Land use in ensi ica ion ans o ms g assland a h opod communi ies: A
ew winne s expand bu mos species disappea
J.P. Reinie de V ies
a,*
, Jos´
e G. an Paassen
a
, Remco Ploeg
a
, Aa Noo dam
b
, Koen Ve hoog
c
,
Da id Kleijn
a
a
Plan Ecology and Na u e Conse a ion G oup, Wageningen Uni e si y, 6708 PB, Wageningen, he Ne he lands
b
Independen esea che , Leiden, he Ne he lands
c
Du ch Bu e ly Associa ion, Mennonie enweg 10, 6702 AD, Wageningen, he Ne he lands
ARTICLE INFO
Keywo ds:
In e eb a es
Insec s
Spide s
Biodi e si y
Be a-di e si y
Ag icul u e
Conse a ion
ABSTRACT
A h opod declines pose exis en ial h ea s o e es ial ecosys ems bu emain poo ly unde s ood. In human-
domina ed habi a s such as managed g asslands, a h opod communi ies a e hough o decline as a esul o
land-use in ensi ica ion. Howe e , s udies ha link land-use in ensi y o a h opod species ichness, abundance
and composi ion ha e ound mixed esul s. These inconsis encies may s em om di e ences in he in es iga ed
in ensi ica ion le els o a h opod communi ies, as di e en species may show di e en esponses o land-use
change. We show ha e ec s o land-use in ensi ica ion on a h opod communi ies can be unde s ood be e
by examining he dis ibu ions o indi idual species along land-use in ensi y g adien s. We sampled mul iple
a h opod g oups along a wide g adien o g assland p oduc i i y as a p oxy o land-use in ensi y. Species
ichness hal ed om low o high p oduc i i y le els, bu o e all abundance emained s able because inc easing
Dip e a numbe s coun e ac ed s ong declines in o he o de s. Dissimila i y me ics ailed o e eal a subs an ial
change in species composi ion ha was shown in indi idual species' dis ibu ions. While low-p oduc i i y
g asslands suppo ed many species ha we e con ined o na ow p oduc i i y anges, high-p oduc i i y g ass-
lands we e domina ed by a ew widely occu ing species. O e all, 56 % o all species declined o disappea ed as
land-use in ensi ied a low- o-medium in ensi y le els (‘lose s’); 35 % p e e ed in e media e in ensi ies, and only
9 % p o i ed om high-in ensi y managemen . These ‘winne s’ p obably sus ain o e all abundance le els in
high-p oduc i i y g asslands. Such la ge changes in species composi ion could ha e signi ican consequences o
ecosys em unc ioning, necessi a ing u he expe imen al s udies. The une en dis ibu ion o biodi e si y losses
implies ha ecological es o a ion should ocus on low-in ensi y g asslands.
1. In oduc ion
A h opod popula ions in human-domina ed landscapes ha e
signi ican ly declined in he ecen pas (Ben on e al., 2002; Hallmann
e al., 2017; Van Klink e al., 2020; Wagne , 2020). As he mos abun-
dan and di e se aunal li e o m (S o k e al., 2015), a h opods a e
cen al o ophic ne wo ks and ecosys em unc ions such as nu ien
cycling and pollina ion (Ca dinale e al., 2012; IPBES, 2019; Po s e al.,
2016). The e o e, he e ela ion o d as ic a h opod declines has been
cause o ala m (Ca doso e al., 2020; Goulson, 2019). Many s udies
ha e a ibu ed hese declines o ongoing land-use in ensi ica ion,
causing habi a loss, declines o c i ical esou ces (e.g. h ough plan
di e si y decline), inc eased dis u bances and landscape simpli ica ion
(Gossne e al., 2016; Ha pole e al., 2016; Ra en and Wagne , 2021;
Seibold e al., 2019). Ye , mos a h opods a e challenging o su ey and
sys ema ic long- e m moni o ing schemes exis o only a ew a h opod
g oups (Van Swaay e al., 2025; Wa en e al., 2021). Ha dly any mul i-
decadal eco ds o o he a h opods exis , and hose a ailable may
e lec al eady deg aded e e ence si ua ions o sampling loca ions ha
a e no ep esen a i e o en i onmen al changes in he wide landscape
(Cooke e al., 2025; Mon gome y e al., 2020; Wagne e al., 2021). As a
esul , he e ec s o land-use in ensi ica ion emain poo ly unde s ood
o a wide ange o di e en a h opod g oups.
S udying a h opod di e si y along spa ial g adien s in land-use
* Co esponding au ho .
E-mail add esses: [email p o ec ed], [email p o ec ed] (J.P.R. de V ies), [email p o ec ed] (J.G. an Paassen), [email p o ec ed]
(R. Ploeg), [email p o ec ed] (A. Noo dam), [email p o ec ed] (K. Ve hoog ), [email p o ec ed] (D. Kleijn).
Con en s lis s a ailable a ScienceDi ec
Biological Conse a ion
jou nal homepage: www.else ie .com/loca e/biocon
h ps://doi.o g/10.1016/j.biocon.2025.111522
Recei ed 5 June 2025; Recei ed in e ised o m 28 Augus 2025; Accep ed 29 Sep embe 2025
Biological Conse a ion 313 (2026) 111522
A ailable online 10 Oc obe 2025
0006-3207/© 2025 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license (
h p://c ea i ecommons.o g/licenses/by/4.0/ ).
in ensi y p o ides a way o accele a e ou unde s anding o
in ensi ica ion-d i en changes o a h opod assemblages (Cooke e al.,
2025; Wagne , 2020). Al e na i e o ime se ies, biodi e si y pa e ns
along spa ial g adien s a e mo e eadily ob ained and can p o ide mo e
di ec accoun s o he d i e s o biodi e si y change (Blü hgen e al.,
2022; Ca dinale e al., 2018). G adien s udies may include ecologically
deg aded loca ions ha a e a ely pa o exis ing long- e m s udies
(Wagne e al., 2021), and can be conside ed p oxies o changes o e
ime, such as he shi om adi ional ex ensi e land managemen o
mode n in ensi e ag icul u e ha occu ed o e he pas cen u y in
Eu ope and o he empe a e egions (Ba dge e al., 2021; Emme son
e al., 2016). Global syn heses ha e indica ed ha land-use in ensi ica-
ion is especially impac ing biodi e si y in g assland habi a s (Newbold
e al., 2015, 2016). While ex ensi ely used g asslands a e known ho -
spo s o a h opod di e si y in empe a e egions (Ba dge e al., 2021;
Habel e al., 2019), s eep declines o species ichness wi h land-use in-
ensi y we e shown among se e al g oups o g assland a h opods, such
as wild bees (Ek oos e al., 2020), A aneae, Hemip e a, Coleop e a and
O hop e a (Allan e al., 2014; Simons e al., 2015).
Howe e , he e ha e also been many s udies ha ound no ela ion
be ween land-use in ensi y and species ichness (Beckmann e al., 2019).
This migh be due o non-linea i y along he ull land-use in ensi y
g adien , as species ichness losses we e shown o be mos p o ound
unde in ensi ica ion a low in ensi y le els ha migh be missed by
s udies in mo e in ensi ely used landscapes (Allan e al., 2014; Kleijn
e al., 2009, 2011). The lack o ela ionships wi h species ichness may
also be in luenced by he limi a ions o using species ichness as he
main esponse a iable, no ably because i does no cap u e changes in
species composi ion (Chase and Knigh , 2013; Hilleb and e al., 2018;
Kuczynski e al., 2023). Mo e de ailed quan i a i e da a o a h opod
communi ies can unco e changes in species assemblages o a h opod
abundances. These me ics a e inc easingly assessed in biodi e si y
s udies, bu also show con as ing esul s. Land-use in ensi ica ion has
been linked o bio ic homogeniza ion (i.e. inc easing simila i y o spe-
cies assemblages be ween si es) ac oss mul iple axonomic g oups in
Eu opean g asslands, ampli ying biodi e si y decline a landscape scales
(Chis ´
e e al., 2018; Gossne e al., 2016). Bu me ics o communi y
dissimila i y ha e also e ealed widesp ead pa e ns o bio ic di e en-
ia ion in esponse o land-use change (Do nelas e al., 2023; Keck e al.,
2025). Fu he mo e, abundances o g assland a h opods ha e shown
s onge declines han species numbe s (Seibold e al., 2019; Van S ien
e al., 2019), bu ha e also been ound o inc ease wi h land-use in-
ensi y (Kleijn e al., 2010; Sil a-Mon ei o e al., 2022). This sugges s
ha he e a e also a h opods ha can cope wi h o e en p o i om
land-use in ensi ica ion.
