animals
A icle
Se Use, Densi y and B eeding Phenology o Badge s in
Medi e anean Ag o-Syl o-Pas o al Sys ems
Ma celo Sil a †, Luís Miguel Rosalino †, Sand a Alcobia and Ma ga ida San os-Reis *


Ci a ion: Sil a, M.; Rosalino, L.M.;
Alcobia, S.; San os-Reis, M. Se Use,
Densi y and B eeding Phenology o
Badge s in Medi e anean
Ag o-Syl o-Pas o al Sys ems.
Animals 2021,11, 2663. h ps://
doi.o g/10.3390/ani11092663
Academic Edi o s: Nicola Ma ples,
Aoibheann Gaugh an and
Enda Mullen
Recei ed: 1 Augus 2021
Accep ed: 9 Sep embe 2021
Published: 10 Sep embe 2021
Publishe ’s No e: MDPI s ays neu al
wi h ega d o ju isdic ional claims in
published maps and ins i u ional a il-
ia ions.
Copy igh : © 2021 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
cE3c—Cen e o Ecology, E olu ion and En i onmen al Changes, Faculdade de Ciências, Uni e sidade de
Lisboa, 1749-016 Lisboa, Po ugal; [email p o ec ed] (M.S.); lm [email p o ec ed] (L.M.R.);
[email p o ec ed] (S.A.)
*Co espondence: mm [email p o ec ed]
† These au ho s con ibu ed equally o his wo k.
Simple Summa y:
Unde s anding ca ni o es social s uc u e a ia ion is pi o al o p ope ly ad-
d essing conse a ion challenges and solu ions. The Eu opean badge s is a social ca ni o e o
which mos o he a ailable in o ma ion ega ding how his species is socially o ganized de i es
om cen al wes popula ions. This a icle desc ibes he g oup composi ion, den use pa e ns and
b eeding phenology o a Medi e anean popula ion o badge s. We showed ha badge li e in
low densi y, in ela i ely small g oups, composed by 2–4 adul animals and ca. 2 cubs, bo n in
win e . These pa e ns, ep esen ing a a ia ion o wha was desc ibed o o he popula ions, show
ha badge s ake ad an age o he landscape con ex , whe e human- ela ed esou ces and mild
en i onmen al condi ions allow badge o each highe densi ies han in many sou he n popula ions,
and o ep oduce ea lie han hei no he n coun e pa s.
Abs ac :
Ca ni o es social o ganiza ion a ies widely, om s ongly social o soli a y p eda o s.
Eu opean badge s a e acul a i e social ca ni o es ha also shows a geog aphical a ia ion in
social s uc u e. These pa e ns de i e mainly om cen al/wes Eu opean egions, wi h an unde -
ep esen a ion o Medi e anean popula ions ha ace di e en conse a ion challenges, especially
ega ding g oup composi ion, se use pa e ns and b eeding phenology. We add essed hese ai s
opics o a popula ion inhabi ing a Po uguese ag o-sil o-pas o al sys em. Based on mon hly
moni o ing o 34 se s and con inuous came a- apping su eys o 12, we showed ha se s su -
ounded by di e si ied ege a ion and loca ed in sandy si es a e mo e used, a pa e n p obably
linked o ood a ailabili y and ease o se exca a ion and main enance, espec i ely. Badge s ol-
lowed a gene al pa e n ega ding g oup size (2–4 adul s), bu showed an in e media e popula ion
densi y (
0.49–0.73 badge s/km2
), wi h alues highe han hose es ima ed o o he Medi e anean
en i onmen s, bu lowe han o cen al-wes e n popula ions. This, oge he wi h he b eeding
(No embe /Janua y) and cub eme gence (1.8 cubs/se ; Ma ch/Ap il) pe iods, indica es an ecolog-
ical adap a ion o he landscape con ex , whe e human- ela ed esou ces and mild en i onmen al
condi ions allow badge o each highe densi ies han in many sou he n popula ions, and o ep o-
duce ea lie han hei no he n coun e pa s.
Keywo ds: came a- apping; Meles meles; densi y; social o ganiza ion; ep oduc ion
1. In oduc ion
Ca ni o es show a wide a ay o social o ganiza ions, anging om soli a y (e.g., wild-
ca s, Felis sil es is) o highly social species, such as he mee ka (Su ica a su ica a) [
1
].
