G aphical Abs ac ( o e iew)
This documen is he Accep ed Manusc ip e sion o a Published Wo k ha appea ed in
inal o m in:
Cu iel Yus e J., Flo es-Ren e ía D., Ga cía-Angulo D., He es A.-M., B aga C., Pe i an A.-M.,
Pe i an I.C. 2019. Cascading e ec s associa ed wi h clima e-change-induced coni e
mo ali y in moun ain empe a e o es s esul in ho -spo s o soil CO 2 emissions. SOIL
BIOLOGY & BIOCHEMISTRY. 133. 50-59. DOI (10.1016/j.soilbio.2019.02.017).
© 2019 Else ie L d
This manusc ip e sion is made a ailable unde he CC-BY-NC-ND 3.0 license h p://
c ea i ecommons.o g/licenses/by-nc-nd/3.0/
Episodes o d ough -induced coni e mo ali y a e becoming mo e equen in
Eu ope
Coni e mo ali y igge s seconda y succession o au och honous b oadlea species.
T ee mo ali y igge s complex cascading e ec s which ul ima ely a ec s Rs
Changes in ine oo speci ic oo leng h (SRL) co ela es nega i ely wi h RH.
*Highligh s ( o e iew)
Cascading e ec s associa ed wi h clima e-change-induced coni e mo ali y in
moun ain empe a e o es s esul in ho -spo s o soil CO2 emissions
Cu iel Yus e J. 1,2*, Flo es-Ren e ía D.3, Ga cía-Angulo D.4A.-M.1,5, B ag C.6,
Pe i an A.-M.6, Pe i an I.C.5
1 BC3 - Basque Cen e o Clima e Change, Scien i ic Campus o he Uni e si y o he
Basque Coun y, 48940 Leioa, Spain;
2 IKERBASQUE - Basque Founda ion o Science, Ma ia Diaz de Ha o 3, 6 solai ua,
48013 Bilbao, Bizkaia, Spain
3 CONACYT- CINVESTAV Unidad Sal illo, G oup o Sus ainabili y o Na u al Resou ces
and Ene gy. A . Indus ia Me alú gica 1062, Pa que Indus ial Ramos A izpe, Ramos
A izpe, 25900, Coahuila, Mexico.
4 Museo Nacional de Ciencias Na u ales (MNCN, CSIC) Se ano 115 dpdo; E-28006
Mad id, Spain.
5 -1,
6 Na ional Ins i u e o Resea ch-
E oilo , Volun a i, 077190, Romania
*Co esponding au ho : Tl : +34 944014690; e-mail= jo ge.cu iel@bc3 esea ch.o g
Keywo ds: Soil espi a ion, he e o ophic espi a ion, seconda y succession, ee
mo ali y, ine oo s unc ional ai s, cascading e ec s
*Manusc ip wi h con inous line numbe ing
Click he e o iew linked Re e ences
Summa y
As a widesp ead phenomenon a ec ing e es ial ecosys ems wo ldwide, he ex en and
spa io- empo al scales a which he inc easing numbe o epo ed e en s o clima e-
change-induced ee mo ali y could a ec he ecology and ca bon (C) sink capaci y o
e es ial soils, emains unknown. We he e s udy how egional-scale d ough -induced ee
mo ali y e en s egis e ed a e a e y d y 2012 yea in he Ca pa hians moun ain ange
(Romania), which a ec ed h ee o he mos widely dis ibu ed coni e species: Sco s pine,
Black pine, and Sil e i , esul ed in ho -spo s o biogenic soil CO2 emissions (soil
espi a ion; Rs). Fou o i e yea s a e he main mo ali y e en , Rs- ela ed soil CO2
emissions unde dead ees we e, on a e age, 21% highe han CO2 emissions unde li ing
ees ( anging om 18 o 35%). To al (Rs) and he e o ophic (RH)- ela ed soil CO2
emissions we e s ongly de e mined by he soil en i onmen al al e a ions ollowing ee
mo ali y (e.g. changes in quan i y and quali y o soil o ganic ma e , mic oclima e, pH o
ine oo demog aphy). Mo eo e , he massi e mo ali y e en o 2012 ul ima ely esul ed
in a s onge dominan ole o successional ege a ion (b oadlea seedlings, sh ubland and
g asses) in con olling hose en i onmen al ac o s ha ei he di ec ly o indi ec ly a ec ed
bio ic soil luxes (Rs and RH). We, he e o e, show ha apa om he well-known di ec
e ec s o clima e change o e soil CO2 emissions, cascading e ec s igge ed by clima e-
change-induced ee mo ali y could also exe a s ong indi ec impac o e soil CO2
emissions, al e ing he magni ude and he en i onmen al con ols o Rs and hence
de e mining ecosys em C budge and hei esponse o clima e.
