Ea lywood essel cha ac e is ics a e ea ly indica o s o d ough -induced
decline in ing-po ous oak species wi hin he Medi e anean Basin
Michele Colangelo
a,*
, An onio Gazol
b
, J. Julio Cama e o
b
, Ma co Bo ghe i
a
,
Raúl S´
anchez-Salgue o
c
, Luis Ma ias
d
, Ma ia Cas ellane a
a
, Paola Nola
e
, F ancesco Ripullone
a
a
Dipa imen o di Scienze Ag a ie, Fo es ali, Alimen a i ed Ambien ali, Uni e si `
a degli S udi della Basilica a, 85100 Po enza, I aly
b
Ins i u o Pi enaico de Ecología (IPE-CSIC), A da. Mon a˜
nana 1005, 50192 Za agoza, Spain
c
Depa amen o de Sis emas Físicos, Químicos y Na u ales, Uni e sidad Pablo de Ola ide, 41013 Se illa, Spain
d
Depa amen o de Biología Vege al y Ecología, Uni e sidad de Se illa, 41012 Se illa, Spain
e
Dipa imen o Scienze della Te a e dell'Ambien e, Uni e si `
a degli S udi di Pa ia, 27100 Pa ia, I aly
HIGHLIGHTS GRAPHICAL ABSTRACT
•Ea lywood ana omy can se e as a
aluable ea ly-wa ning signals o oak
decline.
•G ow h a es di e ed be ween igo
classes 10 o 40 yea s be o e decay.
•Dieback is p eceded by decline o essel
a ea dispe sion decades ea lie .
•Dispe sion and spa ial a angemen
me ic di e en ia es ees based on
igo class
•Di e en esponses be ween igo clas-
ses can explain physiological adjus -
men s e lec ed in wood ana omy.
ARTICLE INFO
Edi o : Manuel Es eban Lucas-Bo ja
Keywo ds:
Clima e change
Que cus
Xylem
Vessel a ea
Vessel dis ibu ion
Ea lywood ana omy
ABSTRACT
Hea and d ough s ess ha e igge ed o es dieback episodes wo ldwide, a ec ing oak o es s, pa icula ly in
ho spo s o clima e change such as he Medi e anean Basin. Howe e , o ecas ing dieback is no s aigh o wa d.
In his s udy, we used he ea lywood ana omy o imp o e dieback o ecas s in i e oak species cha ac e ized by
di e en d ough sensi i i y (i.e. om high o low Que cus obu , Q. ce is, Q. aine o and Q. cana iensis, Q.
humilis, Q. pubescens) ac oss I aly and Spain. We measu ed adial g ow h, exp essed as basal a ea inc emen
(BAI), ea lywood hyd aulic diame e (Dh) and essel a ea o coexis ing non-declining (ND) and declining (D)
ees in each s and. Then, we calcula ed he p oduc be ween he coe icien o a ia ion (CV) o essel a ea and a
spa ial agg ega ion index (AI). High CV ×AI alues indica e egula ly spaced essels wi h a iable a ea o
essels, while low alues co espond o clus e ed essels wi h simila a ea. ND ees showed highe BAI alues
han D ees om 10 o 40 yea s be o e he dieback onse , when ND ees g ew 20–50 % mo e han he D ees.
We obse ed a decline in he essel a ea CV se e al decades p io o dieback in D ees, wi h he excep ion o
* Co esponding au ho .
E-mail add esses: [email p o ec ed] (M. Colangelo), [email p o ec ed] (A. Gazol), [email p o ec ed] (J.J. Cama e o), [email p o ec ed]
(M. Bo ghe i), [email p o ec ed] (R. S´
anchez-Salgue o), [email p o ec ed] (L. Ma ias), [email p o ec ed] (M. Cas ellane a), [email p o ec ed] (P. Nola),
[email p o ec ed] (F. Ripullone).
Con en s lis s a ailable a ScienceDi ec
Science o he To al En i onmen
jou nal homepage: www.else ie .com/loca e/sci o en
h ps://doi.o g/10.1016/j.sci o en .2025.179565
Recei ed 3 Janua y 2025; Recei ed in e ised o m 25 Ap il 2025; Accep ed 28 Ap il 2025
Science o he To al En i onmen 980 (2025) 179565
A ailable online 3 May 2025
0048-9697/© 2025 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY license ( h p://c ea i ecommons.o g/licenses/by/4.0/ ).
Q. ce is. The AI showed highe alues in ND han in D ees. Consequen ly, he CV ×AI p oduc was consis en ly
highe in ND han in D ees. The CV ×AI di e gence be ween ND and D ees was p onounced in he we es
si es, speci ically o Q. obu and Q. humilis. Time se ies o CV ×AI e ec i ely di e en ia ed ees based on hei
igo . Wood ana omy a iables could be used o enhance p edic ions o ulne abili y o d ough -induced
dieback. This s udy can help iden i y ulne able ees be o e he onse o dieback symp oms, se ing as a ool
o suppo he managemen o o es s p one o d ough .
1. In oduc ion
Fo es s expe iencing d ough -induced dieback se e as aluable
na u al labo a o ies o iden i ying po en ial ac o s ha igge g ow h
decline and o de e mining ulne abili y h esholds (Allen e al., 2010;
Cama e o e al., 2015; S´
anchez-Salgue o e al., 2017; DeSo o e al.,
2020; P eisle e al., 2021). The assessmen o o es ulne abili y o
dieback is no s aigh o wa d due o he complex in e ac ions o ac o s
ha in luence he esponse o ees o clima e s ess. Fo his eason,
some s udies ha e used p oxies o dieback such as c own de olia ion,
ee g ow h and wood ana omy da a o assess changes in ee igo (e.g.
Pellizza i e al., 2016; Colangelo e al., 2017a, 2017b; Cama e o e al.,
2021; Vale iano e al., 2021a; Vale iano e al., 2021b; Cas ellane a e al.,
2022).
Radial g ow h da a ha e been used o de ec ea ly-wa ning signals o
impending dieback, bu hey we e no e y success ul in he case o oaks
which show a high capaci y o apidly eco e a e d ough (Caille e
e al., 2019). The a iabili y in xylem essel s uc u e be ween yea s is a
c ucial ac o in unde s anding he po en ial o adap a ion in wood
ana omy esponses o d ough (A güelles-Ma ´
on e al., 2023). Ex ensi e
me a-analyses examining he po en ial o xylem unc ional ai s as in-
dica o s o ee igo and decline in Medi e anean o es s we e p e-
sen ed by Bo ghe i e al. (2017, 2020). They e idenced ha wood
ana omical ea u es se e as ea ly wa ning signals o o es ulne abili y
in d ough -p one egions. I is hus jus i ied o sea ch o di e ences in
wood-ana omical ai s be ween coexis ing declining and non-declining
ees which may subsequen ly be used as po en ial ea ly-wa ning signals
o ee mo ali y.