Howe e , less han 20% a e conside ed social p eda o s [
2
], and e en hose show dis inc
deg ees o sociali y. While wol es (Canis lupus) can o m packs wi h mo e han 40 animals
sha ing hun ing, e i o y pa olling and de ense, cub ea ing, and o he ac i i ies [
3
],
o he s, such as Eu opean badge s (Meles meles), o m clans usually wi h less han se en
Animals 2021,11, 2663. h ps://doi.o g/10.3390/ani11092663 h ps://www.mdpi.com/jou nal/animals
Animals 2021,11, 2663 2 o 15
indi iduals ha sha e he same e uge and e i o y and show some in a-clan in e ac ions,
bu lack he social sophis ica ion showed by highly social species [
4
]. The way badge social
g oups a e o ganized and use he communal e uges hey build (known as se s) is howe e
highly a iable ac oss i s dis ibu ion ange and his a iabili y s ill poses ques ions abou
how he landscape con ex may shape his species’ social pa e ns.
Th oughou Eu ope, he size o badge social g oups anges om singe pai s in
sou he n Spain [
5
] o se en adul indi iduals in Wy ham Woods, UK (al hough eaching
30 adul s/yea ling badge s in unusual si ua ions, such as hose ound in Woodches e Pa k,
UK, in 1989) [
6
]. Such social s uc u e de e mines he species densi y, which peaks in he
UK (38 ind./km
2
) and eaches i s lowes alue in Eas e n Eu ope (e.g., Czech Republic,
0.12 ind./km
2
[
7
]. This a ia ion in densi y is equen ly linked o clima e a ia ion
(e.g., we e clima e a o s ea hwo m’s abundance and, indi ec ly, badge s), abundance
and a ailabili y o po en ial se si es and/o ood, and le el o dis u bance, such as human
popula ion densi y, oad densi y o hun ing p essu e [8].
Ibe ia, co esponding o he species sou h-wes e n ange limi , seems o be a chal-
lenging en i onmen o badge s, which occu a low densi ies (0.13–0.67 ind./km
2
, in
Spain, and 0.36–0.48 ind./km
2
in Po ugal) [
9
–
12
]. This popula ion s uc u e seems o
be de e mined mos ly by he low a ailabili y o ood esou ces [
11
] o se si es [
13
] in
Medi e anean en i onmen s. Howe e , es ima ions o badge densi y a e limi ed o e y
ew egions in Ibe ia—Doñana Na ional Pa k (S) and Pa k o Collse ola (NE), in Spain,
and Se a de G ândola in Po ugal [
10
–
12
]—no ep esen ing he ull ange o en i on-
men al condi ions badge s explo e in Ibe ia. Thus, mo e da a a e needed o allow a be e
unde s anding o how Ibe ian popula ions espond o and a e in luenced by he Ibe ian
landscape con ex , because his egion is expec ed o su e d as ic en i onmen al changes
in he nex u u e due o clima e wa ming [14,15].
Se s play a cen al ole in badge socio-ecology, ac ing no only as e uge si es (whe e
indi iduals spend as much as 70% o hei ime), bu also as ocal poin s o ep oduc ion,
a nu se y, and a social exhibi ion and in e ac ion a ena [
4
]. Fo all hese easons, badge s
densi y is o en linked o se densi y, especially when conside ing he main se s, i.e., la ge
se s (>5 en ances) wi h signs o a egula use h oughou he yea [
16
]. Howe e , in
low densi y a eas, as Ibe ia, his ela ion is weake , and keeps alid only including also
smalle and less equen ly used se s (o en called, seconda y, ou lie s o subsidia y) [
16
].
Wha e e he ype o se s conside ed, hei use pa e n is de e mined by hei in e nal
s uc u e, landscape con ex and dis u bance isk [17–19].
In he Medi e anean egion, human p esence has shaped he landscape o millennia,
con e ing na u al en i onmen s o ag icul u e land, pas u e o li es ock o human
se lemen s/in as uc u es [
20
], a ec ing wildli e ecological pa e ns. In Ibe ia, badge s
show low se use a es, wi h animals equen ly changing be ween se s in consecu i e
days (ca.50% o he occasions), and he e-use a e g ea ly declining a e i e nigh s [
21
].
This pa e n seems o be linked o an h opic dis u bances, such as ca le b eeding and
emo al o o es unde s o y a ound he se s [22].
Se s a e also a pi o al s uc u e in badge ep oduc ion, as bi h occu s in unde -
g ound se chambe s, and cubs s ay inside un il hey can explo e he ou side en i on-
men [
4
]. While ma ing can occu h oughou he yea , mos a ailable bi h eco ds a e
concen a ed in Feb ua y, due o delayed implan a ion, wi h a li e size o 2–3 cubs ( ange
1–5) [
4
]. Howe e , his in o ma ion o igina es mainly om UK popula ions, wi h sca ce
da a a ailable o Medi e anean popula ions, whe e he highly di e en clima e may
p oduce a ia ions o he B i ish pa e n. A ailable da a o he Ibe ian Peninsula indica e
ha in cap i e animals ma ing occu ed be ween Feb ua y and May, wi h he bi h o
1–3 cubs
in Feb ua y [
23
], while 3–4 cubs pe se ha e been epo ed o a wild popula ion
in Po ugal [12].
