This is a sel -a chi ed e sion o an o iginal a icle. This e sion
may di e om he o iginal in pagina ion and ypog aphic de ails.
Au ho (s):
Ti le:
Yea :
Ve sion:
Copy igh :
Righ s:
Righ s u l:
Please ci e he o iginal e sion:
CC BY
h ps://c ea i ecommons.o g/licenses/by/4.0/
Mo emen o o es ‐dependen dung bee les h ough ipa ian bu e s in Bo nean oil
palm plan a ions
© 2021 he Au ho s
Accep ed e sion (Final d a )
G ay, Ross E., J.; Rod iguez, Luisa, F.; Lewis, Owen, T.; Chung, A hu Y., C.;
O askainen, O so; Slade, Eleano , M.
G ay, R. E., Rod iguez, L., Lewis, O., Chung, A. Y., O askainen, O., & Slade, E. (2022). Mo emen
o o es ‐dependen dung bee les h ough ipa ian bu e s in Bo nean oil palm plan a ions.
Jou nal o Applied Ecology, 59(1), 238-250. h ps://doi.o g/10.1111/1365-2664.14049
2022
238
|
J Appl Ecol. 2022;59:238–250.wileyonlinelib a y.com/jou nal/jpe
Recei ed: 14 Ap il 2021
|
Accep ed: 20 Sep embe 2021
DOI: 10.1111/1365-2664.14049
RESEARCH ARTICLE
Mo emen o o es - dependen dung bee les h ough ipa ian
bu e s in Bo nean oil palm plan a ions
Ross E. J. G ay1,2  | Luisa F. Rod iguez3 | Owen T. Lewis1 | A hu Y. C. Chung4 |
O so O askainen3,5,6  | Eleano M. Slade1,7
1Depa men o Zoology, Uni e si y o Ox o d, Ox o d, UK; 2Depa men o Li e Sciences, Impe ial College London, Asco , UK; 3O ganismal and E olu iona y
Biology Resea ch P og amme, Uni e si y o Helsinki, Helsinki, Finland; 4Fo es Resea ch Cen e, Fo es y Depa men , Sandakan, Malaysia; 5Depa men
o Biological and En i onmen al Science, Uni e si y o Jy äskylä, Jy äskylä, Finland; 6Depa men o Biology, Cen e o Biodi e si y Dynamics, No wegian
Uni e si y o Science and Technology, T ondheim, No way and 7Asian School o he En i onmen , Nanyang Technological Uni e si y, Singapo e Ci y, Singapo e
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium,
p o ided he o iginal wo k is p ope ly ci ed.
© 2021 The Au ho s. Jou nal o Applied Ecology published by John Wiley & Sons L d on behal o B i ish Ecological Socie y.
Ross E. J. G ay and Luisa F. Rod iguez sha es i s au ho ship.
O so O askainen and Eleano M. Slade sha es senio au ho ship.
Co espondence
Ross E. J. G ay
Email: oss.g ay1[email p o ec ed]c.uk
Funding in o ma ion
MOE AcRF Tie 1, G an /Awa d Numbe :
RG119/19; Jane ja Aa os E kon Sää iö,
G an /Awa d Numbe : 223257; Na u al
En i onmen Resea ch Council, G an /
Awa d Numbe : NE/K016261/1; H2020
Eu opean Resea ch Council, G an /Awa d
Numbe : 856506; Academy o Finland,
G an /Awa d Numbe : 309581
Handling Edi o : Filipe F ança
Abs ac
1. F agmen a ion o opical o es s is inc easing globally, wi h nega i e impac s o
biodi e si y. In Sou heas Asia, expansion o oil palm ag icul u e has caused wide-
sp ead de o es a ion, o es deg ada ion and agmen a ion.
2. Pe sis ence o o es - dependen species wi hin hese agmen ed landscapes is
likely o depend on he capaci y o indi iduals o mo e be ween o es pa ches. In
oil palm landscapes, ipa ian bu e s along s eams and i e s a e po en ial mo e-
men co ido s, bu hei use by mo ing animals is poo ly s udied.
3. We examined how six dung bee le species a e sed ipa ian bu e s connec ed o
a con inuous o es ese e a ea wi hin an oil palm plan a ion in Sabah, Malaysian
Bo neo. We used a ma k– elease– ecap u e s udy and a new Bayesian Join
Species Mo emen Modelling (JSMM) app oach, ex ended o a con inuous cap-
u e p ocess model.
4. Dung bee le species we e ai ly gene alis in hei habi a use, bu wo species
showed a s a is ically suppo ed p e e ence o ipa ian bu e o es o e oil
palm, and one species showed a s ong p e e ence o o es ese e o e ipa ian
bu e , indica ing he impo ance o o es ed a eas wi hin oil palm landscapes o
some species.