1. In oduc ion
The numbe o episodes o o es de olia ion and mo ali y associa ed wi h clima e change
has inc eased subs an ially du ing he las decades (Allen e al., 2010; Ca nice e al., 2011),
and i is u he expec ed o inc ease e en mo e in u u e decades (IPCC 2014). In his
ega d, unde s anding physiological and ecological causes o ee mo ali y as well as
p edisposi ion o ees o die is nowadays a ho - opic ha has a ac ed many a en ion and
s udies (Allen e al., 2010, 2015; Ande egg e al., 2012; McDowell e al., 2015; Sangüesa-
Ba eda e al. 2015; Roge s B endan e al., 2016; Neumann e al., 2017; Llo e and
Ki zbe ge , 2018). The e is, howe e , a knowledge gap on how ecosys ems a e ac ually
esponding o such pe u ba ions, i.e. whe he and a which ex en ee mo ali y could
a ec ecosys em unc ioning (Ande egg, e al. 2013) and mo e pa icula ly how ee
mo ali y could a ec soil espi a ion (Rs), which ep esen s he o al biogenic CO2
p oduced and emi ed om soils (Va gas e al., 2010), and is he majo ou going lux o
CO2 om ecosys ems o he a mosphe e (Cu iel Yus e e al., 2005; Da idson e al., 2005;
Ba ba e al., 2018). T ees ha e he capaci y o modula e he belowg ound en i onmen
(Flo es-Ren e ía e al., 2015, 2016) igge ing cascading causal-e ec ela ions ha could
esul in subs an ial changes in he biological unc ioning o he soil sys em and in
undamen al al e a ions o he soil nu ien and soil CO2 emissions (Flo es-Ren e ía e al.,
2018). Howe e , da a and e idences on how hese al e a ions occu and a which ex en ee
mo ali y could a ec pa e ns and con ols o CO2 emissions om e es ial soils a e
sca ce.
Besides al e ing soil abio ic condi ions, ee mo ali y limi s he supply o subs a e in he
o m o ca bohyd a es (e.g. exuda es) o nu ien s (li e ) demanded by belowg ound o gans
( ine oo s), symbion s (e.g. myco hiza), and soil biological communi ies om he
hizosphe e and he soil (e.g. Högbe g e al., 2001; Binkley e al., 2006; Ba ba e al., 2016).
The dis up ion o he C low o belowg ound has been di ec ly ela ed o almos immedia e
dec eases in Rs (Högbe g e al., 2001; Binkley e al., 2006; Na e e al., 2011; Le y-Va on
e al., 2014), due o a pa allel dec ease in au o ophic ( ine oo and myco hiza) and
he e o ophic ( espi a ions om mic obes and soil auna) espi a ion. On medium-long
e m, he dea h o ees may ha e c i ical e ec s o e key soil biogeochemical cycling
(Rod iguez e al. 2016), esul ing in ch onical losses o key nu ien s such as ni ogen (N),
wi h unknown consequences o he capaci y o he sys ems o eco e p e-pe u ba ions
pools and unc ions (Ga cía-Angulo e al. in p ep). This happens because he dis up ion o
he low o C om plan s o soils and he changes in he mic oclima ic condi ions
associa ed wi h ee mo ali y may p ominen ly al e he composi ion, s uc u e and
unc ionali y o soil biological communi ies (Cu iel Yus e e al. 2012, A ila e al. 2016)
esul ing in i e e sible losses o key unc ional g oups ha sus ain impo an soil unc ions
such as N ixa ion o mine aliza ion o essen ial nu ien s (Gómez Apa icio e al. 2017).
Few s udies ha e been designed, howe e , o in es iga e in dep h how p ocesses igge ed
by ee mo ali y could a ec biogenic soil CO2 emissions and a which ex en . In his
ega d, ee mo ali y has been associa ed wi h ecosys ems eaching new equilib iums,
esul ing in impo an changes in he di e si y o soil bio a and soil unc ions (Cu iel Yus e
e al., 2012; Llo e e al., 2015, A ila e al., 2016), as well as in he o e all biogenic
emissions o CO2 om soils (Moo e e al. 2013; A ila 2018). Depending on he magni ude
o he ee mo ali y e en and/o legacies om his o ical managemen , o es ecosys ems
a e able o coun e ac po en ial nega i e e ec s o ee mo ali y and eco e p e-
pe u ba ions unc ioning a es ela i ely as (Na e e al., 2011; Gough e al., 2013; Le y-
Va on e al., 2014; Ba ba e al., 2016). This is because ee mo ali y, depending on he
ecosys em's cha ac e is ics and i s ini ial condi ions, igge s a p ocess o ecoloniza ion by
seedlings o he same species ( egene a ion) o o o he species be e i ed o p esen
condi ions (seconda y succession), which slowly eplace he niche le by he dea h o he
ees (e.g. Vay eda e al., 2016; Ruiz Beni o e al., 2017). How hese complex abo eg ound
ecological p ocesses could ac ually impac belowg ound unc ioning, subsequen ly
a ec ing magni ude and con ols o bio ic CO2 emissions emains unknown.
The objec i e o his s udy was o deepen ou mechanis ic unde s anding o he e ec s o
la ge-scale ee die-o e en s on o al soil espi a ion (Rs). Fo ha , we he e show a
egional scale s udy, spanning o wo consecu i e yea s (2016, ; and 2017,
2 ), on a es o Rs in s ands loca ed in
Romania), whe e ecen ex emely d y yea s, such as 2012, ha e esul ed in ex ended
mo ali y a es. These mo ali y e en s ha e mainly a ec ed coni e ee species, especially
Sco s pine (Pinus syl es is L.), Sil e i (Abies alba Mill.), and Black pine (Pinus nig a
A nold)mo ali y obse ed in 2012 and es ima ed o ex end o e la ge a eas
K ons ad ), ollowed a sequence o se e al ex ao dina y d y and ho yea s
egis e ed du ing he i s decade o he cen u y
(h p://www.me eo omania. o/anm/?lang= o_ o). These a ec ed coni e s a e slowly
eplaced by au och honous b oadlea species, especially Que cus obu , Fagus syl a ica,
F axinus o nus, F axinus excelsio , Ca pinus be ulus, Que cus pe aea o Ace campes e
(Table 1). Addi ionally, he s udy collec ed de ailed in o ma ion on a iables po en ially
sensi i e o ee die-o and di ec ly/indi ec ly associa ed wi h Rs, e,g, soil wa e con en
(SWC) and empe a u e (Tsoil), soil C and nu ien s pools, soil he e o ophic espi a ion
(RH), and ine oo biomass and unc ional ai s (e.g. speci ic oo a ea, SRA and oo
leng h, SRL).