Conside ing he speci ic cha ac e is ics o Medi e anean landscapes, and he lack o
in o ma ion on he ecology o badge s in his egion, ou s udy aimed o assess badge :
Animals 2021,11, 2663 3 o 15
(1) densi y; (2) se use pa e ns and unde lying de e minan s; and (3) b eeding phenology
and li e size.
To ul ill hese goals, we es ed eigh hypo heses (Table 1).
Table 1.
Hypo heses es ed a ge ing h ee li e-his o y pa e ns: se use, densi y and b eeding phenology. Fo each, we
de ail he unde lying easoning and he suppo ing e e ences.
Pa e n Hypo hesis Reasoning Suppo ing
Re e ences
Densi y
H1—Badge ’s densi y will be lowe
when compa ed o he species co e ange
a ea, bu wi hin he limi s epo ed o
Ibe ian popula ions
S udies implemen ed in Ibe ia es ima e a
popula ion densi y o 0.13–0.67 ind./km2,
signi ican ly lowe han ha es ima ed o
cen al/wes Eu opean popula ions
(mean = 9.2 ind./km2, SD = 10.5)
[4,9–12]
Se use pa e n
H2—Se use is p omo ed by he
occu ence o dense ege a ion
su ounding he se
In human-shaped landscapes, he occu ence o
dense ege a ion p o ides a mo e p o ec i e
con ex , whe e animals may socially in e ac mo e
sa ely (including cubs)
[24,25]
H3—Se s loca ed in a eas wi h a easily
diggable and well-d ained soil will be
mo e equen ly used
Badge s p e e well d ained and cohesi e soils as
se si es [26,27]
H4—Se use is p omo ed by he
p oximi y o eeding pa ches (e.g., oli e
ya ds, wildli e eeding s a ions)
Se loca ion nea pa ches p o iding ood
esou ces allows badge s o sa e ene gy and ime
in hei o aging bou s
[22,28]
H5—Se s subjec o high dis u bance
(e.g., ca le, oads p oximi y) will be less
used by badge s
An h opogenic dis u bance is known o a ec se
use by badge s [22,29–31]
H6—Se s loca ed in a eas whe e
pe cei ed compe i ion can be high will be
less used
E idence exis ha he p esence o wild boa s
cons ains badge s p esence, due o esou ce
compe i ion. In e -g oup compe i ion is also
conside ed a passi e ange exclusion mechanism.
[32,33]
Rep oduc ion
phenology
H7—Badge ma ing peaks in
Janua y/Feb ua y and cubs will s a
eme ging om he se s in Ap il
Badge ma ing can occu all yea ound, wi h wo
peaks in Win e /Sp ing (main) and
Summe /Au umn; cubs eme ge eigh weeks a e .
[11,34]
H8
—Li e s will be composed o 2–3 cubs
Al hough a ailable da a o Po ugal indica e 3-4
cubs/li e , he a e age alues o mos badge
popula ions is lowe
[12,34]
2. Ma e ials and Me hods
2.1. S udy A ea
The s udy was conduc ed in he “Cha neca do In an ado” (38
◦
48
0
N 8
◦
49
0
W;
Figu e 1
),
a 100 km
2
a ms ead managed by Companhia das Lezí ias S.A., mainly de o ed o he
p oduc ion o co k (co k oak, Que cus sube , woodlands—67.50 km
2
), wood (ma i ime
pine, Pinus pinas e , plan a ions—9.71 km
2
; Eucalyp us, Eucalyp us globulus, plan a ions;
4.76 km
2
), pine nu s (s one pine, Pinus pinea, plan a ions—5.08 km
2
), and ag icul u al goods
(e.g., co n, Zea mays—2.50 km
2
; ice, O yza spp.—2.40 km
2
; oli es, Olea eu opaea—0.70 km
2
).
This mul iuse ag o o es y sys em also includes o he ac i i ies, namely seasonal ca le
aising, wi h ca le g azing in he a ms ead in Au umn/Win e , and hun ing mos ly o
wild boa s (Sus sc o a) and pa idges (Alec o ys u a), o which se e al a i icial eeding
s a ions wi h ce eals and wa e a e a ailable h oughou he a ms ead [35].
The egion has a ypical Medi e anean clima e, wi h ho and d y summe s and
mild bu ainy win e s. Du ing he s udy pe iod he mean empe a u e and p ecipi a ion
we e o 16.8
◦
C and 361 mm, espec i ely. “Cha neca do In an ado” is loca ed in a plain
lowland a ea, wi h ew ela i ely deep alleys o med by empo a y wa e cou ses. Only
Animals 2021,11, 2663 4 o 15
one pe manen s eam c osses he a ms ead (“Ribei a de Vale Cob ão”), al hough se e al
a i icial wa e poin s and empo a y ponds a e sca e ed h oughou he a eas.
