese in asi e and po en ial in asi e spe-
cies o consume s make i easie o be plan ed in pa ks and ga dens
despi e ha i is o bidden. Second, Spain has a b oad ange o clima ic
condi ions ( om Medi e anean and semia id o empe a e and con i-
nen al clima es). We expec ha inland cold condi ions will impose a
s ong il e o non-na i e woody species, while d ough e ec s in some
coas al a eas in sou he n Spain will be amelio a ed by i iga ion du ing
summe . Mo eo e , he d ough pe iods cha ac e is ic o mos clima es
in Spain has e olu iona y limi ed he ai p o ile o na i e lo a o
de elop species wi h a summe lowe ing pe iod and dense oliage
(Godoy e al., 2009), and we would expec his o be e lec ed in
non-na i e species. Consequen ly, we expec a b oade a ia ion o ai
p o iles o non-na i e plan s in coas al owns. Finally, he owns selec ed
a e no equally weal hy ac oss he coun y. In gene al, he weal h o a
own and i s unemploymen a es migh ac as impo an d i e s
di e en ia ing he non-na i e lo as o u ban pa ks (Schwa z e al.,
2006; T en ano i e al., 2013; Vaz e al., 2017).
2. Me hods
2.1. U ban pa ks da ase
We ob ained lo is ic ca alogs o 46 u ban pa ks and ga dens
dis ibu ed in 23 owns ac oss he Spanish peninsula e i o y (Ap-
pendix 1) om he ‘Vi i lospa ques’ da abase (www. i i lospa ques.
es/; las accessed 12 Feb 2021), an online ool by he ‘Spanish Associ-
a ion o Public Pa ks and Ga dens’ (AEPJP, 2010). This da abase p o-
ides he mos upda ed and comple e in o ma ion o he species plan ed
wi hin u ban pa ks in Spain as well as addi ionally impo an de ails
such as o al pa k a ea, he ounding yea and he spa ial coo dina es o
mos pa ks. Missing in o ma ion o pa k a ea was comple ed by using
Google Maps ools and, o he yea he pa k was ounded, by ci y hall
websi es.
F om hese ca alogs, we selec ed he comple e lis o 486 woody
plan species (i.e. ees, sh ubs, a bo eal cac i and palm ees). He bs
we e excluded because hei plan ings can a y adically om yea o
yea . Once he species lis was ob ained, we used The Plan Lis (2013) o
ha monize scien i ic species names (www. heplan lis .o g; las accessed
5 Oc obe 2018). Non-speci ic axa, such as gene a wi h unspeci ied
species epi he s, such as Rosa spp. o hyb ids such as Ci us ×au -
an i olia, we e excluded. In he case o in aspeci ic axa (e.g. subspecies,
a ie ies), such as Ced us a lan ica a . glauca, we only kep he binomial
species names. Then, each species was classi ied, acco ding o hei
o igin, as na i e o non-na i e in Spain, based on Sanz-Elo za e al.
(2004) and Cas o iejo (2012). Fo non-na i e species, we consul ed
hei es ablishmen s a us in Spain in (Sanz Elo za e al., 2004), which
ollows he Richa dson e al. (2000) s a us de ini ions: es ablished and
no es ablished in na u al ecosys ems. We included as es ablished spe-
cies, hose ha can be ound g owing in na u al ecosys ems in Spain.
These include na u alized species ha ha e es ablished pe sis en pop-
ula ions, bu we he e also included non-na i es classi ied as casuals (i.e.
species ha a e equen ly ound in na u al a eas bu do no a ain
pe sis ing popula ions). We included casual non-na i es among he
es ablished ones because he na u alized s casual classi ica ion is no
e y p ecise o in oduced species in Spain, and because many species
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
3
conside ed casual a e known o cause en i onmen al impac s in na u al
a eas (And eu e al., 2009).
2.2. Socioeconomic and clima ic da ase
Clima ic a iables collec ed om Wo ldClim 2 (Fick and Hijmans,
2017) we e calcula ed o each pa k wi h QGIS so wa e (QGIS De el-
opmen Team, 2009). Ras e maps wi h 1 km esolu ion we e ob ained
o all he a iables a ailable in he sou ce: 11 empe a u e a iables
and 8 p ecipi a ion a iables, as well as al i ude (Table 1a). Finally, we
ob ained in o ma ion o nume ous socioeconomic a iables o he o al
a ea o he 23 owns selec ed. Speci ically, we ob ained all he a iables
a ailable a he Spanish S a is ics O ice (INE, 2017), which p o ide
in o ma ion o mean alues be ween 2010 and 2016 (Table 1b) o 11
a iables associa ed wi h popula ion demog aphy, 4 wi h u baniza ion,
and 6 wi h employmen and economy. We decided o ob ain all he in-
o ma ion a ailable because we do no ha e ‘a p io i’ expec a ion o
which a iables a e ele an o ou s udy. Such high dimensionali y was
la e educed when pe o ming s a is ical analyses (see below).
Due o he limi ed esolu ion o he da abase o socioeconomic
Table 1
Clima ic and socioeconomic a iables used o desc ibe 46 u ban pa ks in 23 owns ac oss Spain.
A) Clima ic a iables
Type o a iable Va iable Code Kind o a iable Uni Fi s PCA selec ion
Tempe a u e
Annual mean empe a u e mean_ quan i a i e Celsius deg ee (◦C) *
Mean diu nal ange diu n_ ang quan i a i e Celsius
Iso he mali y iso he m quan i a i e %
Tempe a u e seasonali y _season quan i a i e %
Maximum empe a u e o he wa mes mon h max_ _wm quan i a i e Celsius
Minimum empe a u e o he coldes mon h min_ _cm quan i a i e Celsius *
Annual ange o empe a u e _ann_ g quan i a i e Celsius *
Mean empe a u e o he we es qua e m _we _Q quan i a i e Celsius
Mean empe a u e o he d ies qua e m _d ie_Q quan i a i e Celsius
Mean empe a u e o he wa mes qua e m _wa m_Q quan i a i e Celsius
Mean empe a u e o he coldes qua e m _cold_q quan i a i e Celsius
P ecipi a ion
Annual p ecipi a ion ann_p ecip quan i a i e mm
P ecipi a ion o he we es mon h p ec_we M quan i a i e mm
P ecipi a ion o he d ies mon h p ec_d ieM quan i a i e mm
P ecipi a ion seasonali y p ec_seaso quan i a i e %
P ecipi a ion o he we es qua e p ec_we Q quan i a i e mm
P ecipi a ion o he d ies qua e p ec_d ieQ quan i a i e mm *
P ecipi a ion o he wa mes qua e p ec_wa mQ quan i a i e mm
P ecipi a ion o he coldes qua e p ec_coldQ quan i a i e mm
Al i ude al s quan i a i e m a.s.l.