Hyd aulic ailu e caused by clima ic s ess phenomena p oduce
isible s uc u al changes a he canopy le el, such as desicca ion and
loss o igo o he mos ma ginal po ions o he canopy (Na dini e al.,
2013).
O e all, deciduous ing-po ous oak species a e ega ded as d ough -
ole an , anisohyd ic species which main ain high pho osyn he ic a es
a low wa e po en ial (Ep on and D eye , 1993; Roman e al., 2015).
Addi ionally, hey show he capaci y o ap wa e om deep sou ces
du ing d y pe iods (Skiada esis e al., 2019; Ripullone e al., 2020;
Kahmen e al., 2022). The la ge ea lywood essels o hese species
enable he main enance o high hyd aulic conduc i i y and s oma al
conduc ance a es du ing he ea ly g owing season. (Klein, 2014; Guada
e al., 2019). Howe e , he o ma ion o la ge essels can also inc ease
he ulne abili y o d ough by enhancing he isk o xylem embolism
(Spe y e al., 1994; P´
e ez-de-Lis e al., 2018). In ac , deciduous oaks
end o o m ea lywood essels wi h smalle lumen in d ie si es (Alla
and Cama e o, 2012). Recen ee- ing s udies indica e ha he g ow h
o deciduous oaks is ulne able o d ough , and he e ec s o p e ious
d ough e en s— e e ed o as legacies— a y depending on he species
(Cama e o e al., 2021; Bose e al., 2024). D ough legacies ha e also
been ound o a ec also wood ana omy in he yea ollowing a d ough
e en , esul ing in he o ma ion o smalle essels (Süßel and B ügge-
mann, 2021), due o he impo ance o p e ious-yea clima e condi ions
on ea lywood o ma ion (Ma chand e al., 2021). Thus, in e annual
a ia ions in ing-wid h and essel lumen dimensions o e comple-
men a y in o ma ion on oak pe o mance (Alla and Cama e o, 2012;
Colangelo e al., 2017a; P´
e ez-de-Lis e al., 2018; Guada e al., 2021),
which migh help o o ecas dieback a ee le el.
Mo eo e , no only do he size and densi y o essels a y wi h
en i onmen al condi ions, bu d ough can also al e he dis ibu ion o
essels wi hin he ings, shi ing om a ing-po ous o a semi- ing
po ous s uc u e, as obse ed in some Fagaceae species (Cas agne i
e al., 2020; G anda e al., 2017; Nola e al., 2020; A niˇ
c e al., 2021).
This means ha he pa e ns o spa ial agg ega ion and lumen size o
essels change wi hin he ea lywood. Fa o able g ow h condi ions
du ing he p e ious win e can a o he o ma ion o wide ea lywood
essels (Ga cía-Gonz´
alez and Ecks ein, 2003; Ma chand e al., 2021)
e en unde subsequen d y condi ions (Van de We e al., 2007). Wa m
empe a u es and high mois u e a ailabili y du ing he g owing season
enhance essel enla gemen and la ewood o ma ion (G iˇ
ca e al.,
2013; Guada e al., 2021). Howe e , he ing conduc i e a ea and
la ewood p oduc ion dec ease in esponse o d ough (Alla and
Cama e o, 2012; Cas agne i e al., 2020), and his migh also be a
consequence o he spa ial a angemen o essels and hei a ia ion in
lumen size, which a e p oxies o pheno ypic plas ici y. Thus, yea - o-
yea a ia ions in essel size, densi y, and o e all conduc i e a ea can
p o ide insigh s in o clima e limi a ions and po en ially di e en ia e
be ween declining and non-declining indi iduals (Fon i and Ga cía-
Gonz´
alez, 2008; P´
e ez-de-Lis e al., 2018). Addi ionally, gi en ha
essel size a ies wi hin he ing, he po en ial o wi hin- ing essel size
a ia ion can be in es iga ed as an ea ly-wa ning signal o ee g ow h
decline and o es dieback (G iˇ
ca e al., 2013).
We expec ha d ough will educe he numbe and size o ea ly-
wood essels hus a ec ing wi hin- ing spa ial clus e ing o essels and
hei coe icien o a ia ion in size. This is based on he impo ance o
wa m empe a u e and mois u e a ailabili y o essel o ma ion and
enla gemen (Van de We e al., 2007; G iˇ
ca e al., 2013; Guada e al.,
2021). Besides, we hypo hesizes ha spa ial clus e ing o ea lywood
essels will inc ease and a ia ion in lumen size will dec ease in ees
exhibi ing signs o d ough -induced dieback, such as educed g ow h
a es.
This is undamen ed on he impo ance ha d ough legacies can
ha e on ee g ow h and ing o ma ion (Süßel and B üggemann, 2021).
To es his hypo hesis, we quan i ied and analysed yea - o-yea a ia-
ions in adial g ow h and ea lywood essel a ea o i e deciduous oak
species showing di e en sensi i i y o d ough (Que cus obu L.,
Que cus ce is L., Que cus aine o Ten., Que cus cana iensis Willd.,
Que cus humilis Mill., Que cus pubescens Willd.) in o es s ac oss Spain
and I aly. To assess po en ial di e ences be ween non-declining and
declining ees, he s udy was ca ied ou in si es showing e idence o
d ough -induced dieback and conside ed wo igo classes.
2. Ma e ial and me hods
2.1. S udy si es and ee species
We s udied i e deciduous, ing-po ous oak species widely dis ib-
u ed in empe a e and Medi e anean o es s o Eu ope and showing
ecen d ough -induced dieback. This was done in i e si es loca ed in
I aly and Spain, whe e in ensi e moni o ing o dieback phenomena was
ca ied ou (Table 1, Fig. S1). The selec ed species occupy wide
biogeog aphical and clima ic g adien s om no he n Eu ope o no h-
e n A ica and all a e p esen and common in mixed o es s o he
Medi e anean Basin.
F om a axonomic pe spec i e Q. pubescens in I alian and Q. humilis in
Spanish si es a e conside ed synonym o he same species (Di Pie o
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
2
e al., 2020). Indeed, Q. pubescens is conside ed a alid binomial ac oss
Eu ope, excep o Spain, whe e i is eplaced by Q. humilis Mill. (Ama al
F anco, 1990). Gi en hese egional di e ences in nomencla u e and o
main ain consis ency wi h ou p e ious s udies, we ha e chosen o keep
Q. pubescens ( o I alian si es) and Q. humilis ( o Spanish si es) sepa a e
in ou analysis.
