animals
A icle
Pa e ns and D i e s o Roden Abundance ac oss a Sou h
A ican Mul i-Use Landscape
Bea iz C. A onso 1,* , Lou ens H. Swanepoel 2, Bea iz P. Rosa 1, Tiago A. Ma ques 3,4, Luís M. Rosalino 1,
Ma ga ida San os-Reis 1and Gonçalo Cu ei a-San os 1


Ci a ion: A onso, B.C.; Swanepoel,
L.H.; Rosa, B.P.; Ma ques, T.A.;
Rosalino, L.M.; San os-Reis, M.;
Cu ei a-San os, G. Pa e ns and
D i e s o Roden Abundance ac oss
a Sou h A ican Mul i-Use Landscape.
Animals 2021,11, 2618. h ps://
doi.o g/10.3390/ani11092618
Academic Edi o : Emiliano Mo i
Recei ed: 2 Augus 2021
Accep ed: 31 Augus 2021
Published: 7 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/).
1
cE3c—Cen e o Ecology, E olu ion and En i onmen al Changes, Faculdade de Ciências da Uni e sidade de
Lisboa, Campo G ande, 1749-016 Lisboa, Po ugal; [email p o ec ed] (B.P.R.);
[email p o ec ed] (L.M.R.); mm [email p o ec ed] (M.S.-R.); [email p o ec ed] (G.C.-S.)
2Depa men o Zoology, School o Ma hema ical & Na u al Sciences, Uni e si y o Venda,
Thohoyandou 0950, Limpopo, Sou h A ica; [email p o ec ed]
3Cen e o Resea ch in o Ecological and En i onmen al Modelling, The Obse a o y, Uni e si y o S
And ews, S And ews KY16 9LZ, UK; iago.ma [email p o ec ed]
4Cen o de Es a ís ica e Aplicações, Depa amen o de Biologia Animal, Faculdade de Ciências, Uni e sidade
de Lisboa, 1749-016 Lisboa, Po ugal
*Co espondence: bea izca [email p o ec ed]
Simple Summa y:
Wildli e ecological pa e ns a e d i en no only by en i onmen al and biological
con ex s, bu also by landscape-managemen schemes ha shape hose con ex s. The p esen s udy
aims o de e mine he e ec o di e en en i onmen al ac o s (including managemen schemes)
on he occu ence pa e ns o a sou he n A ican small mammal communi y. Based on a landscape
whe e h ee land-use con ex s ha di e in hei le els o human p esence and/o whe e ac i i ies
coexis (p i a e eco ou ism ese e, mixed a ms and adi ional communal a eas), and by using
a body-size-based app oach (i.e., using wo size-based oden g oups—medium and small—as
models), we ound ha he mean ela i e abundance o medium-sized species did no di e ac oss
he managemen con ex s, bu small species’ mean ela i e abundance was highe in he game ese e.
The o e all a ia ion in oden abundance was nega i ely a ec ed by ungula e p esence (possibly
linked o a dec ease in ood a ailabili y) and by human p esence (inc eased dis u bance). Roden
abundance seems o be in luenced by en i onmen al g adien s ha a e di ec ly linked o a ying
managemen p io i ies ac oss land uses, meaning ha hese communi ies migh no bene i uni o mly
by he inc eased amoun o habi a p omo ed by he comme cial wildli e indus y.
Abs ac :
Sou h A ica’s decen alized app oach o conse a ion en ails ha wildli e ou side o mally
p o ec ed a eas inhabi complex mul i-use landscapes, whe e p i a e wildli e business (eco ou ism
and/o hun ing) co-exis in a human-domina ed landscape ma ix. Unde decen alized conse a ion,
wildli e is pe cei ed o bene i om inc eased amoun o a ailable habi a , howe e i is c ucial
o unde s and how dis inc managemen p io i ies and associa ed landscape modi ica ions impac
noncha isma ic axa, such as small mammals. We conduc ed ex ensi e ink- acking- unnel su eys
o es ima e he e ogenei y in oden dis ibu ion and in es iga e he e ec o di e en en i onmen al
ac o s on abundance pa e ns o wo size-based oden g oups (small- and medium-sized species),
ac oss h ee adjacen managemen con ex s in NE KwaZulu-Na al, Sou h A ica: a p i a e eco ou ism
game ese e, mixed a ms and adi ional communal a eas (consis ing o small clus e s o houses
in e spe sed wi h g azing a eas and semina u al ege a ion). Ou hypo heses we e o mula ed
ega ding he (1) a ea ypology, (2) ege a ion s uc u e, (3) ungula e p essu e and (4) human
dis u bance. Using a boos ed- eg ession- ee app oach, we ound conside able di e ences be ween
oden g oups’ abundance and dis ibu ion, and he unde lying en i onmen al ac o s. The mean
ela i e abundance o medium-sized species did no di e ac oss he h ee managemen con ex s,
bu small species mean ela i e abundance was highe in he game ese es, con i ming an in luence
o he a ea ypology on hei abundance. Va ia ion in oden ela i e abundance was nega i ely
co ela ed wi h human dis u bance and ungula e p esence. Roden abundance seems o be in luenced
by en i onmen al g adien s ha a e di ec ly linked o a ying managemen p io i ies ac oss land
Animals 2021,11, 2618. h ps://doi.o g/10.3390/ani11092618 h ps://www.mdpi.com/jou nal/animals
Animals 2021,11, 2618 2 o 18
uses, meaning ha hese communi ies migh no bene i uni o mly by he inc eased amoun o habi a
p omo ed by he comme cial wildli e indus y.
Keywo ds: non-in asi e sampling; ecological modelling; managemen op ions; conse a ion
1. In oduc ion
In Sou h A ica, ag icul u al in ensi ica ion, and o e g azing ha e led o p o ound
land use changes [
1
]. His o ically, mos landscapes we e con e ed in o li es ock a ms and
a mlands, ei he as in ensi e, ex ensi e, o communally managed a eas [
2
], leading o he
des uc ion, deg ada ion and/o agmen a ion o na u al ecosys ems [
3
]. Consequen ly,
such habi a des uc ion led o declines in wildli e popula ions and dis ibu ion in much o
Sou h A ican nonp o ec ed a eas [4].
Howe e , he es ablishmen o na ional policies a ibu ing cus odial igh s o e
wildli e o landowne s, p omp ed a ansi ion in he go e nance o na u al esou ces om
he s a e o p i a es [
5
]. This poli ical op ion led o widesp ead con e sion o angelands,
i.e., a mlands and li es ock a ms, in o a eas dedica ed o comme cial wildli e indus ies,
such as game anching and p i a e game/eco ou ism ese es [
6
]. The posi i e conse a-
ion ou comes o hese policies o economically aluable and cha isma ic species [
7
] is
belie ed o ha e an umb ella e ec on o he axa, mainly h ough he inc eased co e age,
ep esen a i eness and connec i i y o p o ec ed/ es o ed habi a s [
6
,
8
]. Howe e , he e -
ec o such managemen app oaches is unexplo ed o mos o e looked—bu unc ionally
impo an — axa, such as oden s [
9
]. Thus, in o ma ion on he ecological esponses o
less-cha isma ic axa is needed o be e gauge he complemen a y conse a ion ole o
Sou h A ica’s p i a e land.
