A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Land managemen shapes d ough
esponses o dominan soil mic obial axa
ac oss g asslands
J. M. La allee
1,2
,M.Chomel
1,3
,N.Al a ezSegu a
4,5
,F.deCas o
6,7
,
T. Goodall
8
,M.Magil on
6,9
,J.M.Rhymes
1,10
,M.Delgado-Baque izo
11,12
,
R. I. G i fi hs
8,13
,E.M.Baggs
14
,T.Ca uso
15
,F.T.deV ies
1,16
,M.Emme son
6
,
D. Johnson
1
&R.D.Ba dge
1
Soil mic obial communi ies a e domina ed by a ela i ely small numbe o axa
ha may play ou sized oles in ecosys em unc ioning, ye li le is known abou
hei capaci ies o esis and eco e om clima e ex emes such as d ough ,
o how en i onmen al con ex media es hose esponses. He e, we imposed an
in si u expe imen al d ough ac oss 30 di e se UK g assland si es wi h con-
as ing managemen in ensi ies and ound ha : (1) he majo i y o dominan
bac e ial (85%) and ungal (89%) axa exhibi esis an o oppo unis ic
d ough s a egies, possibly con ibu ing o hei ubiqui y and dominance
ac oss si es; and (2) in ensi e g asslandmanagemen dec eases hep opo ion
o d ough -sensi i e and non- esilien dominan bac e ia—likely ia alle ia ion
o nu ien limi a ion and pH- ela ed s ess unde e ilisa ion and liming—bu
has he opposi e impac on dominan ungi. Ou esul s sugges a po en ial
mechanism by which in ensi e managemen p omo es bac e ia o e ungi
unde d ough wi h implica ions o soil unc ioning.
Soil mic obial communi ies media e ecosys em unc ions including
nu ien cycling, o ganic ma e decomposi ion, and pa hogen
con ol1–3, bu hei unc ioning can be impac ed by clima e ex emes4,5
which a e becoming inc easingly common. Recen e idence shows
ha despi e e y high di e si y o soil mic obial axa, a small p opo -
ion can be conside ed dominan , i.e., hey a e ound ac oss mos soils
and a e highly abundan ela i e o o he axa6,7. These dominan axa
may be d i e s o ecosys em esponses o clima e ex emes (i.e., he
mass- a io hypo hesis8), an idea suppo ed by s udies o plan com-
muni ies linking ecosys em esponses o he abundances o dominan
plan species9,10. The e o e, unde s anding how dominan mic obial
axa espond o clima e ex emes and how hese esponses a e shaped
Recei ed: 14 May 2023
Accep ed: 22 No embe 2023
Check o upda es
1
Depa men o Ea h and En i onmen al Sciences, The Uni e si y o Manches e , Ox o d Road, Manches e M13 9PT, UK.
2
En i onmen al De ense Fund, 257
Pa k A e S, New Yo k, NY10010, USA.
3
FiBL F ance, Resea ch Ins i u e o O ganic Ag icul u e, 26400 Eu e, F ance.
4
Ins i u e o Biological and En i onmen al
Sciences, Uni e si y o Abe deen, S Macha D , Old Abe deen, Abe deen AB24 3UL, UK.
5
EURECAT—Cen e Tecnològic de Ca alunya, C/ de Bilbao, 72,
08005 Ba celona, Spain.
6
School o Biological Sciences and Ins i u e o Global Food Secu i y,Queen’s Uni e si yo Bel as , 19 Chlo ine Ga dens, Bel as BT9
5DL, UK.
7
Ag iFood & Biosciences Ins i u e, 18a New o ge Ln, Bel as BT9 5PX, UK.
8
UK Cen e o Ecology & Hyd ology Walling o d, Maclean Building, Benson
Lane, C owma sh Gi o d, Walling o d,Ox o dshi e OX10 8BB,UK.
9
School o Li e Sciences, Uni e si y o Lincoln, B ay o d Pool, Lincoln LN6 7TS, UK.
10
Cen e
o Ecology & Hyd ology Bango , En i onmen Cen e Wales, Deiniol Road, Bango LL57 2UW, UK.
11
Labo a o io de Biodi e sidad y Funcionamien o Ecosis-
émico. Ins i u o de Recu sos Na u ales y Ag obiología de Se illa (IRNAS), CSIC, A . Reina Me cedes 10, E-41012 Se illa, Spain.
12
Unidad Asociada CSIC-UPO
(BioFun). Uni e sidad Pablo de Ola ide, 41013 Se illa, Spain.
13
School o Na u al Sciences, Bango Uni e si y, Deiniol Rd, Bango LL57 2UR, UK.
14
Global
Academy o Ag icul u e and Food Sys ems, Royal (Dick) School o Ve e ina y S udies, Eas e Bush Campus, Cha nock B adley Building, Uni e si y o
Edinbu gh,Edinbu gh EH259RG,UK.
15
School o Biology and En i onmen al Science, Uni e si y College Dublin, Dublin, I eland.
16
Ins i u e o Biodi e si y and
Ecosys em Dynamics, Uni e si y o Ams e dam, 1090 GE Ams e dam, Ne he lands. e-mail: jla a[email p o ec ed]g
Na u e Communica ions | (2024) 15:29 1
1234567890():,;
1234567890():,;
by en i onmen al ac o s and land managemen will enable be e
p edic ions o ecosys em beha iou in o he u u e11,12.
Soil mic obial axa can be ca ego ised by li e his o y
s a egies13,14 o in o m on hei capaci y o esis and eco e om
clima e ex emes such as d ough 11,15. These li e his o y s a egies a e
hough o eme ge om co ela ed se s o ai s (e.g., ela ed o
esou ce acquisi ion, g ow h yield, and s ess ole ance), which a e
a ou ed unde di e en en i onmen al condi ions14. Fo example,
soil mic obial communi ies subjec ed o mois u e pulses may ha e
g ea e p opo ions o axa exhibi ing a s ess- esis an s a egy,
whe eas hose unde ambien condi ions may ha e highe abun-
dances o d ough -sensi i e axa13. Land managemen may also shi
mic obial li e his o y s a egies by changing esou ce a ailabili y and
plan communi ies—en i onmen al ac o s known o shape mic obial
communi y s uc u e and unc ion16–21. Howe e , he in e ac ing
e ec s o land managemen and clima e ex emes such as d ough
ha e no been s udied in he con ex o mic obial li e his o y s a-
egies. This is a necessa y s ep owa ds using ecological knowledge o
soil mic obes o p edic and unde s and he consequences o land
managemen decisions on soil unc ioning and sus ainabili y in he
ace o clima e change.
He e, we ca ied ou a la ge-scale field expe imen ac oss a b oad
ange o g assland si es o explo e how he ela i e abundances o
dominan mic obial axa wi h di e en d ough - esponse s a egies
a e shaped by soilcondi ions, clima e, and land managemen in ensi y.
We imposed a simula ed d ough on 15 pai s o g asslands unde
con as ing managemen (i.e., in ensi e and ex ensi e) in h ee geo-
g aphically dis inc egions o he UK ep esen ing a ange o soil and
clima ic condi ions (Fig. S1, Table S1). Using an ope a ional app oach,
we iden ified dominan mic obial axa and classified hem in o h ee
b oad d ough - esponse s a egies (i.e., esis an [no de ec able
esponse], oppo unis ic [posi i e esponse], o sensi i e [nega i e
esponse])13. We examined he in e ac ing e ec s o clima e, soil
p ope ies, and his o ical g assland managemen on dominan
mic obial axa by d ough - esponse s a egy immedia ely ollowing
he d ough and a e a 60-day pos -d ough pe iod22, o cap u e bo h
mic obial esis ance (lack o esponse o a pe u ba ion) and esilience
( eco e y o an un-pe u bed s a e) o d ough 23,24.