Animals2021,11,26634o 15

a i icial eedings a ionswi hce ealsandwa e a ea ailable h oughou  he a ms ead
[35].
The egionhasa ypicalMedi e aneanclima e,wi hho andd ysumme sandmild
bu  ainywin e s.Du ing hes udype iod hemean empe a u eandp ecipi a ionwe e
o 16.8°Cand361mm, espec i ely.“Cha necadoIn an ado”isloca edinaplainlowland
a ea,wi h ew ela i elydeep alleys o medby empo a ywa e cou ses.Onlyone
pe manen s eamc osses he a ms ead(“Ribei adeValeCob ão”),al houghse e al
a i icialwa e poin sand empo a ypondsa esca e ed h oughou  hea eas.

Figu e1.S udya ealoca ion(A),andspa ialdis ibu iono de ec edbadge se s(B),highligh ing hoseselec ed o 
in ensi emoni o ingusingcame a‐ aps(C).Di e en colo s ep esen dis inc badge socialg oups,andmainse sa e
ep esen edbybigge ci cles.
2.2.Se Su ey
We i s  e iewedandcompiledallse  eco ds egis e edinp e iouss udies
conduc edin hes udya ea(e.g.,[22,36]).All hesese s(n=33)we e e isi ed ocon i m
hei usebybadge andnewse swe ein ensi elysea chedin hes udya ea.Thea eais
di idedinca leg azingplo sand,be weenSep embe 2016andAp il2017,wesu eyed
20 ha  ul illed hecha ac e is icsconside edp omo e so se emplacemen ( ege a ion
andsoil ea u es;see[4];anda oidingin ensi eag icul u ea eas,suchas iceandco n
ields,wa e channeland ese oi s,e c.).Theseplo s, o alizing15km2(meana ea1.45
km2,sd=1.05, ange:0.05–6.66km2)we esu eyedbasedonse e allinea andpa allel
ansec s,co e ing heen i ecell.Thedis ancebe ween he ansec swasabou 50me e s,
Figu e 1.
S udy a ea loca ion (
A
), and spa ial dis ibu ion o de ec ed badge se s (
B
), highligh ing hose selec ed o
in ensi e moni o ing using came a- aps (
C
). Di e en colo s ep esen dis inc badge social g oups, and main se s a e
ep esen ed by bigge ci cles.
2.2. Se Su ey
We i s e iewed and compiled all se eco ds egis e ed in p e ious s udies conduc ed
in he s udy a ea (e.g., [
22
,
36
]). All hese se s (n= 33) we e e isi ed o con i m hei use
by badge and new se s we e in ensi ely sea ched in he s udy a ea. The a ea is di ided
in ca le g azing plo s and, be ween Sep embe 2016 and Ap il 2017, we su eyed 20 ha
ul illed he cha ac e is ics conside ed p omo e s o se emplacemen ( ege a ion and soil
ea u es; see [
4
]; and a oiding in ensi e ag icul u e a eas, such as ice and co n ields, wa e
channel and ese oi s, e c.). These plo s, o alizing 15 km
2
(mean a ea 1.45 km
2
,
sd = 1.05
,
ange: 0.05–6.66 km
2
) we e su eyed based on se e al linea and pa allel ansec s, co e ing
he en i e cell. The dis ance be ween he ansec s was abou 50 me e s, since we su eyed
a bu e o 25 m in each pe pendicula di ec ion. In mo e closed a eas (woodlands, o
example), he dis ance be ween ansec s was sho ened o ensu e ha he whole a ea was
su eyed, as he isual de ec ion o se s dec eases in dense ege a ion a eas.
Assuming ha main se s a e loca ed in he co e a ea o each g oup e i o y [
24
], and
conside ing ha badge s home ange in a simila Po uguese landscape was 4.46 km
2
[
12
],
Animals 2021,11, 2663 5 o 15
we assumed ha all seconda y se s loca ed wi hin a adius o 1191 m om a main se (see
Sec ion 2.3) will belong o he same social g oup. The only excep ions we e seconda y se s
ei he sepa a ed by an insu moun able geog aphical ba ie (e.g., pe manen s eam) o
loca ed wi hin he bu e bounda ies o wo neighbo main se s (in such si ua ion hey
we e g ouped wi h he closes main se ).
2.3. Se Moni o ing
To de e mine he se use pa e n, all de ec ed se s we e isi ed mon hly o de e mine
badge signs o p esence, and es ima e a se use index, anging om 0 (ne e ac i e) o 7
always ac i e. A se was conside ed ac i e i nea by i s en ances we ound: (1) eshly
exca a ed soil, (2) ecen ly used la ines; (3) esh nes ma e ial (Figu e S1; Supplemen a y
Ma e ial); and/o (4) high abundance o oo p in s.