B) Socioeconomic a iables
Type o a iable Va iable Code Kind o a iable Uni Fi s PCA
selec ion
Basic demog aphy
Numbe o esiden s n_ esid_10000 quan i a i e x 10 000
P opo ion o esiden s ≤14 yea s old n_0.14y quan i a i e %
P opo ion o esiden s 15−64 yea s old n_15.64y quan i a i e %
P opo ion o esiden s ≥65 yea s old n_.64y quan i a i e %
Median age median_age quan i a i e yea s *
P opo ion o na ional esiden s na ionals_pe cen quan i a i e %
P opo ion o o eign esiden s o eign_pe cen quan i a i e %
P opo ion o esiden s na i e om he own na i e_pe cen quan i a i e % *
P opo ion o esiden s bo n ou side he own bo n_ou s_pe cen quan i a i e %
Na ali y a e na ali y_pe _ housand quan i a i e ‰
Mo ali y a e mo ali y_pe _ housand quan i a i e ‰
U banis ics
Numbe o homes n_homes_10000 quan i a i e x 10 000
Numbe o con en ional homes (Ca as al) con _homes_10000 quan i a i e x 10 000
Mean household size (numbe o inhabi an s pe
home) home_size quan i a i e *
P opo ion o single-pe son homes single_pe s_home_pe cen quan i a i e %
Employmen and
economics
Unemploymen a e unemploy_pe cen quan i a i e %
Wo king popula ion in ac i e age (20−64 yea s old) X20.64y_pe cen _ocupa ed_in_ac i e quan i a i e % *
Ac i e popula ion ac i i y_pe cen quan i a i e %
Indus y wo ke s indus y_wo ke s_pe cen quan i a i e %
Se ice wo ke s se ice_wo ke s_pe cen quan i a i e %
A e age money incoming pe home a g_inc_eu o_10000 quan i a i e x 10 000
€
*
A ea o he own a ea_c quan i a i e km
2
C) Pa k-dependen a iables
Type o a iable Va iable Code Kind o a iable Uni
Pa k-dependen A ea o he pa k a ea_m_sq quan i a i e m
2
Founding yea o he pa k yea quan i a i e
Table 1. A i s compila ion o clima ic a iables (A) was pe o med h ough a sys ema ic sea ch om Wo ldClim 2 (Fick and Hijmans, 2017). The i s compila ion o
socioeconomic a iables (B) was pe o med h ough a sys ema ic sea ch om he Spanish S a is ics O ice (INE, 2017). Pa k-dependen a iables (C) we e mainly
ob ained om AEPJP (2010), and seconda y, om Google Map ools (a ea) and ci y hall websi es (yea ). Finally, we made a selec ion using a Pea son co ela ion
analysis (Appendix 2, Fig. A2.1 and A2.2) and PCA (Fig. 1a and b) in which we applied a b oken-s ick c i e ion (Appendix 2, Fig. A2.4 and A2.5), as desc ibed in he
Me hods.
*
Va iables ma ked wi h an as e isk we e inally selec ed by he PCA and b oken-s ick me hod.
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
4
a iables, possible a ia ions below he own le el, such as dis ic s o
neighbo hoods, we e no possible o add ess. The e o e, he socioeco-
nomic da a o pa ks wi hin he same ci y we e he same. We ob ained
da a om a wide ange o own sizes — om e y densely popula ed (e.g.
Ba celona) o e y small owns (e.g. Ciudad Real)— and clima ic con-
di ions — om a we empe a e clima e (e.g. Bilbao) o a ho and d y
Medi e anean clima e (Lo ca)—. The a e age numbe o pa ks wi hin
he same ci y is 2, and only 2 owns ha e mo e han 3 pa ks (Appendix
1).
2.3. Plan ai s da ase
We used se e al sou ces o in o ma ion o ob ain o gan-le el and
whole-plan ai s a he species le el. The majo i y o plan ai s we e
ex ac ed om he TRY Da abase (Ka ge e al., 2011). We pe o med a
second sea ch in se e al in e ne da abases (Table 2). Fo each species,
we ob ained 21 plan ai s (17 con inuous and 4 bina y) ela ed o
whole-plan cha ac e is ics, lea es, ep oduc i e o gans ( lowe s and
dispe sal uni s), plan lowe ing and ui ing phenology, and ole ances
o se e al en i onmen al s esso s such as os (minimum absolu e
empe a u e ole a ed and numbe o os days pe yea ), shade,
d ough , wa e logging and i e, as well as he minimum and maximum
ha diness alues and i s ange (Table 2). Ha diness zones a e ca o-
g aphical s anda diza ions o he a e age annual minimum win e
empe a u es di ided in o 10-deg ee F zones (S´
anchez de Lo -
enzo-C´
ace es, 2004; USDA, 2012). Thus, he ha diness alues o a gi en
species is de ined as he abili y o su i e and g ow in a speci ic ange o
ha diness zones. We conside he lowe ha diness zone alue ha he
species can ole a e as “minimum ha diness”, he highes alue as
“maximum ha diness”, and he di e en be ween bo h alues, he
“ ange o ha diness”. Spain includes ha diness zones 7a (-17.8 ◦C o -15
◦C) o 11b (7.2 ◦C–10 ◦C; Fick and Hijmans, 2017; S´
anchez de Lo -
enzo-C´
ace es, 2004).