We hypo hesised ha : (1) dominan soil mic obial axa la gely
display esis an o oppo unis ic s a egies unde d ough , because a
capaci y o wi hs and a iable mois u e condi ions would pa ly
explain hei ubiqui y and abundance ac oss si es; (2) in ensi e
g assland managemen , cha ac e ized by egula e ilise and lime
applica ion and highe plan p oduc i i y (Table S1), a ou s axa ha
a e maladap ed o low esou ce a ailabili y and s ess and he e o e
sensi i e o d ough ; and (3) in ensi e g assland managemen a ou s
mic obial axa ha eco e a e d ough (i.e., esilien ), because mo e
a ou able soil condi ions allow d ough -a ec ed axa o ebound
quickly wi h ewe ing.
Ou esul s show ha mos dominan soil mic obial axa we e
esis an o d ough , as expec ed. We u he show ha in ensi e
g assland managemen inc eases he p opo ion o dominan bac-
e ial axa ha a e esis an o oppo unis ic in he ace o d ough
ela i e o hose ha a e sensi i e, and inc eases he p opo ion o
dominan bac e ial axa ha a e esilien ela i e o hose ha a e no
esilien . Howe e , in ensi e managemen has he opposi e e ec on
dominan ungal axa, inc easing he p opo ions o sensi i e and
non- esilien axa. Ou finding ha land managemen shapes he
d ough - esponse s a egies o dominan soil mic obial axa has
impo an implica ions o mic obial communi y s uc u e and
unc ion. In ensi e g assland managemen is known o b oadly
a ou bac e ia o e ungi, impac ing key unc ions including soil
ca bon and ni ogen cycling25,26; ou esul s sugges his pa e n may
be exace ba ed as d ough s become mo e equen and in ense wi h
clima e change.
Resul s
Mos dominan soil mic obial axa a e esis an o d ough
We ound ha a ela i ely small numbe o bac e ia and ungi dom-
ina e soils ac oss he g assland si es, and ha hese axa we e highly
esis an o an imposed d ough e en . Fo bac e ia, dominan axa
(defined as p esen ac oss all 15 si es and in he op 10% o ela i e
abundance anked by 16S RNA eads7) ep esen ed 1269 ou o 19224
o al ope a ional axonomic uni s (OTUs), which accoun ed o ~7% o
o al OTUs bu 76% o all eads. Fo ungi, dominan axa (p esen
ac oss all h ee egions and in he op 10% by ITS RNA eads) made up
209 ou o 12837 o al OTUs, accoun ing o ~2% o o al OTUs bu 53%
o all eads. O e all, he majo i y o dominan bac e ial (66%) and
ungal (64%) axa we e classified as displaying a esis an d ough
s a egy, as hey showed no esponse o d ough in ou hie a chical
model using all da a ac oss si es and managemen egimes immedi-
a ely a e he simula ed d ough (Table S2). Oppo unis ic axa, whose
ela i e abundances inc eased in esponse o d ough , ep esen ed
19% o dominan bac e ia and 25% o dominan ungi; sensi i e axa,
whose ela i e abundances dec eased wi h d ough , ep esen ed 12%
o dominan bac e ia and 7% o dominan ungi.
Dominan bac e ial phyla in ou da ase comp ised p ima ily (by
eads) P o eobac e ia (32%), Acidobac e ia (21%), Ve ucomic obia
(13%), Bac e oide es (11%), Fi micu es (9%), Ac inobac e ia (7%), Chlo -
oflexi (3%), and se e al o he globally dis ibu ed axa. O hese phyla,
mos con ained axa ep esen ing each o he h ee d ough - esponse
s a egies (Fig. 1). Howe e , membe s o Fi micu es and Bac e oide es
ended o display esis an o sensi i e d ough - esponse s a egies,
wi h ew o no axa iden ified as oppo unis ic (ze o ou o 47 in Fi -
micu es; fi e ou o 175 in Bac e oide es). Membe s o Acidobac e ia,
Ac inobac e ia,andChlo oflexi ended o display esis an o oppo -
unis ic d ough - esponse s a egies, wi h ew axa iden ified as sen-
si i e (nine ou o 227 in Acidobac e ia, one ou o 118 in Ac inobac e ia,
and one ou o 65 in Chlo oflexi). Dominan ungal phyla comp ised (by
eads) Mo ie ellomyco a (48%), Ascomyco a (22%), Basidomyco a
(15%), Glome omyco a (1%), and se e al o he known and globally dis-
ibu ed o uniden ifiable axa. Membe s o Ascomyco a ended o
display esis an o oppo unis ic d ough - esponses s a egies, wi h
only six o 94 axa iden ified as ha ing a d ough -sensi i e s a egy.
Membe s o Mo ie ellomyco a, Basidiomyco a,andGlome omyco a
ended o display esis an o sensi i e d ough - esponse s a egies,
wi h only one o no axa iden ified as oppo unis ic in each phylum
(Fig. 1). O e all, dominan axa esis an o d ough belonged o di -
e en axonomic g oups dispe sed ac oss e e y majo lineage o he
phylogeny, sugges ing ha his capabili y is no limi ed o specific
phylogene ic g oups o mic obes.
Managemen a ec s dominan bac e ia and ungi di e en ly
We used s uc u al equa ion models o in e po en ial mechanisms
h ough which g assland managemen a ec ed oppo unis ic, sensi-
i e, and esis an dominan mic obial axa ac oss si es (Fig. 2). Excep
o sensi i e bac e ial axa, in ensi e managemen inc eased he ela-
i e abundances o all dominan mic obial d ough - esponse g oups.
Oppo unis ic and esis an bac e ial axa we e posi i ely impac ed by
in ensi e managemen a bo h imepoin s (bo h di ec ly and ia
inc eased pH; Fig. 2a, b), while sensi i e bac e ial axa we e ei he
una ec ed ( ollowing d ough ) o nega i ely a ec ed (a e he
eco e y pe iod). Oppo unis ic and esis an ungal axa we e also
posi i ely a ec ed by in ensi e managemen (ei he di ec ly o ia
inc eased pH; Fig. 2c, d), bu in con as o sensi i e bac e ial axa,
sensi i e ungal axa we e posi i ely and di ec ly a ec ed by in ensi e
managemen a bo h imepoin s. As a esul , he a io o oppo unis-
ic:sensi i e dominan axa inc eased unde in ensi e managemen o
bac e ia bu dec eased o ungi (Fig. 3a).
O he en i onmen al a iables we conside ed in he SEMs ( o al C
and N, empe a u e, ex u e, mois u e, and pH), pH played he mos
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 2
impo an ole. The e we e s ong posi i e indi ec e ec s o man-
agemen in ensi y ia inc eased soil pH o oppo unis ic and esis an
bac e ial axa a bo h imepoin s (Fig. 2a, b). Fu he in es iga ion
e ealed unimodal ela ionships be ween pH and esis an and esi-
lien bac e ial axa ha peaked ca. pH 5.7 (Fig. S4). Fungal axa we e
less impac ed by pH o e all, bu he e was a posi i e e ec on
oppo unis ic ungal axa a e he d ough (Fig. 2c), and a nega i e
e ec on esis an ungal axa a e he eco e y pe iod (Fig. 2d). While
he inclusion o pH did accoun o one mechanism by which man-
agemen impac s mic obial axa, he ac ha di ec pa hs om he
managemen a iable mani es ed in he SEMs indica es ha o he
mechanisms ela ed o managemen (and no cap u ed by o al soil C
and N, soil empe a u e, ex u e, and soil wa e con en ) a e also
impac ing dominan mic obial axa in hese soils. In ensi e manage-
men did impac o he key a iables including abo e-g ound plan
biomass and plan -a ailable N (Fig. 3) ha a e implici ly ep esen ed by
ou managemen a iable in he SEM. In gene al, dominan ungal
g oups we e impac ed mo e s ongly by he managemen a iable in
ou SEMs, while dominan bac e ial g oups we e impac ed mo e
s ongly by pH and o he soil cha ac e is ics ( o al soil C and N, soil
empe a u e, ex u e, and soil wa e con en ).