We, he e o e, hypo hesized ha abio ic/bio ic changes p omo ed by ee mo ali y will
igge a cascade o causal-e ec ela ions ha could ha e esul ed in subs an ial changes in
he biological unc ioning o he soil sys em, subsequen ly al e ing he soil CO2 emissions.
In pa icula , we hypo hesized ha in hese coni e ous o es s, whe e ee mo ali y is
gi ing way o o es s domina ed by na i e ha dwood species, he p ocess o seconda y
succession associa ed wi h mo ali y o coni e s would be associa ed wi h p o ound
ans o ma ions o he mic oclima e, biology and chemis y o he soil, which ul ima ely
will a ec he magni ude and con ols o soil CO2 emissions.
2. Ma e ials and Me hods
2.1. S udy si es
Fo his s udy, we selec ed a o al numbe o 9 coni e s ands (3 o Sil e i , 3 o Sco s
pine, and 3 o Black pine), all o hem a ec ed by ecen mo ali y e en s (i.e., ollowing
he 2012 d ough ) (Table 1). These o es s we e loca ed in he T ansyl anian side o he
Eas e n Romanian Ca pa hians Moun ain ange ( Coun y). As we wan ed o
a oid/limi as much as possible o he dis u bance ac o s (e.g. managemen ), o es s we e
selec ed ei he in p o ec ed a eas o in a eas whe e managemen in ensi y has been minimal
o he las decades. All s ands we e loca ed on sloppy e ains (slopes anging om 17 a o
37 º). Bo h pine species s ands we e loca ed be ween 450 and 700 m a.s.l., while he
Sil e
almos pu e s ands, ha we e a i icially egene a ed 100 yea s ago, whe eas he Sil e i
s ands a e une en, na u ally egene a ed, mixed s ands wi h Fagus syl a ica (up o 35% in
he o es composi ion) o mo e han 150 yea s old . Soil ype in he
Sil e i s ands is mainly Eu icambisols, while Rendzina is he main soil ype o he pine
s ands (Table 1). Bo h mean annual p ecipi a ions (MAP) and mean annual empe a u es
(MAT) (Clima e Resea ch Uni Time Se ies, CRU TS3.10; ia h p://climexp.knmi.nl)
we e ela i ely low and no e y a iable among loca ions, anging, espec i ely om 593
o 693 mm and om 3.7 o 6.6 ºC (Table 1). Ou s udy also shows a na u al unde s o y
g adien ( ee saplings and seedlings) and g ass co e , om si es wi h e y sca ce
unde s o y/g ass co e (7%) o si es wi h unde s o y/g ass co e a e aging up o 70%
(Table 1). The d ough -induced mo ali y a e was es ima ed o ound 19-23% o Sil e
i , 16-27% o Black pine, and 17-22% o Sco s pine (Table 1).
2.2.Field measu emen s
2.2.1. Expe imen al design and ee age es ima ion
A each o he 9 coni e s ands a ec ed by mo ali y, 5 pai s o s anding adul dead and
li ing ees we e sampled (see below) along a ansec pe pendicula o he slope. We used
a pai ed sampling design (Bigle and Bugmann 2004), in which he selec ed li ing ees had
simila size (diame e a b eas heigh , DBH), compe i ion le el, and mic osi e condi ions
wi h he dead ones. T ees no iceably a ec ed by biological agen s (e.g. pa hogens, ungus),
wind, o human in luences we e a oided du ing he sampling. The sampling o he 5 pai s
was ca ied ou in ansec s s a ing a a andom poin wi hin each s and and main aining a
cons an al i ude, and hus simila humidi y condi ions, un il he equi ed numbe o ees
S uc u al equa ion models (SEMs) we e inally used o es he di ec and indi ec
in luence o he di e en bio ic and abio ic ac o s on Rs. Since he sample size was
ela i ely small (n = 88), he numbe o p edic o s included in he model was ho oughly
limi ed, as ecommended by Shipley (2002). Ou models conside ed a comple e se o
hypo hesis based on li e a u e, p e ious explo a o y analyses (K uskal-Wallis, co ela ions,
e c.), and ou own p e ious expe ience (Flo es-Ren e ía e al., 2016, 2018; Pé ez-Izquie do
e al., 2017, see Fig. S2). The model assumed ha abo eg ound ege a ion s uc u e (size o
he sampled li ing and dead ees, ee compe i ion index, and he % o unde s o y and
g ass co e ) would a ec he abio ic (mic oclima e, pH) and he bio ic ( oo demog aphy)
soil en i onmen , as well as he nu i ional s a us (PC1) and he o ganic ma e con en
(PC2) in soils unde ees. O e all, bo h RH and Rs, would be s ongly con olled by
changes in all hese en i onmen al ac o s which a e ul ima ely con olled by he
abo eg ound ege a ion s uc u e.
