padrÕes de ocorrÊncia e coexistÊncia de mamÍferos de … · 2020. 9. 27. · de serrinha dos...
TRANSCRIPT
UNIVERSIDADE FEDERAL DO DIO GRANDE DO NORTE
CENTRO DE BIOCIÊNCIAS
DEPARTAMENTO DE ECOLOGIA
PROGAMA DE PÓS-GRADUAÇÃO EM ECOLOGIA
PADRÕES DE OCORRÊNCIA E COEXISTÊNCIA DE
MAMÍFEROS DE MÉDIO E GRANDE PORTE NA CAATINGA
PAULO HENRIQUE DANTAS MARINHO
NATAL
2020
i
Paulo Henrique Dantas Marinho
PADRÕES DE OCORRÊNCIA E COEXISTÊNCIA DE
MAMÍFEROS DE MÉDIO E GRANDE PORTE NA CAATINGA
Tese apresentada ao Programa de Pós-
graduação em Ecologia da Universidade
Federal do Rio Grande do Norte como parte dos
requisitos para a obtenção do título de Doutor
em Ecologia.
Orientador:
Eduardo Martins Venticinque
Co-orientador:
Carlos Roberto Fonseca
NATAL
2020
ii
iii
iv
À universdiade pública, que abre mentes e portas para o mundo e é capaz de transformar
vidas como nenhuma outra instituição
(Dedico)
v
“Quem não tiver debaixo dos pés da alma, a areia de sua terra, não resiste aos atritos da sua
viagem na vida, acaba incolor, inodoro e insípido, parecido com todos.”
Luís da Câmara Cascudo
vi
AGRADECIMENTOS
Foram quatro longos, e ao mesmo tempo rápidos, anos, durante os quais, apesar dos
percalços e desafios, realizei muito mais do que planejei e desejei, e por isso aproveito esse
espaço para agradecer a todos que de alguma forma contribuíram para eu conseguir encerrar
esse ciclo da melhor forma possível, mesmo em meio à uma pandemia e à desvalorização da
pesquisa por parte de quem mais deveria apoiá-la, o atual governo brasileiro.
Esta tese foi feita a várias mãos, umas mais privilegiadas e acostumadas a interagir com
computadores e equipamentos tecnológicos, como as minhas, e outras, igualmente importantes,
mas mais acostumadas a lidar com ferramentas de lida com a terra, muitas vezes injustiçadas e
esquecidas por aqueles que deveriam valorizá-las. Por isso, eu começo meus agradecimentos
pelos vários auxiliares de campo e amigos que encontrei pelos caminhos da Caatinga
nordestina, e que me ensinaram grande parte do que acho que sei hoje sobre bicho, mato e gente
do interior.
Entre eles está o que mais tempo passou comigo nos campos, ouvindo minhas piadas sem
graça e lamentos, seu João da Boa Vista, de Lajes do Cabugi, ou João Bernardino de Lima, para
os que não tiveram a sorte de conhecê-lo melhor. Comigo desde os primeiros campos do
mestrado até o último do doutorado, seu João, um dia espero conseguir lhe transmitir um terço
da sua importância no meu trabalho, e na pessoa que sou hoje, por isso sou profundamente grato
por ter lhe conhecido e por sua essencial ajuda e amizade. E com seu João, veio um combo de
gente querida que me ajudou em momentos cruciais. Obrigado Joana Darc (Darquinha) pela
sua amizade, cuidado e refeições deliciosas e revigorantes. Bem como, obrigado Jussara,
Pedrinho, Joseane, Gabriel, seu Lourival e dona Dasdores, sempre amáveis e generosos, como
é o povo bom do interior.
Além de seu João, durante esses quatro anos, e antes deles, quando coletei parte dos dados
apresentados aqui também no meu mestrado, contei com a importante ajuda de outros grandes
vii
homens e mulheres, a quem serei eternamente grato. Entre ele estão seu Chico Preto e Aldenora
de Serrinha dos Pintos, Geílson e Geison de Felipe Guerra, Galego do Rela de Luís Gomes, seu
Cícero de Ponta do Mel, Bala, Veio e seu Severino e família de Cerro Corá, Titico, Gilvan e
Liélio de Martins, Alex, Suiane, Marlus, Renato, Rielson e o grande Iatagan de Mossoró e
Baraúna, Rogério Santos, Paulo e Leomar Martins (Babá) de Curaçá, Neto e família do
Boqueirão da Onça, entre vários outros e outras que certamente cometerei a injustiça de não
citá-los por esquecimento. A todos vocês, meu muito obrigado!
Agradeço à minha família, a base de tudo, de onde eu vim e quem me deu o suporte
necessário para eu chegar até aqui. Sou grato à minha mãe e ao meu pai, Maria de Fátima Dantas
e Francisco Marinho, que me proporcionaram algo que nunca tiveram durante sua juventude,
quando estudar era um privilégio ainda maior do que é hoje. À minha mãe, sou mais grato ainda
pela sua força e cuidado imensuráveis e por ser um exemplo de resiliência para seus filhos. À
Izabel e ao Ricardo, meus irmãos, sou grato pelo companheirismo e aprendizado que me dão
sendo quem são. Ao Ricardo, sou grato ainda por ter nos dado um motivo a mais para seguir e
tentar melhorar esse mundo, Zé Arthur.
Nessa reta final, contei com o carinho e suporte de uma pessoa muito especial sem a qual
não teria chegado até aqui, pelo menos não tão são e confiante como (acho que) cheguei.
Obrigado Virgínia Paixão por me aguentar, incentivar e me ensinar a ser uma pessoa melhor a
cada dia, você me inspira e me faz querer evoluir sempre, te quiero mucho!
Ao meu orientador, Dadão (vulgo Eduardo Venticinque), tenho que agradecer por tantos
anos de parceria científica e ensinamentos de vida. Com seu jeitão leve e descontraído, mas
super responsável e dedicado à mensagem que devemos passar nos trabalhos, e para além deles,
você me ensinou muito. Sou grato pela oportunidade de ter sido seu orientando durante esse
tempo, e por você ter acreditado em mim e no meu trabalho quando eu muitas vezes não
acreditava mais...
viii
Ao meu co-orientador, Carlos R. Fonseca, que foi meu primeiro orientador acadêmico,
durante meu TCC da graduação, sou muito grato pela oportunidade que me deu e por ter sido
minha primeira referência acadêmica, me acompanhando desde lá até aqui com seus
comentários e revisões cirúrgicos, colocados de uma forma sempre respeitosa e construtiva, um
grande mestre!
Aos dois, obrigado por me mostrar que existe vida, música e amizade na academia (e
além dela) entre orientador e orientando!
Agradeço imensamente a disponibilidade e valiosa contribuição dos membros da banca,
Mauro Pichorim, Claudia Campos, Rodrigo Massara, Fabiana Rocha e Guilherme Lima
(suplente, mas titular na competência). Vocês são exemplo e inspiração para minha atuação
profissional. Espero que façamos ainda muitas outras parcerias pela ciência e conservação da
biodiversidade!
Aos meus amigos e parceiros da Biologia, da Ecologia, da Consultoria Ambiental, da
Mastozoologia nordestina, do Programa Amigos da Onça, e da vida, sou grato pela amizade,
companheirismo e aprendizado, sem os quais esses quatros anos teriam sido mais difíceis e
menos divertidos: Felipe, Alan, Damião e Daniel, que além da amizade me ajudaram nos
campos; João Paulo, Ivanice, Dellano, Érika, Sean e Laís, amigos mais malucos e melhores não
há, seja de perto ou de longe, sempre lá; Renato, Annie e Paloma, amigos queridos desde a
graduação, os quais cito para representar toda nossa turma Bio 2008.2 e tantos outros amigos
que fiz durante a graduação; Paulo Fernandes, Talles, Adriana Almeida, Andressa Scabin,
Nádia, Andressa Meirelles, Rafael Domingos, Maria Clara, Marina Antongiovanni, Janina
Calado e Carolina Lisboa, amigos e colegas de profissão que a pós-graduação me deu e que
levarei pra vida; Dyego, Raissa, Tony, Raul Sales e Bruno França, que me acompanharam nas
consultorias nordeste a fora, afinal, não se vive só de bolsa; tantos amigos dos tempos de Traíras
e do Colégio Equipe, que ficaram distantes mas continuam nos meus pensamentos; Thalita, a
ix
quem agradeço pelo apoio em boa parte desse período; Liana Sena, Mayara Beltrão, Hugo
Fernandes, Lucas Silva e Robério Filho, que em sintonia temos avançado no conhecimento
sobre os mamíferos do nordeste; Claudia Martins, Claudia Campos (novamente), Carolina
Esteves e Maísa Ziviane, obrigado pela oportunidade de conhecer e contribuir para a
conservação do Boqueirão da Onça; Dulce, Taís, Gustavo, Abhishek, Luis, Lucas e Fany,
amigos que fizeram minha estadia em Portugal bem mais “fixe”. A todos e todas, meu mais
sincero obrigado!
Aqui agradeço aos meus colegas de “laboratório”, por dividirem trabalho, angústias e
alegrias. Eugênia e Juan, serei eternamente grato pela contribuição que vocês deram e dão para
a Caatinga e pelo exemplo e inspiração que representam, obrigado pela amizade e suporte de
sempre. À Virgínia, agradeço aqui pelas ajudas em campo, pela leitura dos meus textos e pelos
conselhos visando o melhor pra mim, você é a melhor “colega de lab” que eu poderia desejar.
Fernanda, obrigado pela parceria e pelo bom humor de sempre, espero que continue trabalhando
com a Caatinga, mas sei que será exitosa em qualquer linha que escolher seguir. Ellen, obrigado
pela sua sensibilidade e por me permitir te ajudar e de alguma forma conhecer mais sobre a
Amazônia e sua biodiversidade, torço muito pelo seu sucesso. Nessa linha, estendo meus
agradecimentos ao Duka, quem primeiro me mostrou a bicharada dessa belezura de bioma. À
Patrícia Ribeiro, agradeço pela amizade e parceria importante desde o mestrado até os dias de
hoje.
À Maria Luiza Falcão e ao Raul Santos, alunos de iniciação científica e amigos, agradeço
pela ajuda essencial que me deram e me dão, e pela paciência e confiança em continuar
trabalhando com alguém tão ocupado e atrapalho quanto eu. Tenho orgulho dos profissionais
que se tornaram e contem comigo para o que precisarem.
Nas pessoas das professoras Ivoneide Ferreira e Elineí Araújo agradeço a todos os
professores e professoras que me incentivaram e ensinaram que a educação é o bem mais
x
precioso que podemos almejar e utilizar para transformar vidas. Ivoneide, ainda no CERU em
Traíras, que é de onde eu venho, incentivou como ninguém o início da minha caminhada para
chegar até aqui. Elineí, já na UFRN, me apoiou nos momentos de maior fragilidade, até mesmo
financeiramente, para que eu não desistisse dessa caminhada.
Agradeço a todos que fazem o Programa de Pós-graduação em Ecologia e o Departamento
de Ecologia da UFRN, entre professores e colegas discentes, por incentivarem o crescimento
profissional da forma mais humana possível, algo cada vez mais raro na academia. Agradeço
especialmente aos queridos e queridas dona Marlene, Víctor, Dimas e Kionara, sempre
disponíveis e sorridentes ao ajudar.
Durante essa jornada, tive a oportunidade e felicidade de viver uma experiência na
Universidade de Aveiro, em Portugal, a qual agradeço imensamente. Particularmente, agradeço
ao meu supervisor durante o intercâmbio, Carlos Fonseca, pela oportunidade e receptividade, e
a toda a equipe da Unidade de Vida Selvagem, onde aprendi muito e fiz bons amigos. Obrigado
João Carvalho, Dário Hipólito, Rita Torres, Ana Figueiredo, Eduardo Ferreira, Joana
Fernandes, Filipa Costa, Raquel Martins, Raquel Crespo e Victor Bandeira, pelas boas
conversas durante o trabalho, os almoços e os finos, mas principalmente pelos campos e eventos
que me permitiram conhecer a biodiversidade e as belas paisagens naturais de Portugal. Ao
Pedro Sarmento, agradeço especialmente pelo suporte em um dos capítulos da tese e por me
proporcionar uma das maiores experiências da minha vida, ver um lince ibérico em vida livre,
e ainda com filhotes. Chorei nesse dia, e não era pra menos!
Sou grato à Camile Lugarini pela confiança e oportunidade de colaborar com o Projeto
Ararinha na Natureza, desenvolvido pelo Centro Nacional de Pesquisa e Conservação de Aves
Silvestres (CEMAVE/ICMBio) e financiado pela Vale através do Funbio. Sou grato ainda pela
amizade e ajuda de Cristine Prates, Sueli Damasceno, Damilys Oliveira, Mércia Milena e
Tatiane Alves, com as quais aprendi muito durante minha passagem pelo projeto!
xi
Agradeço à toda equipe do Parque Nacional da Furna Feia (ICMBio), nas pessoas do
Leonardo, Lúcia, Suiane, Rielson e Iatagan (mais uma vez), pelo importante apoio e parceria
durante o projeto que desenvolvi nesta importante unidade de conservação.
Agradeço ao Gabriel Penido pela parceria em um dos capítulos deste trabalho; à Barbara
Zimbres e Tadeu de Oliveira pela ajuda em análises e identificação de registros de felinos,
respectivamente; e ao Aírton Galvão (Tito) pela amizade e ajuda nos campos do mestrado que
subsidiaram um dos capítulos apresentados aqui.
Para realizar esse trabalho, contei também com o suporte financeiro, gerencial e logístico
de importantes parceiros, entre instituições e financiadores, a quem sou grato: The Mohamed
bin Zayed Species Conservation Fund, Restaurante Camarões (através de Vitor Medeiros e
família), Fundação Grupo Boticário de Proteção à Natureza, Wildlife Conservation Society –
WSC Brasil, Tropical Forest Conservation Act (TFCA) e Fundo Brasileiro para a
Biodiversidade. Ao Santander Universidades agradeço pela bolsa-auxílio para fazer o
intercâmbio. O presente trabalho foi realizado com apoio da Coordenação de Aperfeiçoamento
de Pessoal de Nível Superior - Brasil (CAPES) - Código de Financiamento 001 / This study
was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -
Brasil (CAPES) - Finance Code 001.
Sou grato a todos os moradores locais e proprietários de terras que me receberam e
permitiram realizar esse trabalho nas suas propriedades e vizinhança.
À Caatinga e aos seus moradores (gente, planta e bicho), sou grato por me fazer valorizar
minhas origens e por dar mais sentido a minha existência nesse planeta.