The Ticino si e is a mixed loodplain o es spanning app oxima ely
500 ha wi hin he “Pa co Lomba do della Valle del Ticino” in No hwes
I aly nea he Ticino i e (Pe icolo e al., 2023). The o es is p ima ily
composed o peduncula e oak (Que cus obu L.), ash (F axinus excelsio
L.), black alde (Alnus glu inosa (L.) Gae n.) and elm (Ulmus mino Mill.).
In his a ea, mean oak densi y and basal a eas we e 179 s ems ha
−1
and
18.8 m
2
ha
−1
, espec i ely. Clima e is empe a e wi h a mean annual
empe a u e o 12.8 ◦C and annual o al p ecipi a ion o 1104 mm.
Subs a es a e g a els and sand-clay deposi s. Acco ding o local epo s
(DEPFAR, 2026), oaks showed decline symp oms since he ea ly 2000s
(shoo dieback, lea loss and wi he ing, epico mics shoo s, g ow h
decline and ecen inc ease in mo ali y) (Colangelo e al., 2018a).
The A amendia si e is loca ed in no he n Spain. He e, une en-aged
s ands o Que cus humilis Mill. we e sampled. The s udy o es showed
abundan dieback symp oms since 2017 a ec ing a leas 25 % o he
indi iduals, wi h sca e ed g oups o ecen ly dead ees (Cama e o
e al., 2021). This is a high o es wi h basic soils and unde s o y
domina ed by Buxus sempe i ens L., Junipe us oxyced us L. and C a aegus
monogyna Jacq. In his si e, ee densi y and basal a ea we e 1150 s ems
ha
−1
and 36.9 m
2
ha
−1
, espec i ely. He e, clima e condi ions a e sub-
Medi e anean and d ie han in Ticino, wi h mean annual empe a-
u es o 12.6 ◦C and annual p ecipi a ion o 620 mm. The wo oak
species s udied in hese mesic si es a e shade-in ole an , wi h Q. obu
dominan in alley bo oms wi h we soils, and Q. humilis ound in
moun ain si es wi h d y, basic soils ac oss cen al and sou he n Eu ope
(Damesin and Rambal, 1995).
In he Medi e anean Go goglione si e, loca ed in he moun ainous
Basilica a egion (sou he n I aly), Q. pubescens Willd. and Que cus ce is
L. mixed s ands was sampled. This is a high o es wi h a mean densi y o
600 s ems ha
−1
domina ed by Q. ce is (71 % o basal a ea) ollowed by
Q. pubescens (25 % o basal a ea) and o he b oadlea species.
Q. pubescens occupies he mos xe ic si es (e.g., low-ele a ion si es s eep
slopes wi h sou he n aspec and shallow soils), whils Q. ce is abounds
in mo e mesic loca ions. Soils a e a mix u e o sand, sil and clay ex u es
(Colangelo e al., 2018b). Clima e condi ions co espond o cool-we
win e s and d y-wa m summe s wi h mean annual empe a u e o
11.6 ◦C and annual p ecipi a ion o 722 mm. The wo oak species
showed decline symp oms since 2007 wi h he mos a ec ed a ea
sp eading ac oss ca. 450 ha and including many ees wi h de olia ion
>50 % and pa ches wi h 15 % o dead ees (Ripullone e al., 2020). Soils
a e sandy-loam ex u ed wi h a mean pH o 7.5 and shallow.
The second Medi e anean I alian si e is loca ed nea San Paolo
Albanese in he Basilica a egion, whe e we sampled a pu e high o es
o Que cus aine o Ten. This s and has a densi y o 348 s ems ha
−1
(Colangelo e al., 2017a). Q. aine o is a ligh -demanding oak species
sensi i e o d y condi ions in la e sp ing and summe , bu h i es in
mesic si es wi h deep soils (Gen ilesca e al., 2017). The clima e is
cha ac e ized by wa m, d y summe s, wi h a mean annual empe a u e
o 16.4 ◦C and annual p ecipi a ion o 742 mm. The a ea ea u es sandy
soils, and hese s ands we e adi ionally managed as coppices wi h
li es ock g azing. Dieback symp oms s a ed in he ea ly 2000s and
co esponded o abundan lea loss and wi he ing, g ow h decline and
high mo ali y (Colangelo e al., 2017a, 2017b, Colangelo e al., 2018a,
2018b). In a ec ed s ands, 10–15 % o ees ecen ly died.
The Gami si e is loca ed wi hin “Los Alco nocales” Na u al Pa k,
Andalusia (sou hwes e n Spain). He e, Que cus cana iensis Willd. ees
we e sampled in s ands wi h mean densi y o 54 s ems ha
−1
(S´
anchez-
Salgue o e al., 2020). This Medi e anean oak inhabi s si es wi h acid
soils in humid and wa m a eas o sou he n Spain and no hwes e n A -
ica, whe e i s g ow h is cons ained by d y and wa m sp ing-summe
condi ions (Gea-Izquie do e al., 2012). Clima e in he s udy a ea is
Medi e anean wi h we and mild win e s and d y and wa m summe s,
co esponding o a mean annual empe a u e o 16.4 ◦C and annual
p ecipi a ion o 560 mm. In ense dieback was obse ed he e ollowing
he se e e d ough o 2005. Addi ionally, he in asi e oomyce e Phy-
oph ho a cinnamomi Rands. is widesp ead in he a ea and has con ib-
u ed o oak decline, pa icula ly in poo ly d ained si es wi h high soil
mois u e e en ion (Se ano e al., 2022).
2.2. Field sampling
In s udy si es, a pai s o dominan and neighbo ing oak ees wi h
con as ing igo non-declining (ND) and declining (D) couples we e
selec ed o es ima e hei g ow h ends using dend och onology.
Wi hin each ND–D couple, ees we e 10–15 m apa a maximum. Fi s ,
i e o se en ci cula plo s ( adius o 15 m) we e andomly loca ed in
each si e o cha ac e ize he s and s uc u e (densi y, basal a ea). Sec-
ond, in each plo , he igo o each ma u e oak indi idual was cha ac-
e ized by a isual assessmen o c own de olia ion (Dobbe in, 2005).
Oaks we e classi ied in o wo igo ca ego ies based on c own ans-
pa ency: declining ees (D) had c own anspa ency exceeding 50 %,
while non-declining ees (ND) had anspa ency below 50 %. We
measu ed he diame e a b eas heigh (dbh) o sampled oaks a 1.3 m
using measu ing apes and measu ed ee heigh using a Nikon T uPulse
360 lase ange inde . Pai s o neighbo ing oak ees wi h con as ing
igo (ND–D couples) we e selec ed a each s udy si es o compa e hei
g ow h and wood ana omy se ies.