Se e al SEMs we e un and he bes - i ed ones we e inally selec ed acco ding o he
co a iance p oximi y be ween obse ed and expec ed da a (goodness-o -
gene al model we used mul ig oup SEM o es whe he he s udied ac o s we e linked by
he same causal s uc u e in each ee s a us (dead o li ing) and o iden i y he pa hs ha
did no beha e simila ly in he wo condi ions (Shipley, 2002; Ga cía-Camacho e al.,
2010). Fo his analysis, we used he same hypo he ical model used o each s a us g oup
sepa a ely. A cons ained model in which all ee pa ame e s we e o ced o be equal ac oss
he wo condi ions was buil and con as ed wi h ield da a. Since a lack o i was de ec ed
in he ully cons ained mul ig oup model, a se ies o nes ed models, whe e equali y
cons ain s we e emo ed one a a ime, we e de eloped o de ec which one would
signi ican ly imp o e he model (Shipley, 2002; Ga cía-Camacho e al., 2010). Di e ences
in X2 and AIC s a is ics be ween he ully cons ained model and he models eed om a
cons ain we e used o es o di e ences in pa ame e alues be ween he wo condi ions.
S anda dized pa h coe icien s we e es ima ed by using he maximum likelihood algo i hm
(Shipley, 2002). SEM analyses we e pe o med using SPSS® and SPSS® AMOS 20.0
3. Resul s
Bo h SWC and Tsoil a ied signi ican ly be ween yea s 1 6) wa me and
we e on a e age han 2 7), al hough no such signi ican end was ound o Rs
(Figs. 1 and S3). No di e ences in Rs we e ound be ween he h ee coni e species,
al hough he soils om Sil e i s ands we e signi ican ly colde (lowe a e age Tsoil) and
we e (highe a e age SWC) han he ones om he Black pine and Sco s pine s ands (Fig.
S3). When looking o di e ences in SWC, Tsoil and Rs conside ing ee s a us (li ing o
dead), li ing ees di e ed signi ican ly om he dead ones only in SWC and Rs.
Speci ically, dead ees showed highe SWC and Rs alues han he li ing ones (Figs. 1 and
S3).
O e all, Rs was consis en ly highe unde dead ees wi h espec o li ing ones (Fig. 2,
Table S1). Acco ding o ou esul s, CO2 emissions unde dead ees we e on a e age 21%
highe han unde li ing ones. Fu he mo e, RH, speci ic oo leng h/a ea (SRL/SRA), TOC,
TON, and SWC we e among he en i onmen al a iables ha we e consis en ly highe ,
al hough no always signi ican ly, unde dead ees compa ing wi h he li ing ones (Fig. 2,
Table S1). O he a iables such as pH, C:N a io, o o al ine oo biomass/ olume (FRB
and FRV, espec i ely) showed less sensi i i y o ee mo ali y and/o less consis en
ends ac oss ee species, al hough FRB ended o be lowe unde dead ees (Fig. 2, Table
S1). No clea di e ences we e ound be ween li ing and dead ees when he g ass and
unde s o y co e , o he ee compe i ion index we e conside ed (Table S1). This happened
because co e o unde s o y o b oadlea seedlings and g asses was independen o he
heal h s a us o he sampled ees and e y dependen on si e-speci ics condi ions.
Bo h mean soil empe a u e and quan i y o SOM (PC2) we e, oge he wi h ee s a us
(li ing and dea h) he a iables ha be e explained Rs a iabili y ac oss si es (Fig. 3, Table
S1). As expec ed o ecosys ems gene ally limi ed by empe a u e, Tsoil explained a la ge
po ion o ac oss-si e a iabili y in Rs (Fig. 3, Table S2), whe eas Rs was also s ongly
d i en by SOM quan i y (PC2, Fig. 3, Table S2). No di e ences in RH be ween soil
collec ed unde dead and li ing ees we e ound (Figu e 4). On he o he hand, we ound a
s ong e ec o he nu i ional s a us (PC1) and he quan i y o SOM (PC2) on RH (Fig. 4,
Table S3). Addi ionally, he bes ob ained model also included a signi ican nega i e e ec
o SRL o e RH (Fig. 4, Table S3).
SEMs showed he complex causal-e ec cascade o p ocesses con olling Rs and RH (Fig.
5). Speci ically, his analyses highligh ed how s ongly o es s uc u e ( ee DBH, ee
compe i ion, unde s o y and g ass co e ) in luenced, di ec ly o indi ec ly, he obse ed
a iabili y o soil abio ic (mic oclima e, pH, nu ien con en ) and bio ic (SRL, FRV, RH)
a iables, esul ing in he obse ed a iabili y in Rs ac oss si es. Bo h, ees (i.e. size and
ee compe i ion) and unde s o y co e exe ed a s ong e ec o e soil mic oclima e
(Tsoil), soil pH, and nu ien s (PC1). While coni e s end o acidi y soils, we he e obse e
how an inc easing co e o b oadlea unde s o y was associa ed wi h inc eases in pH,
which, on he o he hand, was also behind he obse ed imp o emen o soil nu i ional
s a us (PC1) and SOM seques a ion (PC2). The p esence o g asses was also di ec ly
associa ed wi h an inc ease in SOM (PC2) and SRA, bu a dec ease in FRV. Mul ig oup
SEM u he showed a igh e con ol o Tsoil o e Rs unde dead han unde li ing ees, as
illus a ed by he signi ican ly highe ML coe icien ob ained. Mo eo e , mul ig oup SEM
also showed how coni e s and successional ege a ion exe ed an opposi e e ec o e he
demog aphy o ine oo s (SRL and FRV). Speci ically, li ing ees exe ed an o e all
nega i e e ec o e SRL, whe eas he p esence o g asses was posi i ely associa ed wi h
SRL. On he o he hand, he FRV was s imula ed in poo soils (high PC1) and unde high
SOM con en s (low PC2), bu was also nega i ely co ela ed wi h he p esence o g asses
and he inc ease in b oadlea unde s o y co e . Fu he mo e, we he e show how,
independen ly o he heal h s a us o he ees, inc ease in he SRA nega i ely a ec ed RH.