D ough ea men and soil mois u e e ec s on dominan
mic obes
D ough ea men was he bes p edic o o soil mois u e immedia ely
a e he simula ed d ough (day 0), wi h la i ude and soil p ope ies
cap u ed in he composi e soil a iable ( o al C and N, empe a u e,
ex u e) also playing impo an oles (Fig. 2a, c). The d ough ea -
men e ec on he di e en mic obial d ough - esponse s a egy
g oups was no ully cap u ed by he field measu emen s o soil
mois u e—which only p o ided a snapsho o soil mois u e condi ions
a he ime o sampling—indica ed by he di ec pa hs om d ough
ea men o se e al mic obial g oups a ha imepoin a day 0
(Fig. 2a, c). A e he 60-day pos -d ough pe iod, he d ough ea -
men no longe p edic ed soil mois u e o mic obial d ough - esponse
s a egy g oups. Ins ead, la i ude was a e y s ong p edic o o soil
mois u e, and soil p ope ies (composi e soil a iable) we e an
impo an p edic o o bac e ial d ough esponse g oups, bu no
ungal d ough esponse g oups (Fig. 2b, d). The absence o d ough
ea men e ec s on sensi i e and oppo unis ic bac e ial o ungal
axa a e he 60-day pos -d ough pe iod indica es g oup-le el
eco e y wi hin ha ime (Fig. 2b, d).
Mos d ough -a ec ed dominan bac e ia and ungi a e esilien
We ca ego ized indi idual oppo unis ic and sensi i e dominan axa
as esilien o no based on hei abundances ela i e o ambien
con ol plo s a e he 60-day pos -d ough eco e y pe iod. While
mos o he 503 d ough -a ec ed (oppo unis ic o sensi i e) axa
we e ound o be esilien a e 60 days, we iden ified 110 axa ha
we e no (Fig. 1). O hese, 34 we e sensi i e bac e ial axa and 8 we e
sensi i e ungal axa ha di e ed om ambien con ol plo le els
a e he 60-day pos -d ough eco e y pe iod. Analyses o esilien
bac e ial and ungal axa g oups in con ol plo s ac oss bo h ime-
poin s e ealed ha he ela i e p opo ion o esilien axa ( a io o
esilien :no esilien axa; Fig. 3b) was highe o bac e ia bu lowe o
ungi unde in ensi e compa ed o ex ensi e g assland managemen .
Discussion
Ou s udy p o ides no el e idence, om a b oad ange o g assland
si es a ying in clima ic and soil condi ions (Table S1), ha dominan
soil mic obial axa a e highly esis an o d ough . Despi e significan
and sizable educ ions in soil mois u e unde expe imen al d ough
ac oss si es (Fig. S2), he majo i y o dominan soil mic obial axa
ei he did no espond o esponded posi i ely. O he axa ha we e
nega i ely impac ed by he d ough ea men (d ough -sensi i e
s a egies), he majo i y we e esilien (i.e., did no di e om ambien
P o eobac e ia
Acidobac e ia
Bac e oide es
Ve ucomic obia
Ac inobac e ia
Fi micu es
Chlo o lexi
Ni ospi ae
Planc omyce es
Resis ance
Bac e ia
Resilience
No esilien
Oppo unis ic
Resilien
Sensi i e
Resis an
Resis ance
Resilience
a
Uniden i ied
Uniden i ied
Chy idiomyco a
Muco omyco a
Ce cozoa
Ascomyco a
Basidiomyco a
Mo ie ellomyco a
Glome omyco a
Resis ance
Resilience
Resis ance
Resilience
No esilien
Oppo unis ic
Resilien
Sensi i e
Resis an
b
Fungi
Fig. 1 | Taxonomic ee showing d ough esponses o dominan soil
mic obial axa. Dominan bac e ial (a) and ungal (b) communi y esponses o
d ough immedia ely ollowing he d ough ea men (“ esis ance”)anda e he
60-day pos -d ough pe iod (“ esilience”), limi ed o he op 500 mos abundan
axa ac oss all samples o eadabili y. The inne ing shows he axonomic ee,
colou ed by phylum. The middle ing displays esponses o each ou immedia ely
ollowing d ough (ligh blue = esis an , black = oppo unis ic, pink = sensi i e).
The ou e ing displays esponses a e he 60-day pos d ough pe iod (g een =
esilien axa ha eco e ed o con ol le els, da k g ey = no esilien ). Taxa
defined as esis an o d ough (ligh blue, inne ing) we e no es ed o esilience.
Sou ce da a and iden i ies o all OTUs a e p o ided on Gi Hub73.
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 3
con ol le els wi hin he 60-day pos -d ough pe iod). The esis ance
and esilience o hese soil mic obial axa o d ough , obse ed he e
ac oss h ee geog aphically dis inc egions o he UK, may in pa
explain why hey a e p esen and highly abundan (i.e., dominan )
ac oss si es7. The use o a dis ibu ed landscape design combined wi h
an in si u expe imen al d ough ea men uniquely demons a es ha
esponses o dominan soil mic obial axa o d ough a e consis en a
a la ge spa ial scale. Though d ough se e i y can be di ficul o
quan i y27, especially a he mic oscale mos ele an o mic obio a28,
we obse ed significan e ec s o he d ough ea men on ecosys em
espi a ion and mic obial communi y s uc u e (including non-
dominan axa) a he plo scale ac oss all egions, indica ing ha
ou d ough ea men was ecologically significan (Fig. S2, Fig. S3,
Table S3). Ou findings align wi h ecen s udies showing ha abun-
dan mic obial axa a e mo e esis an o pe u ba ions29, a e adap ed
o b oade anges o en i onmen al condi ions30,31, and display highe
equencies o genomic ai s associa ed wi h s ess- ole ance and
compe i i e abili ies6 han a e mic obial axa. These esul s sugges
ha dominan mic obial axa in g assland soils a e gene alis s adap ed
o a ying en i onmen al condi ions, allowing hem o wi hs and
pe u ba ions and h i e ac oss a b oad ange o si es.