xii
SUMÁRIO
RESUMO ........................................................................................................................................ 11
ABSTRACT .................................................................................................................................... 13
INTRODUÇÃO GERAL ................................................................................................................ 15
Referências ...................................................................................................................................... 18
MAMÍFEROS DE MÉDIO E GRANDE PORTE DA CAATINGA DO RIO GRANDE DO
NORTE, NORDESTE DO BRASIL ............................................................................................... 22
Resumo ............................................................................................................................................ 23
Abstract ........................................................................................................................................... 24
Introdução ....................................................................................................................................... 25
Material e métodos.......................................................................................................................... 27
Resultados ....................................................................................................................................... 32
Discussão ......................................................................................................................................... 39
Agradecimentos .............................................................................................................................. 49
Literatura citada ............................................................................................................................. 49
TEMPORAL NICHE OVERLAP AMONG MESOCARNIVORES IN A CAATINGA DRY
FOREST .......................................................................................................................................... 61
Abstract ........................................................................................................................................... 63
Introduction .................................................................................................................................... 64
Methods ........................................................................................................................................... 67
Results ............................................................................................................................................. 72
Discussion ........................................................................................................................................ 81
Acknowledgments ........................................................................................................................... 86
References ....................................................................................................................................... 86
Supplementary Material ................................................................................................................. 95
CO-OCCURRENCE PATTERNS BETWEEN OCELOT AND SYMPATRIC
MESOCARNIVORES IN A BRAZILIAN DRY FOREST ......................................................... 102
Abstract ......................................................................................................................................... 104
Introduction .................................................................................................................................. 105
xiii
Material and methods ................................................................................................................... 108
Results ........................................................................................................................................... 114
Discussion ...................................................................................................................................... 122
Acknowledgments ......................................................................................................................... 127
References ..................................................................................................................................... 128
Supplementary Material ............................................................................................................... 138
MULTI-SPECIES OCCUPANCY MODELLING REVEALS MAMMALS’ PREFERENCE
FOR FORESTED HABITATS IN AN OVERGRAZED SEMIARID LANDSCAPE ................ 141
Highlights ...................................................................................................................................... 143
Abstract ......................................................................................................................................... 143
Introduction .................................................................................................................................. 144
Methods ......................................................................................................................................... 147
Results ........................................................................................................................................... 154
Discussion ...................................................................................................................................... 160
Acknowledgments ......................................................................................................................... 166
References ..................................................................................................................................... 167
Supplementary Material ............................................................................................................... 178
CONCLUSÕES GERAIS ............................................................................................................. 185
11
RESUMO 1
As florestas tropicais secas são ecossistemas ameaçados e pouco conhecidos onde a associação 2
entre o clima semiárido e a geralmente elevada degradação ambiental impõe desafios para a 3
persistência da fauna silvestre. Neste contexto, mamíferos de médio e grande porte (MMGP) 4
são especialmente afetados pelas perturbações antropogênicas, entre os quais os carnívoros são 5
intensamente perseguidos, prejudicando seu papel na estruturação das comunidades biológicas 6
através da predação e da competição intraguilda. Nesta tese, nós investigamos os padrões de 7
ocorrência e coexistência de MMGP em diferentes paisagens da Caatinga, a floresta tropical 8
seca brasileira, utilizando dados obtidos com armadilhamento fotográfico. Especificamente 9
nós: 1) realizamos o primeiro levantamento sistemático de MMGP do estado do Rio Grande do 10
Norte (RN) amostrando 10 áreas prioritárias para a conservação; 2) descrevemos os padrões 11
diários e sazonais de atividade e investigamos a sobreposição temporal entre mesocarnívoros, 12
e deles com suas presas potenciais, utilizando estatística circular e análises não paramétricas de 13
sobreposição de atividade; 3) investigamos os padrões de co-ocorrência espacial entre um 14
mesopredador dominante (Leopardus pardalis) e mesocarnívoros simpátricos, considerando a 15
sazonalidade e utilizando modelos de co-ocorrência condicional; e, finalmente, 4) testamos os 16
efeitos relativos de preditores ambientais e antropogênicos na ocupação de MMGP em uma 17
paisagem perturbada pela elevada densidade de gado e durante um período de seca extrema, 18
utilizando modelos Bayesianos de ocupação em uma abordagem multi-espécies. Como 19
resultados principais, nós encontramos 1) uma riqueza de 14 espécies de MMGP na Caatinga 20
do RN, o que representa 50% das espécies deste grupo registradas ao norte do Rio São 21
Francisco, incluindo espécies ameaçadas de extinção como um predador de topo (Puma 22
concolor). 2) Os mesocarnívoros foram principalmente noturnos ao longo das estações seca e 23
chuvosa, sobrepondo grande parte da sua atividade diária, mas segregando os picos de maior 24
atividade, o que pode representar um mecanismo de coexistência. Enquanto isso, Herpailurus 25
12
yagouaroundi foi diurno, evitando o encontro com seus competidores e sincronizando sua 26
atividade com presas potenciais. 3) Espacialmente, os demais mesocarnívoros usaram o habitat 27
independentemente da presença de L. pardalis, com exceção de H. yagouaroundi, que parece 28
preferir os mesmos locais que esse mesopredador dominante, provavelmente por apresentarem 29
melhores condições e recursos, enquanto a segregação temporal diminui os riscos de encontros 30
agressivos. 4) MMGP ocorreram principalmente em manchas de vegetação florestal, as quais 31
representam um habitat chave para a persistência desse grupo em uma paisagem degradada e 32
sob estiagem prolongada, onde muitas espécies apresentaram abundância extremamente baixa. 33
Por isso, esses ambientes devem ser protegidos para garantir a persistência de MMGP na 34
Caatinga. Nossos resultados reforçam a relevância de áreas e habitats prioritários para a 35
conservação de mamíferos na floresta tropical seca brasileira, além de elucidar as estratégias de 36
coexistência intraguilda que mantém a diversidade de mesocarnívoros neste ambiente 37
semiárido. 38
Palavras-chave: Distribuição de espécies; Espécies ameaçadas; Floresta tropical seca; 39
Interação interespecífica; Modelos de ocupação; Padrão de atividade; Semiárido. 40
41
42
43
44
45
46
47
48
49
50
13
Occurrence and coexistence’s patterns of medium to large-sized mammals in Caatinga 51
52
ABSTRACT 53
Dry tropical forests are threatened and little-known ecosystems where the association between 54
the semiarid climate and the generally high environmental degradation imposes challenges for 55
the persistence of wild fauna. In this context, medium to large-sized mammals (MLSM) are 56
especially affected by anthropogenic disturbances, among which, carnivores are intensely 57
persecuted, impairing their role in structuring biological communities through predation and 58
intraguild competition. In this thesis, we investigated the occurrence and coexistence patterns 59
of MLSM in different landscapes of the Caatinga, the Brazilian dry tropical forest, using camera 60
trapping data. Specifically, we 1) carried out the first systematic survey of MLSM in the Rio 61
Grande do Norte state (RN), sampling 10 priority areas for conservation; 2) we described the 62
daily and seasonal activity patterns and estimated the temporal overlap among mesocarnivores 63
using circular statistics and non-parametric analyzes of activity overlap; 3) we investigated the 64
patterns of spatial co-occurrence between a dominant mesopredator (Leopardus pardalis) and 65
sympatric mesocarnivores, considering seasonality and using conditional co-occurrence 66
models; and finally, 4) we tested the relative effects of environmental and anthropogenic 67
predictors on MLSM’s occupancy in a landscape disturbed by high cattle density and during a 68
period of extreme drought, using Bayesian occupancy models in a multi-species approach. As 69
main results, we found 1) a wealth of 14 MLSM’s species in the Caatinga of RN, which 70
represents 50% of the MLSM registered at the north of the São Francisco River, including 71
threatened species as a top predator (Puma concolor). 2) Mesocarnivores were mainly nocturnal 72
throughout the dry and rainy seasons, overlapping most of their daily activity, but segregating 73
the peaks of greater activity, which may represent a coexistence mechanism. Meanwhile, 74
Herpailurus yagouaroundi was diurnal, avoiding encounters with competitors and 75
14
synchronizing its activity with potential prey. 3) Spatially, the other mesocarnivores used the 76
habitat regardless L. pardalis’ presence, with the exception again of H. yagouaroundi, which 77
seems to prefer the same locations as this dominant mesopredator, probably because they have 78
better conditions and resources, while temporal segregation decreases the risk of aggressive 79
encounters. 4) MLSM occurred mainly in patches of forest vegetation, which represent a key 80
habitat for the persistence of this group in a degraded landscape under prolonged drought, and 81
where many species showed an extremely low abundance. Therefore, these environments must 82
be protected to guarantee MLSM’s persistence in Caatinga. Our results reinforce the relevance 83
of priority areas and habitats for mammals’ conservation in the Brazilian dry tropical forest, in 84
addition to elucidating the intraguild coexistence strategies that maintain the mesocarnivores 85
diversity in this semiarid environment. 86
Keywords: Activity pattern; Dry tropical forest; Interspecific interactions; Occupancy 87
models; Semiarid; Species distribution; Threatened species. 88
89
90
91
92
93
94
95
96
97
98
15
INTRODUÇÃO GERAL 99
Mamíferos terrestres de médio e grande porte (MMGP; e.g., > 1 kg, Chiarello et al., 2000) 100
desempenham importantes papeis ecológicos nas florestais tropicais (Bello et al., 2016), mas 101
estão entre os grupos mais afetados pelas pressões antropogênicas presentes nestes ambientes 102
(Benítez-Lopes et al., 2017). Do controle de herbívoros que estruturam a vegetação a dispersão 103
de sementes, a perda de espécies de MMGP pode afetar a estrutura e a capacidade de 104
regeneração dos ecossistemas (Bello et al., 2016; Ripple et al., 2014). Relativamente baixas 105
densidades e taxas reprodutivas tornam mais difícil para populações de espécies de MMGP se 106
recuperarem da caça intensa e da perda e degradação dos seus habitats (Chiarello et al., 2000; 107
Ripple et al., 2014), fazendo desse grupo um importante indicador de integridade biótica de um 108
ecossistema (Cheyne et al., 2016). 109
Entre os MMGP, os grandes carnívoros são especialmente ameaçados principalmente 110
pela perda de habitat, diminuição da disponibilidade de presas e perseguição humana decorrente 111
de conflitos pela predação de animais domésticos (Ripple et al., 2014). Por isso, as populações 112
de predadores de topo de cadeia têm declinado por todo o mundo, gerando, entre outros efeitos, 113
um aumento nas abundâncias de mesopredadores, uma vez que estes ficam liberados do 114
controle top-down (Ripple et al., 2014). Consequentemente, dentre outros efeitos, os 115
mesopredadores pressionam ainda mais suas presas e intensificam as interações intraguilda 116
(Croocks e Soulé, 1999; Jiménez et al., 2019), Neste contexto, espécies dominantes dessa guilda 117
(e.g., maiores) podem se tornar predadores de topo emergentes (Prugh et al., 2009). Desta 118
forma, é essencial investigar os padrões de ocorrência e coexistência de MMGP em 119
ecossistemas tropicais para entender melhor quais fatores determinam a diversidade de 120
mamíferos terrestres e garantem a persistência das comunidades desse grupo em paisagens onde 121
os impactos antropogênicos ameaçam significativamente a biodiversidade. 122
16
As florestas sazonalmente secas estão entre os ecossistemas tropicais menos conhecidos 123
e protegidos (Santos et al., 2011; Banda et al., 2016). A Caatinga brasileira consiste na maior 124
formação desse tipo de bioma na região Neotropical (Silva et al., 2017). Coberta por um 125
mosaico vegetacional que vai da formações mais arbustivas e abertas até florestas secas densas, 126
de acordo com o grau de aridez e degradação (Velloso et al., 2002; Silva e Souza, 2018), a 127
Caatinga é marcada por um clima semiárido, apresentando altas temperaturas, elevada 128
evapotranspiração e baixas precipitações concentradas em poucos meses do ano (Sampaio, 129
2010). Essa ecorregião já perdeu aproximadamente 50% da sua cobertura original 130
(Antongiovanni et al., 2018), e o restante se encontra sob elevados níveis de distúrbio 131
antropogênico crônico (Ribeiro et al., 2015). 132
Entre as maiores ameaças para a biodiversidade da Caatinga estão a conversão da 133
vegetação em culturas agrícolas de ciclo curto, a extração de madeira para usos doméstico, 134
industrial e produção de carvão vegetal, a caça intensa e amplamente difundida, e a criação 135
extensiva de gado bovino, ovino e caprino, que somam mais de 10 milhões de cabeças (Leal et 136
al., 2005; Silva et al., 2017). Essas atividades são desenvolvidas por uma população humana de 137
mais de 28 milhões de habitantes distribuídos ao longo dos 735.000 km² dessa região semiárida 138
(Leal et al. 2005; Silva et al., 2017). As mudanças climáticas representam uma ameaça 139
adicional, uma vez que devem intensificar a escassez de chuvas na região (Seddon et al., 2016). 140
Além disso, a baixa cobertura de áreas integralmente protegidas, que somam menos de 2 % da 141
área da Caatinga (Fonseca et al., 2017), e o ainda escasso, embora crescente, conhecimento 142
sobre as espécies da região, dificultam a elaboração e implementação de estratégias eficientes 143
de conservação e manejo, bem como o avanço no entendimento dos seus padrões ecológicos. 144
Os mamíferos se destacam entre os grupos biológicos menos conhecidos na Caatinga. 145
Reconhecida por muito tempo como pobre em espécies e endemismos desse grupo faunístico 146
(Mares et al., 1981), só nos últimos anos têm surgido esforços significativos de pesquisas sobre 147
17
a ecologia de mamíferos na região, revelandouma riqueza de pelo menos 183 espécies, sendo 148
45 delas de MMGP (incluindo primatas) e 11 endêmicas (Carmignotto e Astúa, 2017). Entre os 149
estados brasileiros, o Rio Grande do Norte sempre foi apontado como uma grande lacuna de 150
conhecimento sobre o grupo (Feijó e Langguth, 2013), o que só nos últimos anos tem sido 151
revertido (Marinho et al., 2018; Vargas-Mena et al., 2018). Um aspecto da ecologia de MMGP 152
que foi pouco investigado até aqui, é como mesocarnívoros coexistem em paisagens onde 153
predadores de topo como onças pardas (Puma concolor) e pintadas (Panthera onca) estão 154
extintos localmente ou funcionalmente, o que representa grande parte da floresta tropical seca 155
brasileira (Feijó e Langguth, 2013; Carmignotto e Astúa, 2017). Além disso, precisamos 156
entender melhor como MMGP lidam com a marcada sazonalidade de ambientes semiáridos 157
como a Caatinga (Stoner e Timm, 2011; Carmignotto e Astúa, 2017), e quais características da 158
paisagem beneficiam a persistência das espécies em um contexto de elevada degradação 159
antropogênica e estiagem prolongada, que tendem a se intensificar no futuro. 160
Para abordar estas temáticas, esta tese está dividida em quatro capítulos. No Capítulo 1 161
descrevemos a riqueza e composição de MMGP da Caatinga do Rio Grande do Norte, a partir 162
de um significativo esforço amostral que cobriu 10 áreas prioritárias para a conservação dessa 163
ecorregião no estado. A partir dos resultados do armadilhamento fotográfico, discutimos a 164
relevância das espécies registradas, as ausências preocupantes e a importância das áreas de 165
estudo para o conhecimento e conservação da biodiversidade local. Optamos por publicar esse 166
capítulo em português por entendemos que ele preenche uma lacuna sobre informações básicas 167
de ocorrência de mamíferos no estado, e que assim deve ser utilizado tanto para pequisa 168