In Sou h A ica, game a ms and p i a e game ese es o en coincide ac oss ela i ely
small scales, oo ed in human-domina ed landscapes (e.g., communal lands) [
10
]. These
land uses ha e con as ing managemen p io i ies and, consequen ly, dis inc impac s on
he landscape s uc u e and wildli e ecological pa e ns. In game a ms, he main objec i e
is o maximize he p oduc ion o ungula es o mea o hun ing, while in p i a e game
ese es he goal is o main ain cha isma ic species, p omo ing eco ou ism-based ac i i-
ies [
11
]. O en, hese wildli e-o ien ed land uses a e su ounded by human-domina ed
a eas wi h high le els o an h opogenic dis u bance. The egional co-exis ence o all hese
land uses gene a es complex mul i- enu ed landscapes, usually di ided by semi-pe meable
wildli e ences, in luencing he biodi e si y suppo ed by each o hese land uses [12].
Managemen ac ions di ec ed o cha isma ic o aluable species may ha e cascading
e ec s on oden s, usually o e looked and handled like pes s [
13
–
15
]. Howe e , i is c ucial
o unde s and he e ec o human-induced land-use changes on oden spa ial pa e ns,
as well as he unde lying ecological mechanisms he eo , since oden s a e undamen al
o some ecosys em unc ions [
16
]. Roden s a e p ima y consume s [
16
] and suppo
a la ge communi y o p eda o s [
17
,
18
], which makes hem a i al link in ood-chain
s uc u ing [
19
]. Mo eo e , hey a e conside ed use ul indica o s o ecosys em unc ioning
as hey a e aluable ools o he desc ip ion and moni o ing o habi a in eg i y. Fo hese
easons, oden s ha e been used as model species o unde s and how land use changes
a ec s wildli e [16].
Se e al ac o s ha e been iden i ied as in luen ial in shaping oden communi y and
popula ion s uc u es, many o which a e o en de e mined by he landscape managemen
op ions [
20
]. Some s udies ha e indica ed ha ege a ion ype and s uc u e a e undamen-
al d i e s o oden occu ence and abundance [
21
–
23
]. Fo example, a eas wi h g ea e
he baceous co e age a o oden s by p o iding shel e agains p eda o s, ood, and ade-
qua e mic oclima ic condi ions [
24
]. S udies ha e shown nega i e e ec s o o e g azing on
small mammals’ abundance, by educing he he baceous s a um, inc easing ampling isk
and eeding compe i ion wi h ungula es [
9
,
24
–
27
]. Rega ding oden dis ibu ion, i ends
Animals 2021,11, 2618 3 o 18
o be uni o m when he habi a is a o able and esou ces a e abundan . Howe e , when
dis u bances inc ease he le el o habi a he e ogenei y, causing landscape agmen a ion,
hei dis ibu ion is mos ly clumped [28,29].
Roden s a e no a homogeneous g oup, since di e en species may es ablish dis inc
ela ionships wi h he en i onmen al and bio ic componen s o he ecosys em. Fo example,
la ge oden s’ ange o e la ge spa ial scales han smalle oden s [
30
] and, he e o e,
a e mo e suscep ible o changes a his landscape le el [31].
Changes in managemen p io i ies ac oss Sou h A ican mul i- enu ed landscapes will
ha e a di ec impac on hese en i onmen al d i e s and, ul ima ely, in he dis ibu ion and
abundance o small mammal species ac oss and wi hin managemen con ex s. Fo ins ance,
when managemen measu es p omo e he abundance o ungula es (e.g., as p ey o la ge
ca ni o e popula ions in eco ou ism ese es, o as hun ing asse s in game a ms), g azing
p essu e will inc ease, nega i ely in luencing he he baceous s a a [
26
]. Alongside wi h
long d y and ho seasons [
32
], hese condi ions may lead o sh ub enc oachmen , known o
educe ood a ailabili y (lea es, seeds, and a h opods) o g ound dwelling oden s [
33
].
Ne e heless, some oden species a e usually conside ed e icien colonize s o human
shaped en i onmen s [
13
,
34
], as hey a e able o use human- ela ed ood esou ces due o
hei omni o e cha ac e [35].
Al hough he p ocesses ha egula e small mammals’ spa ial dis ibu ion a e known
o some landscapes (e.g., woodland [
29
] and mixed o es [
36
]), he e is a lack o in o ma-
ion ega ding he d i e s o oden -abundance pa e ns in A ican sa annas (bu see [
9
,
37
]),
as well as how hese a y ac oss di e en managemen schemes. He e, we e alua ed he
a ia ion in oden abundance ac oss h ee adjacen managemen con ex s, spanning a
p i a e eco ou ism game ese e, mixed a ms and communally owned land, managed
by Zulu ibal au ho i ies [
12
], unde he ollowing wo main objec i es: (1) o es ima e
he e ogenei y in small-mammal-abundance dis ibu ion (mean abundance and pa chiness)
ac oss managemen con ex s (game ese e, mixed a ms and communal lands); and (2) o
de e mine he main, ine-scale en i onmen al ac o s a ec ing small-mammal-abundance
pa e ns ac oss land-use ypes. These objec i es we e es ed in wo size-based oden
g oups, o a mo e de ailed assessmen o ecological esponses.
Linked o hese wo goals, we es ed ou hypo he ical d i e s o oden communi ies:
(i)
An a ea- ypology hypo hesis, i.e., cumula i e e ec o managemen -induced changes
o ege a ion, g azing p essu e, e c., c ea es a ea-speci ic di e ences in oden abun-
dance. Pa chiness will also be es ed o acknowledge in which a ea each g oup is
mo e o less clumped, ega ding hei abundance alues. Al hough he exac e ec o
a ea on oden abundance is no ully p edic able [
37
] (gi en he dis u bance g adien )
we expec ed he communal lands o ha e he lowes alues o abundance and highes
pa chiness (i.e., mo e clumped), ollowed by mixed a ms and he game ese e, wi h
highe abundances and lowe pa chiness;
(ii)
A ege a ion-s uc u e hypo hesis, i.e., a eas wi h highe he baceous co e will ha e
a posi i e in luence on bo h oden size-based g oups, since i shapes he abili y o
he landscape o p o ide p o ec ion agains po en ial p eda o s [21–23,25,27,38];
(iii)
An ungula e-p essu e hypo hesis, i.e., oden species abundance is nega i ely in lu-
enced by he abundance o ungula es, since highe g azing p essu e ends o dec ease
he baceous land co e , inc ease dis u bance due o he ampling e ec , and inc ease
landscape agmen a ion [9,24];
(i )
A human-dis u bance hypo hesis, i.e., oden species’ dis ibu ion is nega i ely in lu-
enced by human dis u bance ac o s, such as he p esence o domes ic animals and
households ha may cons ain species’ p esence [14,39].