The con as ing esul s o hese di e en s udies ela ing changes in
a h opod communi ies o land-use in ensi ica ion migh esul om
changes in species composi ion ha he mos commonly used commu-
ni y indica o s ailed o cap u e. To da e, mos s udies agg ega e he
esponses o many di e en a h opod species o en i e g oups in o
common me ics o species ichness, o al abundance o communi y
dissimila i y (Cooke e al., 2025). Bu while species ichness does no
e lec changes a communi y and popula ion le els (Hilleb and e al.,
2018), o al abundance may be la gely d i en by a ew abundan
a h opod species (Sho all e al., 2009). Fu he mo e, dissimila i y
(be a-di e si y) me ics agg ega e di e ences ha a ise om he u n-
o e o unique species (Legend e, 2014), bu also om s ochas ic d i
(e.g. dispe sal ou side species' ep oduc i e habi a s) and om inc eased
agmen a ion o species dis ibu ions (A aújo e al., 2022; Chase e al.,
2019; Wayman e al., 2024; Zhou e al., 2022). In eali y, a h opods
exhibi as ecological di e ences, and he e ec s o land-use in ensi i-
ca ion p obably di e widely be ween a h opod g oups and species.
Ce ain species may p o i om he decline o o he species o om
inc eased esou ce a ailabili y in in ensi ely managed sys ems, e.g.
when eeding on plan ma e o manu e (Simons e al., 2014; Simons
and Weisse , 2017), and he p e alence o such species may di e
be ween a h opod g oups (Powell e al., 2024; Simons e al., 2014). I is
c i ical o unde s and which species lose and which species win unde
changing en i onmen al condi ions (Boyes e al., 2019; Do nelas e al.,
2023; Wagne , 2020). Hence, land-use in ensi ica ion-d i en shi s in
a h opod communi ies can p obably be be e unde s ood by assessing
he esponses o indi idual species.
In his s udy, we quan i ied he changes in g assland a h opod
communi ies along a land-use in ensi y g adien in he Ne he lands.
Ranging om low-p oduc i e un e ilized hay meadows o highly p o-
duc i e and hea ily e ilized silage ields, his g adien e lec ed he
en i e ange o g assland managemen in ensi ies wi hin one s udy e-
gion and a guably co e s he a ia ion in g assland managemen in-
ensi y h oughou much o no hwes e n and cen al Eu ope. We asked
whe he he ela ionships be ween land-use in ensi y and a h opod
abundance, species ichness and communi y dissimila i y di e be ween
he main abo e-g ound a h opod o de s. Fu he mo e, we asked
whe he esponses o a h opod communi ies o land-use in ensi ica ion
can be be e unde s ood by examining he esponses o indi idual
species han by examining agg ega e indices such as species ichness o
be a di e si y. We add essed his ques ion by analyzing he dis ibu ions
o indi idual species along he sampled land-use in ensi y g adien .
2. Me hods
2.1. S udy si es
Da a we e collec ed in a s udy a ea si ua ed in he alley o he i e
Geul in he sou h o he Ne he lands (be ween 50◦48'09"N, 5◦49'46"E
and 50◦53'30"N, 5◦56'24"E) (Fig. S1). This a ea is a egional ho spo o
g assland biodi e si y, suppo ing a ious g assland ypes on ine clay
and loess soils wi h locally mo e lime- ich condi ions (Willems, 2001;
W¨
os en e al., 2013). He e, di e ences in land-use in ensi y a e mainly
he esul o managemen a ge ing ei he na u e conse a ion,
biodi e si y- iendly ag icul u e, o op imizing ag icul u al p oduc ion.
We selec ed 41 g asslands ha ep esen ed he wides possible g adien
in land use in ensi y (see Table S1 o speci ica ions). S udy si es
included i) un e ilized nu ien -poo semi-na u al g asslands ha we e
mown once pe yea ; ii) hay meadows in p o ec ed a eas and iii) hay
meadows unde ag i-en i onmen schemes, bo h o which we e cu
wice a yea o had limi ed g azing a es (<2 LSU.ha
-1
) and no e ilize
inpu s; i ) pas u es used o g azing o odde p oduc ion o suckle
cows o o ganic dai y cows, wi h in e media e e ilize inpu s (50-150
kg N
e
.ha
-1
.y
-1
); and ) con en ional dai y cow odde p oduc ion
g asslands wi h up o six cu s pe yea and high e ilize inpu s (up o
330 kg N
e
.ha
-1
.y
-1
). This g adien ep esen s he en i e ange in
g assland managemen in ensi y ha is p esen in la ge pa s o
empe a e Eu ope whe e his o ically, g assland habi a s e ol ed as
semi-na u al sys ems main ained by ex ensi e human managemen
(mowing o g azing), which has been s eadily eplaced by high-inpu
managemen o maximum p oduc i i y (Emme son e al., 2016;
Habel e al., 2013). The sampling si es we e on a e age 778 ±428 m
apa and showed no spa ial pa e n in p oduc i i y (Fig. S1). We
quan i ied annual p oduc i i y in e ms o ne me abolic ene gy o
li es ock [GJ.ha
-1
.y
-1
] as an indica o o land use in ensi y ha e lec s
bo h biomass p oduc ion and o age quali y, acili a ing he di ec
compa ison o di e en managemen egimes. Annual biomass p oduc-
ion (d y yield in ons) was collec ed h ough in e iews wi h a me s
and land owne s, and was mul iplied by he s anda d ene gy con en
alues o di e en o age ypes (4.1 GJ.
-1
o low-quali y hay; 5.2 o
high-quali y hay; 5.9 o ex ensi e g azing; 6.2 o con en ional silage;
6.4 o con en ional silage, i s cu ; 6.6 o con en ional g azing) (CVB,
2022; Schippe s e al., 2012).
2.2. A h opod sampling
We used sweep-ne sampling o collec a b oad ange o a h opod
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
2
g oups along a 150-me e ansec in each g assland. T ansec s we e p e-
de ined and an om edge o cen e o he g asslands. We conduc ed 70
1-m-wide single-sweeps pe pa o 50 m in June and July, bu 160 in
May, because we expec ed ha his would p oduce a oughly simila
o al a h opod abundance in each mon h. Sweep-ne ing is a sui able
me hod o ep esen a i e ield-scale sampling o ege a ion-dwelling
a h opods, including he bi o es and p eda o s, bu unde es ima es
he p esence o as - lying pollina o s (Simons e al., 2014; Spa o d and
Lo ie, 2013). The e o e, addi ional coun s o bees (An hophila) and
ho e lies (Sy phidae) we e ob ained along he same ansec du ing 15-
minu e pollina o walks (5 min pe 50 m), su eying a wid h o 1 m
(Schepe e al., 2015). Pollina o su eys we e done on he same day bu
p io o sweep-ne ing o minimize dis u bance. Bo h me hods we e
conduc ed be ween 10:00 and 17:30 and unde sui able wea he con-
di ions ( empe a u e >15 ◦C, wind <3 B , sunshine >50 %) (Wes phal
e al., 2008). Pollina o s we e coun ed in he ield and collec ed only
when hey could no be iden i ied o species le el on he wing, while
sweep-ne samples we e collec ed in zip-lock bags and s o ed ozen.