5. A land- use change simula ion indica ed ha he loss o ipa ian bu e s in oil palm
will esul in educed mo emen by o es - dependen species.
6. Syn hesis and applica ions. Ou esul s p o ide e idence o he use o ipa ian
bu e s in oil palm plan a ions o o es - dependen dung bee le species, s eng h-
ening he case o hei e en ion, es o a ion and e- es ablishmen . Fu he mo e,
|
239
Jou nal o Applied Ecology
GRAY e Al.
1 | INTRODUCTION
Th ea s o opical o es biodi e si y a e unp eceden ed due o he
unpa alleled a es o o es deg ada ion, agmen a ion and con e -
sion o ag icul u e (Ba low e al., 2016; Hansen e al., 2013; Newbold
e al., 2014; Phalan e al., 2013). Sou heas Asia has some o he
highes a es o o es loss, wi h only ~19% o i s in ac o es s e-
maining (Acha d e al., 2014; Es oque e al., 2019; Lewis e al., 2015;
Sodhi e al., 2010). The apid expansion o oil palm (Elaeis guineensis
Jacq.) has made he egion he wo ld's p ima y sou ce o ege able
oil (Fi zhe be e al., 2008; Ga eau e al., 2014; Tu ne e al., 2008).
Al hough biodi e si y loss in oil palm plan a ions is well documen ed
(Meijaa d e al., 2020; Meijaa d e al., 2018), oil palm is now an in eg al
pa o he economies o Sou heas Asian coun ies accoun ing o
3.82% o he g oss domes ic p oduc in Malaysia alone (~$896 mil-
lion pe yea ; Mahidin, 2018). The e o e, i is impo an o unde s and
how bes o manage and design he inc easingly common mosaic
landscapes ha inco po a e bo h na u al o es and oil palm ag icul-
u e, in o de o suppo he emaining biodi e si y o he egion.
Remaining o es pa ches wi hin mosaic oil palm landscapes a e
o en ound in he o m o s ips o na u al o es o ed ege a ion by
he side o wa e ways known as ipa ian bu e s (also called ipa ian
ese es o ipa ian s ips; Luke e al., 2019). These i e ine o es
a eas a e p ima ily se aside o educe un- o in o s eams (Sweeney
e al., 2004) bu can also imp o e wa e quali y (Maye e al., 2007)
and bene i aqua ic and o es - dependen e es ial auna (G ay
e al., 2014, 2019b; Ma czak e al., 2010; Ricke s, 2004). In addi-
ion, ipa ian bu e s ha e he po en ial o se e as mo emen co -
ido s be ween o es agmen s and con inuous o es (Beie &
Noss, 1998; Tewksbu y e al., 2002). The impo ance o ipa ian
bu e s in oil palm landscapes has ecen ly esul ed in hei addi ion
as a equi emen o Round able on Sus ainable Palm Oil (RSPO) ce -
i ica ion (Ba clay e al., 2017; Lucey e al., 2017). The equi emen
s ipula es a minimum o es bu e o 5 o >200 m on each side o
he i e , wi h he minimum bu e wid h depending on i e wid h,
bu e placemen and pe cei ed use (Ba clay e al., 2017; Lucey
e al., 2017). Howe e , he legal minimum wid h a ies ac oss coun-
ies; in Sabah, Malaysia i is 20 m on each side o i e s >3 m (Sabah
Wa e Resou ces Enac men , 1998), bu can be inc eased whe e he
bu e s a e hough o ep esen impo an co ido s o wildli e
(En i onmen P o ec ion Enac men , 2002).
As o es pa ches wi hin oil palm landscapes become inc eas-
ingly agmen ed and isola ed, he pe sis ence o species wi hin o -
es pa ches may become c i ically dependen on he connec i i y
be ween agmen s (Ewe s & Didham, 2006; Hanski, 1999; Lucey
& Hill, 2012). The e o e, unde s anding how animal species mo e
h ough he landscape sepa a ing agmen s has become a key con-
side a ion in conse a ion and managemen s a egies o human-
modi ied landscapes (Dohe y e al., 2021; G ay e al., 2019b).
Mo emen abili y wi hin a agmen ed landscape can be in luenced
by a species’ beha iou al esponses o habi a bounda ies (Jain
e al., 2020; Kallioniemi e al., 2014; Lucey & Hill, 2012), he physical
cos s o mo emen (Bon e e al., 2012) and he pe meabili y o he
ma ix (Ewe s & Didham, 2006; Sc i en e al., 2017). Fu he mo e,
species- speci ic li e- his o y ai s can impac mo emen (O askainen
e al., 2019). Species mos ulne able o agmen a ion a e o es -
dependen axa (i.e. hose ha need o es o suppo iable pop-
ula ions). These species ypically ha e es ic ed anges and a e
eluc an o c oss o es bounda ies, esul ing in small, isola ed
popula ions which can su e local ex inc ions wi h li le p ospec o
ecolonisa ion (Sc i en e al., 2015; Sodhi e al., 2010). Howe e , de-
spi e he impo ance o connec i i y o agmen s o conse a ion
planning, ela i ely ew s udies ha e in es iga ed he mo emen be-
ha iou o opical o es - associa ed axa (Boucha d & B ooks, 2004;
B ouwe s & New on, 2009; G ay e al., 2019b; Khazan, 2014; Lucey
& Hill, 2012; Sc i en e al., 2017).