cientídica quanto por técnicos de órgãos ambienatsi e tomadores de decisão. 169
Nos Capítulos 2 e 3 investigamos os mecanismos de coexistência de mesocarnívoros 170
através da análise de sobreposição dos eixos temporal e espacial do nicho das espécies, 171
respectivamente, em uma área prioritária para conservaçãoda Caatinga do Rio Grande do Norte 172
18
onde predadores de topo estão virtualmente ausentes, e onde a jaguatirica (Leopardus pardalis) 173
atua como mesocarnívoro dominante. Além disso, consideramos potenciais efeitos da 174
sazonalidade e de fatores como disponibilidade de presas e interferências antrópicas nesses 175
padrões espaço-temporais. 176
No Capítulo 4 testamos os efeitos relativos de preditores ambientais e antropogênicos na 177
ocupação de MMGP em uma paisagem da Caatinga de alta relevância para conservação, mas 178
que se encontra sob sobrepastejo no norte da Bahia. Neste trabalho tentamos identificar quais 179
mecanismos permitem a persistência das espécies em um ambiente perturbado durante um 180
período de seca extrema através de modelos Bayesianos de ocupação baseados em uma 181
abordagem multi-espécies. 182
Com os resultados desta tese, esperamos melhorar o entendimento sobre aspectos 183
ecológicos de MMGP em uma região semiárida sob forte interferência humana, de forma que 184
essas informações auxiliem na tomada de decisão e na elaboração e execução de estratégias de 185
conservação e manejo das espécies e paisagens da Caatinga. Além disso, esperamos que nossos 186
resultados estimulem o desenvolvimento de novas pesquisas que avancem nossa capacidade de 187
entendimento de como mamíferos lidam com ecossistemas semiáridos. 188
189
Referências 190
Antongiovanni M, Venticinque EM, Fonseca CR. 2018. Fragmentation patterns of the Caatinga 191
drylands. Landscape Ecology 33:1353-1367. 192
Banda RK, Delgado-Salinas A, Dexter KG et al. 2016. Plant diversity patterns in neotropical 193
dry forests and their conservation implications. Science 353:1383-1387. 194
Bello C, Galetti M, Pizo MA et al. 2015. Defaunation affects carbon storage in tropical forests. 195
Science Advances 1:e1501105. 196
19
Benítez-López A, Alkemade R, Schipper AM, Ingran DJ, Verweij PA, Eikeboom JAJ, 197
Huijbregts MAJ. 2017. The impact of hunting on tropical mammal and bird populations. 198
Science 356:180-183. 199
Carmignotto AP, Astúa D. 2017. Mammals of the Caatinga: diversity, ecology, biogeography, 200
and conservation. In: Silva MC, Leal IR, Tabarelli M (eds). Caatinga: The largest tropical 201
dry forest region in South America. Springer, Cham, pp. 211-254. 202
Cheyne SM, Sastramidjaja WJ, Muhalir, Rayadin Y, Macdonald DW. 2016. Mammalian 203
communities as indicators of disturbance across Indonesian Borneo. Global Ecology and 204
Conservation 7:157-173. 205
Chiarello AG. 2000. Density and population size of mammals in remnants of Brazilian Atlantic 206
Forest. Conservation Biology 14:1649-1657. 207
Fonseca CR, Antongiovanni M, Matsumoto M, Bernard E, Venticinque EM. 2017. 208
Conservation opportunities in the Caatinga. In: Silva MC, Leal IR, Tabarelli M (eds). 209
Caatinga: The largest tropical dry forest region in South America. Springer, Cham, pp. 429-210
443. 211
Jiménez J, Nuñez-Arjona JC, Mougeot F et al. 2019. Restoring apex predators can reduce 212
mesopredator abundances. Biological Conservation 238:108234. 213
Mares MA, Willig MR, Steilein KE, Lacher-Jr TE. 1981. The mammals of Northeastern Brazil: 214
a preliminary assessment. Annals of the Carnegie Museum 50:81-137. 215
Marinho PH, Bezerra D, Antongiovanni M, Fonseca CR, Venticinque EM. 2018. Mamíferos 216
de médio e grande porte da Caatinga do Rio Grande do Norte, nordeste do Brasil. 217
Mastozoología Neotropical 25:345-362. 218
Prugh LR, Stoner CJ, Epps CW, Bean WT, Ripple WJ, Laliberte AS, Bashares JS. 2009. The 219
rise of the mesopredator. Bioscience 59:779-791. 220
20
Ribeiro E, Arroyo-Rodríguez V, Santos BA, Tabarelli M, Leal IR. 2015. Chronic anthropogenic 221
disturbance drives the biological impoverishment of the Brazilian Caatinga vegetation. 222
Journal of Applied Ecology 52:611-620. 223
Ripple WJ, Estes JA, Beschta RL et al. 2014. Status and ecological effects of the world’s largest 224
carnivores status and ecological effects of the world’s largest carnivores. Science 225
343:1241484. 226
Sampaio EVSB. 2010. Características e potencialidades. In: Gariglio MA, Sampaio EVSB, 227
Cestaro LA, Kageyama PY (eds). Uso sustentável e conservação dos recursos florestais da 228
Caatinga. Serviço Florestal Brasileiro, Brasília, pp 29-48. 229
Santos JC, Leal IR, Almeida-Cortez JS, Fernandes GW, Tabarelli M. 2011. Caatinga: the 230
scientific negligence experienced by a dry tropical forest. Tropical Conservation Science 231
4:276-286. 232
Seddon AWR, Macias-Fauria M, Long PR, Benz D, Willis KJ. 2016. Sensitivity of global 233
terrestrial ecosystems to climate variability. Nature 531:229-232. 234
Silva AC, Souza AF. 2018. Aridity drives plant biogeographical sub regions in the Caatinga, 235
the largest tropical dry forest and woodland block in South America. Plos One 13:e0196130. 236
Silva JMC, Leal IR, Tabarelli M (eds). 2017. Caatinga: the largest tropical dry forest region in 237
South America. Springer, Cham. 238
Stoner K, Timm RM. 2011. Seasonally dry tropical forest mammals: Adaptations and seasonal 239
patterns. In: Dirzo R, Hillary S, Young S, Mooney HA, Ceballos G (eds). Seasonally dry 240
tropical forests: Ecology and conservation. Island Press, Washington. 241
Vargas-Mena JC, Alves-Pereira K, Barros MAS et al. 2018. Te bats of Rio Grande do Norte 242
state, northeastern Brazil. Biota Neotropica 18:e20170417. 243
21
Velloso AL, Sampaio E, Pareyn F (eds). 2002. Ecorregiões propostas para o bioma Caatinga. 244
Associação de Plantas do Nordeste/Instituto para Conservação Ambiental/The Nature 245
Conservancy do Brasil, Recife. 246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
22
Capítulo 1 270
271
MAMÍFEROS DE MÉDIO E GRANDE PORTE DA CAATINGA DO RIO GRANDE 272
DO NORTE, NORDESTE DO BRASIL 273
274
275
276
277
Capítulo publicado na revista Mastozoología Neotropical (https://www.sarem.org.ar/wp-278
content/uploads/2019/01/SAREM_MastNeotrop_25-2_08_Marinho.pdf). 279
23
Mamíferos de médio e grande porte da Caatinga do Rio Grande do Norte, 280
nordeste do Brasil 281
282
Paulo H. Marinho¹, Daniel Bezerra¹, Marina Antongiovanni¹, Carlos R. Fonseca¹ e 283
Eduardo M. Venticinque¹ 284
285
¹Programa de Pós-graduação em Ecologia, Departamento de Ecologia, Centro de Biociências, 286
Universidade Federal do Rio Grande do Norte, Lagoa Nova, Natal, RN, Brasil 287
288
Resumo. Mamíferos de médio e grande porte são especialmente afetados pela caça e perda de 289
habitat, ao mesmo tempo que desempenham importantes funções no ecossistema. O 290
conhecimento sobre esse grupo na Caatinga, a floresta tropical sazonalmente seca do nordeste 291
do Brasil, contudo, ainda é escasso. Neste trabalho realizamos o primeiro levantamento 292
sistemático de mamíferos de médio e grande porte no estado do Rio Grande do Norte, através 293
de armadilhamento fotográfico em 188 pontos distribuídos em 10 áreas prioritárias para a 294
conservação da biodiversidade da Caatinga. Com um esforço amostral de 7271 câmeras-dias, 295
obtivemos 1607 registros de 14 espécies nativas, distribuídas em seis ordens e 10 famílias: oito 296
espécies da ordem Carnivora, duas espécies da ordem Cingulata, e uma espécie para cada uma 297
das demais ordens (Artiodactyla, Didelphimorphia, Pilosa e Primates). A riqueza encontrada 298
representa 31% das 45 espécies de mamíferos de médio e grande porte que ocorrem na 299
Caatinga, e 50% das espécies deste grupo registradas no domínio da Caatinga ao norte do Rio 300
São Francisco, ampliando significativamente o conhecimento sobre o grupo na região. Entre as 301
espécies registradas estão três felinos ameaçados de extinção, incluindo um grande predador de 302
topo (Leopardus tigrinus, Herpailurus yagouaroundi e Puma concolor). O baixo número de 303
registros de algumas espécies e a ausência de outras indicam o estado crítico da mastofauna do 304
24
estado e sugerem a importância e urgência do estabelecimento de novas unidades de 305
conservação na Caatinga do Rio Grande do Norte. 306
Palavras chave: Armadilha fotográfica. Conservação da Biodiversidade. Espécies ameaçadas. 307
Floresta sazonalmente tropical seca. Riqueza de espécies. 308
309
Abstract. Medium to large-sized mammals are the most affected by habitat loss and hunting, 310
while at the same time playing important roles in the ecosystem. The knowledge about this 311
group in the Caatinga, the dry tropical forest of northeastern Brazil, however, is still scarce. In 312
this study, we carried out the first systematic survey of medium to large-sized mammals in the 313
Rio Grande do Norte state, with camera traps installed in 188 points distributed in 10 priority 314
areas for the conservation of the Caatinga biodiversity. With a sampling effort of 7271 camera-315
days, we obtained 1607 records of 14 native medium to large-sized mammals, distributed in six 316
orders and 10 families: eight species of the order Carnivora, two species of the order Cingulata, 317
and one species for each of the other orders (Artiodactyla, Didelphimorphia, Pilosa e Primates). 318
The richness recorded represents a third of the 45 large and medium-sized mammal species 319
found in the Caatinga, and half of the large and medium-sized mammal species recorded in 320
northern Caatinga dominium, significantly increasing the knowledge about the group in the 321
region. Among the recorded species are three threatened felids, including a large top predator 322
(Leopardus tigrinus, Herpailurus yagouaroundi, and Puma concolor). The low number of 323
records of some species and the absence of others indicate the critical status of the mammal 324
fauna in the state and highlight the importance and urgency of the creation of new protected 325
areas in the Caatinga of the Rio Grande do Norte. 326
Key words: Biodiversity conservation. Camera trap. Seasonally dry tropical forest. Species 327
richness. Threatened species. 328
329
25
Introdução 330
O conhecimento regional sobre a presença e distribuição das espécies é essencial para 331
planejar e avaliar estratégias de conservação da biodiversidade (Tobler et al. 2008). Esse tipo 332
de informação ajuda a definir a distribuição espacial das espécies, monitorar a diversidade ao 333
longo do espaço e do tempo e avaliar o impacto de atividades humanas sobre as espécies 334
vulneráveis (Tobler et al. 2008; Cheyne et al. 2016; Rosas-Ribeiro et al. 2017). Mamíferos de 335
médio e grande porte (> 1 kg; e.g. Chiarello 1999) são especialmente afetados por ameaças 336
como a caça e a perda e fragmentação dos seus habitats, uma vez que apresentam baixas 337
densidade e fecundidade, e geralmente elevados requerimentos tróficos e de área de vida 338
(Chiarello 1999; Cardillo et al. 2005; Peres 2001). Ao mesmo tempo, esses grandes vertebrados 339
são responsáveis por importantes funções e serviços ecossistêmicos, como o controle de 340
populações de herbívoros, a modulação do ciclo de nutrientes através do consumo da biomassa 341
vegetal e a dispersão de sementes grandes (Chiarello 1999; Terborgh et al. 2001; Cardillo et al. 342
2005; Galetti & Dirzo 2013; Sobral et al. 2017). Além disto, muitas vezes mamíferos de médio 343
e grande porte, devido a questões éticas, estéticas e culturais, são importantes espécies guarda-344
chuva ou bandeira em ações de conservação (Linnell et al. 2000). 345
As florestas tropicais secas abrigam uma relevante diversidade de mamíferos e estão entre 346
os ecossistemas tropicais mais ameaçados e pouco protegidos do mundo (Banda et al. 2016). A 347
Caatinga, localizada no nordeste do Brasil, é a maior floreta tropical seca das Américas (Banda 348
et al. 2016) e já foi considerada uma das 37 grandes regiões selvagens do planeta (Mittermeier 349
et al. 2002). Apesar do déficit histórico de estudos (Santos et al. 2011), uma revisão recente 350
evidencia que a Caatinga abriga uma elevada diversidade de animais e plantas (Silva et al. 351
2017a). Por outro lado, 45,5% da sua cobertura vegetal original já foi perdida (MMA 2016a) e 352
áreas integralmente protegidas cobrem menos de 2% do seu território (Fonseca et al. 2017). 353
Além disso, os fragmentos de Caatinga remanescentes, embora relativamente conectados 354
26
(Antongiovanni et al. 2018), encontram-se sob forte distúrbio antrópico crônico, resultado de 355
atividades como a criação extensiva de rebanhos, a retirada de madeira e a caça ilegal (Alves et 356
al. 2016; Marinho et al. 2016; Ribeiro et al. 2015), todas amplamente praticadas por uma 357
população humana de 28.6 milhões de habitantes (Silva et al. 2017b). E todas essas ameaças 358
devem ser agravadas pela intensificação na aridez da região resultante das mudanças climáticas 359
(Seddon et al. 2016). 360
A Caatinga abriga 183 espécies de mamíferos, sendo 11 delas endêmicas e 45 de médio e 361
grande porte (Carmignotto & Astúa 2017). Embora o número de inventários e investigações 362
ecológicas tenha crescido nos últimos anos, especialmente com a disseminação de técnicas 363
como o armadilhamento fotográfico (e.g. Dias & Bocchiglieri 2016; Delciellos 2016; Astete et 364
al. 2017), uma grande porção desse bioma brasileiro permanece desconhecida ou subamostrada 365
(Santos et al. 2011; Albuquerque et al. 2012). Ao comportamento elusivo, hábitos noturnos e 366
baixas densidades, que contribuem para o desconhecimento dos mamíferos na região, soma-se 367
o historicamente baixo investimento em pesquisa (Santos et a. 2011) e a visão inicial 368
equivocada de que a mastofauna da Caatinga seria depauperada (Mares et al. 1981). 369
No estado do Rio Grande do Norte, extremo nordeste do Brasil e limite da distribuição de 370
muitas espécies, a lacuna de conhecimento sobre mamíferos é ainda mais evidente (Brito et al. 371
2009; Feijó & Langguth 2013), indo de marsupiais (Melo & Sponchiado 2012) até médios e 372
grandes mamíferos (Feijó & Langguth 2013). A lacuna para quirópteros era igualmente grande 373
até levantamentos recentes (e.g. Vargas-Mena et al. 2018). Diferente dos estados vizinhos, o 374
Rio Grande do Norte não foi visitado por expedições históricas, que realizadas a partir do século 375
XVII impulsionaram o conhecimento sobre mamíferos no restante do nordeste do Brasil (Feijó 376
& Langguth 2013). Para mamíferos de médio e grande porte os trabalhos no estado são 377
relativamente recentes e representados basicamente por inventários paleontológicos (Araújo-378
Júnior & Porpino 2011), atualizações de distribuição e primeiros registros de ocorrência 379
27
(Ferreira et al. 2009; Dantas et al. 2016; Marinho et al. 2017; Rosas-Ribeiro et al. 2017). Ainda 380
mais escassos são os trabalhos sobre a ecologia de espécies, até o momento focados restritos a 381
Sapajus libidinosus (Ferreira et al. 2010; Emidio & Ferreira 2012) e Leopardus tigrinus 382
(Marinho et al. 2018a; b). 383
A melhoria do conhecimento biológico em áreas prioritárias para a conservação é uma ação 384
relevante para fortalecer a aptidão dessas áreas em conservar a biodiversidade (Silva et al. 385
2004), sobretudo em uma região com baixa cobertura de unidades de conservação como a 386
Caatinga. Neste contexto, o objetivo deste estudo foi caracterizar a composição e riqueza de 387
mamíferos de médio e grande porte da Caatinga do estado do Rio Grande do Norte, Brasil, 388
através de armadilhamento fotográfico em 10 áreas prioritárias para a conservação da 389
biodiversidade do bioma. Com este trabalho esperamos (i) preencher a lacuna existente no 390
estado de registros de mamíferos de médio e grande porte, (ii) caracterizar a qualidade relativa 391
das 10 áreas prioritárias em termos de mamíferos de médio e grande porte e (iii) fornecer 392
subsídios para a criação de novas unidades de conservação no Rio Grande do Norte. 393
394
Material e métodos 395
Área de estudo 396
Nossa área de estudo compreende a Caatinga do estado do Rio Grande do Norte (Fig. 1, 397
Tabela 1), sendo que as amostragens foram realizadas em 10 paisagens disjuntas consideradas 398
áreas prioritárias para a conservação da biodiversidade do bioma tanto em nível estadual 399
(https://brasil.wcs.org/pt-br/Lugares-naturais/Projeto-Caatinga.aspx) quanto em nível nacional 400
(MMA, 2016b; Tabela 1). Todas as áreas estudadas são formadas basicamente por 401
propriedades privadas e não possuem proteção legal, com exceção de Dunas Rosado, que 402
recentemente se tornou uma Área de Proteção Ambiental (APA) estadual (IDEMA 2018) 403
(Tabela 1). 404
28
405
406
Fig. 1. Localização das 10 áreas prioritárias para a conservação da Caatinga do Rio Grande do Norte, 407
nordeste do Brasil, onde foram realizados levantamentos de mamíferos de médio e grande porte (para 408
mais informações sobre as áreas acessar: https://brasil.wcs.org/pt-br/Lugares-naturais/Projeto-409
Caatinga.aspx). Embora Martins e Serrinha dos Pintos componham uma mesma área prioritária, estas 410
áreas foram amostradas de forma independente. 411
412
Mais de 90% do território do Rio Grande do Norte está inserido no domínio da Caatinga 413
(IDEMA 2014). O clima da região é quente e semiárido, com chuvas irregulares e concentradas 414
em poucos meses do ano (e.g. fevereiro a maio), com médias pluviométricas entre 400 e 800 415
mm (Ab’Sáber 1974; Velloso et al. 2002). A vegetação da Caatinga é um mosaico de formações 416
arbustivas, manchas arbóreas e florestas secas (Santos et al. 2011) que variam de acordo com o 417
relevo, solo, clima local e nível de antropização (Velloso et al. 2002). Entre as nossas áreas de 418
estudo, três (Felipe Guerra, Dunas do Rosado e Caiçara do Norte) se encontram em altitudes 419
baixas (média de 20 a 75 m acima do nível do mar) (Tabela 1), onde predomina uma vegetação 420
mais baixa e espaçada, embora existam manchas arbóreas em solos mais ricos e com menos 421
29
pressão antrópica. Já as demais áreas cobrem principalmente ambientes elevados e inclinados 422
(e.g. serras e encostas, respectivamente) (médias entre 250 e 480 metros de altitude) (Tabela 423
1), onde predominam as formações vegetais densas e arbóreas, mas também manchas arbustivas 424
nas áreas mais baixas e pressionadas. Semelhante ao que acontece com o restante do bioma, 425
45% da cobertura original da Caatinga no estado já foi modificada por atividades como 426
agricultura, pecuária e corte de lenha para uso doméstico e industrial (IDEMA 2014). 427
Fruticultura irrigada, exploração de calcário para produção de gesso e expansão de usinas 428
eólicas e solares são vetores adicionais de perda de cobertura vegetal na região (obs. pessoal). 429
430
431
432
433
434
435
436
437
438
439
30
Tabela 1. Esforço amostral empregado para levantar a composição e riqueza de mamíferos de médio e grande porte em 10 áreas prioritárias para a conservação
da Caatinga do Rio Grande do Norte, nordeste do Brasil. A área amostral foi medida a partir do mínimo polígono convexo dos pontos amostrais em cada área e
somados para todo o estudo. São apresentados os códigos dos polígonos prioritários que cobrem totalmente ou parcialmente as áreas de estudo, de acordo com
lei federal (MMA 2016b). Unidade de conservação (UC). DP: desvio padrão.Para mais informações sobre as áreas acessar: https://brasil.wcs.org/pt-br/Lugares-
naturais/Projeto-Caatinga.aspx (Projeto Caatinga Potiguar – Cartograma).