We ound ha en i onmen al con ex and land managemen did
a ec he ela i e abundances o dominan mic obial axa wi h di -
e en d ough esponse s a egies, bu no in ways we expec ed. We
expec ed ha he impac s o g assland managemen on he esis ance
and esilience (i.e., he capaci y o eco e ) o dominan mic obial axa
o d ough would be in e sely ela ed32,33, and ha bac e ial and ungal
communi ies would espond simila ly. Mo e specifically, we hypo he-
sised ha mic obial communi ies in in ensi ely managed g asslands
would be mo e sensi i e o d ough due o lowe s ess- ole ance bu
be mo e esilien due o highe a ailable nu ien s and mo e ideal pH
le els enabling eco e y. Howe e , ou findings sugges ha esis ance
and esilience o dominan soil mic obial axa a e posi i ely ela ed in
he con ex o g assland managemen , and ha bac e ial and ungal
Bac e ia
ab
cd
SWC
Soil
pH
.22
.25
.28
.57
.55
-.7
-.72
-.26
.48
.38
.41
.37
A e eco e y pe iod, day 60 A e d ough , day 0
Fungi
La i ude
Posi i e
DF = 5, χ2 = 2.5, CFI = 1.0, RMSEA = 0.0, P alue = 0.77
Bollen S ine Boo s ap P alue = 0.783
DF = 5, χ2 = 2.5, CFI = 1.0, RMSEA = 0.0, P alue = 0.77
Bollen S ine Boo s ap P alue = 0.788
DF = 5, χ2 = 0.356, CFI = 1.0, RMSEA = 0.0, P alue= 0.996
Bollen S ine Boo s ap P alue = 0.994
DF = 5, χ2 = 0.356, CFI = 1.0, RMSEA = 0.0, P alue = 0.996
Bollen S ine Boo s ap P alue = 0.994
Nega i e
D ough
In ensi e
Managemen
Oppo unis ic Sensi i e Resis an
-.48
Soil
pH
.4
.46
.51
-.42
.67
.67
.74
.40
.5
.65
.68
.55
.57
SWC
Soil
pH
.25
.36
.31
.57
-.72
-.26
.41
.57
.47
.67
.35
.27
.3
R2 = 40%
R2 = 32%
R2 = 37%R2 = 57%R2 = 43%
R2 = 40%
R2 = 33%
R2 = 60%
R2 = 40%R2 = 48%R2 = 66%
R2 = 39%
R2 = 45%
R2 = 36%R2 = 56%
R2 = 59%R2 = 26%R2 = 42%
R2 = 45%
R2 = 36%R2 = 49%R2 = 36%
La i ude
D ough
In ensi e
Managemen
Oppo unis ic Sensi i e Resis an
-.48
.25
.49
.29
-.28
-.29
-.68
-.32
-.38
.42
La i ude
D ough
Oppo unis ic Sensi i e Resis an
-.46
.54
.3
.25
SWC
Soil
pH
.56
.67
.67
.41
.42
La i ude
D ough
In ensi e
Managemen
Oppo unis ic Sensi i e Resis an
-.46
R2 = 39% R2 = 36%
-.37
SWC
In ensi e
Managemen
Fig. 2 | Po en ial mechanisms po en ial mechanisms a ec ing dominan soil
mic obial axa ac oss si es. S uc u al equa ion models (SEMs) o dominan
bac e ia (panels aand b) and ungi (panels cand d) a he wo ime poin s in his
s udy: immedia ely ollowing he d ough ea men (day 0, panels aand c), and
a e he 60-day pos -d ough eco e y pe iod (panels band d). The “soil” a iable
is a composi e ep esen a ion o soil C, N, ex u e, and empe a u e. SWC is soil
wa e con en by olume. The d ough - esponse s a egy (oppo unis ic, sensi i e,
esis an ) o each OTU wasde e mined by he d ough ea men e ec using linea
mixed models. A ow (pa h) hickness co esponds o he s anda dized
coe ficien s, also w i en nex o hei espec i e pa hs. Pa hs o less in e es a e
shaded g ey o imp o e o e all eadabili y. Wald es s we e used o e alua e he
null hypo hesis ha indi idual pa h coe ficien s we e equal o ze o. Solid a ows
indica e indi idual pa h coe ficien s wi h P- alues < 0.05; pa h coe ficien s wi h P-
alues > 0.05 a e no shown. O e all model fi o each model was assessed using
Chi-squa ed es (χ2) wi h associa ed P- alue and deg ees o eedom (DF), Com-
pa a i e Fi Index (CFI),and oo mean squa ee o o app oxima ion (RMSEA). See
Fig. S5 and Supplemen al No e 1 o mo e de ail. Sou ce da a a e p o ided on
Gi Hub73.
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 4
communi ies espond o in ensi e and ex ensi e g assland manage-
men in di e gen ways. Compa ed o communi ies unde ex ensi e
managemen , dominan bac e ial communi ies unde in ensi e man-
agemen shi ed owa d less sensi i e and mo e esilien d ough
s a egies, while dominan ungal communi ies shi ed owa d mo e
sensi i e and less esilien d ough s a egies (Fig. 3). This sugges s
ha ac oss hese g assland si es, dominan bac e ial communi ies
unde mo e in ensi e managemen a e be e able o wi hs and and
eco e om d ough han hose unde ex ensi e managemen , while
dominan ungal communi ies a e no . Again, hese findings we e
appa en when da a we e agg ega ed ac oss all h ee UK egions,
which co e a b oad ange o clima ic and soil condi ions.
The di e gence be ween bac e ial and ungal esponses o mo e
in ensi e managemen may be explained by di e ences in hei sen-
si i i ies o p e ailing condi ions including pH, nu ien s, and plan
p oduc i i y. The in ensi ely managed g asslands used in ou s udy all
ecei e egula inpu s o ino ganic e ilise s o educe nu ien lim-
i a ion along wi h lime, which inc eases pH owa d neu al le els and
leads o inc eased plan p oduc i i y (Fig. 4). Fo he dominan bac-
e ial communi ies a hese si es, liming likely alle ia es pH- ela ed
s ess, allowing oppo unis ic axa o succeed unde he d ough
ea men ela i e o o he axa. These oppo unis ic axa may ha e
ai s ela ed o high g ow h yields o e ficien esou ce acquisi ion34
ha enable hem o ake apid ad an age o abundan esou ces unde
a
−0.5
0.0
0.5
Ex ensi e In ensi e
Oppo unis ic:Sensi i e
Log Ra io, S d. Rel. Abundance
b
−2
−1
0
1
2
Ex ensi e In ensi e
−0.50
−0.25
0.00
0.25
Ex ensi e In ensi e
Managemen
Resilien :No Resilien
Log Ra io, S d. Rel. Abundance
−1
0
1
2
3
Ex ensi e In ensi e
Managemen
Managemen Managemen
Bac e ia
Bac e ia
Fungi
Fungi
* *
* *
Fig. 3 | G assland managemen e ec s on dominan soil mic obial axa. Ra ios
o he s anda dized ela i e abundances o oppo unis ic:sensi i e dominan axa
(a) and esilien :no esilien dominan axa (b) by g assland managemen , wi h
bac e ia on le and ungi on igh . Boxplo s show he median (cen e line), fi s and
hi d qua iles (box limi s), and smalles and la ges alues wi hin 1.5x in e qua ile
ange (whiske s), and all da apoin s a e shown (n= 90 expe imen al plo s o all
boxplo s). Resilien axa we e defined as ha ing simila ela i e abundances o
hose in con ol plo s a e he 60-day pos -d ough eco e y pe iod. Ra ios we e
highe o bac e ia bu lowe o ungi in in ensi ely managed g asslands based on
linea mixed models using da a om con ol plo s ac oss bo h imepoin s (P-
alue < 0.05 indica ed by *). Ou pu o linea mixed models o e ec s o in ensi e
e sus ex ensi e managemen o each esponse a iable: (a), le : (14) = 5.61,
P= 0.0001, e ec size = 0.277, 95% Confidence In e als = 0.267, 0.287; panel a,
igh : (14) = −4.58, P= 0.0004, e ec size = −0.382, 95% Confidence In e als =
−0.542, −0.222; (b), le : (14) = 3.48, P= 0.0036, e ec size = 0.134, 95%Confidence
In e als = 0.058, 0.21; panel b, igh : (14) = −2.20, P= 0.045, e ec size = −0.377,
95% Confidence In e als = −0.717, −0.037. Sou ce da a a e p o ided on Gi Hub73.
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 5

changing condi ions. Indeed, he highe soil pH obse ed in he
in ensi ely managed g asslands (due o lime applica ion) posi i ely
a ec ed esis an and esilien bac e ial axa ela i e o he ex ensi e
g asslands wi h mo e acidic soils (Fig. S4). This finding ag ees wi h
p e ious wo k on simila soils sugges ing ha elie om acidic con-
di ions allows bac e ial communi ies o shi om main enance o
g ow h s a egies35. In ha s udy, he key pH h eshold o shi s in
mic obial s a egies was ound o be pH ca. 6.2, howe e , in ou s udy
he pH in in ensi ely managed fields a ely su passed ha h eshold,
sugges ing he pH h eshold could be lowe o many o ou si es. In
addi ion o highe pH, he highe soil nu ien a ailabili y and plan
p oduc i i y in he in ensi ely managed g asslands likely u he
a ou ed copio ophic o high-yield bac e ial axa36 capable o aking
ad an age o changing condi ions unde d ough , o capi alizing on
flushes o nu ien s upon ewe ing o he d ough ed plo s20,34.Indeed,
Ac inobac e ia had he highes p opo ion o oppo unis ic axa in ou
s udy (consis en wi h a p e ious la ge-scale s udy o d ough e ec s
on mic obial communi ies in g asslands4) and his phylum is hough
o comp ise p ima ily copio ophs o high-yield s a egy axa a ou ed
by N addi ions20,36–38.Fu he ,Ve ucomic obia and Acidobac e ia,
which ha a e hough o be comp ised o mainly oligo ophs ( axa
ha g ow slowly and pe o m well unde nu ien -poo condi ions
ela i e o copio ophs)20,37,38, had he lowes p opo ions o d ough -
sensi i e axa ha we e esilien .