Also, con ols o RH di e ed depending on he coni e heal h s a us (li ing o dead): li ing
ees exe ed a posi i e con ol o e RH, bu when ee dies RH a iabili y was mainly
con olled by nu ien s a us (PC1) and SOM quan i y (PC2) (besides SRL). Finally,
obse ed a iabili y o Rs seemed o be pa ially explained by a iabili y o RH unde dead
ees while unde li ing coni e s no ela ion was ound be ween bo h luxes.
4. Discussion
We he e epo ed la ge e en s o d ough -induced ee mo ali y on s ands domina ed by
h ee o he coni e ee species mos widely dis ibu ed in he Ca pa hian moun ain ange
(Table 1). This d ough -induced mo ali y coincides wi h globally inc easing epo ed
e en s o coni e decline, which a e being gene ally a ibu ed o his o ical managemen
p ac ices (e.g. a o ing/plan ing coni e species ou side hei clima ic and s uc u al
op imum; U bie a e al., 2008; Ruiz Beni o e al., 2012), o o clima e-change induced
inc eases in empe a u e and d ough s in ensi y and equency ha seem o a ec mo e he
coni e -like han he angiospe ms-like ela ed unc ional ai s (Henne Paul e al., 2015;
McIn y e e al., 2015; Ruiz Beni o e al., 2017). Indeed, he obse ed inc easing p esence o
an unde s o y ege a ion o seedlings mainly composed by na i e b oadlea species (e.g.
Fagus syl a ica, F axinus excelsio , F axinus o nus, Ace campes e, Que cus pe aea,
Ulmus glab a) (Table 1) coincides wi h hese men ioned abo e obse a ions and a e,
he e o e, in acco dance wi h his gene al decline in coni e dominance, especially
ubiqui ously obse ed in Eu opean o es s.
D ough induced coni e mo ali y esul ed in la ge inc eases in biogenic soil CO2
emissions (Rs), a e aging a 21% inc ease unde he dead ees compa ing wi h he li ing
ones, and pe sis ing du ing wo consecu i e yea s (2016-2017), ou o i e yea s a e he
mo ali y e en occu ed in 2012. Gi en he obse ed la ge p opo ion o ee mo ali y
obse ed a e he 2012 d ough (Table 1), his ee mo ali y-induced ho -spo o CO2
emissions migh be esponsible o decele a ing he capaci y o hese ecosys ems o eco e
p e-mo ali y le els o C seques a ion (e.g. Moo e e al, 2013). Ou esul s a e no in
acco dance wi h dec eases in Rs obse ed unde expe imen al ee gi dling manipula ions
(e.g. Högbe g e al., 2001; Binkley e al., 2006; Na e e al., 2011; Le y-Va on e al., 2014),
o unde na u al condi ions, when ee mo ali y e en s we e massi ely caused by ba k
bee le a acks (e.g. Moo e e al. 2013) o by in ec ions wi h pa hogens, e.g. Phy oph o a
cinammoni (A ila e al., 2016). Acco dingly, li e a u e gene ally shows how ee dea h
esul s in an almos immedia e and d ama ic dec ease in Rs a es associa ed wi h he
dec ease in he supply o newly plan s- ixed ca bohyd a e o belowg ound me abolic
ac i i y, e.g. au o ophic espi a ion, myco hiza ac i i y and hizosphe e he e o ophic
espi a ion (Subke e al., 2004; Högbe g e al., 2007). This Rs d op associa ed wi h ee
mo ali y may las o decades in monospeci ic o es (e.g. Moo e e al. 2013, A ila e al.,
2016), while o he s udies ha e shown how depending on he le el o he pe u ba ion and
he seconda y successional p ocesses, Rs may eco e p e-pe u ba ion alues a e se e al
yea s (e.g. Le y-Va on e al., 2014; Ba ba e al., 2016), o e en inc ease du ing a o able
seasons in mixed o es (Ba ba e al., 2013). Hence, due o he ini ial physiological collapse
ha ee mo ali y p oduces in a sys em, he unc ional eco e y o his sys em in gene al
and o he Rs in pa icula depends on he deg ee o pe u ba ion bu also on seconda y
successional p ocess igge ed by ee mo ali y (Le y-Va on e al., 2014; Llo e e al.,
2015; Ba ba e al., 2018).
In hese coni e o es s so ep esen a i e o he Ca pa hians' landscape, he obse ed ho -
spo s o CO2 unde dead ees we e s ongly domina ed by RH (Fig. 2 and 5). We he e
pos ula e ha hese mo ali y- igge ed ho -spo s o RH and Rs we e mos ly explained by an
inc ease in he quali y and quan i y o SOM which esul s om bo h he inc ease in
senescen ma e ial and om he successional p ocesses ollowing ee dea h (Fig. 2, 4 and
5). The obse ed inc ease in opsoil SOM unde dead ees (inc ease in TOC, Figu e 2)
could be a ibu ed, a leas pa ially, o he accumula ion o senescen plan ma e ial
(lea es, oo s and b anches) which gene ally accumula e unde dead ees (Moo e e al.