He e, we examine he mo emen beha iou o o es - associa ed
dung bee les (Sca abaeidae: Sca abaeinae) in ipa ian bu e s wi hin
oil palm landscapes in Sabah, Malaysian Bo neo. T opical dung bee-
les a e good indica o s o habi a dis u bance (e.g. Da is e al., 2001;
Ga dne e al., 2008; Nichols & Ga dne , 2011), and he e ha e been
se e al s udies on hei mo emen in opical ag icul u al landscapes
(A ellano e al., 2008; Cul id- Medina e al., 2015; G ay e al., 2016;
da Sil a & He nández, 2015). Fo es - associa ed species in sou he n
Mexico we e ound o mo e h ough highly agmen ed deciduous
o es landscapes, p o ided he e we e co ido s o connec he
agmen s (A ellano e al., 2008); while dung bee les in Andean ag-
icul u al landscapes showed in e speci ic di e ences in mo emen
pa e ns and mo emen dis ances, associa ed wi h wing loading and
habi a p e e ences (Cul id- Medina e al., 2015). Wi hin oil palm
domina ed landscapes, he e has been only one s udy which has
documen ed limi ed ‘spillo e ’ o dung bee les ac oss ipa ian bu e
ou s udy demons a es he wide applicabili y o he Join Species Mo emen
Modelling (JSMM) amewo k o assess mo emen beha iou o species in ag-
men ed landscapes, a i al ool o u u e o es and landscape managemen and
conse a ion p io i isa ion exe cises.
KEYWORDS
Bayesian Join Species Mo emen Modelling, dispe sal, insec s, Malaysia, Ma k– elease–
ecap u e, mo emen co ido , ipa ian ese es, opical o es
240
|
Jou nal o Applied Ecology
GRAY e Al.
bounda ies in o oil palm, bu did no documen o measu e mo e-
men di ec ly (G ay e al., 2016).
We used ma k– elease– ecap u e (MRR) me hods, a common
echnique o he s udy o insec mo emen (G ay e al., 2019b;
Hanski, 1999; Lewis e al., 1997; Slade e al., 2013) and he newly
de eloped Join Species Mo emen Modelling app oach (JSMM—
O askainen e al., 2019). JSMM allows bo h species- and communi y-
le el mo emen pa ame e s o be es ima ed simul aneously. He e,
we ex end he amewo k o accoun o a con inuous cap u e p o-
cess and o model di e en land- use change scena ios. We used
hese me hods o ask how ipa ian bu e s in luence he mo emen
o dung bee les wi hin oil palm landscapes. In pa icula , we es
he ollowing hypo heses: (a) Dung bee les a e mo e likely o p e e
mo ing in ipa ian bu e s han in oil palm plan a ions, and o p e e
mo ing in con inuous o es ese e o ipa ian o es bu e s. (b)
Mo emen abili y and he a es a which dung bee les c oss habi a
bounda ies will di e among species. (c) Con e sion o o es o oil
palm will limi he abili y o dung bee les o mo e wi hin agmen ed
landscapes, educing he numbe o indi iduals cap u ed in oil palm.
2 | MATERIALS AND METHODS
2.1 | S udy si es
Ou h ee si es we e si ua ed wi hin he S abili y o Al e ed Fo es
Ecosys ems (SAFE) P ojec landscape in sou h- eas e n Sabah,
Malaysia (4.72°N, 117.60°E; Ewe s e al., 2011; Figu e 1a). A each
si e he e was a o es ed ipa ian bu e embedded wi hin an oil
palm ma ix, which was connec ed o a la ge a ea (2,200 ha) o con-
inuous o es ese e consis ing o lowland dip e oca p ain o es .
The o es ese e is pa o a >1 million ha a ea o p o ec ed o es
(Ewe s e al., 2011). The palms wi hin each oil palm ma ix su ound-
ing he ipa ian bu e s we e app oxima ely he same age (~8 yea s).