A e wa ds, (sub)o de -le el abundance coun s we e ob ained o adul
Dip e a, Aucheno hyncha, He e op e a, Hymenop e a, Coleop e a,
A aneae and O hop e a (o he a h opod o de s occu ed in low
numbe s bu we e included in o al abundance coun s, while S e no -
hyncha, Thysanop e a, Aca i and Collembola we e no included in he
coun s). Pollina o coun s we e combined wi h he sweep-ne da a o
complemen he assessmen s o bo h pollina o g oups, p ima ily
ega ding species di e si y. Coun da a ep esen ed o e 80 % o all
indi iduals o bo h bees and ho e lies, bu only 6 % and 1.4 % o he
o al abundance o Hymenop e a and Dip e a eco ded.
Fu he mo e, we ob ained species iden i ies o all specimens
belonging o He e op e a, A aneae and O hop e a, and o subg oups o
Hymenop e a (An hophila, Fo micidae), Coleop e a (Can ha idae,
Ch ysomelidae, Coccinellidae, Ela e idae) and Dip e a (Asilidae, Bibio-
nidae, Conopidae, Opomyzidae, Pla ys oma idae, Rhagionidae, Scio-
myzidae, S a iomyidae, Sy phidae, Tabanidae, Teph i idae, Tipulidae,
Ulidiidae). Species we e iden i ied wi h help o he ci izen science
pla o m Waa neming.nl and ele an iden i ica ion keys ( e e ences a e
gi en in Table S2). In a ew cases, iden i ica ion was limi ed o genus o
species-pai le el. The selec ion o subg oups o species iden i ica ion
aimed o co e a b oad ange o axonomic and ecological g oups, while
he inclusion o addi ional species g oups was mainly cons ained by
lack o sui able iden i ica ion keys o axonomic expe ise.
2.3. Da a analyses
The indi idual samples (nine pe si e, e e ing o he h ee ansec
sec ions o e h ee ounds) we e summed (abundance da a) o agg e-
ga ed (species iden i y da a) o ob ain si e-le el es ima es o abundance
and species ichness, wi h abundances being co ec ed o he inc eased
sweep-ne ing e o in May and o one missing sample. We checked
sampling comple eness wi h he Jackni e-1 es ima o o o al ichness,
which was de i ed using R-packages ‘Biodi e si yR’ and ‘ egan’ (Kind
and Coe, 2005; Oksanen e al., 2024). Sampling comple eness anged
be ween 60 and 80 % and showed no end wi h p oduc i i y. This in-
dica es ha mo e species- ich communi ies we e sampled oughly as
comple e as simple communi ies, and we he e o e used he o iginal
da a in ou analyses. Abundance coun s o e all a h opods and o he
se en (sub)o de s, and obse ed ichness o e all iden i ied g oups and
o Dip e a, He e op e a, An hophila, Coleop e a, A aneae, O hop e a
and Fo micidae, we e ela ed o p oduc i i y (me abolic ene gy) as a
single explana o y a iable. We used gene alized linea models wi h
gene alized Poisson o nega i e binomial dis ibu ions o accoun o
o e dispe sion, using R-package ‘glmmTMB’ (B ooks e al., 2017). Non-
linea ela ions we e e alua ed by he s a is ical suppo o quad a ic
and cubic model e ms, which we e kep i no ably imp o ing he model
(ΔAIC
c
>2), using R-package ‘MuMIn’ (Ba o´
n, 2023). Model esidual
pa e ns we e alida ed using R-package ‘DHARMa’ (Ha ig, 2022).
Non-no mal esidual pa e ns we e p e en ed by he selec ed dis ibu-
ions and esiduals we e no co ela ed wi h land use ac o s ou side he
model (silage o hay p oduc ion, mowing o g azing, g ass- o- o b a io),
hence no addi ional co a ia es o andom e ec s we e included in he
models.
We in es iga ed communi y dissimila i y (be a-di e si y) wi hin
classes o low p oduc i i y (4-21 GJ.ha
-1
.y
-1
: 14 si es), medium p o-
duc i i y (21-48 GJ.ha
-1
.y
-1
: 13 si es) and high p oduc i i y (48-91 GJ.
ha
-1
.y
-1
: 14 si es). Be a-di e si y was quan i ied by Jacca d's dissimi-
la i y index and was pa i ioned in o componen s o eplacemen ( ue
niche di e en ia ion) and ichness o abundance di e ences (loss o
gain) (Legend e, 2014), ollowing bo h abundance-based (B ay-Cu is)
and incidence-based app oaches, using R-package ‘BAT’ (Ca doso e al.,
2015). We de i ed all possible unique compa isons o pai s o si es
wi hin one p oduc i i y class. The dissimila i y le els o adjacen classes
we e compa ed wi h Dunn's es s, using Bon e oni's adjus men o p-
alues and a h eshold o p =0.01 o accoun o he la ge numbe o
compa isons, using R-package ‘ s a ix’ (Kassamba a, 2023).
To be e unde s and how he ela ionships be ween land-use in-
ensi y and a h opod abundance, species ichness and communi y
dissimila i y we e shaped by he esponses o indi idual species, we
pe o med addi ional analyses based on he cen e poin s o he dis i-
bu ions o indi idual species along he land-use in ensi y g adien . In
line wi h Chis ´
e e al. (2016), we calcula ed hese cen e poin s as
abundance-weigh ed means, i.e. he mean p oduc i i y (P
mean
) o e all
si es (i-n) o a pa icula species (j) weigh ed by i s abundance a each
si e (x
ij
), using he ollowing o mula:
Pmean(j) = ∑n
i=1xij⋅Pi
∑n
i=1xij
Nex , we excluded single ons (i.e. species wi h single obse a ions
o e all samples), and classi ied all emaining species based on whe he
hei cen e poin was loca ed a low, medium o high p oduc i i y
(using he same classes as he communi y dissimila i y analysis). The
species ha we e cen e ed a low p oduc i i y le els we e ega ded as
species ‘losing’ unde land use in ensi ica ion ( hei numbe s dec eased
as land use in ensi ied), while species cen e ed a medium p oduc i i y
le els we e ega ded as ‘in e media e’ and species cen e ed a high
p oduc i i y le els we e ega ded as ‘winning’ ( hei numbe s inc eased
as land use in ensi ied). We hen calcula ed he pe cen age o losing,
in e media e and winning species in each axonomic g oup. These pe -
cen ages we e compa ed o a simula ion o 10.000 andomized dis i-
bu ions o species along he p oduc i i y g adien (i.e. assuming no
consis en e ec o p oduc i i y on species dis ibu ions). Randomized
dis ibu ions we e cons uc ed by andomly ea anging he eco ded
coun s o each species o e all 41 sampling si es, hus main aining each
species' o iginal coun s. Subsequen ly, we ecalcula ed each species'
cen e poin and de i ed he andomized pe cen ages o losing, in e -
media e and winning species (i.e. he pe cen age o cen e poin s alling
in each p oduc i i y class). The densi y cu es o he 10.000 andomized
pe cen ages hen allowed us o s a is ically es o di e ences wi h he
measu ed pe cen ages o losing, in e media e o winning species a he
communi y le el.
Finally, we assessed how he measu ed and andomized pe cen ages
o losing, in e media e and winning species changed wi h he dis ibu-
ion wid h o species. Fo his he pe cen ages o losing, in e media e
and winning species we e calcula ed o subse s o species ha had been
eco ded a an inc easing minimum numbe o si es ( ange om 1 o 33,
as abo e his numbe , oo ew species (<14) emained o eliably es i-
ma e pe cen ages). We used sa u a ion cu es o model how he pe -
cen ages o losing, in e media e and winning species changed while
inc easing he minimum numbe o occu ences, and assessed he di -
e ences be ween he measu ed and andomized da a. All da a analyses
we e pe o med in R 4.5 (R co e eam, 2025) and can be ound in a
Zenodo eposi o y (De V ies e al., 2025).