2. Ma e ials and Me hods
2.1. S udy A ea
This s udy was implemen ed in he Mapu aland–Pondoland–Albany Biodi e si y
Ho spo [
40
] in no he n KwaZulu-Na al, Sou h A ica. Ou speci ic s udy a ea is cha -
Animals 2021,11, 2618 4 o 18
ac e ized by a spa ial g adien o human in e en ion, anging om he Mun-ya-wana
p i a e game ese e (less subjec o human associa ed ac i i ies), o mixed game a ms
and o communally managed lands, whe e wo dis inc Zulu communi ies a e se led
(Figu e 1b). The Mun-ya-wana p i a e game ese e (27
◦
40
0
S–27
◦
55
0
S; 31
◦
12
0
E–32
◦
26
0
E)
ep esen s he union o se e al p ope ies wi hou in e nal ences, managed by p i a e
owne s whose goal is o explo e eco- ou is ic p oduc s, he e o e p omo ing wildli e and
habi a conse a ion. Those managemen objec i es a e commonly ela ed wi h a mo e
sus ainable use o wildli e, ypically wildli e- iewing ou ism [
41
]. The ese e is su -
ounded, o he Sou h, by a mosaic o comme cial game anches o he p oduc ion o wild
ungula e species, occasionally mixed wi h domes ic ca le [
42
] (he ea e mixed a ms) and
ep esen s la ge expanses o na u al habi a wi h low human densi y. Communal lands o
he eas a e composed o households, in e spe sed wi h pas u e a eas and semi-na u al
ege a ion. The egion is cha ac e ized by a wa m- empe a u e clima e, wi h a humid
and ho summe (Oc obe o Ap il), acco ding o he Köppen–Geige classi ica ion. Mean
mon hly empe a u es ange om 19
◦
C in July o 31
◦
C in Janua y, and he a e age annual
p ecipi a ion is 800 mm [
43
,
44
]. Ele a ion anges om 3 m o 304 m abo e sea le el [
45
],
domina ed by a simila mix u e o ege a ion h oughou he a ea (bush eld, woodland
and g assland) [
46
] (Figu e 1b). Ne e heless, he game ese e hos s a highe di e si y and
abundance o p is ine habi a s, such as indigenous o es s, while mixed a ms a e mainly
composed o pas u e a eas (low sh ubland and g assland–Figu e 1). Con a ily, communal
lands ha e he lowes p opo ion o ege a ion and he highes co e o u ban– illage
occupa ion (Figu e 1).
2.2. Roden Sampling
Roden s we e sampled be ween Oc obe and No embe 2017 ( he sou he n hemi-
sphe e’s sp ing) using ink- acking unnels [
42
], le ac i e in he ield o ou consecu i e
nigh s (open ci cles in Figu e 1c). Ink- acking unnels we e made o obus co uga ed
plas ic (55
×
10
×
10 cm), open on bo h ends o allow oden s o en e . Bo h en ances o
he unnel a e equipped wi h an adhesi e pape wi h he glue side up, and an ink pad
(
12 ×10 cm
) was placed in he loo cen e [
47
] (Figu e S1B). In he middle o he unnel,
a small PVC-pipe sec ion, hanging om he ceiling, was ins alled, and con ained bai
composed o a mix u e o peanu bu e , oa meal and sun lowe oil [
46
]. The pipe was
used o p e en he consump ion o he bai by he animals en e ing/c ossing he unnel.
The ink unnels we e placed on he g ound, g ouped in clus e s o nine, in a Y o ma ion,
10 m apa om each o he (Figu e 1c). The a ms o he Y o ma ion we e 120 deg ees
apa (
Figu e 1c
). This design p o ided an adequa e spa ial co e age in ela ion o he
home- anges o he oden species, also ensu ing some le el o independence be ween
sampling uni s, conside ing he mean dis ance be ween si es (see below). A e he ou -
day sampling pe iod, he pla es o each ink unnel (con aining oo p in s and acks) we e
pho og aphed indi idually, always a he same dis ance and wi h a e e ence scale.
The oo p in da a was used o es ima e oden ela i e abundance, using he p opo -
ion o he unnels wi h eco ds ( ack index; TI– o mo e de ails see Supplemen a y Ma e i-
als) [
48
]. To ensu e ha his app oach cap u ed spa ial he e ogenei y in ela i e abundance,
we conduc ed a small ial, compa ing he abundance indices de i ed om ink unnels
o hose ob ained om li e- apping (see Supplemen a y Ma e ials, PART A). As ack
iden i ica ion a he species le el is e y ime consuming and no iable in la ge-scale
s udies, and as dis inguishing oo p in s om simila -sized species is e y di icul and
bias p one, we op ed o di iding acks in o g oups based on ack size ( o mo e de ails
see Supplemen a y Ma e ials PART A; Figu es S1A and S2A, Table S1A). Roden oo p in s
we e g ouped in o h ee di e en size-based g oups pe body leng h/weigh , assuming
a ela ion be ween oden body leng h/weigh and oo p in sizes [
49
,
50
]: small (body
leng h: 50–100 mm), medium (100–150 mm) and la ge oden s (150–200 mm) (
Figu e S2B
).
Sampling in en ionally ook place ou side he b eeding season (which peaks in he we
season, [
51
]), in o de o a oid g ouping ju eniles in he w ong size-based g oup. Howe e ,
Animals 2021,11, 2618 5 o 18
conside ing he low numbe o de ec ions o la ge oden s in ink- acking unnels, we only
analyzed he da a om small- and medium-sized oden s (see Resul s). The mos common
species cap u ed du ing li e apping and linked o each g oup we e Mus minu oides and
Dend omus melano is o small oden s, Mas omys na alensis and Saccos omus campes is o
medium oden s and O omys angoniensis and Ra us a us o la ge oden s (Table S2A).