Selec ion o he h ee ocal si es (RR03, RR10 and RR18) was made
o ensu e an app oxima e s anda d con igu a ion and s uc u e o
he landscape. Ripa ian bu e o es was composed o emnan old-
g ow h and seconda y o es . The o es wi hin each si e had all
ees (some >40 m), high canopy co e and simila mean ipa ian o -
es wid hs (48, 58 and 41 m in RR03, RR10 and RR18 espec i ely;
G ay e al., 2019b; Williamson e al., 2020). Mino a ia ions we e in-
e i ably ound among si es in he p ecise con igu a ion o landscape
elemen s (Figu e 1b– d). Fieldwo k ook place be ween No embe
2016 and Ap il 2017.
2.2 | Species selec ion
Six dung bee le species we e selec ed based on a p e ious s udy
(G ay e al., 2016) o span a ange o body sizes (~1– 5 cm), include
species ep esen ing he wo main dung bu ial modes ( unnelle s and
olle s), and bo h diu nal and noc u nal species (Table 1). All six spe-
cies we e chosen as hey occu commonly in o es habi a and ha e
been obse ed in oil palm plan a ions a much lowe abundances
(G ay e al., 2016). In addi ion, he six species had high abundance
in p e ious s udies indica ing hey we e e ec i e candida es o a
MRR s udy. As only he males o Ca ha sius enaudpauliani (Ochi &
Kon, 1996) and Ca ha sius dayacus (Lansbe ge, 1886) could be dis-
inguished eliably o species le el in he ield, hese wo species
we e pooled o he analysis. Howe e , Ca ha sius enaudpauliani is
he mo e abundan Ca ha sius species in ipa ian bu e s in his s udy
a ea (G ay e al., 2016), and we assume ha he majo i y o indi idu-
als we e o his species.
2.3 | Dung bee le mo emen
A each si e 17– 18 li e- cap u e bai ed pi all aps we e se , spaced a
minimum o 50 m apa , ollowing s anda d me hods o dung bee le
sampling (Figu e 1b– d; G ay e al., 2016; La sen & Fo sy h, 2005;
da Sil a & He nández, 2015). Six aps we e placed in he oil palm
ma ix, se en o eigh in he ipa ian bu e and ou in he o es e-
se e. One ap was emo ed om RR10 as i could no be success-
ully es ablished. Each ap consis ed o a 1.5 L plas ic bo le wi h
i s op emo ed and in e ed o o m a unnel (~92 mm diame e ).
T aps we e bai ed wi h 25 g o human aeces, w apped in muslin
and suspended 5 cm abo e he unnel (Pa e e al., 2019). Small
holes we e made in he base o he ap o allow ainwa e o d ain,
and a hand ul o lea es was placed in he bo om o p o ide shel e
o apped bee les. A S y o oam pla e p o ec ed he ap om ain
(Pa e e al., 2019).
T aps we e checked and e- bai ed e e y second day o a pe-
iod o 14 days and all indi iduals o he ocal dung bee le spe-
cies ma ked. Two pen ypes we e used o ma k bee les (Mi subishi
Uni Pain Ma ke PX- 21 Fine Bulle Tip in O ange/G een/Pink,
and A line 999XF Sil e Me allic Bulle Tip Ma ke 0.8 mm).
Pilo s udies showed ha ma ks on he bee les las ed o a leas
14 days, and p e ious s udies ha e shown no e ec on bee le lon-
ge i y (Ba es e al., 2006). Each indi idual o he ocal species was
gi en a unique code using a se ies o do s on he ely a (A ellano
e al., 2008; La sen & Fo sy h, 2005; No iega & Acos a, 2011).
A e ma king, he aps we e closed and bai emo ed o allow he
bee les 24 h o dispe se be o e he bai was eplaced he ollow-
ing day. Bo h newly ma ked and ecap u ed bee les we e eco ded
and eleased in o he ege a ion a he poin o cap u e. Simila
MRR me hods ha e been used p e iously o assess dung bee le
mo emen (A ellano e al., 2008; Cul id- Medina e al., 2015; da
Sil a & He nández, 2015).
2.4 | Da a and s a is ical analysis
All s a is ical analysis was conduc ed in (Ve sion 4.0.3 - R
De elopmen Co e Team, 2021). We applied he JSMM amewo k
o O askainen e al. (2019) o analyse he cap u e– ecap u e da a-
se s o e he h ee ocal si es (Figu e 1b– d). As he p ocess model
|
241
Jou nal o Applied Ecology
GRAY e Al.
we assumed he mo emen model o O askainen (2004), ha is,
spa ially explici di usion supplemen ed wi h mo ali y and habi-
a selec ion a edges be ween habi a ypes. Fo each species s
(Table 1), we deno ed he di usion coe icien by
Ds
(uni m2/day;
measu ing he mo emen a e) and he mo ali y a e by
ms
(uni /
day). As he da a we e no su icien o es ima e hese pa ame e s
sepa a ely o each habi a ype, we assumed hem o be cons an
o e he en i e s udy a ea. We se he habi a selec ion pa ame e
o ipa ian bu e s as
kRB
s
=
1,
and hus measu ed habi a selec ion
o he oil palm
(
k
OP
s)
and o es ese e
(
k
FR
s)
habi a s ela i e o
he ipa ian bu e s. Fo he model, we se he i e as pa o he
ipa ian bu e habi a .