Localidade Elevação
média
(DP)
Período Nº pontos Esforço
(câmeras-dias)
Área
(km²)
Coordenadas
principais
Área prioritária /
UC
Serra de Santana 401,89
(199,50) 05 Mai - 18 Jun
19 703 13452.0 6°10’ e 6°0’S;
36°50’ e 36°35’W CA045 /
Não
Lajes 355,9
(101,09)
26 Mai -
06 Jul 20 793 20184.5
5°57’ e 5°43’S,
36°5’ e 36°17’W
CA078 /
Não
Cerro Corá 468,05 (83,25)
01 Jun - 08 Jul
19 654 7476.6 6°13’ e 6°2’S;
36°15’ e 36°22’W CA096 /
Não
Martins 303,2
(83,48)
22 Jun - 30
Jul 20 742 6074.1
6°0’ e 5°4’S;
37°50’ e 38°0’W
CA045 /
Não Serrinha dos
Pintos
345,42
(68,25)
28 Jun - 05
Ago 20 740 4035.1
6°15’ e 6°5’S;
37°50’ e 38°00’W
CA063 /
Não
Felipe Guerra 75,25
(18,57) 09 Jul - 16
Ago 20 736 30116.7
5°38’ e 5°23’S; 37°30’ e 37°43’W
CA078 / Não
Caiçara do Norte 20,13
(9,34)
19 Jul - 22
Set 15 845 7497.4
5°13’ e 5°5’S;
36°1’ e 36°10W
CA102 /
Não
Luis Gomes 482,23 (95,69)
02 Ago - 23 Set
17 846 2976.5 6°25’ e 6°20’S;
38°19’ e 38°26’W CA088 /
Não
Dunas do Rosado 62,26
(33,81)
08 Ago -
04 Set 19 499 5979.0
5°1’ e 5°8’S;
36°48’ e 36°55’W
CA087 /
Sim
Coronel Ezequiel 351,684
(67,25)
18 Ago -
25 Set 19 713 7439.5
6º28’ e 6°15’S;
36°5’ e 36°14’W
CA088 /
Não
Amostragem total 05 Mai -
25 Set 188 7271 105231
31
Levantamento de dados
De maio a setembro de 2014, na transição entre as estações chuvosa e seca, 20 armadilhas
fotográficas (Bushnell® Trophy Cam ™) foram instaladas em cada uma das 10 áreas
prioritárias, permanecendo ativas 24 h por dia durante um a dois meses (média de 38 dias)
(Tabela 1). Essas câmeras acionadas por calor e movimento foram posicionadas a 30-40 cm do
solo e programadas para registrar a data e hora de cada registro fotográfico, com intervalos
mínimos de cinco minutos entre dois disparos consecutivos. Nós estabelecemos uma distância
mínima de um km entre câmeras de forma a minimizar a dependência espacial entre os pontos
amostrais, exceto em locais que as dificuldades de acesso não permitiram esta padronização.
No geral, as câmeras foram posicionadas a uma distância média de 1,6 km (0,8 – 3 km) uma da
outra, ao longo de trilhas feitas por pessoas ou de animais de criação, estradas abandonadas e
leito de rios intermitentes. Nenhuma isca foi utilizada para atrair os animais.
O banco de imagens foi triado com auxílio do programa Camera Base v.1.6 (Tobler 2007).
A nomenclatura e taxonomia das espécies seguiu Kitchener et al. (2017) para os felinos e
Carmignotto e Astúa (2017) para as demais espécies. O estado de conservação global e nacional
das espécies está de acordo com a IUCN (2017) e o MMA (2014), respectivamente.
Análise dos dados
A suficiência amostral foi investigada através de uma curva de acumulação de espécies
(randomizada 1000 vezes) agrupando os dados de todas as áreas. A riqueza de espécies
estimada para a Caatinga do Rio Grande do Norte foi obtida através do estimador Jackknife de
primeira ordem, considerado ideal para dados de médios e grandes mamíferos a partir de
armadilhas fotográficas (Tobler et al. 2008). Todas as análises foram realizadas usando o
programa livre EstimateS 9.1.0 (Colwell 2013). Nas análises, para uma determinada espécie
em um mesmo ponto, consideramos apenas registros consecutivos com intervalos maiores do
que uma hora para minimizar problemas de dependência temporal. Apresentamos o índice de
32
abundância relativa (uma estimativa do sucesso de detecção) para cada espécie registrada,
calculado como a razão entre o número de registros independentes (registros da mesma espécie
no mesmo local com mais de uma hora) e o esforço amostral total (em câmeras-dias)
multiplicado por 100 (Rovero et al. 2014). Embora esse índice deva ser interpretado com cautela
por sofrer influência de fatores ambientais e espécie-específicos que dificultam generalizações
(ver Sollman et al. 2013), ele pode servir como parâmetro populacional inicial para estudar
espécies (Rovero et al. 2014) em um contexto de escassez de informação.
Resultados
Um total de 14 espécies de mamíferos silvestres de médio e grande porte foi registrado na
Caatinga do estado do Rio Grande do Norte (Fig. 2, Tabela 2), sendo que o número médio de
espécies por área foi 8,2 (DP = 1,23), variando de seis a 10 (Tabela 2). No total, 1607 registros
independentes foram obtidos a partir de um esforço de 7.271 câmera-dias distribuído em 188
pontos amostrais (12 equipamentos apresentaram problemas ou foram roubados) que cobriram
uma área total de 1052 km² (Tabela 1). As 14 espécies pertencem a seis ordens e 10 famílias,
sendo a ordem Carnivora a mais representativa, com oito espécies (57% do total), seguida pela
ordem Cingulata com duas espécies, e as demais ordens apresentaram uma espécie cada. Dentre
as famílias, Felidae foi a que apresentou o maior número de espécies, com quatro ao todo (28%
do total). Entre as espécies registradas, três (21%) estão sob algum nível de ameaça (nacional
ou global), sendo todos felinos (Tabela 2).
33
Fig. 2. Mamíferos de médio e grande porte registrados através de armadilhamento frotográfico
em 10 áreas prioritárias para a conservação da Caatinga do Rio Grande do Norte, extremo
34
nordeste do Brasil; (a) Cerdocyon thous, (b) Procyon cancrivorus, (c) Conepatus amazonicus,
(d) Galictis cuja, (e) Leopardus tigrinus, (f) Leopardus pardalis, (g) Herpailurus
yagouaroundi, (h) Puma concolor, (i) Mazama gouazoubira, (j) Tamandua tetradactyla, (k)
Euphractus sexcinctus, (l) Dasypus novemcinctus, (m) Sapajus libidinosus e (n) Didelphis
albiventris.
Além dos mamíferos silvestres, registramos sete espécies de mamíferos domésticos ou
asselvajados (caso de algumas populações de Equus asinus) nas nossas áreas estudo: vaca (Bos
taurus, 469 registros), cabra (Capra hircus, 440), burro (E. asinus, 74), ovelha (Ovis aries, 65),
cavalo (Equus caballus, 25), cão (Canis lupus familiaris, 39) e gato-doméstico (Felis catus,
12). Esses registros (1124) representam 41% de todos os registros de médios e grandes
mamíferos (considerando silvestres e domésticos). Obtivemos ainda 83 registros de pessoas,
algumas acompanhadas por cães e com apetrechos utilizados em atividades de caça.
A curva de acumulação de espécies para o Rio Grande do Norte, ou seja, considerando
todas as amostras conjuntamente, apresentou uma tendência à estabilização com pouco mais da
metade das unidades amostrais (Fig. 3A). Contudo, o estimador de riqueza Jacknife de primeira
ordem indicou que cerca de duas espécies ainda poderiam ser registradas com um maior esforço
amostral (Fig. 3A). Isto é mais evidente para as áreas com maior riqueza como Lajes e Serrinha
dos Pintos, com exceção de áreas como Martins, Caiçara do Norte e Dunas do Rosado, as curvas
das demais áreas permanecem crescentes (Fig. 3B).
As espécies com o maior número de registros e consequentemente maiores índices de
abundância relativa foram Cerdocyon thous, D. albiventris, Mazama gouazoubira e L. tigrinus
(respectivamente, Fig. 4, Tabela 2). Por outro lado, Puma concolor e Galictis cuja foram
registrados uma única vez cada, exibindo assim os menores índices de abundância relativas,
35
seguidos por Leopardus pardalis e H. yagouaroundi (Fig. 4, Tabela 2). C. thous, D. albiventris,
L. tigrinus e Euphractus sexcinctus foram registrados em todas as 10 áreas (Tabela 2).
No presente trabalho não consideramos registros de Callithrix jacchus como mamífero
de médio porte, enquanto que D. albiventris foi incluído neste grupo por seu maior tamanho
(500-2700 g) e hábito escansorial (Paglia et al. 2012) que facilita sua detecção por
armadilhamento fotográfico.
36
Tabela 2. Mamíferos de médio e grande porte registrados em 10 áreas prioritárias para a conservação da Caatinga do Rio Grande do Norte, nordeste do Brasil.
N – número de registros. Dieta ou grupo funcional de acordo com Paglia et al., 2012: Ca – Carnívoro, Fr – Frugívoro, Hb – Herbívoro pastador, In – Insetívoro,
Myr – Mirmecófago, On – Onívoro. Estado de conservação segundo a IUCN (2017) ou MMA (2014): LC (pouco preocupante), VU (vulnerável), EN (em
perigo), NC (não consta na lista). Os símbolos para as áreas amostradas são: Serra de Santana (SS), Lajes (LA), Cerro Corá (CC), Martins (MA), Serrinha dos
Pintos (SP), Felipe Guerra (FG), Caiçara do Norte (CN), Luís Gomes (LG), Dunas do Rosado (DR), Coronel Ezequiel (CE). Para mais informações sobre as
áreas acessar: https://brasil.wcs.org/pt-br/Lugares-naturais/Projeto-Caatinga.aspx.
Táxon Nome comum Áreas prioritárias
N Dieta IUCN/
MMA SS LA CC MA SP FG CN LG DR CE
ORDEM CETARTIODACTYLA
Família Cervidae
Mazama gouazoubira
(G, Fischer [von Waldheim], 1814) Veado-catingueiro X X X X X X X X 176 Fr/Hb LC/NC
ORDEM CARNIVORA
Família Canidae
Cerdocyon thous (Linnaeus, 1766)
Cachorro-do-mato X X X X X X X X X X 693 In/On LC/NC
Família Felidae
Leopardus tigrinus
(Thomas, 1904)
Gato-do-mato-
pintado X X X X X X X X X X 157 Ca VU/EN
Leopardus pardalis (Linnaeus, 1758)
Jaguatirica X 4 Ca LC/NC
Herpailurus yagouaroundi
(É, Geoffroy Saint-Hilare, 1803) Gato-mourisco X X X 7 Ca LC/VU
Puma concolor (Linnaeus, 1771)
Onça-parda X 1 Ca LC/VU
Família Procyonidae
37
Procyon cancrivorus
(G,[Baron] Cuvier, 1798) Mão-pelada X X X X X X X X 29 Fr/On LC/NC
Família Mephitidae
Conepatus amazonicus
(Lichtenstein, 1838) Jaritataca X X X X X X 20 In/On LC/NC
Família Mustelidae
Galictis cuja (Molina, 1782)
Furão X 1 Ca LC/NC
ORDEM CINGULATA
Família Dasypodidae
Dasypus novemcinctus
Linnaeus, 1758 Tatu-galinha X X X X X 22 In/On LC/NC
Euphractus sexcinctus
(Linnaeus, 1758) Tatu-peba X X X X X X X X X X 90 In/On LC/NC
ORDEM PILOSA
Família Myrmecophagidae
Tamandua tetradactyla (Linnaeus, 1758)
Tamanduá-mirim X X X X X 10 Myr LC/NC
ORDEM PRIMATES
Família Cebidae
Sapajus libidinosus (Spix, 1823)
Macaco-prego X X X X 38 Fr/On LC/NC
ORDEM DIDELPHIMORPHIA
Família Didelphidae
Didelphis albiventris
(Lund, 1840)
Gambá-de-orelha-
branca X X X X X X X X X X 359 Fr/On LC/NC
38
Fig. 3: Curva de acumulação de espécies observada (linha preta) com intervalo de confiança de 95%
(linhas tracejadas cinzas) e curva de riqueza estimada por Jacknife 1 (círculos pretos) para a comunidade
de mamíferos de médio e grande porte de 10 áreas prioritárias da Caatinga do Rio Grande do Norte,
nordeste do Brasil (A); também são apresentadas individualmente as curvas aleatorizadas de
acumulação de espécies para cada uma das 10 áreas prioritárias para a conservação estudadas (B).
39
Fig. 4: Índice de abundância relativa (ou sucesso de captura) das 14 espécies de mamíferos de médio e
grande porte registrados em 10 áreas prioritárias para a conservação da Caatinga do Rio Grande do
Norte, nordeste do Brasil. Fonte das ilustrações: De Angelo et al. 2008.
Discussão
As 14 espécies de mamíferos silvestres de médio e grande porte registradas neste trabalho
compreendem cerca de um terço das 45 espécies que ocorrem em toda Caatinga (Carmignotto
& Astúa 2017), e metade das 28 espécies registradas na Caatinga dos estados de Alagoas,
Pernambuco, Paraíba e Ceará (Feijó & Langguth 2013). É preciso ressaltar, contudo, que
Carmignotto e Astúa (2017) incluem espécies com ocorrência atual bastante restrita no bioma
como Panthera onca e Tapirus terrestris; enquanto Feijó e Langguth (2013) abrangem espécies
restritas a brejos de altitude como Coendou baturitensis e Nasua nasua. Nossos resultados são
fruto de um dos maiores investimentos em armadilhamento fotográfico já realizados até o
40
momento na Caatinga, tanto em termos de esforço quanto em abrangência espacial (Tabela 3).
Além disto, este trabalho contribui para diminuir a lacuna de dados mastozoológicos existente
no Rio Grande do Norte, evidenciada por Feijó e Langguth (2013). Inventários de mamíferos
realizados em outros estados registraram de cinco a 25 espécies de médios e grandes mamíferos
(ver Tabela 3), com as maiores riquezas geralmente registradas em ambientes mésicos como
brejos de altitude e melhor protegidos como unidades de conservação (Tabela 3). Contudo,
diferenças no esforço amostral, abrangência espacial e métodos de levantamento também
devem ser considerados. Além disso, variações na riqueza e composição entre as áreas
amostradas neste estudo devem ser consideradas.
41
Tabela 3. Riqueza de mamíferos de médio e grande porte registrados em levantamentos em áreas de Caatinga (clima semiárido), brejos de altitude (ambientes
mésicos relacionados a florestas tropicais úmidas) e outras formações associadas dentro do domínio da Caatinga, com seus respectivos métodos empregados
(armadilhamento fotográfico - af; armadilhas de queda ou captura viva - aq; busca ativa por espécimes e vestígios aleatória ou através de transectos - ba;
entrevistas - en; espécimes de coleção ou museu - ec; parcelas de pegadas - pp; registros oportunistas – ro). Para o método armadilhamento fotográfico é
apresentado (entre parênteses) o respectivo esforço amostral em câmeras-dias. Levantamentos realizados em Unidades de Conservação estão indicados na coluna
(UC). Não consideramos registros de Callithrix jacchus, enquanto Didelphis albiventris foi avaliado como mamífero de médio e grande porte. Dias et al. (2014)
é focado somente em carnívoros.
Estado Local, município (s) Ambiente UC Riqueza Métodos e esforço Referência
Ceará PARNA de Ubajara, Ubajara Brejo de altitude Sim 19 ba, en Guedes et al. (2000)
Ceará, Pernambuco e Paraíba
RPPN Serra das Almas (CE), RPPN
Maurício Dantas (PE), Parque Estadual Pedra da
Boca (PB), vários municípios
Caatinga Sim 13 aq, en, op Cruz et al. (2005)
Bahia PARNA da Chapada Diamantina e
arredores, vários municípios
Formações de Caatinga, Cerrado e
Floresta Atlântica
Sim 25 aq, ec, en, op Pereira & Geise (2009)
Sergipe Fazenda São Pedro, Porto da Folha Caatinga stricto
sensu Não 5 ba, ni Freitas et al. (2011)
Sergipe, Alagoas e Bahia
14 locais amostrais, vários municípios
Caatinga stricto sensu
Não 5 aq, bc, en Bezerra et al. (2014)
Sergipe Serra dos Macacos, Tobias Barreto Caatinga stricto sensu
Não 7 ap, ba, en Dias et al. (2014)
Alagoas Serra do Mamão, Traipu Brejo de altitude Não 18 aq, ba Silva & Palmeira (2014)
Ceará Maciço de Baturité, vários
municípios Brejo de altitude Não 18 ec, aq, ro, en
Fernandes-Ferreira et al.