In con as o dominan bac e ial communi ies, dominan ungal
communi ies unde mo e in ensi e managemen gene ally displayed
lowe esis ance and esilience od ough hanin ex ensi ely managed
g asslands. Fungal communi ies a e known o be less sensi i e o pH
han bac e ia39, and we didn’ obse e s ong pH e ec s on esis an o
esilien dominan ungal axa in his s udy (Fig. S4), sugges ing ha
alle ia ion o pH- ela ed s ess was no as ele an a mechanism o
ungi in his case. Ins ead, o he local-scale impac s o managemen
such as inc eased plan biomass and a ailable nu ien s (Fig. 4)we e
he likely d i e s o ungal esponses, as sugges ed by he ac ha he
managemen a iable in ou SEMs gene ally a ec ed dominan ungal
g oups mo e s ongly han pH o p e ailing soil condi ions (Fig. 2).
Fungal communi ies ha e been shown o espond s ongly o
e ilisa ion20,40 and a e o en supp essed ela i e o bac e ia unde
mo e in ensi e g assland managemen 17,41,42, consis en wi h ou
obse a ion o lowe ungal:bac e ial a ios unde in ensi e compa ed
o ex ensi e managemen ac oss si es (Fig. 4). Fu he mo e, we
ecen ly showed in a sub-se o he g assland si es s udied he e ha
in ensi e managemen educes he flux o ecen pho osyn ha e o soil
ood webs including a buscula myco hizal ungi, indica ing impo -
ance o his pa hway o d i ing ungal ac i i y43. I is he e o e pos-
sible ha his pa hway o educed ene gy flux could con ibu e o he
inc eased sensi i i y o dominan ungal communi ies o d ough
(which u he educes he flux o ecen pho osyn ha e below-
g ound44) in in ensi ely managed g asslands. The opposing espon-
ses o dominan bac e ia and ungi o g assland managemen in e ms
o hei esis ance and esilience o d ough may help o explain
widesp ead obse a ions o dec easing ungal:bac e ial biomass a ios
wi h g assland in ensifica ion17,42,45.
O e all, he alignmen o esis ance and esilience in he con ex
o g assland managemen in ensi y o bo h bac e ia and ungi was
unexpec ed, as o he s udies ha e ound ade-o s be ween esis ance
and esilience in soil mic obial communi ies33,46,47.Howe e ,inou
s udy encompassing a ela i ely b oad ange o soils, pH was an
impo an d i e o bo h esis ance and esilience in dominan bac-
e ial communi ies, while e ilisa ion may ha e d i en bo h esis ance
and esilience o dominan ungal communi ies, which would help o
explain he alignmen in bo h esponses wi h managemen . While
consis encies in axon-le el esponses o sepa a e d ough and ni o-
gen addi ion ea men s has been obse ed p e iously48, he se s o
ai s de e mining esponses o soil wa e a ailabili y e sus nu ien
a ailabili y o pH may no always align and a mul i-dimensional a-
mewo k may be necessa y o conside ing mic obial li e his o y
s a egies49 and p edic ing mic obial esponses o clima e ex emes.
Fungal:bac e ial
a io
Mic obial biomass
pH
Plan −a ailable N
Soil C:N a io
Soil C
Abo eg ound
plan biomass
−1 0 1 2
Log Response Ra io, In ensi e s. Ex ensi e
Region
De on
No h Yo kshi e
Abe deenshi e
Fig. 4 | G assland managemen a ec s a ange o en i onmen al a iables. Log
esponse a io o key a iables ela ed o g assland in ensifica ion (C is ca bon, N is
ni ogen). Log esponse a io is calcula ed as he na u al log o he a io o he alue
o a gi en a iable in an in ensi ely managed field o he co esponding alue in he
pai ed ex ensi ely managed field. Fi een pai s o in ensi e and ex ensi e
g asslands, 5 pe egion, a e shown he e wi h each poin ep esen ing he log a io
o wi hin-field means o one pai (si e). Boxplo s show he median (cen e line),
fi s and hi d qua iles (box limi s), and smalles and la ges alues wi hin 1.5x
in e qua ile ange (whiske s), and all da apoin s a e shown. Sou ce da a a e p o-
ided on Gi Hub73.
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 6
We obse ed con as ing phylum-le el esponses o d ough in
soil dominan bac e ial and ungal communi ies, sugges ing ha ce -
ain phyla may be inhe en ly mo e esis an and esilien o d ough
han o he s. Ac inobac e ia con ained a high p opo ion o esis an
and oppo unis ic axa, wi h only one axon iden ified as sensi i e,
consis en wi h p e ious obse a ions ha Ac inobac e ia a e p e-
alen in d y en i onmen s50 and a e highly esis an o inc ease in
esponse o d ough 4,49. Membe s o Fi micu es and Bac e oide es we e
gene ally mo e sensi i e o he d ough ea men , and while membe s
o Bac e oide es ha e been shown o dec ease in ela i e abundance in
d ie soils, membe s o Fi micu es ha e p e iously shown he opposi e
esponse51. I is possible ha hese p e ious obse a ions may ha e
been d i en p ima ily by one o a ew axa, which may no ha e been
p esen (o defined as dominan ) he e. Con ex -dependen d ough
esponses ha e been p e iously obse ed o o he phyla including
P o eobac e ia and Planc omyce es51, and we also obse ed ela i ely
high numbe s o axa wi h di e en d ough - esponse s a egies in
hose phyla.
Dominan membe s o Ascomyco a we e pa icula ly oppo unis-
ic unde d ough , which ag ees wi h findings ha Ascomyco a a e
dominan globally and a e gene alis s ha a e adap ed o a wide ange
o condi ions6,50.Wi hinGlome omyco a, dominan axa ha espon-
ded o ou d ough ea men we e sensi i e, in ag eemen p e ious
findings ha bo h communi y composi ion4and unc ionali y52,53 o
his g oup o ungi espond o d ough in o he sys ems. Howe e , he
majo i y o dominan Glome omyco a in his s udy we e ound o be
esis an o he d ough , sugges ing again ha he esul s om hese
o he s udies my la gely be d i en by only a ew dominan membe s o
Glome omyco a,o by axa ha we eno defined as dominan he e.
Two membe s o Basidiomyco a indica ed sensi i i y o d ough , and
one axon did no eco e a e he 60-day pos -d ough pe iod. Gi en
ha Basidiomyco a a e impo an decompose s and ec omyco hizal
symbion s in o es s54, mic obial communi ies in o es ed sys ems (o
unde o es expansion) may be sensi i e o d ough wi h po en ial
implica ions o o es g ow h and ecosys em unc ioning55,which
dese es u he s udy.
O e all, ou findings om a b oad ange o g assland si es ac oss
he UK indica e ha mos o he dominan soil mic obial axa a e
highly esis an o d ough , which may explain hei p e alence ac oss
a di e se ange o g assland soils. We u he show ha g assland
managemen , along wi h clima e and soil p ope ies, shapes he ela-
i e abundances o dominan soil mic obial axa wi h di e ing
d ough - esponse s a egies. Mo e in ensi e g assland managemen ,
which c ea es mo e op imal pH and highe ni ogen a ailabili y com-
pa ed o ex ensi e managemen , p omo es oppo unis ic and esilien
bac e ial axa ha may employ copio ophic o as - esponse s a e-
gies and a e able o ake ad an age o changing condi ions. Howe e , i
has he opposi e e ec on dominan ungal axa which may help o
explain inc eases in bac e ial p e alence o e ungi wi h g assland
in ensifica ion17,42,45,56. Ou esul s sugges he pa e n o bac e ial
p e alence o e ungi unde in ensi e managemen may be ein o ced
o exace ba ed as d ough s become mo e equen and in ense wi h
clima e change, and po en ially con ibu e o less e ficien ca bon and
ni ogen cycling in hese sys ems25,26.