Animals 2021, 11, x 5 o 18
2. Ma e ials and Me hods
2.1. S udy A ea
This s udy was implemen ed in he Mapu aland–Pondoland–Albany Biodi e si y
Ho spo [40] in no he n KwaZulu-Na al, Sou h A ica. Ou speci ic s udy a ea is cha ac-
e ized by a spa ial g adien o human in e en ion, anging om he Mun-ya-wana p i-
a e game ese e (less subjec o human associa ed ac i i ies), o mixed game a ms and
o communally managed lands, whe e wo dis inc Zulu communi ies a e se led (Figu e
1b). The Mun-ya-wana p i a e game ese e (27°40′ S–27°55′ S; 31°12′ E–32°26′ E) ep e-
sen s he union o se e al p ope ies wi hou in e nal ences, managed by p i a e owne s
whose goal is o explo e eco- ou is ic p oduc s, he e o e p omo ing wildli e and habi a
conse a ion. Those managemen objec i es a e commonly ela ed wi h a mo e sus aina-
ble use o wildli e, ypically wildli e- iewing ou ism [41]. The ese e is su ounded, o
he Sou h, by a mosaic o comme cial game anches o he p oduc ion o wild ungula e
species, occasionally mixed wi h domes ic ca le [42] (he ea e mixed a ms) and ep e-
sen s la ge expanses o na u al habi a wi h low human densi y. Communal lands o he
eas a e composed o households, in e spe sed wi h pas u e a eas and semi-na u al eg-
e a ion. The egion is cha ac e ized by a wa m- empe a u e clima e, wi h a humid and ho
summe (Oc obe o Ap il), acco ding o he Köppen–Geige classi ica ion. Mean mon hly
empe a u es ange om 19 °C in July o 31 °C in Janua y, and he a e age annual p ecip-
i a ion is 800 mm [43,44]. Ele a ion anges om 3 m o 304 m abo e sea le el [45], domi-
na ed by a simila mix u e o ege a ion h oughou he a ea (bush eld, woodland and
g assland) [46] (Figu e 1b). Ne e heless, he game ese e hos s a highe di e si y and
abundance o p is ine habi a s , such as indigenous o es s, while mixed a ms a e mainly
composed o pas u e a eas (low sh ubland and g assland–Figu e 1). Con a ily, commu-
nal lands ha e he lowes p opo ion o ege a ion and he highes co e o u ban– illage
occupa ion (Figu e 1).
Figu e 1.
Loca ion o he s udy a ea in Sou h A ica, wi h he black do ep esen ing he loca ion
o he s udy a ea in he Mapu aland egion o no he n KwaZulu-Na al (
a
); landscape composi ion
o he h ee s udied a eas wi h dis inc managemen schemes–Mun-ya-wana p i a e game ese e,
mixed a ms and communal land (Zulu ibal land)–wi h he loca ion o he sampling poin s and he
numbe o sampling poin s pe a ea (in pa en hesis) (
b
); each sampling poin included a came a ap
in he cen e and nine ink unnels, dis ibu ed in a Y shape (open ci cles ep esen ink unnels) (c).
2.3. En i onmen al Va iables
Vege a ion s uc u e a iables we e collec ed using wo di e en app oaches: ield
measu es and emo e-sensed p oduc s [
52
]. All a iables collec ed ha e been p e iously
de ec ed as in luen ial o oden p esence elsewhe e (e.g., ege a ion co e ) [
21
,
23
]. Sh ub-
and-g ass co e we e isually es ima ed and assigned he co esponding Edwa ds classi i-
ca ion ca ego y [
53
] (see Table 1 o de ails), wi hin a 30 m adius bu e , cen e ed on he ink
unnel’s Y o ma ion. Rega ding he land use, he p edominan ca ego ies we e selec ed
( hicke , g assland, sand o es and u ban illages) and, o each bu e , was assigned
he ca ego y wi h he highes co e . Acco ding o he ype o c ops p esen in he s udy
a ea, he ha es ing season occu s mos ly be ween Ap il and June [
54
], no coinciding
wi h he s udy pe iod. The e o e, we assumed ha he e would be no in luence o c op
p oduc i i y on he dis ibu ion/abundance o oden s in ou s udy. The pe cen age o
ee co e was assessed based on he Global Fo es Wa ch da abase (Table 1). We also
selec ed he No malized Di e ence Vege a ion Index (NDVI), widely used as a ege a ion
p oduc i i y p oxy, collec ed om Landsa 8 Images [55].

Animals 2021,11, 2618 6 o 18
Va iables o ungula e p essu e and human dis u bance we e collec ed om Cu ei a-
San os e al. [
12
] came a- ap su eys. Came as, loca ed in he cen e o he Y o ma ion,
we e ac i e o 60–90 days, and a ached o a ee o me al s ake, 30 cm abo e he g ound,
wi hou any bai and se o pho og aph a minimum delay (1 s o day ime and 30 s o
nigh - ime) (see [
12
] o de ails). Each o he de ined ink- unnel clus e s (i.e., one clus e
includes nine ink unnels and one came a- ap; Figu e 1c) we e spaced app oxima ely
1.4 km apa (Figu e 1b). In o al, we e sampled 196 poin s: 100 poin s in Mun-ya-wana
eco- ou ism/game ese e, 50 poin s in mixed a ms and 46 poin s in communal lands.
Cap u e a es, exp essed as he numbe o independen came a eco ds (>1 h in e al
be ween pho og aphs o he same species, pe 100 ap-days) o li es ock (cows and goa s),
wild ungula es and human dis u bance, we e used as su oga es o dis u bance in he
modeling p ocedu e (Table 1).
Wild ungula es we e g ouped acco ding o wo c i e ia: weigh , since ampling is
one o he main nega i e impac s o ungula es on oden s [
26
], and/o he ac ha hey
a e ac i ely managed in all s udied a eas (Table S1B). Only ungula es weighing be ween
45–200 kg and ac i ely managed we e used in he analysis, since hey a e mo e abundan
han o he ungula es, as hey a e p esen h oughou he a eas unde s udy, and because
hey ha e a g ea e impac on oden s, due o hei weigh (Table S1B). Li es ock we e also
sepa a ed in wo weigh classes: i.e., goa s and cows.
Table 1.
En i onmen al a iables used in he modeling p ocedu e used o assess he de e minan s o oden abundance,
collec ed in he ield, om came a- apping o based on emo e-sensing da a (GIS-based a iables). The a iable desc ip ion,
ac onym, ange, esolu ion and sou ce, as well he e e ence ha suppo hei in luence on oden p esence/abundance, a e
lis ed. H1—Hypo hesis 1; H2—Hypo hesis 2; H3—Hypo hesis 3, H4—Hypo hesis 4.
Va iable
Ac onym Desc ip ion Mean/Range Resolu ion Sou ce Suppo ing
Re e ences
AREA TYPE (H1)
A ea Managmen con ex
Mixed a ms
Mun-ya-wana
Communal lands
Collec ed a poin - [37]
VEGETATION STUCTURE (H2)
T ee_Co e % T ee Co e 30.80/6–72% 30 ×30 m
Global Fo es Wa ch
h ps://www.
global o es wa ch.o g/
(16 Ap il 2019)
[27,56]
Sh ub_Co e % o Sh ub co e
Con inuous (C)—76–100%
Semi-con inuous
(SC)—51–74%
Mode a ed closed
(MC)—26–50%
Semi-open (SO)—11–25%
Open (O)—0–10%
30 m bu e Visually es ima ed [22,27,39,57–60]
G ass_Co e % o G ass co e
Con inuous (C)—76–100%
Semi-con inuous
(SC)—51–74%
Mode a ed closed
(MC)—26–50%
Semi-open (SO)—11–25%
Open (O)—0–10%
30 m bu e Visually es ima ed [22,25,38,57]
Land_use Land use ca ego ies
Thicke
G assland
Sand Fo es
U ban Villages
30 m bu e
2013–2014 Na ional Land
Co e Sou h A ica-SASDI
h p://www.sasdi.ne /
(16 Ap il 2019)
[21–23]
NDVI
No malized
di e ence ege a ion
index calcula ed om
Landsa images
0.48/0.28–0.67 30 ×30 m
Landsa 8
h ps:
//ea hexplo e .usgs.go /
(18 Ap il 2019)
[60,61]
Animals 2021,11, 2618 7 o 18
Table 1. Con .