(2015)
42
Sergipe e Bahia
Serra da Guia, Poço
Redondo (SE) e Pedro Alexandre
(BA)
Caatinga stricto sensu
Não 8 Ba Rocha et al. (2015)
Paraíba Serra de Santana, vários municípios Caatinga stricto
sensu Não 11 af (475), aq, ba Campos et al. (2016)
Piauí e Pernambuco Ouricuri (CE) e São João do Piauí
(PI)
Caatinga stricto
sensu Não 13 af (721), aq, ba Delciellos (2016)
Sergipe
MONA Grota do Angico, Poço
Redondo e Canindé do São Francisco
Caatinga stricto
sensu Sim 12 af (2.912), pp
Dias & Bocchiglieri
(2016)
Ceará e Piauí RPPN Serra das Almas, Crateús (CE) e Buriti dos Montes (PI)
Caatinga stricto sensu
Sim 17 af (3.600), aq, ba,
ro Dias et al. (2017)
Bahia Serra de Santana, Senhor do Bonfim
e Jaguarari
Caatinga stricto
sensu Não 13 ba, en
Pereira & Peixoto
(2017)
Rio Grande do
Norte
10 áreas prioritárias, vários
municípios
Caatinga stricto
sensu
Sim (1 área),
Não (9 áreas) 14 af (7.221) Este estudo
43
Três dos felinos registrados (L. tigrinus, H. yagouaroundi e P. concolor) encontram-1
se ameaçados de extinção, nacionalmente ou globalmente (Tabela 2), e são alvos de 2
Planos de Ação Nacionais para a Conservação das Espécies Ameaçadas de Extinção 3
(ICMBIO 2017). Como principais ameaças para essas espécies estão a perda de habitat, 4
a perseguição resultante de conflitos com criadores, os atropelamentos e a transmissão de 5
doenças por carnívoros domésticos (Azevedo et al. 2013; Almeida et al. 2013; Oliveira 6
et al. 2013). Aqui optamos por seguir a última classificação dos felinos do mundo 7
(Kitchener et al. 2017) considerando sua relevância taxonômica e de conservação. 8
Contudo, pesquisas recentes sugerem que as populações de L. tigrinus das regiões 9
nordeste, parte do norte e Brasil central constituem uma espécie distinta e endêmica do 10
país, nomeada de Leopardus emiliae (Nascimento & Feijó 2017; Ruiz-García et al. 2017). 11
No caso de L. pardalis, embora tenha saído da lista nacional de espécies ameaçadas, a 12
espécie é classificada como Vulnerável no estado da Bahia (Cassano et al. 2017) e pode 13
se encontrar em estado semelhante em outros estados do nordeste do país (Feijó & 14
Langguth 2013; Marinho et al. 2017). De forma geral, trabalhos que abordam aspectos da 15
ecologia e conservação desses felinos na Caatinga têm surgido apenas nos últimos anos 16
e ainda são muito insipientes (Marinho et al. 2018a; b; Astete et al. 2017; Penido et al. 17
2017), especialmente se considerado o seu importante papel de predadores na 18
estruturação e regulação das comunidades biológicas. 19
Na Caatinga, P. concolor se encontra em estado mais crítico que o nacional, 20
classificada como Em Perigo (Azevedo et al. 2013), o que é reforçado pelo nosso único 21
registro desse predador de topo de cadeia em Luís Gomes. Nessa região a vegetação 22
arbórea é predominante e cobre serras íngremes e de difícil acesso que podem alcançar 23
até 800 m de altitude, podendo fornecer presas e refúgios que favorecem a persistência 24
da espécie. Contudo, a caça dessas presas potenciais e o desmatamento para agricultura 25
44
são comuns na região, e por isso devem ser mitigados juntamente com os abates por 26
conflitos com criadores. O registro mais próximo e mais recente de P. concolor está a 27
aproximadamente 75 km, no estado da Paraíba (Campos et al. 2016), também em uma 28
área de altitude elevada. Na região entre Lajes e Cerro Corá, uma das mais relevantes em 29
termos de habitat e disponibilidade de presas, moradores locais relatam a presença recente 30
da espécie, contudo as evidências mais concretas são de animais abatidos há mais de 20 31
anos (P.H. Marinho obs. pessoal; Pichorim et al. 2014) (Fig. 5A). O avançado declínio 32
de predadores de topo como P. concolor implica em desequilíbrios ecológicos ainda 33
desconhecidos na Caatinga, como um possível aumento na abundância de 34
mesopredadores (Crooks & Soulé 1999), que no presente estudo representaram 50% das 35
espécies e 56% de todos os registros de mamíferos silvestres. 36
37
38
45
Fig. 5. Registros de ameaças para mamíferos de médio e grande porte na Caatinga do Rio Grande 39
do Norte. Pele de Puma concolor abatido há mais de 20 anos na região de Lajes (A) e espécies 40
silvestres criadas como animais domésticos: Herpailurus yagouaroundi (B), Leopardus tigrinus 41
(C), Mazama gouazoubira (D) e Sapajus libidinosus (E). Fotos: P.H. Marinho (A e C), D. Bezerra 42
(B, D e E). 43
44
Entre os mamíferos com maior índice de abundância relativa e detectados em todas 45
as áreas estudadas estão espécies generalistas de hábitat. C. thous e D. albiventris são 46
espécies amplamente distribuídas, onívoras e tolerantes a perturbações antrópicas 47
(Aléssio et al. 2005; Beisiegel et al. 2013), estando entre as mais registradas em outros 48
estudos na Caatinga (e.g. Delciellos 2016; Dias & Bocchiglieri 2016; Dias et al. 2017). 49
E. sexcinctus também foi registrado em todas as 10 áreas e embora seja bastante caçada, 50
essa espécie de hábitos onívoros parece ser abundante e amplamente presente na Caatinga 51
(Feijó & Langgtuh 2013; Alves te al. 2016). Por sua vez, como sugerem nossos 52
resultados, L. tigrinus parece ser o felino mais abundante e amplamente presente em boa 53
parte da Caatinga (Feijó & Langguth, 2013), embora sua presença seja mais esperada em 54
áreas florestadas e com menor interferência antrópica (Marinho et al. 2018a). Outra 55
espécie que merece destaque é M. gouazoubira, não registrado somente em duas áreas 56
(Tabela 2). Embora esse cervídeo seja considerado relativamente tolerante a ambientes 57
perturbados (Duarte et al. 2012), a grande pressão de caça tem levado M. gouazoubira ao 58
declínio ou mesmo à extinção local em áreas mais perturbadas da Caatinga (Bezerra et al. 59
2014). 60
Nossa detecção, contudo, é imperfeita e a ausência de registros de algumas espécies 61
em determinadas áreas estudadas pode não representar a realidade, da mesma forma a 62
abundância relativa também deve ser interpretada com cautela (ver Sollmann et al. 2013). 63
Galictis cuja, por exemplo, se locomove rapidamente e prefere ambientes ripários, o que 64
46
diminui suas chances de detecção (Magioli et al. 2014). No caso de H. yagouaroundi e L. 65
pardalis, embora não tenhamos registrado, provavelmente devido às suas baixas 66
densidades populacionais, esses dois felinos também ocorrem em Serra de Santana e 67
Lajes (Pichorim et al. 2014; Marinho et al. 2017; PH Marinho dados não publicados). 68
Todas as espécies encontradas são de ampla distribuição no Brasil, não endêmicas e 69
registradas para estados próximos como Paraíba e Ceará (e.g. Feijó & Langguth 2013), 70
de forma que sua ocorrência do Rio Grande do Norte já era esperada (e.g. Oliveira 2004; 71
Ferreira et al. 2009; Marinho et al. 2017), embora não documentada formalmente na 72
maioria dos casos. Contudo, algumas espécies com ocorrência confirmada para a 73
Caatinga do estado não foram registradas aqui. Sylvilagus brasilensis ocorre em uma área 74
a aproximadamente 35 km a sudeste de Caiçara do Norte (Dantas et al. 2016). Já Lontra 75
longicaudis, espécie semi-aquática, ocorre na faixa de transição da Caatinga com a 76
Floresta Atlântica (Rosas-Ribeiro al. 2017), região não amostrada neste trabalho. Por fim, 77
embora não tenha sido foco do presente estudo pelo seu menor porte (< 1 kg), Kerodon 78
rupestris merece nota por estar em estado Vulnerável no Brasil (MMA 2014) e ser 79
endêmico da Caatinga. Encontramos vestígios da presença desse roedor em afloramentos 80
rochosos de quase todas as áreas, com exceção de Dunas do Rosado e Caiçara do Norte, 81
onde essas formações são raras. 82
Nossos resultados revelam, por outro lado, o relativo grau de empobrecimento da 83
mastofauna do extremo nordeste da Caatinga. Entre as espécies com distribuição prevista 84
para o estado segundo os mapas de distribuição da IUCN (2017) e registros em estados 85
vizinhos (Feijó & Langguth 2013; Feijó et al. 2015), Dasyprocta prymnolopha, Pecari 86
tajacu e Tolypeutes tricinctus, por exemplo, não foram registrados neste estudo, mesmo 87
com o nosso significativo esforço amostral e abrangência espacial. Embora existam 88
relatos sobre a presença dessas espécies na Caatinga do estado, baseados principalmente 89
47
em entrevistas e registros fósseis (Oliveira 2004; Ferreira et al. 2009; Araújo-Júnior & 90
Porpino 2011; Lucena & Freire 2012; Barboza et al. 2016), não obtivemos vestígios ou 91
mesmo informações seguras com os moradores locais da sua presença nas nossas áreas 92
de estudo (P.H. Marinho, dados não publicados). Relatos históricos sobre a extinção local 93
dessas espécies há mais de 50 anos em algumas regiões do estado citam a caça intensa 94
como principal causa (Faria 1961). Contudo, não podemos descartar a presença dessas 95
espécies na Caatinga do Rio Grande do Norte, uma vez que elas podem ocorrer em locais 96
mais bem protegidos e de difícil acesso, mas provavelmente em densidades extremamente 97
baixas e em declínio populacional, o que compromete a efetividade das funções 98
ecológicas desempenhadas por delas (Galetti & Dirzo 2013). 99
Entre as principais ameaças para os mamíferos de médio e grande porte da Caatinga 100
estão aquelas globalmente comuns ao grupo como a caça e a perda e degradação dos 101
habitats (Alves et al. 2016; Feijó & Langguth 2013). A degradação da vegetação da 102
Caatinga para fins domésticos e industriais associada à pecuária extensiva compromete a 103
conservação e regeneração dos habitats. O grande número de animais domésticos 104
registrados por nós reforça a gravidade desse problema mesmo em áreas prioritárias para 105
a conservação. Da mesma forma, a caça vem promovendo o declínio de médios e grandes 106
mamíferos no bioma há séculos (Faria 1961; Fernandes-Ferreira 2014), seja para fins de 107
alimentação ou por conflitos com predadores (Alves et al. 2016; Barboza et al. 2016) 108
(Fig. 5A). A caça está muito presente na Caatinga do Rio Grande do Norte, vide nossos 109
registros fotográficos e a grande quantidade de vestígios de caça encontrados durante as 110
amostragens. Além disso, a criação de animais silvestres como domésticos é 111
relativamente comum na região (Fernandes-Ferreira et al. 2015; Alves te al. 2016; 112
Delciellos 2016) e atinge espécies como H. yagouaroundi, L. tigrinus, M. gouazoubira e 113
S. libidinosus (Pichorim et al. 2014; P.H. Marinho obs. pessoal) (Fig. 5). Adicionalmente, 114
48
obras de infraestrutura e estradas elevam o grau de perda e fragmentação dos habitats da 115
Caatinga (Silva et al. 2017b). No Rio Grande do Norte parques eólicos e linhas de 116
transmissão de energia, por exemplo, tem se multiplicado nos últimos anos, com os seus 117
impactos sobre a fauna sendo geralmente subestimados (Bernard et al. 2014). 118
Relatamos aqui o primeiro amplo e intensivo inventário de mamíferos de médio e 119
grande porte para o estado do Rio Grande do Norte, nordeste do Brasil, uma região até 120
então subamostrada para o grupo. As áreas estudadas abrigam uma porção importante da 121
diversidade de mamíferos da Caatinga incluindo espécies ameaçadas, mas também 122
indicam ausência relevantes. Nossos resultados reforçam a importância e urgência da 123
criação de áreas protegidas nas áreas prioritárias estudadas e a efetivação e 124
implementação de iniciativas já em andamento como o Monumento Natural das Cavernas 125
de Martins (em discussão) e a Área de Proteção Ambiental Dunas do Rosado (recém-126
criada) (IDEMA 2018). Ações como fiscalização efetiva, políticas abrangentes de 127
educação ambiental e incentivo a práticas de manejo que busquem compatibilizar a 128
exploração de áreas privadas com a conservação da biodiversidade da Caatinga devem 129
ser priorizadas nas áreas sem proteção legal. 130
As informações aqui apresentadas devem estimular a realização de pesquisas 131
ecológicas sobre as espécies registradas e novos levantamentos no estado, especialmente 132
em áreas não cobertas neste estudo como a região centro-sul, conhecida como Seridó e a 133
faixa leste de transição com a Floresta Atlântica, além das suas unidades de conservação. 134
Esperamos ainda que esses resultados subsidiem avaliações regionais do estado de 135
conservação das espécies, instrumento essencial para considerar as particularidades das 136
populações locais (Cassano et al. 2017). Por fim, dada a escassez de conhecimento, as 137
informações apresentadas aqui podem auxiliar a avaliação de empreendimentos com 138
potencial de impactar a mastofauna terrestre da região. 139
49
Agradecimentos 140
Agradecemos a Wildlife Conservation Society - Brasil pela parceria e assistência 141
geral. Também somos gratos ao Tropical Conservation Act (TFCA) através do Fundo 142
Brasileiro de Biodiversidade (FUNBIO) (chamada 04/2012) e à Fundação Grupo 143
Boticário de Proteção à Natureza (projeto 0982-20132) pelo o apoio financeiro; ao Centro 144
de Pesquisas Ambientais do Nordeste (CEPAN) pela assistência institucional. A A.F. 145
Oliveira, D. Valdenor, M.C. Bezerra, A. Galvão, F.P. Marinho e W. Pessoa somo gratos 146
pela assistência de campo; a T.G. Oliveira pela assistência na identificação dos pequenos 147
felinos pintados; e a C. Lisboa pelos comentários que ajudaram a melhorar o texto. A 148
João B. de Lima (seu João) e vário outros moradores locais da Caatinga seremos 149
eternamente gratos por sua assistência no campo e hospitalidade, essenciais para o 150
sucesso deste trabalho. PHM (130648 / 2013-2), CRF (305304 / 2013-5) e EMV 151
(308040/2017-1) foram financiados pelo Conselho Nacional de Desenvolvimento 152
Científico e Tecnológico (CNPq) e PHM, DB e MAF foram apoiados pela Coordenação 153
de Aperfeiçoamento de Pessoal de Nível Superior -Brasil (CAPES) - Código de 154
Financiamento 001. 155
156
Literatura citada 157
AB’SÁBER, N. A. 1974. O domínio morfoclimático semi-árido das caatingas brasileiras. 158
Geomorfologia 43:1-36. 159
ALBUQUERQUE, U. P. ET AL. 2012. Caatinga revisited: ecology and conservation of 160
an important seasonal dry forest. The Scientific World Journal 2012. 161
ALÉSSIO, F. M., A. R. M. PONTES, & V. L. SILVA. 2005. Feeding by Didelphis 162
albiventris on tree gum in the northeastern Atlantic Forest of Brazil. Mastozoología 163
Neotropical 12:53-56. 164
50
ALMEIDA, L. B., D. QUEIROLO, B. M. BEISIEGEL, & T. G. OLIVEIRA. 2013. 165
Avaliação do risco de extinção do gato-mourisco Puma yagouaroundi (É. Geoffroy Saint-166
Hilaire, 1803) no Brasil. Biodiversidade Brasileira 1:99-106. 167
ALVES, R. R. N. ET AL. 2016. Game mammals of the Caatinga biome. Ethnobiology 168
and Conservation 5:1-51. 169
ANTONGIOVANNI, M., E. M. VENTICINQUE, & C. F. FONSECA. 2018. 170
Fragmentation patterns of the Caatinga drylands. Landscape Ecology 33:1353-1367. 171
ARAÚJO-JÚNIOR, H. I., & K. O. PORPINO. 2011. Assembleias fossilíferas de 172
mamíferos do Quaternário do Estado do Rio Grande do Norte, Nordeste do Brasil: 173
diversidade e aspectos tafonômicos e paleoecológicos. Pesquisas em Geociências 38:67-174
83. 175
ASTETE, S. ET AL. 2017. Living in extreme environments: modeling habitat suitability 176
for jaguars, pumas, and their prey in a semiarid habitat. Journal of Mammalogy 98:464-177
474. 178
AZEVEDO, F. C. ET AL. 2013. Avaliação do risco de extinção da onça-parda, Puma 179
concolor (Linnaeus, 1771) no Brasil. Biodiversidade Brasileira 3:107-121. 180
BANDA, K. ET. AL. 2016. Plant diversity patterns in neotropical dry forests and their 181
conservation implications. Science 353:1383-1387. 182
BARBOZA, R. D., S. F. LOPES, W. M. S. SOUTO, H. FERNANDES-FERREIRA, & 183
R. R. N. ALVES. 2016. The role of game mammals as bushmeat in the Caatinga, 184
northeast Brazil. Ecology and Society 21:2. 185