By demons a ing ha land managemen shapes he d ough -
esponse s a egies o dominan mic obial axa ac oss g asslands, ou
findings imp o e ou unde s anding o how soil mic obial commu-
ni ies espond o d ough . Mo eo e , by iden i ying consis en man-
agemen - and d ough -induced esponses o dominan mic obial axa,
ou findings pa e he way o u u e s udies ha in e oga e hei
unc ional a ibu es and links o key ecosys em unc ions57.Gi en he
eno mous complexi y o soil mic obial communi ies and hei
dynamics in space and ime, ou app oach o ocusing on he d ough
esponse s a egies o dominan axa is one way o make his ask mo e
easible in he u u e.
Me hods
Field si es
The field expe imen was ca ied ou be ween May and Sep embe o
2016 ac oss a se ies o meso ophic g asslands in he Uni ed Kingdom,
concen a ed in h ee egions: De on in sou hwes England, No h
Yo kshi e in no he n England, and Abe deenshi e in no heas Sco -
land (Fig. S1, Table S1). P io o he s a o he expe imen , we iden-
ified 15 pai s o fields on wo king a ms wi h con as ing managemen
and classified hem as ei he in ensi ely o ex ensi ely managed based
on obse a ions o plan communi ies and in e iews wi h a me s and
land manage s. Ex ensi ely managed fields ecei ed e y low o no
syn he ic e ilise and lime, had mo e di e se plan communi ies,
we e gene ally no cu o hay o silage, and we e g azed a low
s ocking densi ies by sheep o ca le. In ensi ely managed fields
ecei ed egula applica ions o e ilise and lime (as deemed neces-
sa y by he a me ), had less di e se plan species mix u es, we e cu
o hay o silage, and we e g azed a highe s ocking densi ies. Di -
e ences in managemen had been main ained o a leas 10 yea s, and
ypically longe (Table S1). Whe e e possible, we iden ified pai ed
in ensi e and ex ensi e fields ha we e adjacen , o minimize di e -
ences in in insic en i onmen al a iables such as opog aphy, wea he
pa e ns, and soil ype. I fields we e no immedia ely adjacen , we
chose fields no mo e han 0.5 km apa and used a me and land
manage in e iews o ensu e minimal di e ences be ween pai ed
fields aside om managemen .
Expe imen al design
This s udy employed a andomized comple e block design wi h sub-
sampling. In each egion, 5 si es we e iden ified ha each had wo di -
e en ly managed fields wi hin 0.5 km o a o al o 15 si es and 30 pai ed
fields. In each o he 30 fields, h ee pai s o d ough and ambien
con ol plo s we e es ablished and enclosed in encing o p o ec ion
om la ge mammals and machine y. A field d ough was simula ed by
placing a anspa en oo (1.5 m * 1.3 m) on each d ough plo alongside
i s pai ed delimi ed con ol plo o 60 days be ween May and July o
2016, which equa es o a >100-yea d ough o hese si es58.In o al,
he e we e 90 pai s o d ough ed and ambien con ol plo s, and he
h ee wi hin-field eplica es o each we e ea ed app op ia ely in all
s a is ical models by ei he including si e and field as andom e ec s o
by agg ega ing he da a a he field scale whe e andom e ec s could
no be modelled. A he end o he d ough pe iod, d ough shel e s
we e emo ed and an ini ial (“day 0”) sampling and measu emen o soil
unc ions (in bo h d ough and con ol plo s) was ca ied ou o assess
he impac o he d ough ela i e o he ambien con ol condi ions.
Sampling and measu emen we e done in he cen e o he plo s,
lea ing a 15 cm bu e o minimize edge e ec s. Immedia ely ollowing
his sampling e en , d ough ed plo s we e wa e ed (amoun s we e
based on a e age July ain e en s om 2007-2011 o he nea es Me
O fice om each egion59) o s imula e he s a o he pos -d ough
pe iod. Sampling o d ough and ambien con ol plo s was epea ed
60 days a e he emo al o he shel e s o cap u e eco e y du ing he
pos -d ough pe iod ( esilience).
Soil sampling
A all imepoin s, mul iple soil samples we e collec ed o 10 cm dep h
and composi ed o measu emen s o soil nema ode communi ies
(6 * 1.3-cm diame e co es) soil mic oa h opod communi ies (4 * 2.5-cm
diame e co es), and soil mic obio a and chemical analysis (3 * 2.5-cm
diame e co es). Soil samples we e immedia ely composi ed in plas ic
sample bags and ans e ed o coole s o anspo o labo a o ies
wi hin 24–48 h. Samples in ended o soil auna analysis we e kep open
o allow o gas exchange. A each sampling e en , soil mois u e and
empe a u e we e measu ed using We Senso p obes (WET-2, Del a-T
De ices, Camb idge, UK). Bulk densi y was measu ed using he co e
echnique a he ime o he d ough ea men es ablishmen , using
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 7
one co e pe plo o a o al o 6 co es pe field, and he a e age alue
o each field was used h oughou he s udy.
Soil biogeochemical analysis
Samples o analysis o mic obial communi ies, ex u e, and C and N
analyses we e anspo ed o he Uni e si y o Manches e and s o ed
a 4 °C o a maximum o 3 days un il u he p ocessing and analysis.
All samples we e sie ed o 4 mm o homogeniza ion and emo al o
isible plan ma e ial and ocks, a e which samples we e di ided o
u he analyses. One subsample was immedia ely ozen a −80 °C
awai ing mic obial DNA sequencing. A second subsample was
weighed, placed in a pape bag, and d ied o cons an weigh a 40 °C
o calcula e soil mois u e. This subsample was used o u he ana-
lyses o o al C and N concen a ions using a Va io Cube (Elemen a
Ame icas Inc., Ronkonkoma, NY, USA), and soil ex u e analysis by
lase g anulome y using a Mal e n Mas e size 2000 (Mal e n
Ins umen s L d, Mal e n, Wo ces e shi e, UK) ollowing emo al o
o ganic ma e wi h H
2
O
2
a 50 °C o e nigh . Soil pH was measu ed on
field mois subsamples in slu ies o 1:2.5 soil:deionized wa e using a
pH me e (Se en2GO Me le Toledo, Columbus, Ohio, USA). Fu he
analyses a e desc ibed in Supplemen a y Me hods.
16S and ITS amplicon sequencing and da a analysis
Amplicon sequencing and bioin o ma ic and s a is ical analyses o
sequencing da a we e done ollowing he me hods o De V ies e al.60
DNA was ex ac ed om 0.16 g o soil using he MoBIO Powe Soil-h p
96-Well DNA Isola ion ki (Ca lsbad, CA, USA) acco ding o he man-
u ac u e ’s p o ocols and he DNA quali y was checked by aga ose gel
elec opho esis. Bac e ial 16S RNA sequencing ollowed he dual
indexing p o ocol o Kozich e al.61 o he MiSeq pla o m (Illumina,
San Diego, CA, USA). Each p ime consis ed o he app op ia e Illumina
adap o , 8-n index sequence, a 10-n pad sequence, a 2-n linke , and
he amplicon specific p ime . The V3–V4 hype a iable egions o he
bac e ial 16S RNA gene we e amplified using p ime s 341 F62 and
806 R63, CCTACGGGAGGCAGCAG, and GCTATTGGAGCTGGAATTAC,
espec i ely. Amplicons we e gene a ed using high-fideli y DNA poly-
me ase Q5 Taq (M0491L, New England Biolabs, Ipswich, USA), p e-
mixed dNTPs (BIO-39053, Me idian Bioscience, Ohio, US), and using
Eppedo Mas e cycle Nexus PCR machines (Hambu g, Ge many).