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
3
3. Resul s
3.1. Abundance e sus species ichness
We collec ed nea ly 170,000 specimens in o al and iden i ied
27,200 indi iduals (16 %) o species le el, ep esen ing 520 species.
O e all o de s, o al abundance did no change signi ican ly wi h p o-
duc i i y, whe eas obse ed species ichness ac oss all iden i ied sub-
g oups declined exponen ially (p =0.01) (Fig. 1A, C; Table S3). To al
ichness did no change ma kedly in he medium- o-high p oduc i i y
ange (±45-90 GJ.ha
-1
.y
-1
), bu was app oxima ely hal ed om he
lowes o medium p oduc i i y le els. Richness declines we e ound o
all subg oups excep Coleop e a (no end; p <0.01 o all o he sub-
g oups) and we e exponen ial o He e op e a and A aneae (Fig. 1D,
Fig. S2B). Abundance declined signi ican ly wi h p oduc i i y o
Aucheno hyncha (unimodal), He e op e a (exponen ial), A aneae
(cubic), Coleop e a and O hop e a, bu Hymenop e a abundance
emained s able, and Dip e a abundance inc eased signi ican ly wi h
p oduc i i y (p <0.01) (Fig. 1B, Fig. S2A, Table S3). Being he mos
nume ous g oup (46 % o all specimens), he inc ease in Dip e a was
almos as la ge as he decline in abundance o all o he o de s combined.
Howe e , only subg oups o Dip e a and Hymenop e a we e iden i ied
o species le el and he abundance o hese subg oups did no ollow he
whole-o de ends bu declined wi h p oduc i i y (exponen ial o he
iden i ied Dip e a) (Table S3).
3.2. Communi y dissimila i y
Species assemblages o di e en g asslands di e ed ma kedly,
pa icula ly when di e ences in abundance we e aken in o accoun .
To al abundance-based be a-di e si y es ima es o e all species and o
mos g oups we e abo e 0.8, while incidence-based be a-di e si y es i-
ma es we e mos ly be ween 0.6 and 0.8 (Table S4; class a e ages).
Howe e , dissimila i ies be ween si es we e gene ally high in bo h low-,
medium- and high-p oduc i i y classes (4-21, 21-48 and 48-91 GJ.ha
-1
.
y
-1
, espec i ely). The majo i y o he be a-di e si y di e ences be-
ween adjacen p oduc i i y classes we e no signi ican and when di -
e ences we e signi ican , hey did no show a consis en inc ease o
Fig. 1. Rela ions o a h opod abundance and obse ed species ichness wi h g assland p oduc i i y [GJ.ha
-1
.y
-1
]. Signi ican ela ions a e shown in solid lines wi h
95 % con idence in e als ( ibbons). To al a h opod abundance was no ela ed o p oduc i i y (A), bu di e en (sub)o de s showed con as ing ela ions including
an inc ease o Dip e a abundance (B). To al obse ed species ichness was de i ed o e subg oups wi hin each o de ha we e iden i ied o species le el, and
declined exponen ially wi h p oduc i i y (C). This decline is e lec ed in all analyzed g oups excep Coleop e a (D).
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
4
dec ease wi h g assland p oduc i i y (Table S4; class compa isons).
O e all species, eplacemen a e aged 0.5 in bo h abundance-based and
incidence-based app oaches, and was mo e impo an han ichness o
abundance di e ences (a e age 0.17 esp. 0.35). A g oup le el,
abundance-based dissimila i ies we e domina ed by abundance di e -
ences in An hophila, He e op e a, Fo micidae and O hop e a ( he las
wo g oups being a he species-poo ), while incidence-based dissimi-
la i ies we e mos ly d i en by species eplacemen . Nei he eplacemen
no ichness o abundance di e ences showed a consis en end wi h
g assland p oduc i i y.
3.3. Dis ibu ions along he land-use in ensi y g adien
Mos species had he cen e o hei dis ibu ion in low p oduc i i y
g asslands (Fig. 2), meaning ha hey a e declining unde land use
in ensi ica ion om low o medium in ensi y le els. These ‘losing’
Fig. 2. Species dis ibu ions along he land use in ensi y g adien exp essed in e ms o g assland p oduc i i y [GJ.ha
-1
.y
-1
]. The cen e poin o each species'
dis ibu ion was calcula ed as he p oduc i i y a which he abundance-weigh ed mean o all specimens occu ed (colo ed do s). Species a e o de ed on he y-axis
om high o low by inc easing cen e poin s, wi h single ons (i.e. one eco d o e all si es) indica ed as c osses. Non-single on species we e classi ied by cen e poin
in o low, middle and high p oduc i i y classes ( ep esen ing losing, in e media e and winning species, colo ed in blue, g een and yellow espec i ely; classes
ep esen 14, 13 and 14 si es espec i ely): The ba plo shows he ac ion o species cen e ed in each p oduc i i y class o e all species. G ey do s show all si es
whe e a species was eco ded (single eco ds in ligh g ey, highe numbe s in da k g ey). (Fo in e p e a ion o he e e ences o colou in his igu e legend, he
eade is e e ed o he web e sion o his a icle.)
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
5
species ep esen ed 56 % o he 380 species ha we e obse ed wi h wo
o mo e indi iduals, whe eas 35 % we e in e media e species (cen e ed
a medium p oduc i i y) and 9 % we e winning species (cen e ed a high
p oduc i i y). These pe cen ages con as ma kedly wi h he andom-
ized species dis ibu ions ha assumed no ela ion wi h p oduc i i y, in
which on a e age 62 % (95 %-Con idence In e al ±6) o all species
we e classi ied as in e media e species, while winning species ep e-
sen ed 22 % (±5) and losing species ep esen ed 16 % (CI ±5) o all
species (Fig. 3A-B). The obse ed pe cen age o losing species was 1.8
imes highe han his mean andomized pe cen age o Coleop e a (29
% o species), 3.6 imes highe o all species, and nea ly i e imes
highe o Hemip e a, Fo micidae and O hop e a (each >70 % o spe-
cies) (p <0.0001 o all g oups). On he o he hand, o all g oups excep
Coleop e a he empi ically obse ed pe cen age o winning species was
2-6 imes lowe han wha would ha e been expec ed based on an-
domized dis ibu ions (p <0.0001 o all o he g oups) (Fig. 3B).
Fu he mo e, mos losing species we e en i ely es ic ed o low
p oduc i i y le els o we e only inciden ally eco ded a highe p o-
duc i i y le els (Fig. 2: e lec ed by a lack o poin s in he uppe - igh
a ea o each subg oup). Fo 110 species (29 % o all species
excluding single ons) mo e han 90 % o he specimens we e obse ed a
low-p oduc i i y si es, and o 242 species (64 %) his was he case o
si es o low o medium p oduc i i y. 57 species (15 %) we e eco ded
p opo iona ely (i.e. a leas 33 % o specimens) a high-p oduc i i y
si es. High-p oduc i i y si es we e domina ed by species ha we e
eco ded widely ac oss he p oduc i i y g adien (Fig. 2: e lec ed by a
b oad sp ead o poin s in he lowe a ea o each subg oup) and sup-
po ed e y ew ange- es ic ed species: Only i e species we e
Fig. 3. The measu ed pe cen ages o losing, in e media e and winning species (i.e. species cen e ed a low, medium o high p oduc i i y le els espec i ely)
compa ed o andomized simula ions ha assume no consis en e ec o land-use in ensi y on species dis ibu ions. A: We andomized he dis ibu ions o species on
he p oduc i i y g adien by andomly ea anging he eco ded coun s o each species (n =380 excluding single ons) o e all 41 sampling si es, and e-calcula ing
hei cen e poin s (i.e. he abundance-weigh ed mean p oduc i i y o each species). O de ed by p oduc i i y, he andomized cen e poin s show an S-shaped cu e
as mos species a e classi ied as in e media e species (g een), while small pe cen ages a e classi ied as losing o winning species (blue esp. yellow). B: These
andomized pe cen ages we e de i ed 10.000 imes and plo ed as densi y cu es o losing species (le side o he g aph) and winning species ( igh side o he
g aph). Compa ed o hese andomized simula ions, he measu ed pe cen ages o losing and winning species di e ed signi ican ly o nea ly all axonomic g oups
(plo ed as ho izon al lines). C: The pe cen ages o losing, in e media e and winning species (blue, g een esp. yellow) changed among mo e widely occu ing species
(i.e. species eco ded a an inc easingly high minimum numbe o si es). Howe e , he changes in measu ed pe cen ages (solid lines) con as wi h he change o he
andomized dis ibu ions (100 i e a ions, shown wi h opaque do s and dashed lines). Randomized simula ions p ojec ha wi hou he in luence o land-use in ensi y,
he pe cen age o in e media e species would app oach 100 % among he mos widely dis ibu ed species while he pe cen ages o losing and winning species would
dec ease owa ds ze o. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he web e sion o his a icle.)