This comp ehensi e sea ch did un o una ely no ende ai in o -
ma ion o all species. The e o e, we excluded o u he analyses i e
ai s wi h da a missing o mo e han 50 % o he species (Table 2). Fo
he emaining ai s, missing da a was impu ed. P e ious wo k has
shown ha impu ing missing alues is be e han simply emo ing
missing-da a a iables o missing-da a species, as emo ing da a could
bias he esul s (Penone e al., 2014). We applied nonpa ame ic missing
alue impu a ion using andom o es s implemen ed in he R package
‘missFo es ’ (S ekho en and Buhlmann, 2012), which uses a andom
o es ained on he obse ed alues o p edic he missing alues. This
impu a ion me hod can handle mul i a ia e da a consis ing o con in-
uous and ca ego ical a iables simul aneously, as was he case in ou
s udy. O e all, a e se en i e a ions, he ou -o -bag no malized oo
mean squa ed e o was low (1.1608; PFC 0.0295), indica ing a good
impu a ion pe o mance (S ekho en and Buhlmann, 2012).
2.4. S a is ical analyses
2.4.1. O e iew
Due o he high numbe o a iables ob ained, and some o hem
showing high collinea i y, we decided o educe he numbe o a iables
using s a is ical ools wi h he inal objec i e o a oid o e i ing he
s a is ical models (King and Jackson, 1999; Zhou e al., 2010). We i s
buil a da abase o he plan - ai p o ile o he pa ks, calcula ing he
median alue o each quan i a i e ai and he mean alue o each bi-
na y ai ( equency) o each pa k. We conside ed each pa k as a
eplica e. The educ ion o a iables was done on each o he h ee main
g oups o a iables selec ed (i.e. clima ic a iables, socioeconomic
a iables, and pa k ai p o ile) ollowing h ee complemen a y s eps.
Fi s , we pe o med a co ela ion ma ix o iden i y he mos co ela ed
a iables (Appendix 2). I wo a iables we e closely co ela ed
( -Pea son >0.70), we emo ed one o hem. Second, we pe o med a
p incipal componen analysis (PCA), in which we used a b oken s ick
c i e ion (Simon e al., 2011) on he i s and second dimensions (Ap-
pendix 2) o iden i y he main a iables explaining he ob ained PCA.
Thi d, we pe o med Lasso eg ession models (F iedman e al., 2010;
Simon e al., 2011; Tibshi ani, 1996) o exclude non-in o ma i e
explana o y a iables (Hesamian and Akba i, 2019). Wi h he comple-
ion o hese h ee s eps, we ob ained he inal se o a iables, which
we e used o bes p edic a ia ion in he p opo ion o plan o igin
(na i e s. non-na i e) and in he in asion s a us (es ablished s. no
es ablished in na u al ecosys ems). This s a is ical p ocess is ully
explained in Appendix 3.
Acco ding o his p ocedu e, he i s and second axes o he PCA on
clima ic a iables explained 39 % and 35 % o he a ia ion, espec i ely
(Fig. 1a). We selec ed i e clima ic a iables (annual mean empe a u e,
p ecipi a ion in d ies qua e , minimum empe a u e in coldes mon h,
and annual ange o empe a u es; Table 1a). Fo he PCA on socio-
economic a iables, he i s and second p incipal axes explained 37 %
and 25 % o he a ia ion, espec i ely (Fig. 1b). We selec ed i e so-
cioeconomic a iables (mean household size, median age o inhabi an s,
pe cen o na i e popula ion, pe cen o wo king popula ion be ween 20
and 64 yea s old in ac i e, and a e age incoming money pe home;
Table 1b). Finally, he i s and second p incipal axes o he PCA
in ol ing pa k ai p o iles explained 28 % and 19 % o he a ia ion,
espec i ely (Fig. 1c). We selec ed i e plan ai s (minimum, maximum
and ange o ha diness, ole ance o shade, and ha ing conspicuous
lowe s; Table 2).
Wi h he inal se o a iables selec ed, we hen pe o med gene al-
ized linea models (GLMs), using he ‘glm’ unc ion om he R package
‘s a s’ (R Co e Team, 2019) wi h a linea combina ion o p e iously
selec ed a iables as p edic o s in each case, he species ichness o he
pa ks as p io weigh s, and a ea and ounding yea o he pa k as
co a ia es because bigge and olde pa ks ha e had mo e space and ime
o accumula e a wide di e si y o species.
2.4.2. Rela ionships be ween plan - ai p o ile and clima ic and
socioeconomic cha ac e is ics
To answe he i s ques ion, namely whe he clima ic and socio-
economic a iables explain a ia ion in he unc ional plan - ai p o ile
ac oss pa ks, we conside ed in he Lasso eg essions all he clima ic and
socioeconomic a iables p eselec ed as independen a iables, while we
ea ed each p eselec ed ai (i.e. minimum, maximum and ange o
ha diness, ole ance o shade and ha ing conspicuous lowe s) as
dependen a iable in each model (see ‘O e iew’).
A e he educ ion o a iables, we pe o med i e GLMs (Poisson
dis ibu ion wi h he quasipoisson se ing o accoun o o e -
dispe sion), one o each ai selec ed o he pa k p o ile. P edic o s
we e a linea combina ion o clima ic and socioeconomic ac o s p e-
iously selec ed in he Lasso eg ession (i.e. minimum empe a u e in
he coldes mon h, annual empe a u e ange, and household size;
Table A4.1 om Appendix 4), and a ea and ounding yea o he pa k
we e included as co a ia es.