2.3. Clima e da a and d ough index
To cha ac e ize he clima ic condi ions o s udy si es, mon hly
maximum and minimum empe a u es and p ecipi a ion da a we e ob-
ained om he 0.25◦-g idded E-OBS da abase (h ps://climexp.knmi.
nl/s a .cgi) o he pe iod 1950–2018 (Co nes e al., 2018). We used
his da a se o a oid p oblems in homogenei y o gaps in local se ies,
which usually do no co e pe iods longe han 30 yea s. In addi ion, he
S anda dized P ecipi a ion–E apo anspi a ion Index (SPEI) was used o
assess he se e i y and du a ion o d ough s. The SPEI is a mul iscala
d ough index ha accoun s o bo h p ecipi a ion and e apo anspi a-
ion, calcula ed o e a ious ime scales (Vicen e-Se ano e al., 2010).
Posi i e and nega i e SPEI alues indica e we and d y condi ions,
espec i ely. Mon hly SPEI da a we e downloaded o ime scale o 12
mon hs and a a 0.5◦spa ial esolu ion using he SPEI Global D ough
Moni o webpage h ps://spei.csic.es/map/maps.h ml (h ps://spei.cs
Table 1
Cha ac e is ics o he s udy si es. The a io be ween annual p ecipi a ion and e apo anspi a ion (P / PET, las column) was ob ained om Zome e al. (2022).
Si e Species La i ude (N) Longi ude (
–
W, +E) Ele a ion (m a.s.l.) Slope (%) Aspec Yea o dieback onse P / PET
Ticino Q. obu 45.436 8.825 115 0 –2004 4.9
A amendia Q. humilis 42.716 −2.112 683 9 SE 2017 2.2
Go goglione Q. ce is 40.368 16.174 850 25 E 2007 1.8
Go goglione Q. pubescens 40.368 16.174 800 30 E 2007 1.8
San Paolo A. Q. aine o 40.020 16.341 970 25 SO 2003 1.4
Gami Q. cana iensis 36.570 −5.534 510 15 SE 2005 1.1
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
3
ic.es/map/maps.h ml).
2.4. T ee- ing wid h se ies
Fo each species, we selec ed 12–34 pai ed ees o di e en igo , i.
e., neighbo ing ND and D oaks loca ed 10–15 m apa a maximum
(Table 2). They we e sampled in la e summe o 2017 o quan i y hei
adial-g ow h ends using dend och onology (F i s, 1976). Two co es
we e sampled a 1.3 m om each ee a opposi e di ec ions and
pe pendicula o he maximum slope using a 5-mm P essle inc emen
bo e . The ans e sal su ace o co es was cu using a sledge mic o ome
o di e en ia e he annual ings (G¨
a ne and Nie e gel , 2010). Rings
we e isually c oss-da ed and measu ed wi h p ecision o 0.01 mm using
a binocula mic oscope coupled o a compu e wi h he LINTAB package
(Rinn ech, Heidelbe g, Ge many). The COFECHA p og am was used o
e alua e he s a is ical c oss-da ing o ee- ing wid h da a (Holmes,
1983). To quan i y g ow h ends and o emo e pa o age and size-
ela ed dec ease in ee- ing wid h, we ans o med hese da a in o
annual basal a ea inc emen (BAI) assuming a ci cula shape o s ems
and using he o mula:
BAI =
π
( 2− −12)(1)
whe e
and
-1
ep esen he wid h o he ings in he yea and -1,
espec i ely. Mean BAI se ies we e calcula ed by a e aging he BAI se-
ies o he indi iduals o each igo class sampled in each si e. T ee age
a 1.3 m was es ima ed om inc emen co es ha ei he included he
pi h o we e close o i , i.e. he inne mos ings we e clea ly cu ed.
2.5. Ea lywood ana omy
Fo each igo class (ND and D ees), we selec ed i e ees pe si e
and one adius pe ee o pe o m he analyses o ea lywood ana omy.
Selec ed ees we e hose showing he highes co ela ion be ween hei
BAI se ies and he mean si e BAI se ies. Wood ana omy was analysed o
he common pe iod om 1980 o 2017. Fi s , wood c oss-sec ions we e
ob ained using a sledge mic o ome (G¨
a ne and Nie e gel , 2010).
Co es om each ee we e subdi ided in o pieces app oxima ely 2–5 cm
long. Thin sec ions ( hickness o 10–20
μ
m) we e hen cu om each
piece wi h he mic o ome, s ained wi h sa anin (1 %) and As a blue (2
%), and ixed wi h Euki ®. Second, high- esolu ion images o he sec-
ions we e cap u ed using an Olympus BH2 mic oscope equipped wi h
an Olympus DP73 came a. The images we e s i ched wi h he PT-Gui
so wa e (New House In e ne Se ices BV, Ro e dam, NL) o c ea e
one composi e image o each sample. The images we e analysed using
he image analysis eewa e ImageJ e . 1.54i (Schneide e al., 2012).
This allowed measu ing he a ea and he adial diame e (d), assuming a
ci cula shape, and cen oid posi ion o each ea lywood essel wi hin
each ing. The lowe limi o essel iden i ica ion was se a a diame e
o 50
μ
m. The essels we e measu ed in a angen ial window o 2 mm.
Then, we calcula ed o each ing he ea lywood essel densi y
(Scholz e al., 2013). The ea lywood hyd aulic diame e (Dh) was
calcula ed as he a e age o ∑d
5
/∑d
4
(Spe y e al., 1994). The Dh was
calcula ed conside ing all measu ed ea lywood essels and used o
calcula e mean se ies o each igo class and si e. In addi ion, he po-
si ion o each essel wi hin he ing (x and y coo dina es) was eco ded.
Using he indi idual essel a ea da a, we es ima ed he a e age size
o essel in 10 equally spaced po ions o he ea lywood wi hin each
ing. We ha e di ided he coo dina es o he essels by 10 sec ions in
o de ha he cen oid o each essel could only be included in one
sec ion. Tha is, we di ided he ea lywood in 10 adial sec ions om he
ba k o he pi h and es ima ed he a e age size o essels and he numbe
o essels in each sec ion. Doing so, each ing was ep esen ed by 10
po ions i espec i ely o hei size. In addi ion, we ob ained he coe -
icien o a ia ion o essel a ea (he ea e CV) and a spa ial agg ega-
ion index (he ea e AI) (Cla k and E ans, 1954). The AI is a measu e o
clus e ing calcula ed as he a io be ween he obse ed dis ance o he
nea es neighbo and ha expec ed dis ance unde a Poisson p ocess
equi alen o comple e spa ial andomness. An AI alue >1 indica es
ha objec s a e egula ly dis ibu ed, while a alue <1 indica es clus-
e ing. We used he AI o weigh he CV by calcula ing hei p oduc .