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
6
es ic ed o high p oduc i i y si es and hese all had ewe han i e
obse a ions. These pa e ns a e quan i a i ely illus a ed in Fig. 3C
which shows how he pe cen ages o losing, in e media e and winning
species change o e subse s o inc easingly widely occu ing species.
Randomized dis ibu ions indica e ha wi h an inc easing numbe o
occu ences, i can be expec ed ha he pe cen age o in e media e
species inc eases owa ds 100 % as losing and winning species a e
il e ed ou (dashed lines: do s e lec 100 andomiza ions). The
obse ed pe cen age o losing species indeed d ops sha ply when mo e
na ow- anged species a e excluded (losing species ep esen ed wo-
hi ds o he species ound a <5 si es). Howe e , coun e o expec a-
ions based on andomized dis ibu ions, he obse ed pe cen age o
winning species inc eases among mo e widely occu ing species (solid
yellow line c osses he dashed line) and is signi ican ly highe han
andomized pe cen ages o species occu ing a >10 si es (p =0.02).
20 % o he species occu ing a >20 si es (41 species, i.e. 11 % o he
o al) we e classi ied as winning species.
4. Discussion
We ound ha wi h inc easing land use in ensi y, a h opod species
ichness in g asslands declined bu o e all a h opod abundance
emained s able. The s able abundance o e all a h opods was mainly
d i en by an inc ease in Dip e a numbe s ha coun e ac ed abundance
declines in mos o he o de s. The dis ibu ions o indi idual species
e ealed la ge shi s in a h opod communi ies wi h inc easing land-use
in ensi y. A he landscape scale, a h opod biodi e si y was s ongly
cen e ed in low-p oduc i i y g asslands. High-p oduc i i y g asslands
we e domina ed by a small se o widely dis ibu ed species which pa ly
p o i om inc easing p oduc i i y. A ew winning species hus p obably
sus ained he o e all a h opod abundance a high-in ensi y g asslands
while many losing species caused biodi e si y le els o be e y low.
4.1. Con as ing esponses o species ichness and abundance
The exponen ial decline o a h opod species ichness wi h
inc easing p oduc i i y aligns wi h o he s udies on g assland plan s
(Kleijn e al., 2009) and a h opods (Allan e al., 2014). In line wi h
p e ious la ge-scale s udies on g assland a h opods (Allan e al., 2014;
Ek oos e al., 2020; Simons e al., 2015), we measu ed ichness declines
ac oss mul iple a h opod g oups, wi h he excep ion o Coleop e a.
Al hough hey a e o en used as indica o g oups, his lack o end o
Coleop e a does ma ch wi h o he s udies on Coleop e a ichness (Ba ´
a y
e al., 2007; Mei e al., 2024). The sampled Coleop e a amilies include
many gene alis eede s and appea less sensi i e o land use in ensi y
han o he a h opod g oups in e ms o species ichness. Ou esul s
sugges ha He e op e a ichness is mos sensi i e o land use in ensi y.
The eco ded ichness declines a e p obably mainly d i en by p ocesses
ha e ode niche di e si y, such as declines o lo al di e si y o mic o-
habi a a ia ion (Joe n and Laws, 2013; Woodcock and Pywell, 2010).
The selec ed sampling me hods p o ided a b oad and ecologically
di e se ep esen a ion o ege a ion-dwelling a h opods, bu did no
a ge soil- and su ace-dwelling a h opods. Hence, while ou esul s
hold wi hin he ealm o he species conside ed, his s udy can be
expanded by compa isons wi h di e en a ea-based collec ion me hods
(e.g. suc ion sampling) and by he inclusion o species iden i y da a o
ye mo e a h opod g oups. Such s udies can make impo an con i-
bu ions o ou unde s anding o poo ly s udied g oups ha ne e heless
play cen al oles in many ecosys ems.
The abundance declines wi h inc easing p oduc i i y ha we e
obse ed in mos o de s we e o se by he inc ease in abundance o a
single o de : ue lies (Dip e a). This posi i e ela ion be ween p o-
duc i i y and Dip e a abundance aligns wi h he indings o Sil a-
Mon ei o e al. (2022) who measu ed an inc ease o o al a h opod
abundance wi h p oduc i i y in a c oss-con inen al s udy o we g ass-
lands, whe e Dip e a ep esen ed 80 % o all specimens. The inc ease o
Dip e a numbe s wi h g assland p oduc i i y is p obably d i en by
amilies ha li e om decaying o ganic ma e (Powell e al., 2024).
Likewise, he s eep exponen ial decline o He e op e a numbe s mos
likely e lec s he eliance o many species on speci ic o bs ha a e o en
es ic ed o low-p oduc i i y g asslands, while he unimodal end o
Aucheno hyncha in his s udy was pa ly d i en by high numbe s
occu ing in humid g asslands a low- o-medium p oduc i i y le els.
These di e en ela ions o sepa a e a h opod o de s wi h land-use
in ensi y can la gely explain he con as ing ela ions o o e all
a h opod species ichness and abundance.
4.2. Shi s in species composi ion
Ou esul s con i m ha species ichness pa e ns unde es ima e
biodi e si y declines wi h land use in ensi ica ion because hey igno e
e ec s on species numbe s and communi y composi ion (Chase and
Knigh , 2013; Gossne e al., 2016; Hilleb and e al., 2018). Species
ichness in low-p oduc i i y g asslands was app oxima ely wice as
la ge as in high-p oduc i i y g asslands, bu compa ed o high-
p oduc i i y g asslands six imes mo e species had hei dis ibu ional
cen e -poin in low-p oduc i i y g asslands (Fig. 2). This implies ha on
op o signi ican ichness losses, also he numbe o unique species and
he popula ion size o many species declined wi h p oduc i i y. How-
e e , ou analyses o dissimila i y (be a-di e si y) and eplacemen in
species assemblages did no e eal hese pa e ns. This is p obably
because hese me ics agg ega e all di e ences be ween si es wi hou
conside ing whe e indi idual species we e obse ed. Al hough many
s udies ha e in e p e ed dissimila i y me ics as measu es o bio ic ho-
mogeniza ion o di e en ia ion (Do nelas e al., 2023; Gossne e al.,
2016; Keck e al., 2025), o he mechanisms can also play a ole in
de e mining hem. Fo example, we ound an inc easing dissimila i y o
An hophila assemblages om medium- o high p oduc i i y le els, mos
likely because species showed mo e agmen ed dis ibu ions ac oss
high-p oduc i i y g asslands (Fig. 2 & Table S4: see also A aújo e al.,
2022 and Wayman e al., 2024); a pa e n ha seems o e lec he
dependence o a common se o species on he g ass-clo e mix u es ha
a e ound in some o he high-p oduc i i y g asslands in his a ea.