2.4.3. In luence o plan - ai p o ile, clima ic and socioeconomic
cha ac e is ics on he p opo ion o non-na i e species
To answe he second ques ion, namely, whe he species o igin:
na i e s. non-na i e; and es ablishmen s a us: es ablished s. no
es ablished; can be explained by clima ic and socioeconomic a iables,
and he pa ks’ plan - ai p o ile, we i s ob ained he numbe o na i e
and non-na i e species wi hin each pa k, and o he non-na i e ones,
he numbe o species ha ha e es ablished in na u al ecosys ems (being
in asi e, na u alized o casual, acco ding o Sanz Elo za (S´
anchez de
Lo enzo-C´
ace es, 2004; USDA, 2012)(2004) and o he wo ks (Bay´
on
and Vil`
a, 2019; Gass´
o e al., 2010). We used a pai ed S uden ’s - es (‘ .
es ’ unc ion o he ‘s a s’ basic R package; R Co e Team, 2019) o es
whe he he numbe s o non-na i e and na i e species we e simila .
Simila ly, we es ed whe he he numbe s o no es ablished and
es ablished non-na i e species we e simila .
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
5
Table 2
T ai s used o desc ibe woody plan species in 46 u ban pa ks in 23 owns ac oss Spain. (B¨
a els and Schmid (2014); Bo ´
anica (2014); B ines e al. (1997); Cen aal
Bu eau oo de S a is iek (2003); CSIC (2014); Cze epano (1995); Gallaghe and Leishman (2012); Hin ze e al. (2013); Moles e al. (2004); Mo ales (2002); Plan s Fo
A Fu u e (2010); S´
anchez de Lo enzo-C´
ace es (1999); Uni e sidad de M´
alaga (2011); USDA (2008); Whea ley (2009) and Wyse-Jackson (2006)).
Type o ai T ai Code Kind o
a iable
Uni % eal
da a
% impu a ed
da a
Fi s PCA
selec ion
Da a sou ces beyond
TRY
Whole plan Max heigh heigh _m quan i a i e m 100 % 0 % 1, 2, 3, 4, 5, 6, 7, 9, 10,
11, 13, 14, 16
Max li espan — quan i a i e yea s 44 % Va .Ex. 1, 2, 3, 4, 5, 6, 7, 11
Lea es
A ea lea _a ea_sq_mm quan i a i e sq. mm 54 % 46 % 1, 2, 3, 4, 5, 6, 7, 10,
17
Lamina leng h lea _leng h_cm quan i a i e cm 85 % 15 % 1, 2, 3, 4, 5, 6, 7, 10,
17
Lamina wid h lea _wid h_cm quan i a i e cm 63 % 37 % 1, 2, 3, 4, 5, 6, 7, 10,
17
Lea ype lea _ ype_0 o1 bina y 0 =b oadlea ed 97 % 3 % 1, 2, 3, 4, 5, 6, 7, 10,
18 1 =coni e
Rep oduc i e
o gans
Flo al uni size — quan i a i e mm 9 % Va .Ex. 1, 2, 3, 4, 5, 6, 7, 10
Conspicuous lowe isual_ lowe _0 o1 bina y 0 =No 93 % 7 % * 1, 2, 3, 4, 5, 6, 7, 10
1 =Yes
Pollina ion synd ome pollin_synd_0 o1 bina y
0 =
Anemophilous 93 % 7 % 1, 2, 3, 4, 5, 6, 7, 10,
12 1 =
En omophilous
Gymnospe m lowe s gymnosp_ lowe _0 o1 bina y 0 =No 93 % 7 % 1, 2, 3, 4, 5, 6, 7, 12
1 =Yes
Dispe sal uni mass — quan i a i e g 15 % Va .Ex. 1, 2, 3, 4, 5, 6, 7, 10,
12, 17
Plan Phenology
Rep oduc i e
phenology iming phenol_ iming quan i a i e o dinal mon h 85 % 15 % 1, 2, 3, 4, 5, 6, 7, 8, 9,
10, 11, 14, 16
Lea phenology lea _phen_0 o1 bina y 0 =deciduous 97 % 3 % 1, 2, 3, 4, 5, 6, 7, 8, 9,
11, 15, 18 1 =e e g een
Tole ances
To os empe a u e ol_ os _ emp quan i a i e Celsius 82 % 18 % 1, 2, 3, 4, 5, 6, 7, 9
To os days — quan i a i e num. days 32 % Va .Ex. 1, 2, 3, 4, 5, 6, 7
To shade ol_shade_0 o5 quan i a i e 1−5 92 % 8 % * 1, 2, 3, 4, 5, 6, 7, 9, 11
To d ough ol_d ough _0 o5 quan i a i e 1−5 89 % 11 % 1, 2, 3, 4, 5, 6, 7, 9, 11
To wa e logging ol_wa e logging_0 o5 quan i a i e 1−5 76 % 24 % 1, 2, 3, 4, 5, 6, 7, 9
To i e — quan i a i e 1−5 29 % Va .Ex. 1, 2, 3, 4, 5, 6, 7, 11
Ha diness min. ha diness_min quan i a i e 1−12 67 % 33 % * 1, 2, 3, 4, 5, 6, 7, 8, 9,
10, 11
Ha diness max. ha diness_max quan i a i e 1−12 67 % 33 % * 1, 2, 3, 4, 5, 6, 7, 8, 9,
10, 11
Ha diness ange ha diness_ ange quan i a i e 1−12
Calcula ed di ec ly:
di e ence be ween max.