BEISIEGEL, B. M., F. C. LEMOS, F. C. AZEVEDO, D. QUEIROLO, & R. S. P. 186
JORGE. 2013. Avaliação do risco de extinção do Cachorro-do-mato Cerdocyon thous 187
(Linnaeus, 1766) no Brasil. Biodiversidade Brasileira 3:138-145. 188
51
BERNARD, E., A. PAESE, R. B. MACHADO, & L. M. S. AGUIAR. 2014. Blown in 189
the wind: bats and wind farms in Brazil. Natureza & Conservação 12:106-111. 190
BEZERRA, A. M. R., A. LAZAR, C. R. BONVICINO, & A. CUNHA. 2014. Subsidies 191
for a poorly known endemic semiarid biome of Brazil: non-volant mammals of an eastern 192
region of Caatinga. Zoological Studies 53:1-16. 193
BRITO, D., L. C. OLIVEIRA, M. OPREA, & M. A. R. MELLO. 2009. An overview of 194
Brazilian mammalogy: trends, biases and future directions. Zoologia 26:67-73. 195
CAMPOS, B. A. T. P. ET AL. 2016. Mastofauna da Serra de Santa Catarina. 196
Biodiversidade na Serra de Santa Catarina – PB: uma proposta de criação do Parque 197
Estadual das Águas Sertanejas. (H. F. P. Araujo & A. H. Vieira-Filho, orgs.). Areia. 198
CARDILLO, M. ET AL. 2005. Multiple Causes of High Extinction Risk in Large 199
Mammal Species. Science 309:1239-1241. 200
CARMIGNOTTO, A. P., & D. ASTÚA. 2017. Mammals of the Caatinga: diversity, 201
ecology, biogeography, and conservation. Caatinga: The largest tropical dry forest 202
region in South America. (J. M. C. Silva, I. R. Leal & M. Tabarelli, eds.). Springer, 203
Cham. 204
CASSANO, C. R. ET AL. 2017. Primeira avaliação do status de conservação dos 205
mamíferos do estado da Bahia, Brasil. Oecologia Australis 21:156-170. 206
CHEYNE, S. M., W. J. SASTRAMIDJAJA, Y. RAYADIN, & D. W. MACDONALD. 207
2016. Mammalian communities as indicators of disturbance across Indonesian Borneo. 208
Global Ecology and Conservation 7:157-173. 209
CHIARELLO, A. G. 1999. Effects of fragmentation of the Atlantic forest on mammal 210
communities in south-eastern Brazil. Biological Conservation 89:71-82. 211
52
COLWELL, R. K. 2013. EstimateS: Statistical estimation of species richness and shared 212
species from samples. Version 9. User's Guide and application 213
<http://purl.oclc.org/estimates>. 214
CROOKS, K. R., & M. E. SOULÉ. 1999. Mesopredator release and avifaunal extinctions 215
in a fragmented system. Nature 400:563-566. 216
CRUZ, M. A. O. M. ET AL. 2005. Diversidade de mamíferos em áreas prioritárias para 217
conservação da Caatinga. Análise das Variações da Biodiversidade do Bioma Caatinga: 218
Suporte a Estratégias Regionais de Conservação (F. S. Araujo, M. J. Rodal & M. R. V. 219
Barbosa, eds.). Ministério do Meio Ambiente, Brasília. 220
DANTAS, A. R. C., F. H. MENEZES, K. S. SERRA, E. D. O. BARBOSA, & H. 221
FERNANDES-FERREIRA. 2016. First record of Sylvilagus brasiliensis (Linnaeus, 222
1758) (Lagomorpha: Leporidae) in Rio Grande do Norte state, Northeast Brazil. Check 223
List 12:1856. 224
DE ANGELO. C., A. PAVIOLO, Y. DI BLANCO, M. DI BITETTI, & A. CHIAPPE. 225
2008. Guía de huellas de los mamíferos de Misiones y otras áreas del subtrópico de 226
Argentina. Ediciones del Subtrópico, Tucumán. 227
DELCIELLOS, A.C. 2016. Mammals of four Caatinga areas in northeastern Brazil: 228
inventory, species biology, and community structure. Check List 12:1916. 229
DIAS, D. M., & A. BOCCHIGLIERI. 2016. Riqueza e uso do habitat por mamíferos de 230
médio e grande porte na Caatinga, nordeste do Brasil. Neotropical Biology and 231
Conservation 11:38-46. 232
DIAS, D. M., A. S. RIBEIRO, A. BOCCHIGLIERI, & T. C. PEREIRA. 2014. 233
Diversidade de carnívoros (Mammalia: Carnivora) da serra dos macacos, Tobias Barreto, 234
Sergipe, Brasil. Bioscience Journal 30:1192-1204. 235
53
DIAS, D. M., P. G. GUEDES, S. S. P. SILVA, & L. M. M. SENA. 2017. Diversity of 236
nonvolant mammals in a Caatinga area in northeastern Brazil. Neotropical Biology and 237
Conservation 12:200-208. 238
DUARTE, J. M. B. ET AL. 2012. Avaliação do Risco de Extinção do Veado-catingueiro, 239
Mazama gouazoubira G. Fischer (von Waldheim), 1814, no Brasil. Biodiversidade 240
Brasileira 2:50-58. 241
EMIDIO, R. A., & R. G. FERREIRA. 2012. Energetic Payoff of Tool Use for Capuchin 242
Monkeys in the Caatinga: Variation by Season and Habitat Type. American Journal of 243
Primatology 74:332–343. 244
FARIA, O. L. 1961. A Caça nos sertões do Seridó. Ministério da Agricultura, Serviço de 245
Informação Agrícola, Rio de Janeiro. 246
FEIJÓ, A., & A. LANGGUTH. 2013. Mamíferos de Médio e Grande Porte do Nordeste 247
do Brasil: Distribuição e Taxonomia, com Descrição de Novas Espécies. Revista 248
Nordestina de Biologia 22:3-225. 249
FEIJÓ, A., G. S. T. GARBINO, B. A. T. P. CAMPOS, P. A. ROCHA, S. F. FERRARI, 250
& A. LANGGUTH. 2015. Distribution of Tolypeutes Illiger, 1811 (Xenarthra: Cingulata) 251
with comments on its biogeography and conservation. Zoological Science 32:77-87. 252
FERNANDES-FERREIRA, H. 2014. A caça no Brasil: panorama histórico e atual. Tese 253
de doutorado. Universidade Federal da Paraíba, João Pessoa-PB, Brasil. 254
FERNANDES-FERREIRA, H., N. M. GURGEL-FILHO, A. FEIJÓ, S. V. 255
MENDONÇA, R. R. N. ALVES, & A. LANGGUTH. 2015. Non-volant mammals from 256
Baturité Ridge, Ceará state, Northeast Brazil. Check List 11:1630. 257
FERREIRA, R. G. ET AL. 2009. On the occurrence of Cebus flavius (Schreber 1774) in 258
the Caatinga, and the use of semi-arid environments by Cebus species in the Brazilian 259
state of Rio Grande do Norte. Primates 50:357-362. 260
54
FERREIRA, R. G., R. A. EMIDIO, & L. JERUSALINSKY. 2010. Three stones for three 261
seeds: natural occurrence of selective tool use by capuchins (Cebus libidinosus) based on 262
an analysis of the weight of stones found at nutting sites. American Journal of 263
Primatology 72:270-275. 264
FONSECA, C. R., M. ANTONGIOVANNI, M. MATSUMOTO, E. BERNARD, & E. 265
M. VENTICINQUE. 2017. Conservation opportunities in the Caatinga. The largest 266
tropical dry forest region in South America. (J. M. C. Silva, I. R. Leal & M. Tabarelli, 267
eds.). Springer, Cham. 268
FREITAS, E. B., C. B. DE-CARVALHO, & S. F. FERRARI. 2011. Abundance of 269
Callicebus barbarabrownae (Hershkovitz 1990), (Primates: Pitheciidae) and other 270
nonvolant mammals in a fragment of arboreal Caatinga in northeastern Brazil. Mammalia 271
75:339-343. 272
GALETTI, M., & R. DIRZO. 2013. Ecological and evolutionary consequences of living 273
in a defaunated world. Biological Conservation 163:1–6. 274
GUEDES, P. G. ET AL. 2000. Diversidade de Mamíferos do Parque Nacional de Ubajara 275
(Ceará, Brasil). Mastozoología Neotropical 7:95-100. 276
ICMBIO. 2017. Planos de Ação Nacional. 277
<http://www.icmbio.gov.br/portal/faunabrasileira/planos-de-acao-nacional> 278
IDEMA. 2014. Perfil do Rio Grande do Norte. Instituto de Desenvolvimento Econômico 279
E Meio Ambiente, Natal. 280
<http://adcon.rn.gov.br/ACERVO/seplan/DOC/DOC000000000129527.PDF> 281
IDEMA. 2018. Unidades Estaduais de Conservação Ambiental do RN. 282
<http://www.idema.rn.gov.br/Conteudo.asp?TRAN=ITEM&TARG=334&ACT=null&283
PAGE=0&PARM=null&LBL=Unidades+de+Conserva%C3%A7%C3%A3o> 284
55
IUCN. 2017. The IUCN Red List of Threatened Species. Version 2017-2. 285
<http://www.iucnredlist.org> 286
KITCHENER, A. C. ET AL. 2017. A revised taxonomy of the Felidae. The final report 287
of the Cat Classification Task Force of the IUCN/SSC Cat Specialist Group. Cat News 288
Special, Issue 11:80. 289
LINNELL, J. D. C., J. E. SWENSON, & R. ANDERSEN. 2000. Conservation of 290
biodiversity in Scandinavian boreal forests: large carnivores as flagships, umbrellas, 291
indicators, or keystones?. Biodiversity and Conservation 9:857-868. 292
LUCENA, M. M., & E. M. X. FREIRE. 2012. Environmental perception and use of fauna 293
from a Private Natural Heritage Reserve (RPPN) in Brazilian semiarid. Acta Scientiarum 294
34:335-341. 295
MAGIOLI, M., K. M. P. M. B. FERRAZ, & M. G. RODRIGUES. 2014. Medium and 296
large-sized mammals of an isolated Atlantic Forest remnant, southeast São Paulo State, 297
Brazil. Check List 10:850-856. 298
MARES, M. A., M. R. WILLIG, K. E. STEILEIN, & T. E. LACHER-JR. 1981. The 299
mammals of Northeastern Brazil: a preliminary assessment. Annals of the Carnegie 300
Museum 50:81-137. 301
MARINHO, F. P., G. G. MAZZOCHINI, A. P. MANHÃES, W. W. WEISSER, & G. 302
GANADE. 2016. Effects of past and present land use on vegetation cover and 303
regeneration in a tropical dryland forest. Journal of Arid Environment 132:26-33. 304
MARINHO, P. H., A. FEIJÓ, A. S. GALIVAN, E. O. MOURA, & E. M. 305
VENTICINQUE. 2017. First records of Ocelot Leopardus pardalis (Linnaeus, 1758) 306
(Carnivora: Felidae) from Rio Grande do Norte, northeastern Brazil. Check List 13:2087. 307
56
MARINHO, P. H., D. BEZERRA, M. ANTONGIOVANNI, C. R. FONSECA, & E. M. 308
VENTICINQUE. 2018a. Estimating occupancy of the vulnerable northern tiger cat 309
Leopardus tigrinus in Caatinga drylands. Mammal Research 83:33–42. 310
MARINHO, P. H., D. BEZERRA, M. ANTONGIOVANNI, C. R. FONSECA, & E. M. 311
VENTICINQUE. 2018b. Activity patterns of the threatened northern tiger cat Leopardus 312
tigrinus and its potential prey in a Brazilian dry tropical forest. Mammalian Biology 313
89:30-36. 314
MELO, G. L., & J. SPONCHIADO. 2012. Distribuição geográfica dos marsupiais do 315
Brasil. Os Marsupiais do Brasil, Revisada e Ampliada (N. C. Cáceres, ed.). 2 ed. Editora 316
da UFMS, Campo Grande. 317
MITTERMEIER, R. A. ET AL. 2002. Wilderness: earth’s last wild places. CEMEX, 318
Agrupacion Serra Madre. 319
MMA. 2014. Lista Nacional Oficial de Espécies da Fauna Ameaçadas de Extinção. 320
Ministério do Meio Ambiente, Brasília. 321
<http://pesquisa.in.gov.br/imprensa/jsp/visualiza/index.jsp?jornal=1&pagina=121&data322
=18/12/2014> 323
MMA. 2016a. Caatinga - Monitoramento do Desmatamento dos Biomas Brasileiros por 324
Satélite - Relatório Técnico 2010 – 2011. Ministério do Meio Ambiente, Brasília. 325
<http://www.mma.gov.br/index.php/comunicacao/agencia-326
informma?view=blog&id=2096> 327
MMA. 2016b. Áreas Prioritárias para a Conservação, Utilização Sustentável e Repartição 328
de Benefícios da Biodiversidade do Cerrado, do Pantanal e da Caatinga. Ministério do 329
Meio Ambiente, Brasília. <http://www.mma.gov.br/biodiversidade/biodiversidade-330
brasileira/%C3%A1reas-priorit%C3%A1rias/item/10724> 331
57
NASCIMENTO, F. O., & A. FEIJÓ. 2017. Taxonomic revision of the Tigrina Leopardus 332
tigrinus (Schreber, 1775) species group (Carnivora, Felidae). Papéis Avulsos de Zoologia 333
57:231-264. 334
OLIVEIRA, J. A. 2004. Diversidade de mamíferos e o estabelecimento de áreas 335
prioritárias para a conservação do bioma Caatinga. Biodiversidade da Caatinga: Áreas e 336
ações prioritárias para a conservação. (J. M. C. Silva, M. Tabarelli, M. T. Fonseca & L. 337
V. Lins, eds.). Ministério do Meio Ambiente, Brasília. 338
OLIVEIRA, T. G., M. A. TORTATO, L. B. ALMEIDA, C. B. CAMPOS, & B. M. 339
BEISIEGEL. 2013. Avaliação do risco de extinção do gato-do-mato Leopardus tigrinus 340
no Brasil. Biodiversidade Brasileira 5:56-65. 341
PAGLIA, A. P. ET AL. 2012. Lista Anotada dos Mamíferos do Brasil / Annotated 342
Checklist of Brazilian Mammals. 2ª Edição / 2nd Edition. Occasional Papers in 343
Conservation Biology 6:1-76. 344
PENIDO, G. ET AL. 2017. Mesocarnivore activity patterns in the semiarid Caatinga: 345
limited by the harsh environment or affected by interspecific interactions?. Journal of 346
Mammalogy 98:1732-1740. 347
PEREIRA, C. J., & R. S. PEIXOTO. 2017. Levantamento de mamíferos terrestres em 348
uma área de Caatinga em Senhor do Bonfm, Bahia. Revista Brasileira de Zoociências 349
183:33-44. 350
PEREIRA, L. G., & L. GEISE. 2009. Non-flying mammals of Chapada Diamantina 351
(Bahia, Brazil). Biota Neotropica 9:185-196. 352
PERES, C. A. 2001. Synergistic effects of subsistence hunting and habitat fragmentation 353
on amazonian forest vertebrates. Conservation Biology 15:1490-1505. 354
PICHORIM, M., T. P. F. CÂMARA, T. M. OLIVEIRA-JÚNIOR, D. V. OLIVEIRA, E. 355
P. G. NASCIMENTO, & J. A. MOBLEY. 2014. A population of Blue-winged Macaw 356
58
Primolius maracana in northeastern Brazil: recommendations for a local conservation 357
action plan. Tropical Conservation Science 7:488-507. 358
RIBEIRO, E. M. S., V. ARROYO-RODRÍGUEZ, B. A. SANTOS, M. TABARELLI, & 359
I. R. LEAL. 2015. Chronic anthropogenic disturbance drives the biological 360
impoverishment of the Brazilian Caatinga vegetation. Journal of Applied Ecology 361
52:611-620. 362
ROCHA, P. A., J. RUIZ-ESPARZA, R. BELTRÃO-MENDES, A. S. RIBEIRO, B. A. T. 363
P. CAMPOS, & S. F. FERRARI. 2015. Nonvolant mammals in habitats of the Caatinga 364
scrub and cloud forest enclave at Serra da Guia, state of Sergipe. Revista Brasileira de 365
Zoociências 16:93-103. 366
ROSAS-RIBEIRO, P. F., R. RANULPHO, & E. M. VENTICINQUE. 2017. New records 367
and update on the geographic distribution of Lontra longicaudis (Olfers, 1818) 368
(Carnivora: Mustelidae) in Seasonally Dry Tropical Forests of northeastern Brazil. Check 369
List 13:2108. 370
ROVERO, F., E. MARTIN, M. ROSA, J. A. AHUMADA, & D. SPITALE. 2014. 371
Estimating species richness and modelling habitat preferences of tropical forest mammals 372
from camera trap data. PLoS One 9:e103300. 373
RUIZ-GARCÍA, M., M. PINEDO-CASTRO, & J. M. SHOSTELL. 2017. Small spotted 374
bodies with multiple specific mitochondrial DNAs: existence of diverse and differentiated 375
tigrina lineages or species (Leopardus spp: Felidae, Mammalia) throughout Latin 376
America. Mitochondrial DNA Part A 29:993-1014. 377
SANTOS, J. C., I. R. LEAL, J. S. ALMEIDA-CORTEZ, G. W. FERNANDES, & M. 378
TABARELLI. 2011. Caatinga: the scientific negligence experiencedby a dry tropical 379
forest. Tropical Conservation. Science 4:276-286. 380
59
SEDDON, A. W. R., M. MACIAS-FAURIA, P. R. LONG, D. BENZ, & K. J. WILLIS. 381
2016. Sensitivity of global terrestrial ecosystems to climate variability. Nature 531:229–382
232. 383
SILVA, J. M. C., I. R. LEAL, & M. TABARELLI (eds.). 2017a. Caatinga: the largest 384
tropical dry forest region in South America. Springer, Cham. 385
SILVA, J. M. C., L. C. F. BARBOSA, I. R. LEAL, & M. TABARELLI. 2017b. The 386
Caatinga: understanding the challenges. The largest tropical dry forest region in South 387
America. (J. M. C. Silva, I. R. Leal & M. Tabarelli, eds.). Springer, Cham. 388
SILVA, J. M. C., M. TABARELLI, M. T. FONSECA, & L. V. LINS (orgs.). 2003. 389
Biodiversidade da Caatinga: áreas e ações prioritárias para a conservação. Ministério do 390
Meio Ambiente, Brasília. 391
SILVA, U. G., & C. N. S. PALMEIRA. 2014. Mamíferos de um brejo de altitude, Traipu, 392
Alagoas. Revista Ourici 4:31-59. 393
SOBRAL, M., K. M. SILVIUS, H. OVERMAN, L. F. B. OLIVEIRA, T. K. RABB, & J. 394
M. V. FRAGOSO. 2017. Mammal diversity influences the carbon cycle through trophic 395
interactions in the Amazon. Nature Ecology & Evolution 1:1670–1676. 396
SOLLMANN, R., A. MOHAMED, H. SAMEJIMA, & A. WILTING. 2013. Risky 397
business or simple solution – Relative abundance indices from camera-trapping. 398
Biological Conservation 159:405-412. 399
TERBORGH, J. ET AL. 2001. Ecological meltdown in predator-free forest fragments. 400
Science 294:1923–1926. 401
TOBLER, M. W. 2007. Camera base version 1.6. Atrium biodiversity information 402
system. <http://www.atrium-biodiversity.org/tools/camerabase/> 403
60
TOBLER, M. W., S. E. CARRILHO-PERCASTEGUI, R. LEITEPITMAN, R. MARES, 404
& G. POWELL. 2008. An evaluation of camera traps for inventorying large- and 405
medium-sized terrestrial rainforest mammals. Animal Conservation 11:169-178. 406
VARGAS-MENA, J. C. ET AL. 2018. The bats of Rio Grande do Norte state, 407
northeastern Brazil. Biota Neotropica 18:e20170417. 408