The JSMM model o O askainen e al. (2019) assumed an ins an-
aneous cap u e p ocess ha mimics a esea che isi ing a pa ic-
ula si e and a emp ing o cap u e he ma ked indi iduals du ing
a sho ime pe iod. The p esen da a o igina ed om bai ed aps
ha cap u ed indi iduals con inuously o e he 24- h pe iod o
which he aps we e kep ac i e, and hus he assump ion o ins an-
aneous cap u e would be a poo app oxima ion. We ex ended he
model o O askainen e al. (2019) by implemen ing he al e na i e
FIGURE 1 (a) The le panel highligh s he loca ion o he s udy a ea in Sabah, No he n Bo neo. The igh panel displays a map o he
ipa ian si es used in his s udy. (b, c and d) a e maps o he h ee s udy si es RR03, RR10 and RR18, espec i ely, showing he posi ion o he
dung- bai ed pi all aps (o ange ci cles wi h adius o 20 m). The landscape is cha ac e ised by Ripa ian Bu e (ligh g een), Fo es Rese e
(da k g een), Oil Palm (beige) and Ri e (blue) habi a s. The iangula ion shown in he panels was used o implemen he Join Species
Dis ibu ion Modelling. Mapped dis ances be ween aps (numbe ed 1– 18 a each si e) a e scaled o ep esen ac ual dis ances in he ield,
wi h a minimum o 50 m be ween aps. T aps we e checked and e- bai ed e e y second day o a pe iod o 14 days and indi iduals o he
ocal dung bee le species ma ked. One ap (numbe 8) is missing om he RR10 si e
400 km
Indonesia
Malaysia
0 m1,000 m
117°30'39.82"E
4°37'38.20"N
4°42'40.39"N
117°33'33.24"E
0
m1
,000 m
0 m1,000 m
Oil palm
Ri e
Fo es ese e
Ripa ian bu e
RR03
RR18
RR10
RR03
RR10
RR18
0 m1,000 m
0 m1,000 m0 m1,000 m
(a) (b)
(d)(c)
Species ID (s)Species
Taxonomic
au ho i y
Dung bu ial
mode
Tempo al
ac i i y
1Ca ha sius spp. — Tunnelle Noc u nal
2On hophagus mulle i Lansbe ge, 1883 Tunnelle Diu nal
3On hophagus obscu io Boucomon , 1914 Tunnelle Diu nal
4P oagode us wa anabei Ochi & Kon, 2002 Tunnelle Diu nal
5Pa agymnopleu us
spa sus
Sha p, 1875 Rolle Diu nal
6Sisyphus ho acicus Sha p, 1875 Rolle Diu nal
TABLE 1 Table o species wi h hei
axonomic au ho i y and ai s. Dung
bu ial mode and empo al ac i i y a e
om G ay e al. (2016). Species ID
ep esen s he code used in he Join
Species Mo emen Modelling analysis

242
|
Jou nal o Applied Ecology
GRAY e Al.
obse a ion model o a con inuous cap u e p ocess, whe e he ac-
i e aps cap u e indi iduals a he species- speci ic a e
qs
wi hin a
dis ance o 20 m om he cen e o he ap (see Appendices S2 and
S3 o de ails on he implemen a ion). As aps we e spaced a min-
imum o 50 m apa , a adius o 20 m was chosen o ep esen he
ci cula a ea su ounding a ap ha was dis inc o ha ap.
We combined he mo emen and obse a ion pa ame e s o
each species
s
o he ec o :
whe e he pa ame e s we e log- ans o med o enable a mul i a ia e
no mal model o
Θs
(O askainen e al., 2019). We i ed he model
wi h Bayesian in e ence, compu ing he likelihood o obse ing he
mo emen da a join ly o e he h ee s udy si es. We assumed he
same p io dis ibu ion as O askainen e al. (2019) and ollowed hei
Ma ko chain Mon e Ca lo (MCMC) me hod o sample he pos e io
dis ibu ion. We pe o med 43,000 i e a ions wi h a bu n- in pe iod o
3,000 i e a ions and se a hinning o 10. All he model analyses we e
pe o med o e sa ed samples e e y 10 i e a ions, o p oduce 4,000
pos e io samples (See Appendix S3 o de ails abou he pa ame e
es ima es and con e gence diagnos ics).
We e alua ed model i by gene a ing pos e io p edic i e da a,
whe e we eleased indi iduals in he same loca ions and a he same
imes as hey we e i s obse ed in he eal da a, and assumed he
same spa io- empo al a iance in cap u e e o as in he eal da a.