and min. alues
* Calcula ed
Da a
sou ces
Num. Sou ce e e ence
1 B ines R, Tejuelo I, Bel ´
an P and Balague ´
A. (1997) Guía Ve de. h ps://www.guia e de.com/. Accessed 1 Sep 2017
2 CSIC A bolapp. (2014) In: A bolapp. h p://www.a bolapp.es/. Accessed 1 Sep 2017
3 Mo ales J. (2002) In oja dín. h p://www.in oja din.com/. Accessed 1 Sep 2017
4 S´
anchez de Lo enzo-C´
ace es JM. (1999) ´
A boles O namen ales. h p://a boleso namen ales.es/. Accessed 1 Sep 2017
5 Uni e sidad de M´
alaga. (2011) Ja dín Bo ´
anico de la Uni e sidad de M´
alaga. h p://www.ja dinbo anico.uma.es/ja dinbo anico/index.php. Accessed 1 Sep 2017
6 Wyse-Jackson P. (2006) Missou i Bo anical Ga den. h ps://www.missou ibo anicalga den.o g/. Accessed 1 Sep 2017
7 Bo ´
anica Y Ja dines. (2014) In: Bo ´
anica Y Ja dines. h p://www.bo anicayja dines.com/. Accessed 1 Sep 2017
8 B¨
a els A. and Schmid P. A. (2014) Enzyklop¨
adie de Ga engeh¨
olze, 2nd edn. Ulme E. Ve lag
9 Plan s Fo A Fu u e (PFAF). (2010) In: Plan s Fo A Fu u e. h ps://p a .o g/use /De aul .aspx. Accessed 1 Jun 2017
10 Wea hley, R. (2009) B and T Wo ld Seeds. h p://b-and- -wo ld-seeds.com/. Accessed 1 Jun 2017
11 USDA. (2008) USDA Plan s Da abase. h ps://plan s.usda.go /ja a/. Accessed 1 Jun 2017
12 Hin ze C, Heydel F, Hoppe C, e al (2013) D3: The Dispe sal and Diaspo e Da abase – Baseline da a and s a is ics on seed dispe sal. Pe spec i es in Plan Ecology, E olu ion
and Sys ema ics 15:180–192.
13 Cen aal Bu eau oo de S a is iek BioBase (2003) In: Cen aal Bu eau oo de S a is iek. h ps://www.cbs.nl/nl-nl/onze-diens en/me hoden/classi ica ies/o e ig/bio
base-2003/biobase-2003. Accessed 1 Jun 2017
14 Cze epano SK (1995)Vascula plan s o Russia and adjacen s a es ( he o me USSR). Camb idge Uni e si y P ess, New Yo k
15 Zanne AE, Tank DC, Co nwell WK, e al (2004) Th ee keys o he adia ion o angiospe ms in o eezing en i onmen s. Na u e 506:89–92.
16 Moles AT, Fals e DS, Leishman MR, Wes oby M (2012) Small-seeded species p oduce mo e seeds pe squa e me e o canopy pe yea , bu no pe indi idual pe
li e ime. Jou nal o Ecology 92:384–396.
17 Gallaghe RV, Leishman MR (2012) A global analysis o ai a ia ion and e olu ion in climbing plan s. Jou nal o Biogeog aphy 39:1757–1771.
18 Panchen ZA, P imack RB, No d B, e al (Panchen e al., 2014) Lea ou imes o empe a e woody plan s a e ela ed o phylogeny, deciduousness, g ow h habi and
wood ana omy. New Phy ol 203:1208–1219.
Table 2. A i s compila ion o plan ai s was pe o med h ough a sys ema ic sea ch in he TRY Da abase (Ka ge e al., 2011). A second sea ch was pe o med o
comple e he da abase —see Da a Sou ces. Then, hose a iables o which we had mo e han 50 % o he da a we e selec ed, and, on hem, we made a nonpa ame ic
missing alue impu a ion by andom o es (S ekho en and Buhlmann, 2012). ‘Va .Ex.’ indica es a iables excluded om impu a ion because no enough da a. Finally,
we made a selec ion using a Pea son co ela ion analysis (Appendix 2, Fig. A2.3) and PCA (Fig. 1c) in which we applied a b oken-s ick c i e ion (Appendix 2, Fig. A2.6)
as desc ibed in he Me hods.
*
Va iables ma ked wi h an as e isk we e inally selec ed by he PCA and b oken-s ick me hod.
´
A. Bay´
on e al.

U ban Fo es y & U ban G eening 63 (2021) 127215
6
Fig. 1. PCAs o ela ions o a iables among pa ks.
P incipal componen analyses (PCA) o A) clima ic a iables o
each pa k, B) socioeconomic a iables o he locali ies o each
pa k and C) median alues o he plan ai s in each pa k. See
Appendix 1 o he lis o locali ies and codes o pa ks, and Ta-
bles 1 and 2 o he desc ip ion o ai s and a iables. (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).
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
7
Then, o each pa k we calcula ed he ollowing p opo ions o spe-
cies:
•P opo ion o non-na i e species, pAi=Ai
Ni+Ai
•P opo ion o es ablished non-na i e species, pEi=Ei
NEi+Ei
Whe e Ni is he numbe o na i e species in pa k i; Ai is he numbe o
non-na i e species in pa k i; NEi is he numbe o non-na i e species ha
ha e no es ablished in na u al ecosys ems o pa k i; and Ei is he
numbe o non-na i e species ha ha e es ablished in na u al ecosys-
ems o pa k i. We es ed whe he hese p opo ions a ied signi ican ly
ac oss pa ks using a Pea son’s Chi-squa ed es (‘chisq. es ’ unc ion o
he R package ‘s a s’; R Co e Team, 2019).
A e he educ ion o he numbe o a iables by using co ela ions,
PCA and b oken s ick c i e ia, we conside ed o he lasso eg essions all
he emaining a iables and ai s p eselec ed as independen a iables,
while we ea ed he p opo ion o non-na i e species and he p opo ion
o es ablished species as dependen a iables in each model (see
‘O e iew’). Then we pe o med wo GLMs (binomial dis ibu ion), one
o each p opo ion. P edic o s we e a linea combina ion o ai s, cli-
ma ic and socioeconomic ac o s p e iously selec ed in he Lasso
eg ession (i.e. annual empe a u e ange, household size, median age o
own inhabi an s, pe cen o wo king popula ion in ac i e age (20−64
yea s old wi h job), maximum ha diness and ange o ha diness;
Table A4.2 om Appendix 4), and a ea and ounding yea o he pa k
we e included as co a ia es.