Thus, high alues o CV ×AI show he p esence o egula ly spaced
ea lywood essels which a y in a ea, while lowe alues o CV ×AI
indica e a clus e ed dis ibu ion o essels wi h simila a ea.
2.6. S a is ical analyses
We conduc ed a se ies o s a is ical analyses o compa e ee cha -
ac e is ics and model ea lywood essel pa e ns. The non-pa am ic
Mann-Whi ney U es was used o compa e dbh, heigh , age, BAI, Dh,
essel a ea, and densi y be ween declining (D) and non-declining (ND)
ees wi hin each si e since some a iables (dbh, age) did no ollow
no mal dis ibu ion acco ding o Shapi o-Wilk es s.
Gene alized addi i e mixed e ec models (GAMMs; Wood, 2017)
we e employed o analyze how essel a ea and numbe a ied wi hin he
ea lywood ing and be ween igo classes. The ea lywood was di ided
in o 10 adial sec ions om he ba k o he pi h wi h dis ance class
(1−10), igo class, and hei in e ac ion as explana o y a iables. Each
ing was ep esen ed by 10 po ions i espec i ely o hei size. in o de
o obse e he a ia ion in he a ea and numbe o essels in ela ion o
he dis ance om he inne mos pa o he ea lywood (annula sec ion
1) o he ou e mos pa (annula sec ion 10). T ee iden i y and yea
we e included as andom ac o s o accoun o epea ed measu es.
Linea mixed-e ec s models (LMMs; Pinhei o and Ba es, 2000) we e
used o es i essel size and numbe p og ession di e ed be ween D and
ND ees in ela ion o he d ough index (SPEI). LMMs we e used o es
o di e ences in BAI and ea lywood ana omical cha ac e is ics ( essel
a ea and densi y, Dh, AI, CV, and CV ×AI) be ween igo classes om
Table 2
Cha ac e is ics o he sampled ND and D oak ees. Values a e p esen ed as means ±s anda d e o (SE). The basal a ea inc emen (BAI) da a co espond o he pe iod
1980–2017. Di e en le e s indica e signi ican di e ences (p <0.05; Mann-Whi ney U es s) be ween ee ypes wi hin each si e.
Si e Species Vigo class No. ees Dbh (cm) Heigh (m) Age a 1.3 m (y s.) BAI
(mm
2
y
−1
)
Dh (
μ
m) Vessel densi y (No. mm
−2
)
Ticino Q. obu ND 27 52.4 ±2.4b 25.1 ±0.9b 79 ±3 2199 ±77b 373 ±3.8b 58 ±2a
D 19 38.7 ±1.8a 15.3 ±0.6a 84 ±3 983 ±47a 332 ±3.5a 62 ±2a
A amendia Q. humilis ND 17 27.1 ±1.3a 11.4 ±0.2b 69 ±9 416 ±14b 297 ±3.3a 66 ±3a
D 15 21.7 ±1.1a 9.2 ±0.4a 74 ±10 183 ±8a 287 ±2.9a 62 ±2a
Go goglione Q. ce is ND 23 32.0 ±0.8b 15.0 ±0.4b 102 ±4 539 ±36b 317 ±3.7a 16 ±1a
D 22 28.0 ±0.7a 12.2 ±0.4a 107 ±2 343 ±18a 298 ±3.7a 14 ±1a
Q. pubescens ND 25 26.6 ±1.0b 11.5 ±0.3b 115 ±4 318 ±19b 289 ±2.0a 71 ±3a
D 19 22.2 ±0.7a 8.6 ±0.3a 110 ±3 195 ±8a 274 ±2.5a 69 ±4a
S. Paolo Q. aine o ND 34 33.3 ±0.7a 13.5 ±0.5b 140 ±7 396 ±16b 323 ±2.7a 59 ±2b
D 34 31.9 ±0.9a 10.2 ±0.4a 132 ±5 307 ±11a 316 ±3.5a 40 ±1a
Gami Q. cana iensis ND 12 51.3 ±2.1a 11.7 ±0.2a 107 ±4 1115 ±35b 281 ±3.5b 41 ±1a
D 12 52.3 ±4.3a 11.1 ±0.5a 112 ±5 936 ±38a 221 ±3.5a 44 ±2a
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
4
1980 o 2017. BAI was modeled agains igo class, calenda yea , and
hei in e ac ion, wi h ee iden i y as a andom a iable. To es o
gene ali y ac oss species, we included all species in he model wi h ee
ID nes ed wi hin species as a andom ac o .
GAMMs we e used o model he esponse o Dh and CV ×AI o
d ough se e i y (SPEI) and igo class. Indi idual log- ans o med BAI,
AI, and CV ×AI we e adjus ed as a unc ion o calenda yea using hin-
pla e eg ession splines (maximum 3 deg ees o eedom; Wood, 2003).
In e ac ions be ween calenda yea and igo class we e conside ed o
accoun o di e gences be ween ND and D ees. T ee iden i y nes ed
wi hin species was included as a andom e ec , and a i s -o de au o-
co ela ion s uc u e was applied o accoun o BAI dependency be-
ween consecu i e yea s.
All analyses we e pe o med using R (R Co e Team, 2024), wi h
LMMs and GAMMs i ed using he nlme (Pinhei o e al., 2023) and mgc
(Wood, 2011) packages, espec i ely.
3. Resul s
3.1. T ees ea u es and g ow h pa e ns
Declining (D) ees exhibi ed smalle diame e s compa ed o non-
declining (ND) ees, wi h he excep ion Q. ce is and Q. pubescens in
Go goglione, as well as Q. aine o in S. Paolo A. (Table 2). In mos si es,
ND ees we e alle han D ees, excep o Q. cana iensis in Gami .
Al hough no signi ican di e ences in age we e obse ed be ween ND
and D ees (Table S4), he mean basal a ea inc emen (BAI) alues we e
consis en ly highe o ND ees han o D ees, whe e he ND ees
g ew om 20 % o o e 50 % mo e han he D (Fig. 1, Fig. S3). Ea ly-
wood ana omical a iables showed con as ing esul s depending on he
Fig. 1. Va ia ion in basal a ea inc emen be ween non-declining ees ( ep esen ed by da k g een lines and squa es symbols) and declining ees ( ep esen ed by
b own lines and poin s symbols) o each species a he s udy si es. The poin s and lines indica e mean alues ac oss indi iduals, while he e ical segmen s ep esen
he s anda d e o s o he mean.