Se s ha showed equen and abundan ac i i y signs (se index = 6–7), o whe e e-
p oduc ion was de ec ed, we e conside ed main se s [
12
], all he emaining being classi ied
as seconda y [
24
]. The main se s we e he a ge o an in ense moni o ing p og am, using a
came a apping app oach [
37
]. We ins alled 14 Moul ie
®
m-990i Gen2 (Cale a, AL, USA)
came a- aps (Figu e S2; Supplemen a y Ma e ial), equipped wi h a 16 Gb memo y ca d,
se o ake 30s low esolu ion ideos (854–480 pixels) when he hea /mo emen senso
was ac i a ed, wi h a minimum ime in e al be ween consecu i e ideos o 5s. Came as
we e ins alled 30 cm abo e he g ound, a ached o a ee o a wooden s ick, acing one
o mo e ac i e se en ances. In wo se s wi h mo e han one en ance ac i e, and ha
could no be simul aneously moni o ed wi h a single came a, an addi ional one was se o
ensu e he moni o ing o all ac i e en ances. Came as we e checked weekly o eplace he
ba e ies and memo y ca ds. Se moni o ing occu ed be ween No embe 2016 and Ap il
2017, and each came a was ac i e o 180 days, o an o e all moni o ing pe iod o 2520
came a- apping nigh s.
2.4. Se Cha ac e iza ion
In o de o assess which en i onmen al d i e s, we e explaining he se use pa e n,
we used 21 a iables o cha ac e ize each se (Table 2). Va iables associa ed o ege a ion
composi ion and s uc u e we e assessed, in si u, wi hin he minimum con ex polygon ha
encompassed all he se en ances, plus a bu e o 2 m. The Simpson di e si y index [
38
] was
es ima ed based on he he baceous, sh ub and ba e soil co e in each polygon. La e , we buil
a Geog aphic In o ma ion Sys em (GIS; QGIS
®
e sion 2.8.3—Wien, Bea e on, OR, USA [
39
]),
ha encompassed emo e sensing in o ma ion ega ding he: ype o soil (da a p o ided by
he Companhia das Lezí ia, S.A, Samo a Co eia, Po ugal); land use [
38
]; loca ion o se s,
game eeding s a ions and wa e ing pi o s, pa ed and unpa ed oads and wa e poin s ( o
allow he es ima ion o he dis ance o each se om hese landma ks; da a p o ided by he
Companhia das Lezí ia, S.A.); g azing p essu e pe plo (da a p o ided by he Companhia
das Lezí ia, S.A.); and wild boa occu ence de e mined in a p e ious s udy [40].
2.5. G oup Size and B eeding Phenology
Each ideo was ca e ully iewed and we egis e ed he numbe o indi iduals iden-
i ied, he age class (i.e., adul , cub; based on animal’s size and su ey mon h [
4
], he
in e ac ion be ween he animal and he se (i.e., en e ing o lea ing he se , exca a ing o
handling nes ma e ial), as well o he beha io s o social in e ac ions (e.g., an agonis ic
o b eeding beha io s). To es ima e he numbe o indi iduals pe social g oup we used
he highes numbe o badge s eco ded simul aneously. Howe e , his es ima ion was
only conside ed alid i ha numbe o indi iduals was obse ed in, a leas 1% o he o al
eco ded ideos o each se . This c i e ion was de ined o educe he isk o coun ing, as
membe s o he moni o ed g oup, indi iduals om o he social g oups ha may be isi ing
he main se o ano he g oup. This conse a i e app oach conside s ha all badge s
de ec ed simul aneously belong o he same social g oup.

Animals 2021,11, 2663 6 o 15
Table 2.
Va iables used o cha ac e ize each se du ing he mon hly moni o ing isi , wi h e e ence o he hypo hesis o which hey we e linked, he ype o a iable and i s ange/uni s.