2.5. Da a esou ces
The da a unde pinning he analyses, epo ed in his pape , a e
deposi ed in he Zenodo eposi o y a h ps://doi.o g/10.5281/zen-
odo.4095422 (Bay´
on e al., 2020)
3. Resul s
O e all, he da abase has 486 woody species. On a e age, he e we e
53 ±5.14 (mean ±se) species pe pa k and he maximum numbe was
144 species. Many species appea only in one o a ew pa ks. Each
species is in 5 ±0.26 (mean ±SE) pa ks. Mo e p ecisely, 180 species
(37.04 %) appea in only one pa k each, 68 species (13.99 %) appea in
wo pa ks, and 45 species (9.26 %) in h ee pa ks, whe eas only 11
species (2.25 %) occu in hal o he pa ks o mo e (Appendix 1). The lis
o mos common species and i s s a us we e: Cup essus sempe i ens (32
pa ks; es ablished), Cel is aus alis (26 pa ks; na i e), Robinia pseudoa-
cacia (26 pa ks; es ablished), S yphnolobium japonicum (25 pa ks, non-
na i e), Ced us a lan ica (24 pa ks; es ablished), Ced us deoda a (24
pa ks; es ablished), Magnolia g andi lo a (24 pa ks; non-na i e), Ce cis
siliquas um (23 pa ks; es ablished), Lau us nobilis (23 pa ks; na i e),
Phoenix cana iensis (23 pa ks; non-na i e) and Populus alba (23 pa ks;
na i e). Acco ding o he o igin, we ound 86 na i e species (17.70 %)
and 400 non-na i e species (82.30 %), 130 (26.75 %) o which ha e
es ablished in na u al ecosys ems (Table 3).
3.1. Pa k cha ac e iza ion
The co elog am o clima ic a iables (Fig. A2.1 om Appendix 2)
shows ha he e is a s ong co ela ion be ween di e en ypes o cli-
ma ic a iables. Values o p ecipi a ion, excep seasonali y, showed e y
high co ela ions be ween each o he (minimum Pea son’s =0.49
be ween p ecipi a ion o he d ies qua e and he coldes qua e ). The
i s axis o he clima ic PCA accoun ed o 38.5 % o he o e all clima ic
a ia ion ac oss pa ks, and i was mainly explained by he opposi e end
be ween p ecipi a ion and mean empe a u e in he d ies qua e . The
second PCA axis accoun ed o 34.9 % o he obse ed clima ic a ia ion
ac oss pa ks, and i was explained by he same beha io o he minimum
empe a u e in he coldes mon h and he mean in he we es qua e
(Fig. 1a).
Among socioeconomic a iables, he highes co ela ions (Fig. A2.2
om Appendix 2) was obse ed be ween he numbe o esiden s and
numbe o homes (Pea son’s =1). A e age household sizes in he
owns o he pa ks we e di ec ly ela ed o he p opo ion o kids and
adul inhabi an s (Pea son’s =0.62 o kids unde 14 yea s old, and =
0.60 o adul s be ween 15 and 64 yea s old), and in e sely o he
p opo ion o people o e 64 yea s old ( =-0.70). This p opo ion o
people o e 64 is in addi ion di ec ly ela ed wi h he median age o he
popula ion (Pea son’s =0.97). Mo eo e , we ound ha o he so-
cioeconomic PCA, household size and median age we e bo h dis ibu ed
along he i s axis (36.7 %) bu in opposi e di ec ions. Median age o he
popula ion ollowed a e y simila pa e n o he pe cen age o wo ke s
in he se ice sec o wi hin he i s socioeconomic PCA axis. Finally, he
pe cen age o employed people, as well as he a e age income pe home,
we e bo h dis ibu ed along he second PCA axis (25 %) and in he same
di ec ion (Fig. 1b).
Rega ding he pa ks’ plan - ai p o iles (Table 2), some a iables
we e highly co ela ed (Appendix 2, ig. A2.3). Fo example, en o-
mophilous pollina ion was co ela ed wi h se e al o he ai s such as
he p esence o conspicuous lowe s (Pea son’s =1), maximum heigh
(Pea son’s =-0.82) and he ole a ed minimum empe a u e (Pea -
son’s =0.52). As expec ed, he e was also a high co ela ion be ween
cold ole ance and he minimum alue o ha diness (Pea son’s =0.79).
O he co ela ions wo h men ioning we e hose ound be ween he ype
o lea (coni e needle – b oad lea ) and whe he o no species a e
gymnospe m (Pea son’s =0.98). Lea size a iables we e all closely
co ela ed wi h each o he (Pea son’s =0.76 o wid h s a ea; 0.76 o
a ea s leng h; 0.73 o wid h s leng h). Wi h all his a ia ion, he i s
axis o he plan ai PCA accoun ed o 28 % o he o e all ai a i-
a ion and i was mainly explained by an opposi e end be ween mini-
mum ha diness, ha diness ange and he p esence o conspicuous
lowe s. In addi ion, maximum ha diness and lea wid h was also
dis ibu ed along he i s axis o a ia ion bu o e all had a less s a is-
ical weigh . Shade and d ough ole ance we e pa ly ela ed o he
second axis o he PCA in opposi e di ec ions (Fig. 1c).
3.2. Rela ionships be ween plan - ai p o iles and clima ic and
socioeconomic cha ac e is ics
A e educing model dimensionali y, wo clima ic a iables we e
e ained o explain some a ia ion in he ai p o ile ac oss pa ks. These
a iables we e he minimum empe a u e in he coldes mon h, and he
ange o empe a u es. The emaining p edic o s, hese a e, household
size, ounding yea o he pa k and pa k a ea, did no explain any o he
ai s selec ed as dependen a iables (Table 4). Speci ically, minimum
and maximum plan ha diness as well as he p esence o conspicuous
lowe s we e posi i ely ela ed o minimum empe a u e in he coldes
mon h. On he o he hand, ha diness ange and shade ole ance we e
nega i ely ela ed o minimum empe a u e in he coldes mon h
(Fig. 2a). Simila ly, we obse ed a signi ican posi i e ela ionship be-
ween he annual ange o empe a u es and he minimum and
maximum ha diness (i.e. highe maximum ha diness alues means
plan s a e less ole an o cold), while he ela ionship be ween he
annual ange o empe a u es and he ole ance o shade was nega i e.
Howe e , he e was no signi ican ela ionship be ween he annual
Table 3
Numbe o woody o namen al plan species by s a us and g ow h o m in 46
u ban pa ks in 23 owns o Spain.