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
5
species wi h Dh being highe in ND han in D ees o Q. obu and
Q. cana iensis, and essel densi y being highe in ND Q. aine o
indi iduals.
We obse ed signi ican di e ences in BAI and i s end be ween ND
and D ees in each species (Table S1; Fig. 2). Howe e , he di e gence
in BAI be ween igo classes p io o dieback onse di e ed in du a ion
and magni ude among species and si es. Speci ically, he di e gence was
g ea e ( alues be ween 1249 and 3000 mm
2
) and pe sis ed longe in
Q. obu compa ed o he o he species (Fig. 1). The s onges BAI di -
e ence be ween ND and D ees obse ed du ing he las yea s co e-
sponded o Q. ce is and Q. aine o.
3.2. Va ia ions in ea lywood essel a ea and densi y
A a ia ion in he a ea and numbe o essels was obse ed in ela-
ion o he dis ance om he inne mos pa o he ea lywood (annula
sec ion 1) o he ou e mos pa (annula sec ion 10) (Fig. 2). Fo all
species, he essel a ea dec eased om he pi h o he ba k, as expec ed.
O e all, he essel a ea in he i s po ion o he ea lywood was la ge in
ND ees han in D ees. This was pa icula ly e iden in Q. pubescens
om Go goglione and less e iden in Q. aine o. The numbe o essels
ended o inc ease om he middle (sec ion 4) owa ds he end o he
ea lywood in mos cases, wi h he excep ion o Q. ce is in Go goglione.
ND ees showed mo e essels in he las o med hal o he ea lywood,
which we e o en o lowe a ea han in D ees (Fig. 2). O e all, he SPEI
exe ed li le signi ican in luences on he change in essel a ea along he
ea lywood and his e ec depended on he species (Table 3). A signi i-
can in e ac ion be ween ea lywood sec ion and SPEI was ound in
Q. humilis, Q. ce is and Q. pubescens bu only in declining ees (Table 3).
Fig. 2. The a ia ion in essel a ea (
μ
m
2
), on he le , and numbe , on he igh , as a unc ion o he dis ance om he inne mos pa o he ea lywood ( ing sec ion 1)
o he ou e mos pa ( ing sec ion 10). Non-declining a e ep esen ed by g een lines, while declining ees a e indica ed by ed lines.
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
6
Table 3
T ends in essel a ea assessed wi hin he ea lywood 10 sec ions (wi h dis ance class 1–10) compa ed be ween D and ND igo classes as a unc ion o d ough se e i y
(SPEI). The F s a is ics associa ed wi h he a iables a e shown. Signi ican alues (p. <0.05) a e indica ed wi h*.
Si e Species Declining ees Non-Declining ees
Sec ion SPEI Sec ion ×SPEI Sec ion SPEI Sec ion ×SPEI
Ticino Q. obu 4048.06* 2.60 4.36* 5429.53* 0.47 1.60
A amendia Q. humilis 1770.89* 8.81* 13.52* 3721.24* 6.17* 3.41
Go goglione Q. ce is 964.23* 5.65* 9.022* 2395.39* 1.67 0.67
Q. pubescens 1078.93* 2.79 9.38* 4040.87* 0.85 0.01
S. Paolo Q. aine o 1778.41* 0.33 0.06 2386.80* 1.74 1.40
Gami Q. cana iensis 2269.42* 1.46 0.40 4177.85* 0.24 0.63
Fig. 3. Va ia ion in hyd aulic diame e (Dh, in
μ
m), on he le , and he p oduc (CV ×AI) o i s coe icien o a ia ion (CV) and agg ega ion index (AI), on he igh ,
o non-declining ees ( ep esen ed by g een lines and squa es symbols) and declining ees ( ep esen ed by b own lines and poin s symbols). Solid lines indica e he
es ima ed ela ionships, while dashed lines ep esen he 95 % con idence in e als (CIs) o hese es ima es.
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
7
3.3. Changes in hyd aulic diame e and i s spa ial a iabili y
Di e ences in Dh h ough ime be ween ND and D ees we e
appa en in some species and si es, pa icula ly in he cases o Q. obu
and Q. cana iensis wi h wide essels in ND ees (Fig. 3 Table S2). The D
ees showed a long decline o CV se e al decades p io o dieback in all
si es, whe e he ND showed di e ences up o double he alues
compa ed o D, being pa icula ly s ong in Q. obu , Q. humilis,
Q. pubescens and Q. cana iensis, bu no in he case o Q. ce is (Fig. S2).
Con e sely, he AI showed di e gences be ween ND and D ees only
in Q. humilis and Q. pubescens wi h highe alues in ND ees. Finally, he
CV ×AI showed highe alues in ND ees o all si es, al hough hey we e
lowe in Q. ce is (Fig. 3 Table S3). In his case he di e gence be ween
ND and D ees was p onounced in Q. obu and Q. humilis, wi h di e -
ences in he alues be ween he wo ca ego ies anging om 0.5 o o 2.0
(Fig. 3).
3.4. Rela ionships be ween he d ough index and he essel hyd aulic
diame e
The ela ionship be ween he SPEI and he Dh and CV ×AI a ied
be ween si es and oak species (Table 4). In gene al, he Dh was less
ela ed wi h he SPEI han CV ×AI, pa icula ly in he case o
Q. cana iensis, and also Q. ce is and Q. pubescens om Go goglione. The
CV ×AI allowed sepa a ing ND and D ees in mos si es and species,
excep ing Q. ce is and Q. aine o. These esul s indica ed di e en e-
sponses o ea lywood ana omy o d ough in ND and D ees.
We obse ed dis inc esponses o hyd aulic diame e (Dh) and he
p oduc o coe icien o a ia ion and agg ega ion index (CV ×AI) o
d ough se e i y (SPEI) be ween ND and D ees (Table 4). The ela-
ionship be ween Dh and SPEI was no signi ican o bo h igo classes
(Table 4) as well as when all species we e conside ed in he same model
(Fig. 4). Howe e , he associa ion be ween SPEI and CV ×AI di e ed
be ween ND and D ees. While D ees exhibi ed a linea ela ionship,
ND ees showed a non-linea esponse wi h a sa u a ion e ec . This
sugges s ha ND ees can comple e ea lywood o ma ion in sp ing
unde mo e mesic condi ions (lowe SPEI alues) compa ed o D ees.