Va iable Desc ip ion Type Range/Uni s
H2
T ee T ee co e O dinal 1–5 (1—low; 5—high) 1,2
Sh ubs Sh ub co e Con inuous Pe cen age 1,2
He bs He baceous co e Con inuous Pe cen age 1,2
Sh ub_H Sh ub mean high Con inuous Cm 1,2
Ba e_soil Ba e soil co e Con inuous Pe cen age 1,2
Veg_di Vege a ion di e si y Con inuous 0–1 (Simpson di e si y index)
Unde s o y Dominan unde s o y Ca ego ical 4 Ca ego ies (Ca le pas u es, Sh ubland, Na u al
pas u es, Na u al pas u es wi h sh ubs)
Land use Type o land use Ca ego ical 4 ca ego ies (Co k oak woodland, Mixes wood,
Na u al pas u e, Pine o es )
H3 Soil_ma Main soil ma e ial Ca ego ical 4 ca ego ies (Sand, Rock, Sil /Clay, Roads 3)1
Soil Type o soil 2 Ca ego ies (Podzols, Regosols)
H4
Dis _ol Dis ance o he nea es oli e ya d Con inuous m
Dis _pi Dis ance o he nea es wa e ing pi o Con inuous m
Dis _ eed Dis ance o he nea es game a i icial eede s Con inuous m
Dis _wa Dis ance o he nea es wa e sou ce Con inuous m
H5
G azz_16–17 G azing p essu e be ween 2016–2017 Con inuous G azing in ensi y 4
G azz_07–17 Cumula i e g azing p essu e be ween 2007–2017 Con inuous G azing in ensi y 4
N_g azz Numbe o consecu i e yea s, since 2007, wi hou ca le g azing Con inuous Numbe o yea s
Dis _ oad Dis ance o he nea es pa ed oad Con inuous m
Dis _ oad2 Dis ance o he nea es di oad Con inuous m
H6 Dis _se Dis ance o he nea es badge se Con inuous m
Wildboa Wild boa abundance Con inuous Numbe o wild boa signs [40]
1
Va iables collec ed in si u, wi h he minimum con ex polygon encompassing all he se en ances.
2
Va iables used in he P incipal Componen s Analysis (PCA).
3
Se s loca ed unde di oads.
4
Exp essed as
he numbe o ca le pe day and pe hec a e, in he po whe e each se was loca ed [35].
Animals 2021,11, 2663 7 o 15
Rega ding ep oduc ion, we conside ed ha a ma ing e en occu ed when we de-
ec ed in he ideo a sexual in e ac ion be ween wo adul badge s. When in subsequen
ideos he ma ing e en s ill occu ed, we egis e ed he en i e sequence as a single ma ing
e en . Ma ing e en s less han 2 minu es long we e conside ed sho -du a ion ma ing o
copula ion e en s, and hose ha occu ed o mo e han 2 minu es ( ypically 12 minu es
o mo e; see esul s) we e egis e ed as long-du a ion ma ing e en s [
34
]. Long ma ing
e en s a e expec ed o inc ease b eeding success [41].
2.6. Da a Analysis
2.6.1. D i e s o Se Use
In he modelling p ocedu e applied o assessing he en i onmen al d i e s o se use
by badge s (hypo heses H2–H6), he numbe o mon hly isi s posi i e o signs o ac i i y
( anging om 0 o 7) was included as dependen a iable (Tables 1and 2).
Due o he high numbe o a iables associa e o H2 we conduc ed a P incipal Compo-
nen Analysis (PCA) [
42
], using he a iables ee, sh ub, he baceous and ba e soil co e
and sh ub heigh (Table 2); p incipal componen s ha cumula i ely explained >80% o he
in o ma ion o he o iginal a iables, we e used as candida e d i e s in he subsequen
modelling p ocedu e (see 3.1).
Spa ial au oco ela ion o da a was assessed using he Mo an I Index [
43
], and we e al-
ua ed mul icollinea i y be ween all candida e d i e s (excep hose used in he PCA) using
he Spea man co ela ion coe icien ,
ρ
[
42
]. When wo a iables we e highly co ela ed
(ρ> 0.70), we excluded he one less co ela ed wi h he dependen a iable [44].
Fo each hypo hesis, we un Gene alized Linea Models (GLM) [
42
], using a Poisson
dis ibu ion and a “log” link unc ion, co esponding o all combina ion o he a iables
associa ed o he hypo hesis (Table 2). Resul ing models we e anked acco ding o he
Akaike In o ma ion C i e ion, wi h co ec ion o small sample size (AICc) [
45
]; hose ha
showed a
∆
AICc < 2 (i.e., he di e ence be ween he AICc o a model and he lowes AICc
alue in he model se [
46
]) we e conside ed he bes models in explaining se use a ia ion.
I mo e han one model ul illed his c i e ion, we used a model a e aging p ocedu e [47]
o assess a iables a e age coe icien and he 95% con iden in e al (95% CI). Va iables
wi h a 95% CI ha did no include he 0 (i.e., we can de e mine mo e p ecisely i i s e ec
is posi i e o nega i e), and ha showed a highe ela i e impo ance, we e conside ed
mo e in luen ial [
48
]. Rela i e impo ance was es ima ed as he sum o he Akaike weigh
(w; p obabili y o a model being he bes model) [
45
], o all he models ha included he
a iable o in e es .
The a iables ha , o each hypo hesis, showed a signi ican (p< 0.05) o almos
signi ican (p< 0.1) in luence on se use pa e n, we e included in o e all models (combined
hypo hesis), ha pos ula ed ha he pa e n o se use is de e mined by a combina ion
o d i e s linked o dis inc ecological/en i onmen al p ocesses ( ege a ion, soil, ood,
dis u bance o compe i ion).