Fu he mo e, s ochas ic e ec s such as spillo e o unsui able habi a s
( e lec ed by single eco ds a om species' p e ailing p oduc i i y
anges: see Fig. 2) can enhance dissimila i y, especially among species-
poo communi ies (Zhou e al., 2022). The lack o consis en pa e ns in
ou esul s sugges s ha he adi ionally used measu es o dissimila i y
do no comp ehensi ely cap u e he changes in species communi ies
esul ing om con as ing land-use (Wayman e al., 2024).
The cen e poin analyses p o ided mo e comp ehensi e esul s wi h
which he consequences o land-use change o species communi ies can
be e alua ed. Cen e poin s o species dis ibu ions on he p oduc i i y
g adien indica e whe he indi idual species a e declining o inc easing
in abundance, i.e. losing o winning, unde land use in ensi ica ion.
Some p e ious s udies did no eco d a clea dominance o ei he losing
o winning species in ime se ies, which hey ela ed o a ia ion in he
esponses o species o en i onmen al change (Do nelas e al., 2019;
Ou hwai e e al., 2020). We ound ha as land use in ensi ies, he e a e
many mo e a h opod species ha decline (‘lose’) han he e a e ha
inc ease (‘win’). This inding aligns closely wi h se e al s udies o single
a h opod g oups in managed g asslands, such as g asshoppe s (Chis ´
e
e al., 2016), Aucheno hyncha (Chis ´
e e al., 2018), an s (Heuss e al.,
2019) and mo hs (Mangels e al., 2017). Chis ´
e e al. (2016) and Heuss
e al. (2019) eco ded la gely analogous ela ions wi h land-use in-
ensi y o he O hop e a and Fo micidae species ha o e lap wi h ou
s udy, bu each o hese s udies also eco ded six addi ional losing spe-
cies, and in e media e dis ibu ions we e ound o six ou o he eigh
Fo micidae species ha we e losing in ou s udy. These di e ences
p obably e lec ha land-use in ensi y le els in ou a ea we e highe
han in he Ge man egions whe e hese s udies we e conduc ed, which
may ha e esul ed in mo e nega i e dis ibu ion pa e ns o he same
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
7
species, bu also in egional ex inc ion o he mos sensi i e species.
The e o e, cen e poin s udies om di e en egions should be
compa ed by he absolu e alues de i ed pe species and no by ela i e
sha es o winne s and lose s.
Di e en om he app oach used by Chis ´
e e al. (2016), we
compa ed he ac ual and andomized species dis ibu ions a he com-
muni y le el, a he han es ing whe he indi idual species inc eased o
declined wi h in ensi ica ion. This allowed us o inco po a e he many
species ha we e only obse ed a ew imes (n >2). Ac oss all sampled
g oups, we eco ded much mo e lose s and (wi h one excep ion) ewe
winne s han would be expec ed based on andom dis ibu ions, e en
when o ha g oup species ichness showed no ela ion wi h p oduc-
i i y. Mo eo e , ‘lose s’ (56 % o all species) ep esen only he species
ha declined in he ansi ion om low o medium in ensi y le els. In
ou esul s, ‘in e media e’ species include species ha declined as land
use changed om medium o high p oduc i i y le els, as well as species
ha we e eco ded p opo iona ely ac oss a wide p oduc i i y ange.
Only 15 % o all species occu ed p opo iona ely a high-in ensi y
g asslands, which implies ha he pe cen age o declining species
inc eased e en o 85 % unde he con e sion o high land-use in ensi y
le els.
4.3. F om na ow o b oad niches
Mos losing species we e ully o la gely es ic ed o a na ow ange
o low p oduc i i y le els, whe eas in each axonomic g oup, a subse o
in e media e and winning species occu ed widely along he p oduc-
i i y g adien (see Figs. 2 & 3C). This pa e n esembles he p e ailing
species abundance dis ibu ion o communi ies in na u e wi h many a e
species and a ew common species (McGill e al., 2007). Bu a he han
‘na u al a i y’, species dis ibu ions on he p oduc i i y g adien e lec
hei sensi i i y o ole ance o land-use in ensi ica ion. Species wi h
na ow niches on he p oduc i i y g adien we e o en nume ous in low-
p oduc i i y g asslands bu we e p og essi ely los a highe p oduc-
i i y le els, whe eas species ha appea ole an o land use in ensi i-
ca ion became inc easingly dominan (see also Simons e al., 2015). This
pa e n esul s in he o en-obse ed bio ic homogeniza ion in in en-
si ely used sys ems (McKinney and Lockwood, 1999; Newbold e al.,
2018). Na owly dis ibu ed species may be ha de o sample, bu no
including hem may lead o un ounded conclusions. Fo ins ance, Si-
mons and Weisse (2017) concluded ha u he in ensi ica ion o cu -
en Ge man g assland landscapes is possible wi hou biodi e si y loss in
a h opod communi ies, bu hei analysis was mainly based on he 25 %
o species ha we e mos widesp ead along g adien s o p oduc i i y.
Ob ained a a highly compa able p oduc i i y g adien , ou esul s
indica e ha hese species we e p obably he mos ole an o land-use
in ensi ica ion.
The winning species in ou s udy did no include any species ha
we e only abundan a high-p oduc i i y g asslands. Mos o hese spe-
cies ole a ed a wide ange o p oduc i i y le els. They ep esen ed only
9 % o all species bu 20 % o he mos widely occu ing species. Mos
winning species seemed o p o i om land-use in ensi ica ion because
o inc eased ood a ailabili y. Such species include mobile gene alis
eede s such as he honey bee Apis melli e a (Hymenop e a; eco ded a
33 si es, cen e ed a 54 GJ.ha
-1
), aphid eede s such as Ty haspis sede-
cimpunc a a (Coleop e a; eco ded a 31 si es, cen e ed a 51 GJ.ha
-1
),
and species eeding on oo s and o ganic ma e such as Ag io es gallicus
(Coleop e a; eco ded a 31 si es, cen e ed a 64 GJ.ha
-1
) and Geomyza
ipunc a a (Dip e a; eco ded a 21 si es, cen e ed a 70 GJ.ha
-1
). These
winning species we e no necessa ily gene alis s, bu ended o be well-
adap ed o he speci ic niches and high dis u bance le els ha cha ac-
e ize in ensi ely used g asslands (Boyes e al., 2019; McKinney and
Lockwood, 1999).
The p esence o a small se o well-adap ed species ha inc ease in
abundance wi h inc easing p oduc i i y le els could explain he
pe sis ence o o e all a h opod abundance a high land-use in ensi y
le els. Howe e , a h opod communi ies in he mo e in ensi ely used
g asslands we e inc easingly domina ed by amilies o Dip e a and Hy-
menop e a ha we e no iden i ied o species le el. These amilies
mainly include small and poo ly known species whose la ae li e as
decompose s (Dip e a) o pa asi oids (bo h Dip e a and Hymenop e a).
As ha dly any esea ch exis s on pa e ns in species composi ion o hese
amilies, i emains unknown i hei nume ic dominance a in ensi ely
used g asslands esul s om a ew winning species o also om
inc eased species di e si y. While pa asi oids would p obably ollow he
pa e ns o hei hos s, decompose di e si y may indeed pe sis unde
in ensi ied land-use (Le P o os e al., 2021). Howe e , he Dip e an
axa wi h decompose la ae ha we e iden i ied o species le el (26
species belonging o 7 amilies) showed simila di e si y losses as o he
a h opod g oups: mos species we e la gely con ined o lowe p oduc-
i i y le els whe eas i e species we e winning as land-use in ensi ied
(Fig. S3). This aligns wi h o he s udies ha ound high u no e and
s ong nume ic dominance o single Dip e a species in in ensi ely used
sys ems (Powell e al., 2024; Sho all e al., 2009). I is he e o e likely
ha also among amilies ha emained uniden i ied in his s udy, a
small se o species main ained high o e all abundance le els unde
in ensi ied g assland managemen .