We compa ed he pos e io p edic i e da a o he eal da a in e ms
o he dis ibu ion o days om i s o las cap u e and in e ms o
he dis ibu ion o he o al dis ance mo ed.
To assess he in luence o he habi a composi ion on dung bee le
mo emen , we gene a ed pos e io p edic i e da a o an a i icial
landscape RR18* (Figu e 2a), which ep esen s a modi ied e sion o
he ac ual s udy si e RR18 (Figu e 1d). In landscape RR18*, he ipa -
ian bu e su ounding he aps was emo ed en i ely and eplaced
wi h oil palm habi a . We simula ed he elease o 100 dung bee-
les om he ap loca ions in he con inuous o es ese e wi hin
landscape RR18* and wi hin he s udy RR18 si e. We epea ed he
p ocess wi h 100 simula ed da ase s. The pos e io p edic i e da a
we e compa ed be ween he RR18 mo emen s and he mo emen s
in landscape RR18* (Figu e 2a). We compa ed he p opo ional num-
be o indi iduals caugh in he wo habi a s in landscape RR18* (oil
palm and o es ese e) wi h he h ee habi a s in RR18 (oil palm,
ipa ian bu e and o es ese e).
3 | RESULTS
We ma ked a o al o 8,646 bee les, o which 355 we e ecap-
u ed, gi ing an o e all ecap u e equency ac oss all h ee si es
and all species o 4.11% (Appendix S1: Table 1). On hophagus mul-
le i (Lansbe ge, 1883) had he highes ecap u e equency (6.19%)
and P oagode us wa anabei (Ochi & Kon, 2002) he lowes (3.03%).
O he ecap u es, 29% in ol ed indi iduals caugh in he ap om
which hey had been eleased. O ecap u ed indi iduals, 10% o
dung bee les we e ecap u ed wi hin 24 h o elease, and he e-
mainde a e mul iple days (2 o 12 days). Dung bee le mo emen s
a ied among and wi hin species (Figu es 3 and 4), and we e no ob-
iously linked o in e speci ic di e ences in size o dung bu ial mode
(Table 1), wi h small bee les such as Sisyphus ho acicus (Sha p, 1875)
and On hophagus obscu io (Boucomon , 1914) mo ing as a as he
la ge bee les. Fo he empi ical da a, he mean obse ed mo emen
dis ance was 102.1 ± SE 5.27 m/day, wi h a maximum obse ed
mo emen dis ance o 220 m/day (Appendix S1: Table 2). All spe-
cies we e ound o mo e a leas 100 m along he bu e , and h ee
species mo ed 350 m ( he la ges dis ance wi hin he s udy design).
Indi iduals om each species cap u ed in he o es showed mo e-
men o he u hes poin along he ipa ian bu e , al hough his
showed no dis inc pa e n, a ying by species and si e (Figu e 4).
3.1 | Join species mo emen modelling
Based on a isual inspec ion o he ace plo s (Appendix S3), he
MCMC sampling scheme showed sa is ac o y con e gence, and he
pos e io p edic i e da a ma ched gene ally well wi h he eal da a
(Figu e 5). Howe e , i p edic ed ewe e y long o e y sho mo e-
men dis ances (and hence mo e in e media e mo emen dis ances)
han obse ed in he eal da a (Figu e 5b; Appendix S4: Table 8). This
indica es ha he e is some le el o he e ogenei y wi hin and/o
among indi iduals in hei mo emen a es ha is no cap u ed by
he mo emen model whe e he di usion pa ame e is assumed o
be species- a he han indi idual speci ic, and whe e i is assumed
o emain cons an o e space and ime.
Ou esul s indica e a ia ion among dung bee le species in hei
p e e ence o di e ence habi a s, hei mo ali y a es, hei mo e-
men abili y and hei cap u e p obabili y (Figu e 6; Appendix S3:
Table 1). Fou o he six species showed no p e e ence o ipa ian
bu e s o e oil palm habi a s (Figu e 6a), wi h only O. mulle i and S.
ho acicus showing a signi ican p e e ence o ipa ian bu e habi-
a , and Ca ha sius spp. and P. spa sus showing a sligh p e e ence o
oil palm habi a (Figu e 6a). Only S. ho acicus showed a signi ican
p e e ence o o es ese e o e ipa ian bu e , al hough P. wa a-
nabei, Ca ha sius spp. and O. mulle i also showed a sligh p e e ence
o o es ese es (Figu e 6b). The es ima ed mo ali y a es we e
high and consis en ac oss species, co esponding o an a e age li e
span o c. 5 days ac oss species (expec ed li e span = 1/mo ali y
a e), and a e mo e likely o e lec he indi iduals becoming inac i e
o lea ing he s udy a ea, o he ma ks ubbing o he ely a, han
ac ual mo ali y (Figu e 6c). Species showed simila mo emen a es,
excep P. spa sus which had a lowe mo emen a e han he o he
species (Figu e 6d). This species also had a much lowe cap u e a e
han he o he species (Figu e 6 ).