A e an ini ial dena u a ion a 95 °C o 2 min, PCR condi ions we e:
dena u a ion a 95 °C o 15 s, annealing a 55 °C o 30 s wi h ex ension
a 72 °C o 30 s, epea ed o 30 cycles, ollowed by a final ex ension o
10 min a 72 °C.
Fungal in e nal ansc ibed space (ITS) amplicon sequences we e
gene a ed using a 2-s ep amplifica ion app oach. P ime s GTGARTC
ATCGAATCTTTG and TCCTCCGCTTATTGATATGC64 we e each mod-
ified a he 5’end wi h he addi ion o Illumina p e-adap o and Nex e a
sequencing p ime sequences. A e an ini ial dena u a ion a 95 °C o
2 min, PCR condi ions we e: dena u a ion a 95 °C o 15 s, annealing a
52 °C o 30 s wi h ex ension a 72 °C o 30 s, epea ed o 25 cycles,
wi h a final ex ension o 10 min a 72 °C included. PCR p oduc s we e
cleaned using a DNA Clean-up Ki (ZR-96, Zymo Resea ch Inc., I ine,
US) ollowing manu ac u e ’s ins uc ions. MiSeq adap o s AATGAT
ACGGCGACCACCGAGATCTACAC and 8n dual-indexing ba code
sequences we e added du ing a second s ep o PCR amplifica ion.
A e an ini ial dena u a ion 95 °C o 2 min, PCR condi ions we e:
dena u a ion a 95 °C o 15 s; annealing a 55 °C o 30 s wi h ex ension
a 72 °C o 30 s; epea ed o 8 cycles wi h a final ex ension o
10 min a 72 °C.
Amplicon concen a ions we e no malized using SequalP ep
No maliza ion Pla e Ki (A10510-01, The mo Fishe Scien ific, Wal ham,
US) and amplicon sizes de e mined using an 2200 TapeS a ion (Agi-
len , San a Cla a, US) p io o sequencing each amplicon lib a y
sepa a ely using MiSeq (Illumina, San Diego, US) wi h V3 600 cycle
eagen s (MS-102-3003, Illumina, San Diego, US) a concen a ions o
14 and 7 pM (16S and ITS espec i ely) wi h a 5% PhiX con ol 3 (FC-
110-3001, Illumina, San Diego, US) lib a y.
Sequenced pai ed-end eads we e joined using PEAR65,quali y
fil e ed using FASTX ools (hannonlab.cshl.edu), and leng h-fil e ed o
a minimum leng h o 300 bp. The p esence o PhiX and adap o s we e
checked o and emo ed wi h BBTools (jgi.doe.go /da a-and- ools/
bb ools/), and chime as we e iden ified and emo ed wi h
VSEARCH_UCHIME_REF66 using G eengenes Release 13_5 (a 97%).
Single ons we e emo ed and he esul ing sequences we e clus e ed
in o ope a ional axonomic uni s (OTUs) wi h VSEARCH_CLUSTER66 a
97% sequence iden i y. Rep esen a i e sequences o each OTU we e
axonomically assigned by RDP Classifie wi h he boo s ap h eshold
o 0.8 o g ea e using he G eengenes Release 13_5 ( ull) as he
e e ence. Unless s a ed o he wise, de aul pa ame e s we e used o
all s eps lis ed. The ungal ITS sequences we e analysed using PIPITS67
wi h de aul pa ame e s. B iefly, his in ol ed quali y fil e ing and 97%
clus e ing o he ITS2 egion as indica ed abo e o he 16S p ocessing,
using he UNITE da abase o chime a emo al and axonomic iden i-
fica ion o ep esen a i e OTUs. Bo h bac e ial and ungal OTU abun-
dance ableswe e a ified o a minimum o 9000 eads pe sample,
and samples wi h ze o eads we e emo ed p io o u he analyses.
Plo s showing ci cula ep esen a ions o he axonomic ees
we e c ea ed using he G aPhlAn so wa e ool (h ps://hu enhowe .
sph.ha a d.edu/g aphlan/).
S a is ical analysis
All analyses we e done sepa a ely o bac e ial and ungal axa in R
e sion 4.0.268.Wedefined dominan axa as hose which we e p esen
ac oss all 15 si es (managemen pai s) and ep esen ed he op 10% o
axa when anked by ela i e abundance ( RNA eads). The esponse o
each o hese dominan axa o d ough ea men was iden ified using
a gene alized linea mixed model ac oss all expe imen al plo pai s
wi h d ough ea men as a fixede ec ,and egion/si e/field as nes ed
andom e ec s (R package glmmTMB e sion 1.1.569). Fo each indi i-
dual model, he app op ia e dis ibu ion (poisson, nega i e binomial,
o binomial) was assumed based on diagnos ics o model esiduals,
which we e assessed using R package DHARMa e sion 0.4.670.The
d ough - esponse s a egy o each axon was iden ified as esis an
(no significan esponse o d ough de ec ed), sensi i e (nega i e
esponse), o oppo unis ic (posi i e esponse) using a significance
le el (α) o 0.05. Fu he s a is ical analysis (linea mixed e ec s
models using R package nlme e sion 3.1–14871) was pe o med a he
d ough - esponse g oup le el (i.e., esis an , oppo unis ic, sensi i e,
esilien , no esilien ). G oup-le el indices we e calcula ed as ollows:
o each OTU in a gi en g oup, i s ela i e abundance in a gi en sample
was s anda dized ela i e o i s abundance ac oss all samples; hese
s anda dized abundances we e hen summed ac oss all OTUs in a
gi en g oup esul ing in one alue (index) pe g oup pe sample.
S uc u al equa ion modelling was used o in es iga e e ec s o
his o ical managemen , d ough , and soil p ope ies on ela i e
abundances o oppo unis ic, sensi i e, and esis an axa a he wo
sampling imepoin s. Wi hin-field eps we e a e aged p io o analysis
(n= 180 expe imen al plo s/3 field eplica es = 60 alues pe ime-
poin ). We cons uc ed an a p io i model based on cu en knowledge
o plan -soil-mic obe- unc ioning in e ac ions (see Fig. S5 and Sup-
plemen a y No e 1) and es ed whe he he da a fi hese models using
he s anda d modelling app oach in he la aan R package, e sion 0.6-
1272. We c ea ed a p oxy o soil p ope ies using axis 1 sco es om a
non-me ic mul idimensional scaling plo ha included o al soil ca -
bon, o al ni ogen, and soil empe a u e (see Supplemen a y No e 1).
We used mul iple pa ame e s including oo mean squa e e o o
app oxima ion (RMSEA), compa a i e fi index (CFI), and S anda dized
Roo Mean Squa ed Residual (SRMR) o assess model fi .
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 8
Repo ing summa y
Fu he in o ma ion on esea ch design is a ailable in he Na u e
Po olio Repo ing Summa y linked o his a icle.
Da a a ailabili y
The sequence da a gene a ed in his s udy ha e been deposi ed in he
EMBL Nucleo ide Sequence Da abase (ENA) unde accession code
PRJEB63076. All o he da a gene a ed in his s udy ha e been depos-
i ed on Gi Hub73 .
Code a ailabili y
All code is a ailable om Gi Hub73.
Re e ences
1. Wagg,C.,Bende ,S.F.,Widme ,F.& ande Heijden,M.G.A.Soil
biodi e si y and soil communi y composi ion de e mine ecosys em
mul i unc ionali y. P oc.Na lAcad.Sci.USA111,5266–5270
(2014).