2013). Howe e , SOM accumula ion unde dead coni e s alone canno explain he obse ed
inc ease in RH because, as obse ed, RH was also e y sensi i e o he inc ease in soil
nu ien a ailabili y (PC1) (Figu e 4, Table S3). Besides, we he e obse ed how shi s in
he con ols o Rs a e he massi e mo ali y e en o 2012 ul ima ely esul ed in a s onge
dominan ole o he successional ege a ion (b oadlea seedlings, sh ubland and g asses)
o e he belowg ound en i onmen al ac o s, di ec ly o indi ec ly a ec ing Rs and RH
luxes (Figu e 5). This shi owa ds g ea e unde s o y con ol o e soil unc ions was in
de imen o he o me con ol exe ed by he coni e s which in luenced he mic oclima e
(SWC and Tsoil), he abio ic soil en i onmen (pH), he nu ien quali y (PC1), SOM
(PC2), and he ine oo demog aphy (speci ically SRA/SRL, FRV) (Figs. 2, 3, 4 and 5).
These changes we e u he e lec ed in changes in he magni ude and he con ols o bio ic
soil luxes (Rs and RH) (Figu e 2 and 5).
Fo ins ance, he inc ease in pH and g assland co e esul ing om he shi in he
abo eg ound ege a ion dominance also played a c i ical ole in inc easing SOM (PC2; Fig.
5). This is because, unde coni e in luence, he gene ally low soil pH and he low quali y
o he esidues due o he high p opo ion o ecalci an compounds, e.g. lignin and/o
allelopa hic molecules (e.g. Cu iel Yus e e al. 2005; Fe nández-Alonso e al., 2018) slow
down he b eakdown o he li e and i s inco po a ion o SOM in he mine al soil. On he
o he hand, inco po a ion o li e in SOM occu s gene ally as e in ecosys ems domina ed
by b oadlea species because he gene ally highe pH and highe quali y o he p oduced
esidues s imula es bio u ba ion (F ouz e al., 2009). Indeed, he mul ig oup SEM u he
showed how he inc ease in pH associa ed wi h he inc easing p esence o he unde s o y
ege a ion had a di ec and s ong posi i e e ec o e SOM quali y (PC1), sugges ing ha
on op o he inc ease in SOM unde dead ees, he seconda y successional p ocesses
igge ed by coni e mo ali y was posi i ely a ec ing he quali y o he subs a e. Ou
esul s, he e o e, clea ly indica e how he shi in ege a ion dominance associa ed wi h
coni e mo ali y had a s ong impac o e he quan i y and quali y o SOM, esul ing in
inc eased RH, which subsequen ly a ec ed Rs.
This dominan ole o successional ege a ion a e ee dea h was also e lec ed in a
subs an ial inc ease in he su ace o abso p ion o he adical sys em (inc ease in SRL and
SRA; Fig. 2) which co esponds o a shi owa ds a ine- oo demog aphy op imized o
maximize nu ien s acquisi ion (Roume e al., 2016). This shi , associa ed wi h an inc ease
in he p esence o g asses (Fig. 5), sugges ha he belowg ound niche le by he dea h o
he coni e s c ea es an oppo uni y o he su ounding ea ly successional ege a ion o
ob ain esou ces (nu ien s and mois u e) (Cu iel Yus e e al., 2012; Ba ba e al., 2013)
whose acquisi ion is, o he wise, subjec ed o s ong compe i ion, especially in nu ien -
poo , low pH coni e si es as hose conside ed in his s udy. Indeed, he mul ig oup SEM
showed how he poo nu ien condi ions unde coni e s (PC1), while ali e, p omo ed a
bigge adical sys em (highe FRV), bu wi h ela i ely less e y ine oo s (supp essing
SRL).
The consis en inc ease in he speci ic leng h and su ace (SRL/SRA) o ine oo s unde
dead ees was pa alleled by he obse ed inc ease in Rs (Figu e 2). This was expec ed,
gi en he gene al obse ed linea ela ion be ween, on one hand, SRL and he ine oo s
u no e a es (Sil e and Miya, 2001; Hobbie e al., 2010; Roume e al., 2016), and, on
he o he hand, SRL and a es o oo espi a ion (RA) (Reich e al., 2008; Maki a e al.,
2012; Picon-Cocha d e al., 2012). Al hough RA we e no measu ed in his s udy, we did
no obse e a signi ican inc ease in SOM u no e ( a es o RH pe uni o soil C; da a no
shown) unde dead ees, sugges ing ha i is mos plausible ha he inc ease o SRL was
associa ed wi h a pa allel inc ease in RA. Ra he han s imula ing RH, ou models also
showed a e y consis en nega i e ela ion be ween SRL and RH (Figu e 4 and 5),
sugges ing ha besides his expec ed posi i e e ec o SRL o e au o ophic ac i i y, he
ne e ec o e RH was nega i e. I , he e o e, could be ha by inc easing hei capaci y o
abso b nu ien s (inc ease in SRL), successional ege a ion compe es mo e e icien ly o
he same esou ces wi h he soil he e o ophic communi y (nega i e p iming, Kuzyako
2002), esul ing in he obse ed supp ession o RH. Indeed, an inc ease in compe i ion o
key nu ien s (e.g. N, P, K) be ween oo s and he e o ophs could be maximal in soils when
nu ien s a e gene ally limi ing, he eby esul ing in he supp ession o RH (e.g. Schimel e
al., 1989; Wang and Bakken 1997; Kuzyako 2002).