3.5. Rela ionships be ween g ow h (BAI), he essel hyd aulic diame e
and CV ×AI
O e all, we ound a posi i e ela ionship be ween BAI and Dh and
CV ×AI o (Fig. S4, Tables S5, S6), bu wi h di e ences in hei
magni ude be ween si es and igo classes. The g ea es di e ences
be ween ND and D ees we e obse ed in Q. humilis and Q. cana iensis
bu wi h opposi e pa e ns since ND ees we e mo e s ongly linked in
Q. humilis while in Q. cana iensis he ela ionship was s onge o D
ees. Obse ing he ela ionships be ween BAI and CV ×AI, di e ences
be ween igo classes eme ged in he we es (Q. obu ) and d ies
(Q. pubescens) si es. Las ly, ega ding he associa ions be ween Dh and
CV ×AI, we obse ed posi i e linea ela ionships be ween hese wo
a iables in he wo igo classes, wi h ew di e ences be ween ND and
D excluding he case o Q. obu in Ticino (Fig. S5, Tables S7).
4. Discussion
We show ha he spa ial con igu a ion o ea lywood essels allows
di e en ia ing be ween declining and non-declining ees in deciduous
ing-po ous oaks. No ably, measu es o dispe sion (CV) and spa ial ag-
g ega ion (AI) o ea lywood essel a ea e ec i ely sepa a ed indi iduals
acco ding o hei igo in mos s udied oak species. This inding con-
i ms ou hypo hesis ha wi hin- ing a ia ion in essel lumen a ea is a
c ucial ac o when compa ing he esponses o declining and non-
declining ees o d ough s ess and can se e as an ea ly-wa ning
Table 4
Rela ionship be ween he hyd aulic diame e (Dh) and he p oduc be ween he coe icien o a ia ion in essel size and he agg ega ion index (CV ×AI) wi h he 12-
mon h d ough index (SPEI 12) o declining (D) and non-declining (ND) ees. Fo each species in each si e, he F s a is ic associa ed wi h he igo class and he
deg ees o eedom associa ed wi h he spline (GAM) e ms a e shown. Signi ican alues a e indica ed wi h * (p <0.05) and ** (p <0.01).
Si e Species Va iable Vigo D ×SPEI 12 ND ×SPEI 12
Ticino Q. obu Dh 36.37 ±17.40* 1 0.57 1 0.77 0.14
CV ×AI 0.29 ±0.14* 1.76 1.77 1 0.98 0.04
A amendia Q. humilis Dh 10.82 ±19.68 1 3.02 1.61 3.46 0.01
CV ×AI 0.693 ±0.21** 1.70 1.93 1.00 0.46 0.22
Go goglione Q. ce is Dh 21.64 ±23.21 1.67 1.23 1.86 2.73 0.05
CV ×AI 0.14 ±0.07 1.00 6.76** 1.70 9.43** 0.09
Q. pubescens Dh 17.72 ±10.74 1 0.01 1.41 2.00 0.07
CV ×AI 0.64 ±0.18** 1 23.20** 1.90 22.15** 0.25
S. Paolo Q. aine o Dh −0.64 ±18.57 1 0.57 1 4.46* <0.01
CV ×AI 0.33 ±0.20 1 0.03 1 2.70 0.08
Gami Q. cana iensis Dh 60.22 ±20.90** 1 0.09 1 0.07 0.27
CV ×AI 0.70 ±0.21** 1.70 5.86* 1.80 20.1** 0.32
Fig. 4. Rela ionship be ween he d ough index (12-mon h SPEI) and ea ly-
wood hyd aulic diame e (Dh) (a); p oduc o essel size coe icien o a ia ion
(CV) and spa ial agg ega ion index (AI) (CV ×AI, b), in non-declining (g een
lines) and declining ( ed lines) ees ac oss species. Solid lines ep esen es i-
ma ed ela ionships, while dashed lines indica e 95 % con idence in e als.
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
8
signal o oak decline. Below we discuss how hese ana omical ea u es
can be iewed as esponses o d y condi ions imp o ing essel connec-
i i y and hyd aulic conduc i i y.
Deciduous, ing-po ous oaks adop an anisohyd ic s a egy
(Hochbe g e al., 2018). This s a egy depends on hei abili y o apidly
conduc wa e in he ea ly g owing season, o access deep wa e sou ces
and o s o e ca bohyd a es in hei xylem ays, con ibu ing o hei high
eco e y a e d ough (Ripullone e al., 2020; Zhang e al., 2024).
Consequen ly, oaks exhibi sho e and less p onounced pos -d ough
legacy e ec s compa ed o coexis ing ee species (Ande egg e al.,
2015; Gazol e al., 2020; Bose e al., 2024). This esilience makes o e-
cas ing oak decline and dieback using ee- ing se ies depending on
species-speci ic and con ingen on si e condi ions (Caille e e al., 2019).
Wood ana omy may o e a mo e obus ool o de ec ing ees expe-
iencing a p og essi e loss o igo induced by wa e sho age, i.e.,
declining ees p one o mo ali y.
The obse ed di e ences in ea lywood ana omy be ween declining
and non-declining ees can be a ibu ed o se e al unde lying mech-
anisms. Fo ins ance, ees subjec ed o ch onic wa e sho age may al e
hei xylem s uc u e o imp o e hyd aulic sa e y, a p ocess known as
hyd aulic adjus men (Fon i e al., 2010). This adjus men can lead o
he o ma ion o smalle essels o changes in essel dis ibu ion wi hin
he ea lywood. Addi ionally, d ough s ess can al e ca bon alloca ion
pa e ns, po en ially educing esou ces a ailable o ea lywood essel
o ma ion (McDowell e al., 2011). These shi s in ca bon alloca ion may
esul in ewe o smalle essels being o med du ing pe iods o wa e
s ess. Phenological changes induced by wa e s ess may also play a ole
in he obse ed ana omical di e ences. D ough can a ec he iming
and du a ion o ea lywood o ma ion, which in u n can esul in
changes o essel cha ac e is ics be o e o e all g ow h is impac ed (Sass-
Klaassen e al., 2011). This sugges s ha ea lywood ana omy may be a
mo e sensi i e indica o o d ough s ess han g ow h me ics.
Fu he mo e, legacy e ec s o d ough on cambial ac i i y and s o ed
ca bohyd a es may in luence ea lywood o ma ion in subsequen yea s,
esul ing in he accumula ion o ana omical changes o e ime
(Ande egg e al., 2015).