4.4. Implica ions
Ou indings sugges ha when species om di e en axonomic o
ecological guilds a e pooled oge he , abundance esponses can di e
ma kedly om di e si y esponses. We showed ha land-use in ensi i-
ca ion on g asslands is ela ed o consis en losses in di e si y bu no
abundance. This means ha local g assland in ensi ica ion canno
explain he se e e losses o o e all a h opod abundance ha we e
eco ded o e decadal imescales (Seibold e al., 2019; Van Klink e al.,
2020). I is e en possible ha ex ensi ied managemen esul s in
declining a h opod abundance in ce ain con ex s, such as when
Dip e a a e dominan (Sil a-Mon ei o e al., 2022). The absence o clea
ela ions be ween land-use in ensi y and o e all a h opod abundance a
si e le el sugges s ha landscape-scale e ec s, such as loss o semi-
na u al habi a (Seibold e al., 2019), changes in c opping pa e ns
(Li o ska e al., 2025), inc eased equency o clima ic ex emes (Boggs,
2016; Sohls ¨
om e al., 2022), o inc eased loading o ha m ul chemicals
such as insec icides (Mancini e al., 2020), migh be impo an d i e s o
he eco ded a h opod declines o e ime. Mo eo e , declines o
a h opod biomass (Hallmann e al., 2017) may be ampli ied by losses o
la ge-bodied species and indi iduals in a h opod communi ies
(Ma ínez-Nú˜
nez e al., 2024; Oli ei a e al., 2016).
We also show ha species ichness and dissimila i y me ics unde -
es ima e he ac ual losses o a h opod biodi e si y because species
iden i ies a e no conside ed. The mos p o ound consequence o land
use in ensi ica ion is he p og essi e loss o species ac oss all a h opod
g oups ha a e sensi i e o land-use in ensi ica ion, mos o which ha e
a dis ibu ion ha is con ined o a na ow ange o low p oduc i i y
le els. A h opod abundance seems mo e esilien o in ensi ica ion
because o a small subse o widesp ead winne s ha compensa e o
losses in he majo i y o species (Newbold e al., 2018). Fu he esea ch
is needed o be e unde s and abundance changes ac oss a h opods
wi h di e ing ecologies, and he consequences o hese changes o
ecosys em unc ioning (Cooke e al., 2025; Wagne , 2020). Winning
species can p obably sus ain ce ain ecosys em p ocesses such as ood
p o ision o gene alis insec i o es, bu may no suppo mo e
specialized ophic in e ac ions (Holland e al., 2006). Fu he mo e,
winning species may also ampli y dis-se ices such as c op damage as
hey o en include impo an pes species (Zhang e al., 2007). While
se e e species losses a e expec ed o e ode ecosys em se ice p o i-
sioning, we can only assess which ecosys em se ices a e a s ake when
knowing which a h opods lose o p o i unde en i onmen al change
(Ca dinale e al., 2012; Cooke e al., 2025).
A h opod species losses we e mos p onounced in g asslands a he
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
8
low- o-medium p oduc i i y ange. This inding implies ha biodi e -
si y conse a ion will be mos e ec i e a lowe land-use in ensi y
le els, which aligns wi h o he s udies (Allan e al., 2014; Kleijn e al.,
2009, 2011; Kleijn and Su he land, 2003). Low-in ensi y g asslands play
c i ical oles in p ese ing many a h opod species (Habel e al., 2013;
Van Swaay e al., 2025; WallisDeV ies e al., 2002) and we show he e
ha losing such g asslands a ec s he la ge majo i y o species ac oss
mul iple axa. Howe e , low-in ensi y g asslands ha e become a e o
absen in mode n ag icul u al landscapes whe e p oduc i i y le els
abo e 50 GJ.ha
-1
.y
-1
ha e become he no m (Ba dge e al., 2021;
Emme son e al., 2016). G ass-domina ed hay meadows ha a e com-
mon in ag i-en i onmen schemes and na u e ese es, bo h aimed a
biodi e si y es o a ion, ypically s ill each medium p oduc i i ies o
20-40 GJ.ha
-1
.y
-1
(Schippe s e al., 2012). Al hough mo e species
ole a e hese p oduc i i y le els han hose o high-in ensi y g asslands,
hei alue o biodi e si y conse a ion is incompa able low-
p oduc i e, o b-domina ed g asslands (Habel e al., 2013, 2019).
Res o a ion o mo e low-in ensi y g assland habi a s will he e o e be
c ucial o hal and e e se he widesp ead biodi e si y loss o g assland
a h opods.
CRediT au ho ship con ibu ion s a emen
J.P. Reinie de V ies: W i ing – o iginal d a , Visualiza ion,
In es iga ion, Fo mal analysis, Da a cu a ion, Concep ualiza ion. Jos´
e
G. an Paassen: W i ing – e iew & edi ing, Supe ision, Concep ual-
iza ion. Remco Ploeg: W i ing – e iew & edi ing, In es iga ion, Da a
cu a ion. Aa Noo dam: W i ing – e iew & edi ing, Da a cu a ion.
Koen Ve hoog : W i ing – e iew & edi ing, Da a cu a ion. Da id
Kleijn: W i ing – e iew & edi ing, Supe ision, Funding acquisi ion,
Concep ualiza ion.
Decla a ion o compe ing in e es
The au ho s decla e no con lic s o in e es .
Acknowledgmen s
We hank 26 a me s and h ee na u e o ganiza ions o hei coop-
e a ion in conduc ing his esea ch on hei lands and hei willingness
o sha e in o ma ion on yields. We u he mo e hank Femke an
Bee sum, Abe Kienhuis, Da id Jacobs, Nienke de Jong and Xam Menue
o hei assis ance in da a collec ion, and I o Raemake s o his assis-
ance in wild bee iden i ica ion. This wo k was suppo ed by he Eu o-
pean Union's Ho izon 2020 esea ch and inno a ion p og amme unde
g an ag eemen no. 862480, SHOWCASE p ojec (h ps://showcase-p
ojec .eu). This a icle e lec s only he au ho s’ iews and he Eu o-
pean Commission is no esponsible o any use ha may be made o he
in o ma ion his a icle con ains.
Appendix A. Supplemen a y da a
Supplemen a y da a o his a icle can be ound online a h ps://doi.
o g/10.1016/j.biocon.2025.111522.
Da a a ailabili y
All aw da a and R code wi h documen a ion a e accessible ia he
Zenodo digi al eposi o y. DOI: h ps://doi.
o g/10.5281/zenodo.15585342 (De V ies e al., 2025).
Re e ences
Allan, E., Bossdo , O., Do mann, C.F., P a i, D., Gossne , M.M., Tscha n ke, T., Nico
Blü hgen, N., Bellach, M., Bi kho e , K., Boch, S., B¨
ohm, S., B¨
o schig, C.,
Cha zino as, A., Ch is , S., Daniel, R., Diek¨
o e , T., Fische , M., 2014. In e annual
a ia ion in land-use in ensi y enhances g assland mul idi e si y. P oc. Na l. Acad.
Sci. U. S. A. 111 (1), 308–313. h ps://doi.o g/10.1073/pnas.1312213111.
A aújo, F. de C., Mendes, C.N., de Souza, C.R., Mi anda, K.F., Mo elli, M.C.M., Coelho, P.
A., Figuei edo, M.A.P., Ca alho, W.A.C., de Cas o, G.C., Fon es, M.A.L., Dos
San os, R.M., 2022. F agmen a ion e ec s on be a di e si y o agmen ed and
conse ed landscapes: insigh s abou homogeniza ion and di e en ia ion p ocesses.