2. Ba dge , R. D. & an de Pu en, W. H. Belowg ound biodi e si y
and ecosys em unc ioning. Na u e 515,505–511 (2014).
3. Delgado-Baque izo, M. e al. Mic obial di e si y d i es mul i-
unc ionali y in e es ial ecosys ems. Na . Commun. 7,
10541 (2016).
4. Ochoa-Hueso, R. e al. D ough consis en ly al e s he composi ion
o soil ungal and bac e ial communi ies in g asslands om wo
con inen s.Glob. Chang. Biol. 24,2818–2827 (2018).
5. Oli e io,A.M.,B ad o d,M.A.&Fie e ,N.Iden i ying hemic obial
axa ha consis en ly espond o soil wa ming ac oss ime and
space. Glob. Chang. Biol. 23, 2117–2129 (2017).
6. Egidi, E. e al. A ew Ascomyco a axa domina e soil ungal com-
muni ies wo ldwide. Na . Commun. 10,2369(2019).
7. Delgado-Baque izo, M. e al. A global a las o he dominan bac e ia
ound in soil. Science 359,320–325 (2018).
8. G ime, J. P. Benefi s o plan di e si y o ecosys ems: immedia e,
fil e and ounde e ec s. J. Ecol. 86,902–910 (1998).
9. Hoo e ,D.L.,Knapp,A.K.&Smi h,M.D.Resis anceand esilience
o a g assland ecosys em o clima e ex emes. Ecology 95,
2646–2656 (2014).
10. Smi h, M. D. e al. Mass a io e ec s unde lie ecosys em esponses
o en i onmen al change. J. Ecol. 108,855–864 (2020).
11. Ba dge , R. D. & Ca uso, T. Soil mic obial communi y esponses o
clima e ex emes: Resis ance, esilience and ansi ions o al e -
na i e s a es. Philos.T ans.R.Soc.BBiol.Sci.375,521–633
(2020).
12. G i fi hs, B. S. & Philippo , L. Insigh s in o he esis ance and esi-
lience o he soil mic obial communi y. FEMS Mic obiol. Re . 37,
112–129 (2013).
13. E ans, S. E. & Wallens ein, M. D. Clima e change al e s ecological
s a egies o soil bac e ia. Ecol. Le . 17,155–164 (2014).
14. Malik,A.A.e al.Defining ai -based mic obial s a egies wi h
consequences o soil ca bon cycling unde clima e change. ISME
J. 14,1–9 (2020).
15. Wallens ein,M.D.&Hall,E.K.A ai -based amewo k o p e-
dic ing when and whe e mic obial adap a ion o clima e change
will a ec ecosys em unc ioning. Biogeochemis y 109,
35–47 (2012).
16. Donnison, L. M., G i fi h, G.S.,Hedge ,J., Hobbs, P. J. & Ba dge , R.
D. Managemen influences on soil mic obial communi ies and hei
unc ion in bo anically di e se haymeadows o no he n England
and Wales. Soil Biol. Biochem. 32,253–263 (2000).
17. G ays on, S. J. e al. Assessing shi s in mic obial communi y
s uc u e ac oss a ange o g asslands o di e ing managemen
in ensi y using CLPP, PLFA and communi y DNA echniques. Appl.
Soil Ecol. 25,63–84 (2004).
18. Boeddinghaus, R. S. e al. Plan unc ional ai shi s explain con-
cu en changes in he s uc u e and unc ion o g assland soil
mic obial communi ies. J. Ecol. 107,2197–2210 (2019).
19. de V ies, F. T. e al. Abio ic d i e s and plan ai s explain
landscape-scale pa e ns in soil mic obial communi ies. Ecol. Le .
15,1230–1239 (2012).
20. Le , J. W. e al. Consis en esponses o soil mic obial communi ies
o ele a ed nu ien inpu s in g asslands ac oss he globe. P oc. Na l
Acad.Sci.USA112, 10967–10972 (2015).
21. Lekbe g, Y. e al. Ni ogen and phospho us e iliza ion consis en ly
a o pa hogenic o e mu ualis ic ungi in g assland soils. Na .
Commun. 12,3484(2021).
22. Vilonen, L., Ross, M. & Smi h, M. D. Wha happens a e d ough
ends: syn hesizing e ms and defini ions. N. Phy ol. 235,
420–431 (2022).
23. Shade, A., Pe e , H., Allison, S. D. & Baho, D. L. Fundamen als o
mic obial communi y esis ance and esilience. F on . Mic obiol.3,
85124 (2012).
24. Nimmo, D. G. e al. la ésis ance: e i ing esis ance o 21s cen u y
conse a ion. T ends Ecol. E ol. 30,516–523 (2015).
25. Malik, A. A. e al. Soil ungal: bac e ial a ios a e linked o al e ed
ca bon cycling. F on . Mic obiol. 7,11511–11512 (2016).
26. de V ies, F. T. e al. Ex ensi e managemen p omo es plan and
mic obial ni ogen e en ion in empe a e g assland. PLoS One 7,
e51201 (2012).
27. Sle e,I.J.e al.Howecologis sdefine d ough , and why we should
do be e . Glob. Chang. Biol.25,3193–3200 (2019).
28. Po , V. e al. Th ee-dimensional dis ibu ion o wa e and ai in soil
po es: compa ison o wo-phase wo- elaxa ion- imes la ice-
Bol zmann and mo phological model ou pu s wi h synch o on
X- ay compu ed omog aphy da a. Ad . Wa e Resou . 84,
87–102 (2015).
29. Jiao,S.,Wang,J.,Wei,G.,Chen,W.&Lu,Y.Dominan oleo
abundan a he han a e bac e ial axa in main aining ag o-soil
mic obiomes unde en i onmen al dis u bances. Chemosphe e
235,248–259 (2019).
30. Jiao, S. & Lu, Y. Abundan ungi adap o b oade en i onmen al
g adien s han a e ungiinag icul u alfields. Glob.Chang.Biol.26,
4506–4520 (2020).
31. Wan, W. e al. En i onmen al adap a ion is s onge o abundan
a he han a e mic oo ganisms in we land soils om he Qinghai-
Tibe Pla eau. Mol. Ecol. 30,2390–2403 (2021).
32. Yi, C. & Jackson, N. A e iew o measu ing ecosys em esilience o
dis u bance. En i on. Res. Le .16, 45216 (2021).
33. O win, K. H., Wa dle, D. A. & G eenfield, L. G. Con ex -dependen
changes in he esis ance and esilience o soil mic obes o an
expe imen al dis u bance o h ee p ima y plan ch on-
osequences. Oikos 112,196–208 (2006).
34. Malik, A. A., Puissan , J., Goodall, T., Allison, S. D. & G i fi hs, R. I. Soil
mic obial communi ies wi h g ea e in es men in esou ce acqui-
si ion ha e lowe g ow h yield. Soil Biol. Biochem. 132,
36–39 (2019).
35. Malik, A. A. e al. Land use d i en change in soil pH a ec s mic obial
ca bon cycling p ocesses. Na . Commun. 9,3591(2018).
36. Fie e , N. e al. Compa a i e me agenomic, phylogene ic and phy-
siological analyses o soil mic obial communi ies ac oss ni ogen
g adien s. ISME J. 6,1007–1017 (2012).
37. Rami ez,K.S.,C aine,J.M.&Fie e ,N.Consis en e ec so ni o-
gen amendmen s on soil mic obial communi ies and p ocesses
ac oss biomes. Glob.Chang. Biol. 18,1918–1927 (2012).
38. Fie e , N., B ad o d, M. A. & Jackson, R. B. Towa d an ecological
classifica ion o soil bac e ia. Ecology 88,1354
–1364 (2007).
39. Rousk, J. e al. Soil bac e ial and ungal communi ies ac oss a pH
g adien in an a able soil. ISME J. 4,1340–1351 (2010).
A icle h ps://doi.o g/10.1038/s41467-023-43864-1
Na u e Communica ions | (2024) 15:29 9