Va iable
Ac onym Desc ip ion Mean/Range Resolu ion Sou ce Suppo ing
Re e ences
UNGULATE PRESSURE (H3)
Goa s
Cap u e a e o goa s
(numbe o eco ds
pe 100 days o
apping)
0.16/0–1.88 Collec ed a poin Came a- apping su ey
[9,24,26]
Li es ock
Cap u e a e o cows
(numbe o eco ds
pe 100 days o
apping)
0.20/0–3.17 Collec ed a poin Came a- apping su ey
Wild
Ungula es
Cap u e a e o
ungula es (numbe o
eco ds pe 100 days
o apping)
0.750/0–3.48 Collec ed a poin Came a- apping su ey
DISTURBANCE VARIABLES (H4)
HUMANS Cap u e a e o
humans 0.84/0–10 Collec ed a poin Came a- apping su ey
[39]
DIST Dis ance o houses 2.738/0.031–9.867 km Collec ed a poin Came a- apping su ey
2.4. Da a Analyses/Modelling
2.4.1. Spa ial Pa e ns o Roden Rela i e Abundance Ac oss A eas and Size-Based G oups
Di e ences in mean abundance alues o size-based g oups (small and medium)
be ween s udy a eas (Mun-ya-wana game ese e, mixed a ms and communal lands) we e
es ed using GLM wi h 3-le el a ea co a ia e and binomial e o dis ibu ion. The magni-
ude o pa chiness in each a ea was asce ained by spa ial-poin pa e n analysis o coun
da a using Lloyd’s index o pa chiness [
62
]. A Lloyd’s index o 1 indica es a andom
dis ibu ion, whils one <1 sugges s uni o mi y and >1 pa chiness.
2.4.2. In luence o En i onmen al Va iables on Roden Rela i e Abundance
Due o he high numbe o candida e a iables and o a oid mul icollinea i y bias,
we i s es ima ed he nonpa ame ic Spea man’s co ela ion (
s
) using he “psych” R
package [
63
]. When a high co ela ion be ween wo co a ia es was de ec ed (
s≥
0.7; [
62
]),
he a iable ha was less co ela ed wi h he dependen a iable was excluded om he
analysis [64].
The in luence o all candida e a iables on oden ela i e abundance was es ed
using a boos ed- eg ession- ee (BRT) app oach, implemen ed wi h he “gbm” package [
65
]
in R [
66
,
67
]. This modelling echnique encompasses he ad an ages o eg ession ees
(e.g., p edic o a iables can be o any ype, analysis is insensi i e o ou lie s and can
accommoda e missing da a [
68
]), o e coming hei low p edic i e capaci y h ough he
boos ing algo i hm [
69
]. The inal model is a linea addi ion o se e al eg ession models
in which he simples e m is a ee [68,70].
Boos ed- eg ession- ee models a e esilien o model o e i ing bu , o ha e a be -
e p edic i e pe o mance, we de ined, a p io i, he model’s inpu pa ame e s based on
Ca slaw and Taylo ’s sugges ions [
70
]. In BRT, lea ning a e (l ) is he sh inkage pa ame e
ha con ols he con ibu ion o each ee o he model, and ee complexi y ( c) de e mines
he numbe o nodes in a ee and, consequen ly, i s size. These wo pa ame e s con ol he
numbe o ees in he model, while he bag ac ion (0.5) selec s he p opo ion o da a be-
ing used a each s ep [
61
,
70
,
71
]. All models we e i ed o allow in e ac ions using a en- old
c oss alida ion o de e mine he op imal numbe o ees o each model. The la ges lea n-
ing a e and he smalles ee complexi y we e selec ed o allow a minimum o 1000 ees
in he BRT i ing p ocess (see [
68
]). Non-in o ma i e a iables we e emo ed du ing he
i ing p ocess, allowing he simpli ica ion o he se o a iables [
68
]. This simpli ica ion
consis ed o de ining how many a iables he unc ion can es o emo e, based on ela i e
Animals 2021,11, 2618 8 o 18
in luence and o al numbe o a iables. Then, a g aph was p oduced showing di e ences
in he p edic ed de iance acco ding o se e al scena ios, each one wi h a di e en numbe
o a iables emo ed. Nex , he numbe o a iables o elimina e was decided, and hey
we e emo ed in o de o mino ela i e in luence. We de ined a h eshold alue and only
epo ed he in e ac ions wi h ela i e in luence alues >10%. The inal ela i e in luence
o each a iable was calcula ed by a e aging he numbe o imes a co a ia e is used o
spli ing, weigh ed by he squa ed imp o emen o he model as he esul o each spli .
I is hen scaled, such ha he alues sum o 100 [
72
]. Fi ed alues we e plo ed in ela ion
o he mos impo an p edic o s, e ealing hei e ec s on oden abundance. Explained
de iance was calcula ed using he ollowing o mula om Abea e (2009) [73]
D2=1− esidual de iance
o al de iance 
The 95% con idence in e als o each a iable we e es ima ed o he i ed unc ion
by aking 500 boo s ap samples o he inpu da a, wi h he same size as he o iginal da a.
A BRT was i ed o each sample, and he 5 h and 95 h pe cen iles we e calcula ed o
he poin s o each unc ion. Models we e buil sepa a ely o small- and medium-sized
oden s. Fo each model pe o med, in e ac ions be ween ypology and he o he in luen ial
independen a iables (i.e., ela i e impo ance abo e >10%) we e es ima ed, o e alua e
con ex -dependency in he in luence in he e ec en i onmen al a iable associa ed wi h
he managemen con ex . All analyses we e implemen ed in R ia R S udio Ve sion
1.1.463 [66,67].
3. Resul s
3.1. Spa ial Pa e ns o Roden Abundance Ac oss A eas and Size-Based G oups
F om he 192 sampling poin s moni o ed, 85% p esen ed small oden acks, while
76% de ec ed he occu ence o medium oden s, wi h an o e lap in 35% o si es and
in e -a ea a ia ion in de ec ion (i.e., numbe o unnels wi h signs/ o al numbe o unnels,
Table S2B). Mean abundance in Mun-ya-wana game ese e was 0.52
±
0.26 (mean
±
SD)
o small oden s and 0.43
±
0.34 o medium oden s; in mixed a ms, 0.31
±
0.21 o small
oden s and 0.52
±
0.32 o medium oden s; and in communal lands was 0.26
±
0.23 o
small oden s and 0.36
±
0.24 o medium oden s (Figu e 2). Rega ding he GLM esul
o size-based g oups, i e ealed signi ican di e ences in ela i e abundances only o
small oden s, be ween Mun-ya-wana game ese e and he emaining a eas (Table S3B,
Supplemen a y Ma e ials). No signi ican di e ences we e de ec ed in ela i e abundances
o medium oden s be ween a eas (Figu e 3). Be ween g oups, signi ican di e ences we e
only ound in mixed a ms (Table S3B, Supplemen a y Ma e ials), wi h medium oden s
being mo e abundan (0.52
±
0.37) han small-size oden s (0.31
±
0.26) (Figu e 3). Based
on hese esul s, he e ec o en i onmen al d i e s on oden abundance was e alua ed
sepa a ely o each o he size-based g oups.