Finally, we compa ed he AICc o he bes model o each hypo hesis and conside ed
he hypo hesis wi h mo e suppo as ha showing he lowes AICc alue. The goodness
o i o he bes o e all model was assessed by es ima ing he R
2
, which iden i ies he
p opo ion o a iabili y o he o iginal da a explained by he model [
47
]. Finally, we also
es ed he au oco ela ion o models esiduals [46].
2.6.2. Badge Densi y Es ima ion
We used he maximum numbe o di e en animals de ec ed in each main se (see
abo e) o es ima e he mean numbe o animals pe social g oup o he en i e s udy a ea.
We hen es ima ed badge densi y as:
Mean numbe o adul badge s pe main se es ima ed o he s udy a ea ×Es ima ed numbe o social g oups
Su eyed a ea 43 km2
Animals 2021,11, 2663 8 o 15
We assumed ha he s udied badge popula ion ollowed he beha io al pa e n
desc ibed o o he a eas o i s dis ibu ion ange, wi h each social g oup ha ing one main
se , used all yea ound and whe e ep oduc ion occu s [49].
3. Resul s
O a o al o 33 se s iden i ied in he a ea in he ame o p e ious s udies, only 22 s ill
occu ed (i.e., we e no des oyed). Twen y- h ee new se s we e added as a esul o he
ield wo k, o alizing 45 se s o moni o ing pu poses (Figu e 1B), esul ing in a se densi y
o 1.05 se s/km2.
3.1. D i e s o Se Use Pa e n
F om he 45 se s we de ec ed in he s udy a ea, we only used 34 in ou modeling p o-
cedu e, since 11 we e only de ec ed close o he end o he s udy no p o iding obus da a
ega ding hei use by badge s. No signi ican spa ial au oco ela ion was de ec ed (Mo an
I =
−
0.053; p= 0.822). F om all he candida e co- a ia es we excluded “G azz_07–17” and
“G azz”, since hey we e signi ican ly co ela ed wi h “G azz_16–17 ( G ass_07–17–G azz
= 0.77, p< 0.001; G ass_07–17–G azz_16–17 = 0.77, p< 0.001;
G azz–G azz_16–17 = 0.79
,
p< 0.001), and he la e was mo e co ela ed wi h he dependen a iable ( s = 0.258).
Fu he mo e, we also excluded “Dis _pi ” as i was signi ican ly co ela ed o “Dis _ol”
( = 0.97, p< 0.001), and he “Dis _ol” was mo e co ela ed o he equency o se use
( s = −0.237).
The i s wo compomen s o PCA analysis p oduced o he ege a ion associa ed
a iables eached a cumula i e explained a iance o 80.1% and hus we e used as candi-
da e a iables in models p ocuced o es Hypo hesis 2 (Vege a ion d i e s). The associa ion
be ween bo h componen s and he o iginal a iables is p esen ed in Table 3.
Table 3.
Co ela ion be ween he o iginal a iables and he PCA componen s used in he modeling
p ocedu e. Va iable desc ip ion is de ailed in Table 2.
O iginal Va iable Componen 1 Componen 2
T ee −0.431 -
Sh ubs −0.501 −0.398
He bs −0.507 −0.132
Sh ub_H −0.547 −0.271
Ba e_soil −0.114 0.865
We p oduced a o al o 91 models, bu only a ew we e conside ed he bes models
(i.e.,
∆
AICc < 2) o each hypo hesis: ou o H2, one o H3, wo o H4, h ee o
H5 and wo o H6 (Table 4). F om he a iables included in hese bes models, only
“Vege a ion di e si y”, “Dis ance o he nea es oli e ya ds”, ”Main soil ma e ial”, “Dis ance
o he nea es pa ed oad”, “Dis ance o he nea es di oad” and “G azing p essu e in
2016–2017
” had a signi ican in luence in he dependen a iable (p< 0.001) and hus we e
used as candida e a iables o he combined model (Table 4).
Models o he combined hypo hesis showed he highes i (i.e., lowe AICc and
O e all
∆
AICc < 2). Bes models included he a iables “Main soil ma e ial”, “Vege a ion
di e si y”, “Dis ance o he nea es di oad” and “G azing p essu e in 2016–2017”. Based
on he model a e aging p ocedu e we can s a e ha se s ha we e mo e used by bad-
ge s du ing he s udy pe iod we e loca ed in si es wi h a highe di e si y o ege a ion
(i.e., highe Simpson di e si y index), highe le els o g azing in he 2016–2017, and so e
soils (e.g., sil /clay soils we e a oided) espec o poo ly used se s (Table 5).
The bes model was able o explain ca. 40% o he a iance in he equency o se use
(R
2
= 0.397), indica ing an accep able goodness o i [
42
], and no spa ial au oco ela ion o
esiduals was de ec ed (Mo an I = −0.09; p= 0.55).