5. Conclusions
We he e collec ed compelling e idences o suppo ou ini ial hypo heses: cascading
mechanisms igge ed by selec i e ee mo ali y and a subsequen seconda y successional
p ocess played a c i ical ole in egula ing soil unc ioning and soil CO2 emissions du ing
ansi ional s a es. Speci ically, we he e show how coni e mo ali y esul ed in an a e age
inc emen o biogenic emissions o 21%, 4-5 yea s a e he la ge mo ali y e en o 2012,
which migh be u he esponsible o decele a ing he capaci y o hese ecosys ems o
eco e p e-mo ali y le els o C seques a ion. These ansi ional s a es a e ee dea h
esul ed in a s imula ion o he he e o ophic ac i i y (RH), a o ed by he inc ease in
senescen ma e ial bu also by changes in he soil mic oen i onmen (e.g. clima e, pH and
SOM) pa ially con olled by successional ege a ion. A shi owa ds a mo e e icien
esou ce-acquisi i e s a egy o ine oo s (inc ease in SRL), igge ed by ee mo ali y and
also associa ed wi h he inc easing dominance o he successional ege a ion, was also
behind he obse ed changes in he magni ude and con ols o RH and Rs. Ou esul s,
hence, call he a en ion on how abo e-belowg ound ecological p ocesses igge ed by ee
mo ali y may subs an ially de e mine dynamics o key biogeochemical cycles (e.g. C and
N) a local and egional scales. One o he d awbacks o his s udy migh be he ac ha he
e ec s o ee-mo ali y we e only e alua ed du ing a ela i ely sho - e m (2 yea s), a sub-
decadal ime scale (4-5 yea s a e he main mo ali y e en ), and in a limi ed numbe o
si es (9). Despi e i s limi a ions, his is one o he i s s udies e idencing he complexi y o
he con ols o e Rs in clima e-change-induced ee mo ali y scena ios, and as such, i
migh se e as a base o de elop u he , mo e ex ended s udies on his opic. In a changing
wo ld whe e episodes o ee mo ali y associa ed wi h clima e change a e subs an ially
inc emen ing, mo e s udies should, he e o e, be designed o deepen he obse ed po en ial
impac s o ee mo ali y and subsequen successional p ocesses a la ge empo al and
spa ial scales
Acknowledgemen s
The au ho s hanks o he s a o Fo es Dis ic s: K ons ad , Rasno , Codlea, Teliu, and
Sacele who adminis a e he 9 o es s conside ed o his s udy. This wo k was suppo ed
by he Spanish Minis y o Economy and Compe i i eness (MINECO) wi h he p ojec s
VERONICA (CGL2013-42271-P) and he p ojec IBERYCA (CGL2017-84723-P), and by
he Romanian Minis y o Educa ion and Scien i ic Resea ch h ough UEFISCDI wi h he
p ojec s TREEMORIS (PN-II-RU-TE-2014-4-0791), NATI E (PN-III-P1-1.1-PD-2016-
0583), and BIOCARB (PN-III-P1-1.1-TE-2016-1508). This esea ch was also suppo ed by
he Basque Go e nmen h ough he BERC 2018-2021 p og am, and by he Spanish
Neumann, M., Mues, V., Mo eno, A., Hasenaue , H., Seidl, R., 2017. Clima e a iabili y
d i es ecen ee mo ali y in Eu ope. Global Change Biology 23, 4788-4797.
Pé ez-Izquie do, L., Zabal-Agui e, M., Flo es-Ren e ía, D., González-Ma ínez, S.C.,
Buée, M., Rincón, A., 2017. Func ional ou comes o ungal communi y shi s d i en by
ee geno ype and spa ial- empo al ac o s in Medi e anean pine o es s. En i onmen al
Mic obiology 19, 1639-1652.
Picon-Cocha d, C., Pilon, R., Ta oux, E., Pagès, L., Robe son, J., Dawson, L., 2012.
E ec o species, oo b anching o de and season on he oo ai s o 13 pe ennial g ass
species. Plan and Soil 353, 47-57.
Reich Pe e , B., Tjoelke Ma k, G., P egi ze Ku , S., W igh Ian, J., Oleksyn, J., Machado,
J.L., 2008. Scaling o espi a ion o ni ogen in lea es, s ems and oo s o highe land
plan s. Ecology Le e s 11, 793-801.
Roge s B endan, M., Jan z, P., Goe z Sco , J., 2016. Vulne abili y o eas e n US ee
species o clima e change. Global Change Biology 23, 3302-3320.
Roume , C., Bi ous e, M., Picon Cocha d, C., Ghes em, M., Osman, N., V ignon B enas,
S., Cao, K. ., S okes, A., 2016. Roo s uc u e unc ion ela ionships in 74 species:
e idence o a oo economics spec um ela ed o ca bon economy. New Phy ologis 210,
815-826.
Ruiz Beni o, P., Gómez Apa icio, L., Za ala, M.A., 2012. La ge scale assessmen o
egene a ion and di e si y in Medi e anean plan ed pine o es s along ecological g adien s.
Di e si y and Dis ibu ions 18, 1092-1106.