Roden Pa chiness
Lloyd’s Index o Pa chiness e ealed ha o e e y a ea and size-based g oup, all abun-
dance alues we e agg ega ed (
γ
> 1; Table 2). Bo h medium and small oden s a e he e oge-
neously dis ibu ed wi hin he h ee s udy a eas (Figu e 3), demons a ing a he e ogenei y
g adien . Acco ding o Table 2, we can obse e ha he highes alues o small oden s a e
in communal lands, ollowed by mixed a ms and inally, he game ese e. Fo medium
oden s, he e is a g ea e clus e ing pa e n in he game ese e, ollowed by communal
lands and mixed a ms. Wi h hese esul s, i is possible o s a e ha he abundance pa e ns
di e be ween he size-based g oups, and wi hin each a ea.
Animals 2021,11, 2618 9 o 18
Animals 2021, 11, x 10 o 18
Figu e 2. Map o he s udy a ea showing oden dis ibu ions: small-size oden s a e in o ange and medium-size oden s
in yellow. The size o each poin is equi alen o abundance alue, as indica ed in he espec i e legend.
Figu e 3. Boxplo o medium and small oden s’ ela i e abundance in he h ee managemen - ype
zones moni o ed: game (mixed) a ms, Mun-ya-wana game ese e and communal lands. Based on
he GLM es , * indica es a signi ican di e ence be ween size-based g oups in mixed a ms (p =
0.011), + indica es a signi ican di e ence be ween Mun-ya-wana game ese e and emaining a eas
o small oden s (p = 0.016).
Figu e 2.
Map o he s udy a ea showing oden dis ibu ions: small-size oden s a e in o ange and medium-size oden s in
yellow. The size o each poin is equi alen o abundance alue, as indica ed in he espec i e legend.
Animals 2021, 11, x 10 o 18
Figu e 2. Map o he s udy a ea showing oden dis ibu ions: small-size oden s a e in o ange and medium-size oden s
in yellow. The size o each poin is equi alen o abundance alue, as indica ed in he espec i e legend.
Figu e 3. Boxplo o medium and small oden s’ ela i e abundance in he h ee managemen - ype
zones moni o ed: game (mixed) a ms, Mun-ya-wana game ese e and communal lands. Based on
he GLM es , * indica es a signi ican di e ence be ween size-based g oups in mixed a ms (p =
0.011), + indica es a signi ican di e ence be ween Mun-ya-wana game ese e and emaining a eas
o small oden s (p = 0.016).
Figu e 3.
Boxplo o medium and small oden s’ ela i e abundance in he h ee managemen - ype
zones moni o ed: game (mixed) a ms, Mun-ya-wana game ese e and communal lands. Based
on he GLM es , * indica es a signi ican di e ence be ween size-based g oups in mixed a ms
(
p= 0.011
), + indica es a signi ican di e ence be ween Mun-ya-wana game ese e and emaining
a eas o small oden s (p= 0.016).
Animals 2021,11, 2618 16 o 18
22.
Layme, V.M.G.; Lima, A.P.; Magnusson, W.E. E ec s o i e, ood a ailabili y and ege a ion on he dis ibu ion o he oden
Bolomys lasiu us in an Amazonian sa anna. J. T op. Ecol. 2004,20, 183–187. [C ossRe ]
23.
Holland, G.J.; Benne , A.F. Di e ing esponses o landscape change: Implica ions o small mammal assemblages in o es
agmen s. Biodi e s. Conse . 2009,18, 2997–3016. [C ossRe ]
24.
Ho mann, A.; Zelle , U. In luence o a ia ions in land use in ensi y on species di e si y and abundance o small mammals in he
Nama Ka oo, Namibia. Belg. J. Zool. 2005,135, 91–96.
25.
Bond, W.; Fe guson, M.; Fo sy h, G. Small mammals and habi a s uc u e along al i udinal g adien s in he sou he n Cape
moun ains. A . Zool. 1980,15, 34–43. [C ossRe ]
26.
Keesing, F. Impac s o ungula es on he demog aphy and di e si y o small mammals in cen al Kenya. Oecologia
1998
,116,
381–389. [C ossRe ]
27.
Delc os, G.; Taylo , P.J.; Schoeman, M.C. Ecological co ela es o small mammal assemblage s uc u e a di e en spa ial scales in
he sa annah biome o Sou h A ica. Mammalia 2015,79, 1–14. [C ossRe ]
28.
Massawe, A.W.; Lei s, H.; Rwamugi a, W.P.; Makundi, R.H. E ec o land p epa a ion me hods on spa ial dis ibu ion o oden s
in c op ields. In Book Ra s, Mice and People: Roden Biology and Managemen , 1s ed.; Aus alian Cen e o In e na ional Ag icul u al
Resea ch: Cambe a, Aus alia, 2003.
29.
Mo elli i, A.; Amo i, G.; Annesi, F.; Boi ani, L. Tes ing o he ela i e con ibu ion o pa ch neighbo hood, pa ch in e nal
s uc u e, and p esence o p eda o s and compe i o species in de e mining dis ibu ion pa e ns o oden s in a agmen ed
landscape. Can. J. Zool. 2009,87, 662–670. [C ossRe ]
30.
Su he land, G.D.; Ha es ad, A.S.; P ice, K.; Le zman, K.P. Scaling o na al dispe sal dis ances in e es ial bi ds and mammals.
Conse . Ecol. 2000,4, 16. [C ossRe ]
31.
Paci ici, M.; Rondinini, C.; Rhodes, J.R.; Bu bidge, A.A.; C is iano, A.; Wa son, J.E.M.; Woina ski, J.C.Z.; Di Ma co, M. Global
co ela es o ange con ac ions and expansions in e es ial mammals. Na . Commun. 2020,11, 2840. [C ossRe ] [PubMed]
32.
Roques, K.G.; O’conno , T.G.; Wa kinson, A.R. Dynamics o sh ub enc oachmen in an A ican sa anna: Rela i e in luences o
i e, he bi o y, ain all and densi y dependence. J. Appl. Ecol. 2001,38, 268–280. [C ossRe ]
33.
Blaum, N.; Rossmani h, E.; Jel sch, F. Land use a ec s oden communi ies in Kalaha i sa annah angelands. A . J. Ecol.
2007
,45,
189–195. [C ossRe ]
34.