T ees Sh ubs Palms A bo eal Cac i To al
Na i e 59 26 1 0 86
Non-na i e no es ablished 154 85 30 1 270
Non-na i e es ablished 95 30 4 1 130
To al 308 141 35 2 486
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
8
ange o empe a u es and he ha diness ange, no wi h he p esence o
isual and conspicuous lowe s (Fig. 2b).
In luence o plan - ai p o iles, clima ic and socioeconomic cha -
ac e is ics on he p opo ion o non-na i e species
O e all, he e is a g ea e p esence o non-na i e (39.2 ±4.15; mean
±se) han o na i e species (14.2 ±1.27; pai ed - es ; =7.475; p- alue
<0.01; Fig. 3a) ac oss pa ks. Fo he case o non-na i e species, he
numbe o species ha a e no known o ha e es ablished in na u al
ecosys ems (21.15 ±2.06) is signi ican ly la ge han he numbe o
es ablished species (18.02 ±2.27; pai ed - es ; =2.4393; p- alue =
0.0187; Fig. 3b). The p opo ion o non-na i e species and he p opo -
ion o es ablished species a e bo h signi ican ly he e ogeneous ac oss
pa ks (Chi-squa ed =132.98, and 148.17, espec i ely; p <0.01 in bo h
cases). A e educing model dimensionali y, only one clima ic a iable
(annual ange o empe a u e), wo socioeconomic a iables (mean o
household size, and median age o inhabi an s) and one plan ai o he
pa k p o ile ( ange o ha diness) explained he p opo ion o non-na i e
species. The emaining p edic o s (pe cen age o ac i e popula ion,
maximum ha diness, ounding yea o he pa k, and pa k a ea) did no
explain any o he p opo ions selec ed as dependen a iables.
The GLM analysis indica es (Table 5) a signi ican nega i e ela-
ionship be ween annual ange o empe a u e and he p opo ion o
es ablished species bu no wi h he p opo ion o non-na i e species
(Fig. 4a). Rega ding socioeconomic a iables, mean o household size
and median age bo h signi ican ly explained he p opo ion o non-
na i e species bu no he p opo ion o es ablished species (Fig. 4b,
c). Finally, a ia ion in ha diness ange o he pa k’s p o ile nega i ely
explained he p opo ion o non-na i e species, bu i did no explain he
p opo ion o es ablished species (Fig. 4d).
4. Discussion
We ha e a poo unde s anding o he combina ion o plan ai s,
clima e a iables and socioeconomic ac o s guiding he plan ing on
non-na i e woody plan species in u ban pa ks. This knowledge is
essen ial o iden i y how di e en d i e s in luence he success o non-
na i e species om in oduc ion o in asion. He e, we p esen insigh s
om a la ge ep esen a ion o u ban pa ks ac oss Spain showing ha
hey ha bo a mo e non-na i e han na i e plan species. In e es ingly,
his high p opo ion o non-na i e woody species (82.3 %) is no asso-
cia ed wi h a pa icula se o axa. Ra he , Spanish pa ks p esen a high
species u no e , meaning ha he as majo i y o woody plan species
appea in one o e y ew pa ks, while e y ew species a e p esen in he
majo i y o pa ks. The a e age pe cen age o non-na i e woody species
in Spanish u ban pa ks is much highe han in Cen al Eu opean u ban
lo as (Pyˇ
sek, 1998), and many o he egions (Dangulla e al., 2019; de
F ei as e al., 2019; Jha e al., 2019; Temple on e al., 2019), whe e he
a e age ep esen a ion o non-na i e species is be ween 40 % and 50 %.
In line wi h his inding, he pe cen age o non-na i e species a ailable
in Spanish nu se ies is 76 % app oxima ely (Bay´
on and Vil`
a, 2019). The
small di e ence be ween he pe cen age in pa ks and nu se ies is
p obably no signi ican , and can be due o he ac ha he s udy o
Spanish nu se ies includes all ypes o plan g ow h o ms, while his
s udy only ocus on woody species. Mo eo e , many plan s in pa ks a e
no longe sold in Spanish nu se ies (Bay´
on and Vil`
a, 2019) such as
Buddleja madagasca iensis, Casua ina cunninghamiana, Elaeagnus pungens,
F axinus pennsyl anica, Pa kinsonia aculea a, Pi ospo um undula um,
Ricinus communis and Schinus molle.
We ha e ound ha he plan ai composi ion o u ban pa ks can be
explained by a combina ion o socioeconomic and clima ic a iables. We
obse ed a ia ion in he p opo ion o species wi h conspicuous lowe s
(which is a di ec p oxy o he pollina ion synd ome) ac oss u ban pa ks,
and his a ia ion was explained by minimum empe a u e in he coldes
mon h. Species wi h en omophilous pollina ion end o be poo ly
ole an o os , p obably due o he dependence o hei pollina ing
insec s on milde empe a u es. Meanwhile, wind-pollina ed plan s ha e
a wide ange o ole ance o os . Fo ins ance, oak species a e known
o be os ole an , whe eas o he species like palms do no show he
same adap a ions. The la e a e p edominan ly ound in owns wi h
oceanic clima e, whe e os is a a e phenomenon. The pollina ion
Table 4
Gene al Linea ized Models on plan ai s.
A) Minimum ha diness Es ima e S d. E o alue P (>|
|)
(In e cep ) 1.5667 0.3541 4.4244 0.0001 ***
Minimum empe a u e
o he coldes mon h
0.0456 0.0071 6.4694 <
0.0001
***
Annual ange o
empe a u e
0.0191 0.0048 3.9906 0.0003 ***
Mean household size −0.0271 0.0448 −0.6033 0.5497
Founding yea o he
pa k
−0.0002 0.0002 −1.1550 0.2549 .