The simula ion o land- use change demons a ed simila pa -
e ns in dung bee le p e e ence o di e en habi a s (Figu e 2b). S.
ho acicus showed a lowe cap u e a e in he non- elease habi a
ollowing he emo al o he ipa ian bu e , which esul is in line
Θs
=
(
log
(
k
RB
s)
, log
(
k
FR
s)
, log
(
k
OP
s)
, log
(
m
s)
, log
(
D
s)
, log
(
q
s)),
|
243
Jou nal o Applied Ecology
GRAY e Al.
wi h he species p e e ence o ipa ian bu e o e oil palm habi a
(Figu e 6a). Simila ly, O. mulle i showed a lowe median cap u e a e
in he non- elease habi a (Figu e 2b). Only Ca ha sius spp., showed
a highe median cap u e a e in he non- eleased habi a ollowing
he emo al o ipa ian bu e bu his di e ence was no signi ican
(Figu e 2b). None o he species displayed di e ences in hei cap-
u e a es in he o es ese e ollowing he simula ed emo al o
he ipa ian bu e , indica ing a p e e ence o emaining in o es ed
habi a (Figu e 2b).
4 | DISCUSSION
Habi a bounda ies a e o en ba ie s o he dispe sal and mo e-
men beha iou o insec s (Jain e al., 2020), especially o o es -
dependen species (G ay e al., 2016, 2019b; Sc i en e al., 2017),
sugges ing ha ag icul u al habi a s bo de ing o es habi a s may
limi mo emen (A ellano e al., 2008; Me ckx e al., 2010; Slade
e al., 2013). Ou da a p o ide new in o ma ion on he mo emen
beha iou o unc ionally impo an insec s, dung bee les, in ipa ian
o es bu e adjacen o an oil palm ma ix.
Low ecap u e a es we e simila o p e ious s udies on dung
bee les (5% in Mexico - A ellano e al., 2008; 18% in Colombia—
Cul id- Medina e al., 2015; 3% in B azil— da Sil a & He nández, 2015)
and on o he in e eb a es in ma ix landscapes in Sabah, Malaysia
(31.6% o bu e lies— Sc i en e al., 2017; 33.7% o mo hs— G ay
e al., 2019b). These low ecap u e a es (due o dung bee les be-
coming inac i e, lea ing he a ea, o because he ma ks ubbed o
he ely a) may lead o unde es ima es o mo emen o ecap u e
a es o e longe ime pe iods, despi e species being ecap u ed
up o 12 days la e . Mo emen showed in e speci ic a ia ion and
al hough mean dis ances mo ed we e no mally less han 50 m in
24 h , ou esul s showed ha dung bee les can mo e much g ea e
maximum dis ances in a day: ypically >100 m and up o 220 m
wi hin 24 h (Figu e 4; Appendix S1). The e o e, p e ious s anda ds
o dung bee le apping which assumed daily mo emen dis ances
o <100 m and so conside ed aps spaced 50 m apa as inde-
penden (La sen & Fo sy h, 2005), a e p obably no su icien o
Sou heas Asian dung bee les. We ecommend ha aps be spaced
150– 200 m apa o assu e independence be ween aps in u u e
s udies. Simila ecommenda ions ha e been sugges ed ollowing
s udies on Sou h Ame ican o es dung bee les showing mo emen s
FIGURE 2 E ec o simula ed land use change ( emo al o ipa ian bu e ) on he p opo ion o indi iduals caugh in each habi a . (a)
Shows an a i icial landscape RR18* ha was modi ied om he ac ual RR18 si e (Figu e 1d) by eplacing he ipa ian bu e wi h oil palm
plan a ion habi a . Fo his simula ion, on he i s day, 100 dung bee les we e eleased o each species om each ap in he o es ese e
(9,10,11,12) o RR18 and RR18*, and hen hei nex cap u e loca ion de e mined. The p ocess was epea ed o gene a e 100 da ase s ha
assumed he same cap u e– ecap u e design as in he ield s udy. The pa ame e s o each da ase we e sampled om he pos e io . (b) The
panel shows he p opo ion o indi iduals o each species whose las cap u e loca ion was he o es ese e— same as he elease (g een
lines), and hose whose las cap u e loca ion was di e en o he elease habi a (i.e. wi hin oil palm in RR18* o wi hin oil palm o ipa ian
bu e in RR18; g ey lines). Poin s ep esen he pos e io median and he ba s he 95% c edible in e al. Solid and dashed lines ep esen
he si es RR18 and RR18* espec i ely
Oil palm
Ri e
Fo es ese e
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
Cap u ed in
RR18
RR18*
Fo es ese e
Fo es ese e
Non- elease habi a
Non- elease habi a
P. spa sus P. wa anabei O. obscu io O. mulle i S. ho acicusCa ha sius spp.
Species
P opo ion o indi iduals caugh
0 m1,000 m
0 m 1,000 m
(b)
(a)
RR18*
244
|
Jou nal o Applied Ecology
GRAY e Al.
be ween 20 and 500 m, depending on species and habi a (A ellano
e al., 2008; Cul id- Medina e al., 2015; No iega & Acos a, 2011; da
Sil a & He nández, 2015).