Hu s , Z.M.; McClee y, R.A.; Collie , B.A.; Sil y, N.J.; Taylo , P.J.; Monadjem, A. Linking changes in small mammal communi ies
o ecosys em unc ions in an ag icul u al landscape. Mamm. Biol. 2014,79, 17–23. [C ossRe ]
35.
Mulungu, L.S.; Mahlaba, T.A.; Massawe, A.W.; Kennis, J.; C auwels, D.; Eiseb, S.; Monadjem, A.; Makundi, R.H.; Ka akweba,
A.A.S.; Lei s, H.; e al. Die a y di e ences o he mul imamma e mouse, Mas omys na alensis (Smi h, 1834), ac oss di e en
habi a s and seasons in Tanzania and Swaziland. Wildl. Res. 2011,38, 640–646. [C ossRe ]
36.
Dalmag o, A.D.; Viei a, E.M. Pa e ns o habi a u iliza ion o small oden s in an a ea o A auca ia o es in Sou he n B azil.
Aus al Ecol. 2005,30, 353–362. [C ossRe ]
37.
Ca o, T.M. Species ichness and abundance o small mammals inside and ou side an A ican na ional pa k. Biol. Conse .
2001
,98,
251–257. [C ossRe ]
38. Monadjem, A. Habi a p e e ences and biomasses o small mammals in Swaziland. A . J. Ecol. 1997,35, 64–72. [C ossRe ]
39.
Duns an, C.; Fox, B. The e ec s o agmen a ion and dis u bance o ain o es on g ound-dwelling small mammals on he
Robe son Pla eau, New Sou h Wales, Aus alia. J. Biogeog . 1996,23, 187–201. [C ossRe ]
40.
Balme, G.A.; Slo ow, R.O.B.; Hun e , L.T. Edge e ec s and he impac o non-p o ec ed a eas in ca ni o e conse a ion: Leopa ds
in he Phinda–Mkhuze Complex, Sou h A ica. Anim. Conse . 2010,13, 315–323. [C ossRe ]
41.
K ug, W. P i a e supply o p o ec ed land in sou he n A ica: A e iew o ma ke s, app oaches, ba ie s and issues. In Wo kshop
Pape , Wo ld Bank/OECD In e na ional Wo kshop on Ma ke C ea ion o Biodi e si y P oduc s and Se ices, Pa is; OECD: Pa is, F ance,
2001; Volume 25.
42.
King, C.M.; Edga , R.L. Techniques o apping and acking s oa s (Mus ela e minea); a e iew, and a new sys em. N. Z. J. Zool.
1977,4, 193–212. [C ossRe ]
43.
Fick, S.E.; Hijmans, R.J. Wo ldClim 2: New 1km spa ial esolu ion clima e su aces o global land a eas. In . J. Clima ol.
2017
,37,
4302–4315. [C ossRe ]
44.
Ha is, I.; Jones, P.D.; Osbo n, T.J.; Lis e , D.H. Upda ed high- esolu ion g ids o mon hly clima ic obse a ions—The CRU TS3.10
Da ase . In . J. Clima ol. 2014,34, 623–642. [C ossRe ]
45. ASTER GDEM Da a Ve sion 2. A ailable online: h ps://sea ch.ea hda a.nasa.go /sea ch/ (accessed on 27 Ap il 2019).
46.
Depa men o En i onmen al A ai s (DEA). 2013-14 SA Na ional Land-Co e —B oad Pa en Classes. A ailable online:
h p://www.sasdi.ne /me a iew.aspx?uuid=20c6336b75031c73041ec8b88372c3b0 (accessed on 27 Ap il 2019).
47.
Glennon, M.J.; Po e , W.F.; Deme s, C.L. An al e na i e ield echnique o es ima ing di e si y o small-mammal popula ions.
J. Mammal. 2002,83, 734–742. [C ossRe ]
48.
Hughes, J.J.; Wa d, D.; Pe in, M.R. P eda ion isk and compe i ion a ec habi a selec ion and ac i i y o Namib Dese ge bils.
Ecology 1994,75, 1397–1405. [C ossRe ]
49.
Wilkinson, E.B.; B anch, L.C.; Mille , D.L.; Go e, J.A. Use o ack ubes o de ec changes in abundance o beach mice. J. Mammal.
2012,93, 791–798. [C ossRe ]

Animals 2021,11, 2618 17 o 18
50.
Palma, A.R.T.; Gu gel-Gonçal es, R. Mo phome ic iden i ica ion o small mammal oo p in s om ink acking unnels in he
B azilian Ce ado. Re . B as. Zool. 2007,24, 333–343. [C ossRe ]
51.
G eene, D.U.; Oddy, D.M.; Go e, J.A.; Gillikin, M.N.; E ans, E.; Gann, S.L.; Leone, E.H. Di e en ia ing oo p in s o sympa ic
oden s in coas al dune communi ies: Implica ions o impe iled beach mice. J. Fish Wildl. Manag. 2018,9, 593–601. [C ossRe ]
52. Delany, M.J. Ecology o small oden s in A ica. Mammal. Re . 1986,16, 1–41. [C ossRe ]
53.
QGIS De elopmen Team. QGIS Geog aphic In o ma ion Sys em. Open Sou ce Geospa ial Founda ion P ojec . A ailable online:
h p://qgis.osgeo.o g (accessed on 12 Ma ch 2019).
54. Edwa ds, E. A b oad-scale s uc u al classi ica ion o ege a ion o p ac ical pu poses. Bo halia 1983,14, 705–712. [C ossRe ]
55.
FAO Coun y Analysis (GIEWS). A ailable online: h p://www. ao.o g/giews/coun yb ie /coun y.jsp?code=ZAF&lang= u&
bclid=IwAR3 djZN3 -NmB1RDGa5Z7Q01OdiKy4nd 3xuFmG WsHHiYisK_pWEdY30Y (accessed on 24 Augus 2021).
56. Ea h Explo e (Landsa 8 Images). A ailable online: h ps://ea hexplo e .usgs.go / (accessed on 18 Ap il 2019).
57.
F ei as, S.R.; Ce quei a, R.; Viei a, M.V. A de ice and s anda d a iables o desc ibe mic ohabi a s uc u e o small mammals
based on plan co e . B az. J. Biol. 2002,62, 795–800. [C ossRe ]
58.
Ma in, G.H.G.; Dickinson, N.M. Small mammal abundance in ela ion o mic ohabi a in a d y sub-humid g assland in Kenya.
A . J. Ecol. 1985,23, 223–234. [C ossRe ]
59. Duese , R.D.; Shuga , H.H., J . Mic ohabi a s in a o es - loo small mammal auna. Ecology 1978,59, 89–98. [C ossRe ]
60.
Monadjem, A. Geog aphic dis ibu ion pa e ns o small mammals in Swaziland in ela ion o abio ic ac o s and human land-use
ac i i y. Biodi e s. Conse . 1999,8, 223–237. [C ossRe ]
61.