Animals 2021,11, 2663 9 o 15
Table 4.
Bes models o each es ed hypo hesis (
∆
AICc < 2). Fo each model we include he deg ee o eedom (d ), he
Akaike In o ma ion C i e ion, co ec ed o small samples (AICc), he AICc di e ence o he lowes AICc in each hypo hesis
(
∆
AICc), he Akaike weigh (Weigh ), and he AICc di e ence o he lowes AICc o all p oduced models (O e all
∆
AICc).
Va iables’ ac onyms a e desc ibed in Table 2.
Model d LogLik AICc ∆AICc Weigh O e all
∆AICc
Null 1 −87.721 169.6 0 - 11.0
H2—Vege a ion
Unde s o y + Veg_di + PCA1 6 −73.950 163.0 0 0.209 4.4
Unde s o y + Veg_di 5 −75.574 163.3 0.3 0.182 4.7
Veg_di 2 −70.981 164.3 1.3 0.107 5.7
Veg_di + PCA1 3 −78.977 164.8 1.8 0.088 6.2
H3—Soil Soil_ma 4 −75.027 159.4 0 0.781 0.8
H4—Food Dis _ol 2 −82.114 168.6 0 0.346 10.0
Dis _ol + Dis _wa 3 −81.902 170.6 2.0 0.128 12.0
H5—Dis u bance
Dis _ oad2 + G azz_16–17 3 −78.930 164.7 0 0.255 6.1
Dis _ oad2 + G azz_16–17 + N_g azz
4−78.370 166.1 1.4 0.123 7.5
G azz_16–17 2 −81.036 166.5 1.8 0.104 7.9
H6—Compe i ion Dis _se 2 −83.464 171.3 0 0.223 12.7
Wildboa 2 −83.715 171.8 0.5 0.174 13.2
Combined
hypo hesis
Soil_ma + Veg_di 5 −73.204 158.6 0 0.151 0
Dis _ oad2 + G azz_16-17 + Veg_di 4 −74.792 159.0 0.4 0.123 0.4
Soil_ma + G azz_16-17 + Veg_di 6 −72.095 159.3 0.7 0.104 0.7
Soil_ma 4 −75.027 159.4 0.8 0.097 0.8
Veg_di + G azz_16-17 3 −76.340 159.5 0.9 0.095 0.9
Table 5.
Va iables included in he a e age model o he bes models p oduced o explain he a ia ion in se use equency
(Combined hypo hesis—Table 4). Fo each a iable he coe icien (
β
), s anda d e o (SE), z- alue and signi icance (p), 95%
con idence in e als (95% CI) and ela i e impo ance a e p esen ed. (See Table 2 o a iables’ desc ip ion).
Va iable βSE z-Value p95% CI Rela i e
Impo ance
In e cep 0.719 0.601 1.175 0.240 [−0.480, 1.919] -
Soil_ma (Sil /clay) −0.740 0.244 2.919 0.004 [−1.237, −0.243] 0.62
Soil_ma (Rock) −0.115 0.246 0.450 0.653 [−0.618, 0.388] 0.62
Soil_ma (Roads) 0.264 0.403 0.629 0.530 [−0.559, 1.087] 0.62
Veg_di 1.989 0.862 2.122 0.034 [0.145, 3.651] 0.83
Dis _ oad2 −0.157 0.091 1.662 0.096 [−0.342, 0.028] 0.22
G azz_16-17 0.194 0.092 2.041 0.041 [0.007, 0.381] 0.56
3.2. Social G oups Size and Densi y Es ima es
We iden i ied 12 se s ha showed a highe and mo e con inuous ac i i y pa e ns
(Figu e 1C) and we e he e o e conside ed main se s. On a e age main se s included
3.85 ±3.32
( anging om 1–7) en ances, 3
±
2.2 o which we e ac i e simul aneously
(i.e., wi h oo p in s, la ines, e ol ed soil, bedding ma e ial; ange 1–6). Seconda y se s
had only an a e age o 1.41
±
1.66 ac i e en ances, anging om 0 o 4 (o en none showed
signs o ac i i y), om an o e all 2.93
±
2.80 o al en ances ( ange: 1–9). The index o se
use was highe o main se s (6.14 ±1.22) han o seconda y se s (3.37 ±2.23).
O e all, we manage o egis e badge s ac i i ies in main se s on a e age
150 ±26 nigh
pe se , ep esen ing 18,966 ideos o 30 seconds (i.e., 158 hou s o ideo). In 46% o he
ideos (n= 8654) we de ec ed a leas one badge , and in 7% (n= 1281) o he mammal
species: ca le, wild boa s, wild abbi (O yc olagus cuniculus), common gene (Gene a
gene a), s one ma en (Ma es oina), ed ox (Vulpes ulpes) and Egyp ian mongoose (He -