Ruiz Beni o, P., Ra cli e, S., Za ala, M.A., Ma ínez Vilal a, J., Vilà Cab e a, A., Llo e ,
F., Mad igal González, J., Wi h, C., G eenwood, S., Kändle , G., Leh onen, A., Ka ge, J.,
Dahlg en, J., Jump, A.S., 2017. Clima e and successional ela ed changes in unc ional
composi ion o Eu opean o es s a e s ongly d i en by ee mo ali y. Global Change
Biology 23, 4162-4176.
Sangüesa-Ba eda, G., Lina es, J.C., Cama e o, J.J., 2015. Reduced g ow h sensi i i y o
clima e in ba k-bee le in es ed Aleppo pines: Connec ing clima ic and bio ic d i e s o
o es dieback. Fo es Ecology and Managemen 357, 126-137.
Shipley, B., 2002. Cause and co ela ion in biology: a use 's guide o pa h analysis,
s uc u al equa ions and causal in e ence. Camb idge Uni e si y P ess, Camb idge, UK.
Sil e , W.L., Miya, R.K., 2001. Global pa e ns in oo decomposi ion: compa isons o
clima e and li e quali y e ec s. Oecologia 129, 407-419.
Subke, J.-A., Hahn, V., Ba ipaglia, G., Linde , S., Buchmann, N., Co u o, M.F., 2004.
Feedback in e ac ions be ween needle li e decomposi ion and hizosphe e ac i i y.
Oecologia 139, 551-559.
Tang, J., Baldocchi Dennis, D., Xu, L., 2005. T ee pho osyn hesis modula es soil
espi a ion on a diu nal ime scale. Global Change Biology 11, 1298-1304.
U bie a, I.R., Za ala, M.A., Ma añón, T., 2008. Human and non human de e minan s o
o es composi ion in sou he n Spain: e idence o shi s owa ds co k oak dominance as a
esul o managemen o e he pas cen u y. Jou nal o Biogeog aphy 35, 1688-1700.
an de Heijden, M.G.A., Ba dge , R.D., Van S aalen, N.M., 2008. The unseen majo i y:
soil mic obes as d i e s o plan di e si y and p oduc i i y in e es ial ecosys ems.
Ecology Le e s 11, 296-310.
Va gas, R., De o, M., Baldocchi Dennis, D., Allen Michael, F., 2010. Mul iscale analysis
o empo al a iabili y o soil CO2 p oduc ion as in luenced by wea he and ege a ion.
Global Change Biology 16, 1589-1605.
Vay eda, J., Ma ínez-Vilal a, J., Vilà-Cab e a, A., 2016. El In en a io Ecológico y Fo es al
de Ca aluña: una he amien a pa a la ecología uncional. Re is a Ecosis emas 25, 70-79.
Fig. cap ions
Fig. 1. E olu ion o soil wa e con en (SWC), soil empe a u e (Tsoil) and soil espi a ion
(Rs) as a unc ion o ime and ees' heal h s a us (li ing o dead). E o ba s ep esen he
s anda d e o o he mean.
Fig. 2. Changes in soil abio ic (mic oclima e, pH) and bio ic ( ine oo demog aphy, Rs, RH)
a iables, as well as soil o al o ganic ca bon (TOC) and o al o ganic ni ogen (TON)
alues e alua ed unde dead ees ela i e o he alues e alua ed unde li ing ees.
The e o e, posi i e alues ( igh hand side o he e ical ba ) ep esen an inc ease in ha
pa icula a iable unde dead wi h espec o unde li ing ees. E o ba s ep esen
s anda d e o o he mean. As e isk ep esen signi ican om ze o di e ences (p alue >
0.05; - es ).
Fig. 3. Linea mixed-e ec s models o soil espi a ion (Rs). Solid lines ep esen modeled
Rs esponses unde dead ees, whe eas do ed lines ep esen modeled Rs esponses unde
li ing ees. Whe e, Ts = soil empe a u e a 5 cm dep h; PC2 = second dimension o he
PCA, he e ep esen ing SOM.
Fig. 4. Linea mixed-e ec s models o he e o ophic espi a ion (RH). Solid lines ep esen
modeled Rs esponses unde dead ees, whe eas do ed lines ep esen modeled Rs
esponses unde li ing ees. SRL = speci ic oo leng h; PC1= i s dimension o he PCA,
he e ep esen ing soil nu ien s; PC2 = second dimension o he PCA, he e ep esen ing
SOM.
Fig. 5. Mul ig oup SEM ep esen a ion. Pa h diag ams ep esen ing hypo hesized causal
ela ionships be ween abo eg ound ege a ion, bio ic and abio ic a iables, soil espi a ion
(Rs) and soil he e o ophic ac i i y (RH) unde li ing (a) and dead (b) coni e ees. A ows
depic causal ela ionships: posi i e and nega i e e ec s a e indica ed by solid and dashed
lines espec i ely, wi h numbe s indica ing s anda dized es ima ed eg ession weigh s
(SRW). A ow wid hs a e p opo ional o he signi icance alues acco ding o he legend.
Pa hs wi h non-signi ican coe icien s a e ep esen ed in g ay. Coe icien s in bold
cha ac e s ep esen hose causal ela ionships whe e he s eng h o he ela ion di e ed
be ween soils unde li ing (g een) and unde dead (o ange) ees. 2= 86.81, NFI= 0.83 y
RMSEA= <0.0001, d = 122, p= 0.99
Fig.1
Click he e o download high esolu ion image
Fig.2
Click he e o download high esolu ion image
Fig.3
Click he e o download high esolu ion image
Fig.4
Click he e o download high esolu ion image
Fig.5
Click he e o download high esolu ion image