T ees o di e en igo o de olia ion classes exhibi ed con as ing
alues o CV ×AI, excep in Q. ce is and Q. aine o, which showed
ecen di e gences in his a iable be ween non-declining (ND) and
declining (D) ees om he 2000s onwa ds. In o he cases, ND and D
ees displayed di e en CV ×AI alues be ween 10 and 30 yea s be o e
he onse o decline/dieback. Gene ally, g ow h (BAI) di e gences be-
ween ND and D ees we e longe -las ing (a leas 30–40 yea s) in he
we es si es (Q. obu in Ticino, Q. humilis in A amendia). G ow h
di e gence be ween ND and D ees inc eases when g ow h begins o
decline in esponse o se e e d ough condi ions. Fu he mo e,
ega ding hese di e gences, in he we es si es he me ics CV ×AI
we e mo e p onounced. This di e en ial beha io sugges s ha CV ×AI
is a mo e e ec i e and unc ional measu e o he ecen loss o igo han
ee- ing wid h. In bo h cases, he esponses co esponded o wa m-d y
s ess condi ions linked o d ough s occu ing in he ea ly 21s cen u y
(Colangelo e al., 2017a, 2018a; Gen ilesca e al., 2017; Cama e o e al.,
2021).
The s onge di e en ia ion obse ed in mo e mesic si es (e.g.,
Q. obu in Ticino, Q. humilis in A amendia) compa ed o d ie si es
sugges s ha ees g owing unde ch onic wa e de ici may ha e less
capaci y o ana omical adjus men . This obse a ion aligns wi h he
concep ha ees in mo e a o able en i onmen s may ha e a g ea e
ange o ana omical esponses o d ough s ess, while hose in ch oni-
cally d y condi ions may al eady be ope a ing nea hei physiological
limi s (B ´
eda e al., 2006; Bose e al., 2024).
Ea lywood essel a ea e lec s changes in heo e ical hyd aulic
conduc i i y and ee igo , linked o dec eases in hyd aulic diame e
(Dh) and g ow h a es a e d ough (Co cue a e al., 2004). Gene ally,
declining ees exhibi lowe g ow h a es and smalle ea lywood essel
lumens han non-declining ees (Cama e o e al., 2015; Colangelo e al.,
2017a; Caille e e al., 2019). Howe e , his pa e n is no uni e sal; in
some ins ances, as -g owing, wa e -spending ees o ming wide essels
may be mo e ulne able o d ough due o being hyd aulically unde -
buil o ole a ing wa e sho age, as obse ed in Q. obu (Le aniˇ
c
e al., 2011) and discussed in Bo ghe i e al. (2020). This scena io does
no appea o apply o ou s udy si es, whe e igo ous ees ended o
o m wide essels.
Disc epancies be ween s udies may also a ise om changes in essel
diame e wi hin he ea lywood ela ed o cambial phenology, as illus-
a ed by Q. pubescens om Go goglione si e. He e, he ND ees o med
wide essels in he ea ly ea lywood, co esponding o he sp ing pe iod
when soil wa e con en , pho osyn he ic a es and g ow h a es ypically
peak in deciduous ees (Michelo e al., 2012). These ees hen p o-
duced smalle essels in he la e ea lywood, which a e less p one o
embolism, compa ed o declining ees.
We also acknowledge ha due o he p ac ical cons ain s o ou
sampling me hodology, i was no easible o selec declining (D) and
non-declining (ND) ees o iden ical size ac oss all s udy si es. Conse-
quen ly, we canno en i ely ule ou he po en ial in luence o on oge-
ne ic e ec s on essel size, as highligh ed by Olson e al. (2014) in hei
la ge-scale in es iga ion o angiospe m species.
The lack o di e en ia ion capaci y o CV ×AI in cases such as
Q. ce is could be explained by he subjec i e classi ica ion o ees in o
wo igo classes, which may change o e ime in si es subjec ed o
ch onic d ough s ess. The non-linea and linea esponses o CV ×AI o
he d ough index in non-declining and declining ees, espec i ely,
sugges di e en imings and a es o ea lywood o ma ion and ee- ing
g ow h be ween hese g oups.
A mo e comp ehensi e unde s anding o essel lumen a ia ions
wi hin g ow h ings can be achie ed by linking hese changes o luc-
ua ions in g ow h a e and wa e -use e iciency. A p omising app oach
could combine in a-annual g ow h moni o ing h ough xylogenesis o
dend ome e s wi h he use o iso opic p oxies (δ
13
C, δ
18
O) o ack wa e
and gas exchange. In eg a ing hese me hods can p o ide deepe in-
sigh s in o he physiological mechanisms unde lying he adjus men o
ascula s uc u e in esponse o en i onmen al changes. This app oach
may e eal c ucial in o ma ion abou he esilience o d ough o oak
species.
The measu emen o essel clus e ing is based on he no ion ha
once a xylem essel becomes embolized, essel connec ions can p o ide
pa hway con inui y (Ca lquis , 1977). Two- o h ee-dimension analyses
o xylem essel dis ibu ion ha e ecei ed a en ion in wood ana omy
s udies and measu es such as essel g ouping, agg ega ion, and con-
nec i i y ha e been ca ied ou (Scholz e al., 2013). In his sense, a
simple essel g ouping index was de eloped by Ca lquis (2001), co -
esponding o he o al numbe o essels di ided by he o al numbe o
essel g oupings. Howe e , ha index did no ake in o accoun he
diame e o essels. Al e na i e g ouping indices based on essel con-
nec i i y in ans e se sec ions and poin pa e n analyses we e de el-
oped by Mencuccini e al. (2010) and Ma ínez-Vilal a e al. (2012), who
ocused on condui densi y-packing ela ionships. He e we go a s ep
u he by in es iga ing how a ia ions in essel clus e ing a y be ween
yea s and conside ing ee i ali y. As a measu e o clus e ing, we op ed
o use he Cla k E ans index (Cla k and E ans, 1954) ha does no ac-
coun o he a ia ions in essel size. Howe e , we weigh ed he ag-
g ega ion index wi h he coe icien o a ia ion accoun ing o wi hin
ee- ing ea lywood essel size di e ences ha can a ise due o
clima e cons ain s (Ga cía-Gonz´
alez and Ecks ein, 2003; Van de We
e al., 2007; G iˇ
ca e al., 2013; Cas agne i e al., 2020; Guada e al.,
2021; Ma chand e al., 2021; Süßel and B üggemann, 2021). In line wi h
ou hypo hesis, a ia ions in he p oposed index sugges ha spa ial
clus e ing o ea lywood essels inc ease while a ia ion in lumen size
dec eases in oaks exhibi ing signs o d ough -induced dieback. This
syn he ic measu e o e s in o ma ion on yea - o-yea a ia ions in
ea lywood essel clus e ing and size asymme y, a me ic ha can
complemen exis ing measu es o ee ing wid h and hyd aulic
M. Colangelo e al.
Science o he To al En i onmen 980 (2025) 179565
9