A ea o he pa k −1.9019E-
07
1.2108E-
07
−1.5708 0.1241
B) Maximum ha diness Es ima e S d. E o alue P (>|
|)
(In e cep ) 2.1459 0.1665 12.8890 <
0.0001
***
Minimum empe a u e
o he coldes mon h
0.0180 0.0033 5.4110 <
0.0001
***
Annual ange o
empe a u e
0.0082 0.0023 3.6049 <
0.0001
***
Mean household size −0.0216 0.0209 −1.0313 0.3086
Founding yea o he
pa k
−0.0001 0.0001 −0.9207 0.3628
A ea o he pa k −1.8047E-
08
5.3817E-
08
−0.3353 0.7391
C) Range o ha diness Es ima e S d. E o alue P (>|
|)
(In e cep ) 1.2575 0.4182 3.0068 0.0045 **
Minimum empe a u e o
he coldes mon h
−0.0288 0.0084 −3.4244 0.0014 **
Annual ange o
empe a u e
−0.0080 0.0057 −1.3926 0.1714
Mean household size −0.0304 0.0513 −0.5929 0.5566
Founding yea o he pa k 0.0001 0.0002 0.6964 0.4902
A ea o he pa k 6.0555E-
08
1.2795E-
07
0.4733 0.6386
D) Tole ance o shade Es ima e S d. E o alue P (>|
|)
(In e cep ) 2.0327 0.5283 3.8473 0.0004 ***
Minimum empe a u e
o he coldes mon h
−0.0376 0.0107 −3.5268 0.0011 **
Annual ange o
empe a u e
−0.0320 0.0072 −4.4644 0.0001 ***
Mean household size 0.0829 0.0687 1.2063 0.2348
Founding yea o he
pa k
−0.0003 0.0002 −1.3355 0.1893
A ea o he pa k −1.2227E-
07
1.6867E-
07
−0.7249 0.4727
E) P esence o
conspicuous lowe s
Es ima e S d. E o alue P (>|
|)
(In e cep ) −3.3205 1.1389 −2.9156 0.0058 **
Minimum empe a u e
o he coldes mon h
0.0815 0.0220 3.7109 0.0006 ***
Annual ange o
empe a u e
0.0273 0.0146 1.8666 0.0693 .
Mean household size 0.1330 0.1388 0.9586 0.3435
Founding yea o he
pa k
0.0008 0.0005 1.5061 0.1399
A ea o he pa k −3.0225E-
07
3.8541E-
07
−0.7842 0.4375
Gene alized linea models o he i e plan ai s selec ed as dependen a iables,
e sus clima ic and socioeconomic a iables p eselec ed by lasso eg ession
(Table A4.1 om Appendix 4) as independen a iables.
**
0.001 <p ≤0.01.
***
p ≤0.001, o he wise non signi ican . G aphical ep esen a ion in Fig. 2.
´
A. Bay´
on e al.
U ban Fo es y & U ban G eening 63 (2021) 127215
9
synd ome can change in u ban lo as depending on hei na i e and non-
na i e o igin. Fo ins ance, A onson e al. (2007) documen ed ha
conspicuous lowe s, ha ends o be beau i ul, di e in hei pollina ion
equi emen s; howe e , we did no ind his pa e n in ou s udy. Di -
e ences be ween esul s migh be ela ed o he ac ha he s udy o
A onson e al. (2007) was conduc ed in si es wi h simila en i onmen al
condi ions, while ou s udy includes owns wi h a wide ange o clima es
om con inen al o Medi e anean, and om inland o oceanic clima ic
condi ions.
Besides cold ac ing as a il e o ep oduc ion synd omes, we ound
ha he ole ance o shade inc eased wi h highe minimum empe a u e
in he coldes mon h and wi h wide he mal ange. In egions wi h an
oceanic clima e, such as he no he n coas , win e s a e milde and
cloudiness is mo e equen and las s longe , especially in he summe
mon hs, han in he es o Spain. Howe e , in egions whe e he clima e
is mo e con inen al, as in he pla eaus, win e s a e colde and summe s
a e no only ho e bu also e y sunny wi h high e apo anspi a ion
a es (AEMET, 2019). These di e ences a e likely o cing plan species
selec ed o be plan ed in oceanic clima es o be mo e ole an o inso-
la ion, educing lea a ea o a oid wa e loss, while plan s wi h la ge
lea es end o be dis ibu ed in a eas wi h highe wa e a ailabili y and
lowe i adiance (Holmg en e al., 2012; Ma kes eijn and Poo e ,
2009). Finally, he minimum and maximum ha diness alues o he
plan s we e di ec ly ela ed o bo h minimum empe a u e in he coldes
mon h and highe annual anges o empe a u e. This is indeed an
ob ious ela ionship, as ha diness is a p oxy o he ole ance o low
empe a u es (S´
anchez de Lo enzo-C´
ace es, 2004; USDA, 2012), and i
ep esen s a p ima y ac o in selec ing which species a e plan ed.
Speci ically, hose species wi h high ole ance o cold, a e usually no
plan ed in wa m places e en i hey could su i e he e. This explana-
ion is suppo ed by he ac ha he ange o ha diness dec eases wi h
he minimum empe a u e in he coldes mon h. These esul s sugges
he e is a ade-o be ween species ha ole a e shade and hose ha
ha e conspicuous lowe s, as well as wi h hose ha ha e highe
ha diness alues.
We expec ed ha di e ences ac oss pa ks be ween he p opo ion o
Fig. 2. Rela ionship o i e plan ai s o (A) he
minimum empe a u e in he coldes mon h and (B) he
annual ange o empe a u e.
Gene alized linea models (GLMs) o plan ai s e sus
minimum empe a u e in coldes mon h and annual
ange o empe a u e. The con inuous ai a iables (i.
e. minimum ha diness, maximum ha diness, ange o
ha diness [scale 1–12], and shade ole ance [scale
1–5]) is shown on he main y axis (le side o he
g aph), while he scale o he unique bina y ai a i-
able (i.e. p esence o conspicuous lowe ), con e ed o
p opo ion o species ha ing pe pa k, is shown on he
seconda y y axis ( igh side). * 0.01 <p ≤0.05, **
0.001 <p ≤0.01, *** p ≤0.001, o he wise non-
signi ican . P- alues in Table 4. (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).
´
A. Bay´
on e al.