VELLOSO, A. L., E. V. S. B. SAMPAIO, & F. G. C. PAREYN. 2002. Ecorregiões 409
propostas para o Bioma Caatinga. Associação Plantas do Nordeste, Instituto de 410
Conservação Ambiental, The Nature Conservancy do Brasil, Recife. 411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
61
Capítulo 2 430
431
TEMPORAL NICHE OVERLAP AMONG MESOCARNIVORES IN A 432
CAATINGA DRY FOREST 433
434
435
436
437
438
Capítulo publicado na revista European Journal of Wildlife Research (link: 439
https://link.springer.com/article/10.1007/s10344-020-1371-6). 440
62
Temporal niche overlap among mesocarnivores in a Caatinga dry forest 441
442
Paulo Henrique Marinho1,2 (ORCID: 0000-0001-7205-3089), Carlos Roberto Fonseca1, 443
Pedro Sarmento2, Carlos Fonseca2 and Eduardo Martins Venticinque1 444
445
1 Departamento de Ecologia, Centro de Biociências, Universidade Federal do Rio Grande 446
do Norte, Natal, Rio Grande do Norte, Brazil 447
2 Departamento de Biologia & Centro de Estudos do Ambiente e do Mar (CESAM), 448
Universidade de Aveiro, Campus Universitário Santiago, Aveiro, Portugal 449
450
Corresponding author: Paulo Henrique Marinho ([email protected]) 451
452
453
454
63
Abstract 455
Agonistic encounters among carnivores can be potentially avoided or minimized by shifts 456
in daily activity patterns. Here, we investigated the temporal ecology of mesocarnivores 457
which co-occur in a semiarid area where top predators are virtually absent. More 458
specifically, we (i) describe the daily activity patterns of six mesocarnivore species, (ii) 459
evaluate possible seasonal changes in their daily activity patterns, (iii) examine patterns 460
of temporal overlap among mesocarnivores, and (iv) test the overlap in daily activity 461
between mesocarnivores and their potential prey. Using camera trapping data (13,976 462
camera-days) and circular and overlapping analyzes, we studied six out of the seven 463
mesocarnivore species recorded. Striped hog-nosed skunk and the crab-eating raccoon 464
were nocturnal, crab-eating fox and northern tiger cat were nocturnal-crepuscular, ocelot 465
was mainly nocturnal, and jaguarundi was diurnal. With the exception of jaguarundi, we 466
fail to find strong temporal segregation among mesocarnivore and interspecific 467
interactions did not vary seasonally, but we observed separation in their activity peaks 468
and significant difference in their activity distributions. This partial temporal segregation 469
can potentially contribute to interspecific coexistence, reducing the chances of 470
interspecific killing, mainly in relation to the dominant species (ocelot). Mesocarnivores 471
did not exhibit a significant synchrony of their activity with any of the preys evaluated, 472
with the exception of jaguarundi, which significantly overlapped its distribution of 473
activity with some preys. Temporal segregation contributes but does not seem to be the 474
only mechanism behind the coexistence of mesocarnivores in Caatinga dry forest, thus 475
other strategies such as spatial and dietary segregation should be considered. 476
Keywords: activity pattern; intraguild interaction; predator-prey interaction; seasonal 477
activity; semiarid; temporal segregation. 478
64
Introduction 479
Patterns of daily activity are a crucial component of mammalian ecology and 480
behavior, through which species respond to variations in biotic, abiotic, and 481
anthropogenic factors (Schoener 1974; Bennie et al. 2014; Gaynor et al. 2018). Many 482
species avoid being active in periods of intense heat or cold to avoid hypo- or 483
hyperthermia (Terrien et al. 2011), or during full moon nights, which can make predators 484
and prey more exposed to mutual detection (Prugh and Golden 2014). Predators can 485
adjust their daily activity in order to maximize the chances of prey encounter thus 486
minimizing the energy expenditure, while pressing the prey to avoid encounters with 487
potential predators (Foster et al. 2013; Monterroso et al. 2013). In addition, the negative 488
effects of interspecific competition can also be minimized through temporal segregation, 489
which is an important mechanism of coexistence among species with similar ecology and 490
morphology (Lucherini et al. 2009; Di Bitetti et al. 2010; Monterroso et al. 2014). This 491
strategy contributes to the avoidance of agonistic encounters which minimize interference 492
competition levels and reduce the chances of intraguild predation and interspecific killing 493
among mammalian carnivores (Carothers and Jaksic 1984; Polis et al. 1989; Palomares 494
and Caro 1999). 495
Mesocarnivores (<15 kg) usually occupy a trophic position just below the top 496
predators (Roemer et al 2009; Ritchie and Johnson 2009), being submitted to top-down 497
control by apex predators while competing for habitat and food resources (Ritchie and 498
Johnson 2009). So, the reduction in numbers or even the absence of top-predators can 499
contribute to increase the abundance of mesocarnivores, which can expand their 500
ecological niche intensifying predation and intraguild competition effects (Crooks and 501
Soulé 1999; Prugh et al. 2009). Under these circumstances, larger and generally dominant 502
mesocarnivores (Oliveira and Pereira 2014) may emerge as apex predators affecting 503
65
subordinate species (Prugh et al. 2009). However, there are only a few studies on the 504
ecology of sympatric neotropical mesocarnivores in areas where top predators were 505
locally extinct, which usually correspond to unprotected areas (Woodroffe and Ginsberg 506
1998). This type of knowledge is essential to the conservation of biological communities 507
(Bu et al. 2016; Wang et al. 2015) in a world with increasingly anthropogenic 508
disturbances. 509
The Caatinga, a seasonally dry tropical forest in northeastern Brazil, is a naturally 510
stressful environment which has been suffering high anthropogenic impacts (Silva et al. 511
2017; Antongiovanni et al. 2018) that can be expected to modulate species ecology 512
(Stoner and Timm 2011; Astete et al. 2017a, 2017b). In most of the Caatinga, top 513
predators such as the jaguar Panthera onca Linnaeus, 1758 and the puma Puma concolor 514
Linnaeus, 1771 are declining or absent (Azevedo et al. 2013; Feijó and Langguth 2013; 515
Morato et al. 2013; Marinho et al. 2018a). In such sites, mesocarnivores coexist while 516
being released from top down regulation. In Caatinga, the mesocarnivore assemblage is 517
most commonly formed by seven more widely distributed species (Feijó and Langguth 518
2013; Marinho et al. 2018a): three omnivorous species, represented by crab-eating 519
raccoon Procyon cancrivorus (G.[Baron] Cuvier, 1798) (adult body mass = 5.4-8.8 kg), 520
crab-eating fox Cerdocyon thous Linnaeus, 1766 (5.7-6.5 kg), and striped hog-nosed 521
skunk Conepatus amazonicus Lichtenstein, 1838 (2.4 kg), and four species that are 522
strictly carnivorous, represented by the felids ocelot Leopardus pardalis Linnaeus, 1758 523
(8-11 kg), jaguarundi Herpailurus yagouaroundi É, Geoffroy Saint-Hilare, 1803 (3-6 kg), 524
and northern tiger cat Leopardus tigrinus Thomas, 1904 (1.5-3 kg), along with the 525
mustelid lesser grison Galictis cuja Molina, 1782 (1-3 kg) (Paglia et al. 2012; Oliveira 526
and Pereira 2014). This overlap in terms of ecological niche, associated with the virtual 527
absence of top-predators, can result in the increase of competition effects, which, in turn, 528
66
can be lowered by deviations in the use of trophic, spatial and temporal resources 529
(Schoener 1974). For the symtopic mesocarnivores, especially with analogous 530
morphology and hunting strategies, alterations in activity patterns can improve 531
coexistence. In Brazilian semiarid regions, studies regarding mesocarnivore activity 532
patterns and temporal interactions are focused on a single species or in a sub-set of the 533
mesocarnivore guild, while they usually don't consider the seasonal effect (Dias and 534
Bocchiglieri 2016; Dias 2017; Penido et al. 2017; Dias et al. 2018, 2019; Marinho et al. 535
2018b). 536
In this study we use camera trapping data to examine the temporal ecology of a 537
syntopic mesocarnivore guild in a Caatinga dry forest area virtually free of top predators. 538
Puma seems absent from the area for over 10 years (Marinho et al. 2018a), while no 539
mention is provided on current or previous presence of the jaguar, the other Caatinga top 540
predator. Our objectives were: (i) to describe the daily activity patterns of the 541
mesocarnivore species, (ii) to evaluate possible seasonal changes in their daily activity 542
patterns, (iii) to examine patterns of segregation or temporal overlap among 543
mesocarnivores, and (iv) to test the overlap in daily activity between mesocarnivores and 544
their potential prey. According to previous information, we expected mesocarnivores to 545
be nocturnal or mainly nocturnal (Dias 2017; Penido et al. 2017; Marinho et al. 2018b; 546
Dias et al. 2018, 2019), except for jaguarundi, which has a well-documented diurnal habit 547
throughout its distribution (Giordano 2016). 548
We expected that mesocarnivores with closely related ecology and morphology, 549
particularly the felids, would exhibit greater temporal segregation, at least during their 550
peaks of activity (Penido et al. 2017; Dias et al. 2019), to avoid aggression risk and 551
competition for resources (exploitative competition), while omnivorous species would 552
differentiate their activity mainly in relation to the ocelot, the largest mesocarnivore 553
67
known in the area. An alternative hypothesis is that the activity patterns of the 554
mesocarnivores are more synchronized with their preferential prey, especially to the 555
hypercarnivorous felids (Linkie and Ridout 2011; Foster et al. 2013; Porfirio et al. 2016). 556
Considering that the Caatinga dry forest is a highly seasonal environment, we 557
hypothesized that temporal niche segregation would be larger during the dry season, when 558
water and food resources are more scarce and concentrated and the visual contact among 559
the mesocarnivores is potentially more common due to the lower understory leaf density, 560
which may increase interference competition (Valeix et al. 2007; Vanak et al. 2013). In 561
general, we expected a possible lower daytime activity in the dry season related to the 562
avoidance of high daytime temperatures (Pita et al. 2011), considering the low humidity 563
and the poor protection of vegetation cover in this period. 564
565
Methods 566
Study site 567
This study was developed in a seasonally dry tropical forest (Caatinga) area between 568
the Feiticeiro and Bonfim mountains in Lajes municipality, Rio dos Ventos, Rio Grande 569
do Norte state, northeastern Brazil (5º44’-5º51’S latitude, 36º11’-36º06’W longitude, Fig. 570
1). The region is considered a priority for the conservation of the Caatinga domain (MMA 571
2016; Fonseca et al. 2017). The climate is semiarid, with high temperatures and low 572
rainfall. The higher precipitations usually occur between March and May, with average 573
annual rainfall varying from 400 to 650 mm. The vegetation varies from open shrub 574
formations, generally more disturbed and occurring at lower altitudes, to more forested 575
and dense patches, mainly covering higher and steeper areas (Velloso et al. 2002). 576
577
68
578
Fig. 1 Study area and distribution of the 50 camera trap stations (A) used to investigate the 579
temporal niche and coexistence of a mesocarnivore guild in a Caatinga dry forest in Ljes 580
municipality, Rio Grande do Norte state (B), northeastern Brazil (C). 581
582
Despite being the largest continuous remnant of Caatinga vegetation in the Rio 583
Grande do Norte state (Projeto Caatinga Potiguar 2015), the area has no legal protection 584
and is under several anthropogenic disturbances developed in private properties, such as 585
extensive livestock, cutting wood for charcoal production and conversion of natural 586
habitats to temporary agriculture. Poaching is another intense activity in the area 587
(Marinho et al. 2018a). The last reliable reports of presence of puma in the region are 588
more than 10 years old (Marinho et al. 2018a), being probably locally extinct (at least 589
functionally) due to hunting and conflicts with cattle ranchers. Finally, the region is the 590
target of mineral exploration, and is being prospected for installation of wind power 591
plants. 592
Sampling design 593
69
Mesocarnivore occurrences were recorded through camera trapping in an area of 594
approximately 70 km2, between May 2016 and February 2019, totaling a sampling effort 595
of 13,976 camera-days divided into three sampling periods and along a total of 50 596
sampling points (Table S1). In the first survey (24 stations from May to June 2016), we 597
used stations with double cameras aiming to estimate the density of the wild cats, while 598
in the remaining two surveys we used stations with a single camera (43 stations from 599
January to October 2017 and 37 stations from October 2018 to February 2019, Table S1). 600
The sampling effort was divided in wet (7,337 camera-days, January - June) and dry 601
season (6,639 camera-days, July - December) (adapted from Tomasella et al. 2018). 602
We used camera traps with heat and motion sensor (model Bushnell® Trophy Cam 603
™ HD). For the installation of the cameras we prioritized trails used by people and/or 604
livestock, dirt roads, and temporary streams. The distance between cameras was in 605
average 929.5 m (SD = 262.1). We set the camera traps to record the date and time of the 606
detections, with a minimum interval of 5 minutes between consecutive records, and we 607
programmed them to take three photos per shot or one photo and one video of 10 seconds, 608
remaining active 24 hours every day. We considered the records of the same species at 609
the same sampling point with more than 1 hour apart as independent (Goulart et al. 2009). 610
We did not use any bait to attract the animals. 611
Potential prey 612
The diet of mesocarnivores in the Caatinga is poorly known, but few existing studies 613
can give an idea of their potential prey (Ximenez 1982; Olmos 1993; Dias and 614
Bocchiglieri 2015, 2016). The striped hog-nosed skunk feeds mainly on arthropods and 615
lizards preyed on burrows, and occasionally on fruits and small mammals such as 616
marsupials (Olmos 1993). The crab-eating fox diet in the region is based mainly on 617
arthropods, but also fruits, lizards, birds and some small rodents such as Spix's cavy Galea 618
70
spixii (Wagler, 1831) (Olmos 1993; Dias and Bocchiglieri 2016). The crab-eating raccoon 619
can consume mainly arthropods, followed by fruits, and small vertebrates such as lizards, 620
birds and small rodents (Dias and Bocchiglieri 2015). In the Serra da Capivara National 621
Park, northern tiger cat fed on large number of lizards such as Ameiva ameiva (Linnaeus, 622
1758) and Tropidurus hispidus (Spix, 1825), as well as on arthropods, birds and a smaller 623
proportion of unidentified small rodents (Olmos 1993). Ximenez (1982) also reported the 624
presence of lizard scales in the stomach contents of the northern tiger cat. With respect to 625
jaguarundi diet in Caatinga, there are reports of predation of common marmoset Callithrix 626
jacchus (Linnaeus, 1758) (Ximenez 1982), rock cavy Kerodon rupestris F. Cuvier, 1825, 627
Spix's cavy, punaré rat Thrichomys laurentius Thomas, 1904 and other small rodents, as 628
well as, birds, lizards and arthropods (Olmos 1993; Dias and Bocchiglieri 2015). Finally, 629
birds and lizards were more present in the diet of ocelot in Caatinga, although punaré rat 630