Ou esul s indica e ha some dung bee le species mo e eely
be ween o es a eas and oil palm, while o he s demons a e s ong
speci ici y o o es ed a eas. Ripa ian bu e s did no display a con-
s ain on mo emen , wi h no ob ious dis ance decay; howe e ,
he e we e no species- o si e- speci ic pa e ns in he mo emen o
bee les om he o es down he ipa ian bu e . The ocal species
in ou s udy we e chosen o ep esen a ange o sizes and unc ional
g oups o dung bee les ound in he a ea and so he pa e ns we
see likely ep esen he mo emen pa e ns o dung bee les wi hin
his egion. Ou indings highligh he signi icance o se - aside o -
es a eas wi hin oil palm landscapes and suppo he g owing e -
idence base ha con iguous ipa ian o es bu e s a e impo an
o main aining biodi e si y o a wide ange o in e eb a e axa
(Ba low e al., 2010; G ay e al., 2016, 2019b; Luke e al., 2017;
Sc i en e al., 2017; Williamson e al., 2020).
4.1 | Habi a p e e ence o dung bee le species
Dung bee le ecap u es la gely consis ed o wi hin- habi a mo e-
men s wi hin ipa ian bu e s and con inuous o es . Indi idual spe-
cies di e ed in habi a speci ici y. Two species (S. ho acicus and O.
mulle i) showed high and medium speci ici y, espec i ely, o bo h
o es ese e and ipa ian bu e s ips, and a ely c ossed in o he
oil palm ma ix. Rolle s, such as S. ho acicus, can be pa icula ly a -
ec ed by o es con e sion o oil palm, and a e a ely ound in oil
palm plan a ions (G ay e al., 2016). P e ious s udies ha e shown
FIGURE 3 Focal dung bee le
mo emen ajec o ies o he h ee s udy
si es (a) RR03, (b) RR10 and (c) RR18.
Lines ep esen mo emen be ween
a ap he dung bee le was eleased
om and ecap u ed in. The in ensi y
o he colou o each do ep esen s he
p opo ion o cap u es ha occu ed in
he co esponding dung- bai ed pi all
ap du ing he empi ical s udy. Numbe s
indica e ap numbe a he si e
(a) (b)
(c)
|
245
Jou nal o Applied Ecology
GRAY e Al.
changes in communi y composi ion o dung bee le communi ies be-
ween o es , ipa ian bu e s and oil palm (G ay e al., 2014), and
weak spillo e e ec s om ipa ian bu e s o o es - dependen
dung bee les (G ay e al., 2016). P e e en ial use o o es o e oil
palm may be because highe mammal abundance in o es ed a eas
inc eases dung a ailabili y (Ba low e al., 2010; Dee e e al., 2018).
Al e na i ely, habi a bounda ies may ac as ba ie s o dung bee le
mo emen (G ay e al., 2016), o example i bee les pe cei e s uc-
u al changes o he ege a ion ha migh inc ease exposu e o p ed-
a o s (Ba low e al., 2010), o i bee les ha e di e ences in he mal
ole ance ha ac as a il e be ween habi a mic oclima es (Bi ke
e al., 2017; Roslin e al., 2009; Williamson e al., 2020). Edge e ec s
FIGURE 4 The p opo ion o
indi iduals as a unc ion o he minimum
Euclidean dis ance (m) om he Fo es
ese e bounda y o he cen e o a ap
loca ed in he Ripa ian bu e . (a) RR03,
(b) RR10 and (c) RR18. The unc ion
shows he p opo ion o animals whose
las cap u e was a a ap loca ed in
he Ripa ian bu e ( aps 1 o 8 in si es
RR03 and RR18, and aps 1 o 7 in
si e RR10) gi en ha hey we e i s
cap u ed in he Fo es ese e (x = 0).
This ep esen s a subse o he indi iduals
caugh . Poin s and da k egions ep esen
he median and 95% c edible in e al
espec i ely. Ligh colou s ep esen
he co esponding ex eme alues. The
p opo ion o cap u ed indi iduals used
o pa ame isa ion a e ep esen ed by
open poin s