Ch is ie, J.E.; Wilson, P.R.; Taylo , R.H.; Ellio , G. How ele a ion a ec s ship a (Ra us a us) cap u e pa e ns, M Mise y,
New Zealand. N. Z. J. Ecol. 2017,41, 113–119. [C ossRe ]
62.
Mi anda, C.S.; Gama a, R.M.; Mio o, C.L.; Sil a, N.M.; Conceição Filho, A.P.; Po , A. Analysis o he landscape complexi y and
he e ogenei y o he Pan anal we land. B az. J. Biol. 2018,78, 318–327. [C ossRe ]
63. Lloyd, M. Mean c owding’. J. Anim. Ecol. 1967,36, 1–30. [C ossRe ]
64.
Re elle, M.W. Package ‘psych’: The Comp ehensi e R A chi e Ne wo k. Ve sion 1.9.12. 2015. A ailable online: h ps:
//c an. -p ojec .o g/web/packages/psych/psych.pd (accessed on 16 Ap il 2019).
65.
Zuu , A.; Ieno, E.N.; Walke , N.; Sa elie , A.A.; Smi h, G.M. Mixed E ec s Models and Ex ensions in Ecology wi h R; Sp inge Science
& Business Media: New Yo k, NY, USA, 2009.
66.
Filipe, A.F.; Cowx, I.G.; Colla es-Pe ei a, M.J. Spa ial modelling o eshwa e ish in semi-a id i e sys ems: A ool o
conse a ion. Ri e Res. Appl. 2002,18, 123–136. [C ossRe ]
67. Package ‘gbm’. A ailable online: h ps://c an. -p ojec .o g/web/packages/gbm/index.h ml (accessed on 16 Ap il 2019).
68.
R Co e Team. R: A Language and En i onmen o S a is ical Compu ing; R Founda ion o S a is ical Compu ing: Vienna, Aus ia;
A ailable online: h ps://www.R-p ojec .o g/ (accessed on 12 Ma ch 2019).
69.
RS udio Team. RS udio: In eg a ed De elopmen o R; RS udio, Inc.: Bos on, MA, USA. A ailable online: h p://www. s udio.com
(accessed on 12 Ma ch 2019).
70.
Eli h, J.; Lea hwick, J.R.; Has ie, T. A wo king guide o boos ed eg ession ees. J. Anim. Ecol.
2008
,77, 802–813. [C ossRe ]
[PubMed]
71. De’A h, G. Boos ed ees o ecological modeling and p edic ion. Ecology 2007,88, 243–251. [C ossRe ]
72.
Ca slaw, D.C.; Taylo , P.J. Analysis o ai pollu ion da a a a mixed sou ce loca ion using boos ed eg ession ees. A mos. En i on.
2009,43, 3563–3570. [C ossRe ]
73.
Williams, G.J.; Aeby, G.S.; Cowie, R.O.; Da y, S.K. P edic i e modeling o co al disease dis ibu ion wi hin a ee sys em. PLoS
ONE 2010,5, e9264. [C ossRe ]
74.
Colin, B.; Cli o d, S.; Wu, P.P.; Ra hmanne , S.; Menge sen, K. Using boos ed eg ession ees and emo ely sensed da a o d i e
decision-making. Open J. S a . 2017,7, 859–875. [C ossRe ]
75.
Abea e, S. Compa isons o Boos ed Reg ession T ee, GLM and GAM Pe o mance in he S anda diza ion o Yellow in Tuna
Ca ch-Ra e Da a om he Gul o Mexico. Mas e ’s Thesis, Louisiana S a e Uni e si y, Ba on Rouge, LA, USA, 2009.
76.
A e beck, C.; Apio, A.; Pla h, M.; W onski, T. En i onmen al pa ame e s and an h opogenic e ec s p edic ing he spa ial
dis ibu ion o wild ungula es in he Akage a sa annah ecosys em. A . J. Ecol. 2009,47, 756–766. [C ossRe ]
77.
Swiha , R.K.; A wood, T.C.; Goheen, J.R.; Scheiman, D.M.; Mun oe, K.E.; Geh ing, T.M. Pa ch occupancy o No h Ame ican
mammals: Is pa chiness in he eye o he beholde ? J. Biogeog . 2003,30, 1259–1279. [C ossRe ]
78.
Lagesse, J.V.; Thondhlana, G. The e ec o land-use on small mammal di e si y inside and ou side he G ea Fish Ri e Na u e
Rese e, Eas e n Cape, Sou h A ica. J. A id En i on. 2016,130, 76–83. [C ossRe ]
79. Keesing, F. C yp ic consume s and he ecology o an A ican sa anna. BioScience 2000,50, 205–215. [C ossRe ]
80.
Iwuala, M.O.; B aide, E.I.; Maduka, N. Obse a ions on he ood habi s o some A ican oden s. Re . Biol. T op.
1979
,28, 227–236.
81.
To e, I.; Díaz, M.; Ma ínez-Padilla, J.; Bonal, R.; Vinuela, J.; Fa gallo, J.A. Ca le g azing, ap o abundance and small mammal
communi ies in Medi e anean g asslands. Basic Appl. Ecol. 2007,8, 565–575. [C ossRe ]
82.
Tew, T.E.; Macdonald, D.W. The e ec s o ha es on a able wood mice Apodemus syl a icus. Biol. Conse .
1993
,65, 279–283.
[C ossRe ]
83.
Pounds, C.J. Niche o e lap in sympa ic popula ions o s oa s (Mus ela e minea) and weasels (M. ni alis) in No heas Sco land.
Ph.D. Thesis, Uni e si y o Abe deen, Abe deen, UK, 1981.
Animals 2021,11, 2618 18 o 18
84.
Ca ilho, M.; Teixei a, D.; San os-Reis, M.; Rosalino, L.M. Small mammal abundance in Medi e anean Eucalyp us plan a ions:
How sh ub co e can eally make a di e ence. Fo . Ecol. Manag. 2017,391, 256–263. [C ossRe ]
85.
Razali, N.M.; Wah, Y.B. Powe compa isons o Shapi o-Wilk, Kolmogo o -Smi no , Lillie o s and Ande son-Da ling es s. J. S a .
Model. Anal. 2011,2, 21–33.
86.
Ali, Z.; Bhaska , S.B. Basic s a is ical ools in esea ch and da a analysis. Indian J. Anaes h.
2016
,60, 662–669. [C ossRe ] [PubMed]
87.
Os e ago á, E.; Os e ag, O.; Ko áˇc, J. Me hodology and Applica ion o he K uskal-Wallis Tes . Appl. Mech. Ma e .
2014
,611,
115–120. [C ossRe ]
88.
Kingdon, J.; Happold, D.; Bu ynski, T.; Ho mann, M.; Happold, M.; Kalina, J. Mammals o A ica; A&C Black: London, UK, 2013;
Volume 1.
89.
Ha licek, L.L.; Pe e son, N.L. Robus ness o he Pea son Co ela ion agains Viola ions o Assump ions. Pe cep . Mo . Skills
1976
,
43, 1319–1334. [C ossRe ]