Ac a Bo . B as. 36, e2020abb0261. h ps://doi.o g/10.1590/0102-
33062020abb0261.
Ba dge , R.D., Bullock, J.M., La o el, S., Manning, P., Scha ne , U., Os le, N.,
Chomel, M., Du igan, G., F y, E.L., Johnson, D., La allee, J.M., Le P o os , G.,
Luo, S., Png, K., Sanka an, M., Hou, X., Zhou, H., Ma, L., Ren, W., Shi, H., 2021.
Comba ing global g assland deg ada ion. Na . Re . Ea h En i on. 2 (10), 720–735.
h ps://doi.o g/10.1038/s43017-021-00207-2.
Ba o´
n, K., 2023. MuMIn: mul imodal in e ence. R package e sion 1.47.5. h ps://c an.
-p ojec .o g/package=MuMIn.
Ba ´
a y, P., B´
aldi, A., Sz´
el, G., Podluss´
any, A., Rozne , I., E d¨
os, S., 2007. Responses o
g assland specialis and gene alis bee les o managemen and landscape complexi y.
Di e s. Dis ib. 13 (2), 196–202. h ps://doi.o g/10.1111/j.1472-4642.2006.00309.
x.
Beckmann, M., Ge s ne , K., Akin-Fajiye, M., Ceaușu, S., Kambach, S., Kinlock, N.L.,
Phillips, H.R.P., Ve hagen, W., Gu e i ch, J., Klo z, S., Newbold, T., Ve bu g, P.H.,
Win e , M., Seppel , R., 2019. Con en ional land-use in ensi ica ion educes species
ichness and inc eases p oduc ion: a global me a-analysis. Glob. Chang. Biol. 25 (6),
1941–1956. h ps://doi.o g/10.1111/gcb.14606.
Ben on, T.G., B yan , D.M., Cole, L., C ick, H.Q.P., 2002. Linking ag icul u al p ac ice o
insec and bi d popula ions: a his o ical s udy o e h ee decades. J. Appl. Ecol. 39
(4), 673–687. h ps://doi.o g/10.1046/j.1365-2664.2002.00745.x.
Blü hgen, N., S aab, M., Achu y, R., Weisse , W.W., 2022. Un a elling insec declines:
can space eplace ime? Biol. Le . 18, 20210666. h ps://doi.o g/10.1098/
sbl.2021.0666.
Boggs, C.L., 2016. The inge p in s o global clima e change on insec popula ions. Cu .
Opin. Insec Sci. 17, 69–73. h ps://doi.o g/10.1016/j.cois.2016.07.004.
Boyes, D.H., Fox, R., Sho all, C.R., Whi ake , R.J., 2019. Bucking he end: he
di e si y o An h opocene “winne s” among B i ish mo hs. F on . Biogeog . 11 (3),
e43862. h ps://doi.o g/10.21425/F5FBG43862.
B ooks, M.E., K is ensen, K., an Ben hem, K.J., Magnusson, A., Be g, C.W., Nielsen, A.,
Skaug, H.J., M¨
achle , M., Bolke , B.M., 2017. glmmTMB balances speed and
lexibili y among packages o ze o-in la ed gene alized linea mixed modeling. R J.
9 (2), 378–400. h ps://doi.o g/10.32614/ j-2017-066.
Ca dinale, B.J., Du y, J.E., Gonzalez, A., Hoope , D.U., Pe ings, C., Venail, P.,
Na wani, A., MacE, G.M., Tilman, D., Wa dle, D.A., Kinzig, A.P., Daily, G.C.,
Lo eau, M., G ace, J.B., La igaude ie, A., S i as a a, D.S., Naeem, S., 2012.
Biodi e si y loss and i s impac on humani y. Na u e 486 (7401), 59–67. h ps://doi.
o g/10.1038/na u e11148.
Ca dinale, B.J., Gonzalez, A., Alling on, G.R.H., Lo eau, M., 2018. Is local biodi e si y
declining o no ? A summa y o he deba e o e analysis o species ichness ime
ends. Biol. Conse . 219, 175–183. h ps://doi.o g/10.1016/j.biocon.2017.12.021.
Ca doso, P., Rigal, F., Ca alho, C., 2015. BAT – biodi e si y assessmen ools, an R
package o he measu emen and es ima ion o alpha and be a axon, phylogene ic
and unc ional di e si y. Me hods Ecol. E ol. 6, 232–236. h ps://doi.o g/10.1111/
2041-210X.12310.
Ca doso, P., Ba on, P.S., Bi kho e , K., Chicho o, F., Deacon, C., Fa mann, T.,
Fukushima, C.S., Gaighe , R., Habel, J.C., Hallmann, C.A., Hill, M.J., Hochki ch, A.,
Kwak, M.L., Mammola, S., No iega, J.A., O inge , A.B., Ped aza, F., P yke, J.S.,
Roque, F.O., Samways, M.J., 2020. Scien is s’ wa ning o humani y on insec
ex inc ions. Biol. Conse . 242, 108426. h ps://doi.o g/10.1016/J.
BIOCON.2020.108426.
Chase, J.M., Knigh , T.M., 2013. Scale-dependen e ec sizes o ecological d i e s on
biodi e si y: why s anda dised sampling is no enough. Ecol. Le . 16, 17–26.
h ps://doi.o g/10.1111/ele.12112.
Chase, J.M., McGill, B.J., Thompson, P.L., An ˜
ao, L.H., Ba es, A.E., Blowes, S.A.,
Do nelas, M., Gonzalez, A., Magu an, A.E., Supp, S.R., Win e , M., Bjo kman, A.D.,
B uelheide, H., By nes, J.E.K., Cab al, J.S., Elahi, R., Gomez, C., Guzman, H.M.,
Isbell, F., O’Conno , M., 2019. Species ichness change ac oss spa ial scales. Oikos
128 (8), 1079–1091. h ps://doi.o g/10.1111/oik.05968.
Chis ´
e, M.N., Mody, K., Gossne , M.M., Simons, N.K., K¨
ohle , G., Weisse , W.W.,
Blü hgen, N., 2016. Lose s, winne s, and oppo unis s: how g assland land-use
in ensi y a ec s o hop e an communi ies. Ecosphe e 7 (11), e01545. h ps://doi.
o g/10.1002/ecs2.1545.
Chis ´
e, M.N., Mody, K., Kunz, G., Gunczy, J., Blü hgen, N., 2018. In ensi e land use
d i es small-scale homogeniza ion o plan - and lea hoppe communi ies and
p omo es gene alis s. Oecologia 186 (2), 529–540. h ps://doi.o g/10.1007/s00442-
017-4031-0.
Cooke, R., Ou hwai e, C.L., Bladon, A.J., Milla d, J., Rodge , J.G., Dong, Z., Dye , E.E.,
Edney, S., Mu phy, J.F., Dicks, L.V., Hui, C., Jones, J.I., Newbold, T., Pu is, A.,
Roy, H.E., Woodcock, B.A., Isaac, N.J.B., 2025. In eg a ing mul iple e idence
s eams o unde s and insec biodi e si y change. Science 388 (45), eadq2110.
h ps://doi.o g/10.1126/science.adq2110.
CVB, 2022. CVB ee oede abel 2022: Chemische samens ellingen en nu i ionele
waa den an oede middelen. In: Wageningen Li es ock Resea ch, Wageningen, he
Ne he lands & ILVO. Belgium, Me elbeke (in Du ch).
De V ies, J.P.R., an Paassen, J.G., Ploeg, R., Noo dam, A., Ve hoog , K., Kleijn, D., 2025.
Da a and codes om: land use in ensi ica ion ans o ms g assland a h opod
communi ies: a ew winne s expand bu mos species disappea . Zenodo, da ase .
h ps://doi.o g/10.5281/zenodo.15585342.
J.P.R. de V ies e al.
Biological Conse a ion 313 (2026) 111522
9