and white-eared opossum Didelphis albiventris (Lund, 1840) also had been predated 631
(Dias and Bocchiglieri 2015). However, due to its larger body size, the ocelot can also 632
feed on larger prey such as armadillos (Wang 2002; Moreno et al. 2006). 633
Data analysis 634
We divided the independent records of each species in one-hour intervals over the 24 635
h circadian cycle. The uniformity of species records throughout the circadian cycle was 636
tested using the Rayleigh test in the Oriana v.4 program (Kovach Commuting Services, 637
Wales, UK). In order to describe the activity patterns, we classified the records in diurnal 638
(between 1 h after sunrise and 1 h before sunset), nocturnal (between 1 h after sunset and 639
1 h before sunrise), and crepuscular (± 1 h of sunrise or sunset) (Porfirio et al. 2016). To 640
define the exact time of sunrise and sunset we used the software Tropsolar 5.0 (Cabus 641
2004). Since the samplings were carried out throughout the year, we estimated the 642
monthly variations in the sunrise and sunset times from the 15th day of each month to 643
71
classify the species records. For sunrise, the average value was 5:20 h, with an annual 644
variation from 04:57 h to 05:35 h, while for sunset the average value was 17:18 h, with a 645
variation of 17:09h to 17:40 h throughout the year. For these analyzes we used legal time 646
as a reference, since annual variations in the daytime period are lower at low latitudes 647
(Nouvellet et al. 2012). 648
We tested the existence of temporal segregation between the activity patterns of 649
species pairs, within the mesocarnivore guild and between mesocarnivores and their 650
potential prey species, with the Mardia-Watson Wheeler test (MWW test) in the software 651
Oriana v.4. For each species, the same method was performed to test for shifts in daily 652
activity patterns between the dry and wet season. In addition, we used the non-parametric 653
Kernel density function to estimate the activity overlap coefficient (Δ) between 654
mesocarnivores, as well as with their prey and between seasons (Ridout and Linkie 2009; 655
Linkie and Ridout 2011). This coefficient, defined as the area under the curves formed 656
by the two density functions in each time unit, ranges from 0 (no overlap) to 1 (total 657
temporal overlap of activity) (Ridout and Linkie 2009). We obtained the 95% confidence 658
intervals of the overlap coefficients estimated through 1000 bootstrap samples (Linkie 659
and Ridout 2011; Meredith and Ridout 2018). According to Ridout and Linkie (2009) we 660
used the coefficient Δ1 for small samples (<75 records for at least one of the pairs 661
compared) and the coefficient Δ4 for large samples (> 75 records). These analyzes were 662
done with the R package overlap (Meredith and Ridout 2018; R Development Core Team 663
2012). Finally, we classified the activity overlap between each comparison as follows: 664
low overlap (Δ ≤ 0.5), moderate overlap (0.5 < Δ ≤ 0.75), and high overlap (Δ > 0.75) 665
(Monterroso et al. 2014). We performed the analysis of activity overlap only with pairs 666
of data (intra or between species) with more than 10 records in each pair evaluated. 667
668
72
Results 669
We recorded all seven species of mesocarnivores expected for the study area (Table 670
1). The number of records ranged from 1133 for the crab-eating fox to just five records 671
for the lesser grison. Due to the low number of records of the lesser grison, this species 672
was excluded from all posterior analyses. 673
The species activity patterns were not homogenous throughout the circadian cycle 674
according to Rayleigh test (Fig. 2; Table 1). The striped hog-nosed skunk was nocturnal, 675
starting its activities during dusk but avoided the dawn (Fig. 2a). The crab-eating fox was 676
nocturnal-crepuscular, this canid started its activity at dusk, remained active throughout 677
the night and reached the peak of activity at dawn, with some residual activity during the 678
first part of the day (Fig. 2b). The crab-eating raccoon was nocturnal, but had a peak 679
activity during dusk and another just before dawn (Fig. 2c). The ocelot was mostly 680
nocturnal, but it initiated strongly its activities during dusk and slowed down continuously 681
until the first hours of the morning (Fig. 2d). The northern tiger cat was nocturnal-682
crepuscular, but it was relatively generalist, performing a fair amount of its activities 683
during day hours (Fig. 2e). Finally, the jaguarundi was diurnal, but it also presented a 684
high activity during the dawn and dusk periods (Fig. 2f). 685
686
73
687
Fig. 2 Daily activity pattern of mesocarnivores in a Caatinga dry forest, northeastern Brazil. Each 688
circular histogram is divided into 24 intervals of 60 minutes, and their bars represent the 689
percentage of the total number of camera-trap detections in each interval. Day (on average, 06:20 690
h - 16:18 h), night (after 18:18 h - 04:20 h), and twilight (± 1 before and after 5:20 h and 17:18 h) 691
correspond to white, black and gray, respectively. Species are: (a) striped hog-nosed skunk, (b) 692
crab-eating fox, (c) crab-eating raccoon, (d) ocelot, (e) northern tiger cat, and (f) jaguarundi. 693
Origin of the specie’s images: De Angelo et al. (2015). 694
695
Daily activity distribution patterns were very similar between dry and wet periods for 696
all mesocarnivore species; differences in activity distributions between seasons being 697
non-significant (Fig. 3; Table 1). The crab-eating fox (Δ = 0.94 [0.90-0.97]) showed the 698
highest activity overlap between seasons, while jaguarundi exhibited the lowest (Δ = 0.78 699
[0.63-0.90]), with some decrease in its sunset peak of activity during dawn in dry season 700
(Fig. 3; Table 1). 701
702
74
703
Fig. 3 Density estimates of daily activity patterns and extension of overlap within mesocarnivore 704
species between dry and wet season in a Caatinga dry forest, northeastern Brazil. Overlap is 705
represented by the shaded grey area. The dashed vertical lines represent the average legal time of 706
sunrise [5:20 h] and sunset [17:18 h]) during the study period. The time of the records is shown 707
as ticks in the bottom of the figures.708
75
Intraguild interactions
Overlap of daily activity patterns among mesocarnivore species varied widely, from 0.14 to 0.92
(Fig. 4). The activity distribution test indicated that there were significant differences in the activity
patterns for 10 of the 15 mesocarnivore contrasts (Table 2). The lower overlap coefficients (0.14-
0.48) appeared between the diurnal jaguarundi and all other species, showing a strong and significant
temporal niche differentiation (Fig. 4; Table 2). The higher overlap coefficients occurred between
crab-eating raccoon and all other mesocarnivores (0.76-0.83), except jaguarundi, and between crab-
eating fox and northern tiger cat (0.92) (Fig. 4; Table 2), which indicate temporal niche similarity.
The remaining species contrasts had moderate to high overlap coefficient (0.70-0.79), but always
showing significant temporal niche differences through their activity distributions (Fig. 4; Table 2).
It should be highlighted that we found no evidence of change in intraguild interactions between dry
and wet seasons (Table S2).
Predator-prey temporal overlap
Camera trapping detected several mammal (N = 1184 records), bird (N = 1031), and lizard (N =
179) species that could be potential preys of mesocarnivores (Table S3). Among the mammals, we
recorded: Spix's cavy, punaré rat, rock cavy, white-eared opossum, yellow armadillo Euphractus
sexcinctus (Linnaeus, 1758), and nine-banded armadillo Dasypus novemcinctus (Linnaeus, 1758). In
relation to the ground-dwelling and ground-foraging birds, we recorded several species of doves,
including Columbina spp., Zenaida auriculata (Des Murs, 1847), and Leptotila verreauxi (Bonaparte,
1855), as well as tinamou species of the genera Crypturellus and Nothura. Among the lizards we
recorded A. ameiva, T. hispidus, Tropidurus semitaeniatus (Spix, 1825), and Ameivula ocellifera
(Spix, 1825) (Table S3). Due to identification uncertainties, in the following analyses, doves,
tinamous, and lizards were analyzed as groups (Table S3).
Most potential mammalian preys exhibited nocturnal or mainly nocturnal activity, with the
exception of rock cavy and yellow armadillo which were cathemeral and mostly diurnal, respectively
(Table S3; Fig S1). For birds, doves were diurnal while tinamous were crepuscular (Table S3; Fig.
76
S1). Finally, lizards were diurnal (Table S3; Fig. S1). Most preys had similar daily activity patterns
throughout the year, with the exception of yellow armadillo and the doves which exhibited significant
difference between the dry and wet season (Table S3).
Among the omnivorous mesocarnivores (Table 3; Fig. S1), the daily distributions of records of
the striped hog-nosed skunk differed significantly from the pattern exhibited by most potential preys,
except from punaré rat and white-eared opossum (Table 3; Fig. S1). The distribution of activity of
the crab-eating fox was significantly different from all potential preys, although it had a reasonable
temporal overlap with some small mammals (Table 3; Fig. S1). Finally, the activity pattern of crab-
eating raccoon was similar to Spix's cavy and punaré rat (Table 3; Fig. S1), and to the white-eared
opossum during dry season (Table S4), but differed from the other potential preys.
Regarding the felids (Table 3; Fig. S1), the ocelot and the northern tiger cat differed significantly
their activity distributions from all potential preys, although they had a reasonably high activity
overlap with some small mammal species such as Spix’s cavy and white-eared opossum (Table 3;
Fig. S1). In fact, in the wet season, the distribution of activity of the ocelot did not differ from the
white-eared opossum (Table S4). Finally, the activity pattern of jaguarundi was similar to that
exhibited by rock cavy and tinamous in both seasons (Table S4), but differed from other potential
prey (Table 3; Fig. S1). Except to the ocelot, predator-prey temporal interactions were extremely
conserved across seasons (Table S4).
77
Table 1 Daily activity patterns of mesocarnivores in a Caatinga dry forest, northeastern Brazil. Rayleigh Z tested the homogeneity of the species daily activity using
the total number of records. Species were classified into activity categories based on the percentage of records falling in different day periods (day, night and twilight).
N is the number of records total and in each season. ∆ measures the species overlap in their temporal niche between the dry and wet seasons (with the respective 95%
confidence intervals). Mardia-Watson-Wheller (MWW) tested the temporal niche segregation between dry and wet seasons. Significant results are presented in bold.
The few records of lesser grison prevented any analysis.
Mesocarnivores N
(total) Rayleigh Z
(p) Activity
(% day/night/twilight) N
(dry/wet) ∆dry-wet
(95% CI) MWWdry-wet
(p)
Striped hog-nosed skunk 288 147.4
(<0.001)
Nocturnal
(0.3 / 94.1 / 5.6) 154 / 134
0.91
(0.83-0.97)
1.27
(0.53)
Crab-eating fox 1133 214.6
(<0.001)
Nocturnal-crepuscular
(10.3 / 61.9 / 27.8) 640 / 493
0.94
(0.90-0.97)
2.02
(0.36)
Crab-eating raccoon 36 12.0
(<0.001)
Nocturnal
(2.8 / 88.9 / 8.3) 10 / 26
0.79
(0.57-0.96)
1.57
(0.46)
Ocelot 143 44.1
(<0.001)
Mostly nocturnal
(7.0 / 79.7 / 13.3) 79 / 64
0.83
(0.73-0.92)
0.02
(0.99)
Northern tiger cat 524 78.5
(<0.001) Nocturnal-crepuscular
(16.2 / 60.1 / 23.7) 252 / 272
0.92 (0.86-0.97)
0.78 (0.68)
Jaguarundi 84 11.2
(<0.001)
Diurnal
(52.4 / 3.6 / 44.0) 33 / 51
0.78
(0.63-0.90)
0.92
(0.63)
Lesser grison 5 - Unclassified
(100 / 0 / 0) 1 / 4 - -
78
Fig. 4 Density estimates of daily activity patterns and extension of their overlap among pairs of mesocarnivores in a Caatinga dry forest, northeastern Brazil. Overlap
is represented by the shaded grey area. The dashed vertical lines represent the average legal time of sunrise [5:20 h] and sunset [17:18 h]) during the study period.
79
Significant differences in MWW test are indicated by an asterisk after overlap coefficient values (∆). The time of the records is shown as ticks in the bottom of the
figures. Origin of the specie’s images: De Angelo et al. (2008).
Table 2 Temporal niche overlap values of mesocarnivores in a Caatinga dry forest, northeastern Brazil. Overlapping coefficient values (∆) with their respective
95% confidence intervals (between parentheses) are represented above the diagonal while the Mardia-Watson-Wheller test values (W), with its respective statistical
significance (p, between parentheses) are represented below the diagonal. Statistically significant values are shown in bold.
Mesocarnivores Striped hog-nosed
skunk
Crab-eating
fox
Crab-eating
raccoon Ocelot
Northern tiger
cat Jaguarundi
Striped hog-nosed skunk - 0.70 0.83 0.79 0.71 0.14
(0.66-0.74) (0.70-0.93) (0.71-0.85) (0.66-0.75) (0.10-0.23)
Crab-eating fox 100.1
- 0.77 0.78 0.92 0.46
(<0.001) (0.65-0.87) (0.72-0.84) (0.88-0.94) (0.38-0.53)
Crab-eating raccoon 3.7 4.3
- 0.79 0.76 0.29
(0.16) (0.11) (0.64-0.92) (0.64-0.87) (0.18-0.40)
Ocelot 12.5 25.1 4.2
- 0.77 0.31
(<0.01) (<0.001) (0.12) (0.70-0.84) (0.23-0.40)
Northern tiger cat 83.1 0.08 5.2 25.8
- 0.48
(< 0.001) (0.96) (0.07) (<0.001) (0.40-0.54)
Jaguarundi 170.4 115.8 62.9 117.5 104.6
- (<0.001) (<0.001) (<0.001) (<0.001) (<0.001)
80
Table 3 Temporal niche overlap coefficient (∆), with its respective 95% confidence intervals (between parentheses), among mesocarnivores and potential prey in a
Caatinga dry forest, northeastern Brazil. The ∆ values with asterisk indicate significant difference (p <0.05) in pairwise activity distribution according to Mardia-
Watson-Wheller test. We considered yellow-armadillo, which is a medium-sized mammal (5.4 kg, Paglia et al. 2012), a potential prey only for larger mesocarnivore
in the area, the ocelot.
Potential prey Striped hog-
nosed skunk
Crab-eating
fox
Crab-eating
raccoon Ocelot
Northern tiger
cat Jaguarundi
Mammals
Spix's cavy 0.76*
(0.70-0.82) 0.81*
(0.77-0.85) 0.83
(0.72-0.93) 0.73*
(0.66-0.80) 0.80*
(0.76-0.85) 0.32*
(0.24-0.40)
Punaré rat 0.93
(0.86-0.98)
0.69*
(0.63-0.74)
0.82
(0.69-0.92)
0.73*
(0.64-0.81)
0.70*
(0.64-0.76)
0.16*
(0.09-0.23)
Rock cavy 0.34*
(0.24-0.44)
0.53*
(0.42-0.62)
0.45*
(0.33-0.57)
0.50*
(0.38-0.61)
0.58*
(0.47-0.68)
0.71
(0.61-0.81)
White-eared opossum 0.92
(0.85-0.97)
0.68*
(0.62-0.73)
0.78*
(0.65-0.90)
0.81*
(0.73-0.88)
0.67*
(0.61-0.72)
0.14*
(0.07-0.21)
Yellow armadillo - - - 0.37*
(0.31-0.44) - -
Birds
Dove 0.04*
(0.02-0.06)
0.26*
(0.24-0.29)
0.10*
(0.03-0.19)
0.14*
(0.09-0.21)
0.27*
(0.24-0.31)
0.70*
(0.62-0.78)
Tinamous 0.19*
(0.07-0.31) 0.45*
(0.32-0.57) 0.26*
(0.10-0.43) 0.27*
(0.14-0.41) 0.46*
(0.32-0.59) 0.75
(0.56-0.91)
Reptiles
Lizard <0.01*
(<0.01-0.02) 0.08*
(0.05-0.11) 0.03*
(<0.01-0.10) 0.08*
(0.04-0.13) 0.14*
(0.10-0.17) 0.41*
(0.31-0.51)
81
Discussion
This study provides a broad description of the temporal ecology of a complete guild of
mesocarnivores, including intraguild and predator-prey temporal interactions, in an area of the
Brazilian Caatinga where top predators are locally extinct or functionally absent. The observed
activity patterns are qualitatively similar to those described in the literature: striped hog-nosed skunk
is classified as nocturnal (Cavalcanti et al. 2014; Dias 2017), crab-eating fox as nocturnal-crepuscular
(Bianchi et al. 2016; Dias and Bocchiglieri 2016; Penido et al. 2017), crab-eating raccoon as nocturnal
(Gómez et al. 2005; Bianchi et al. 2016), ocelot as mainly nocturnal (Di Bitetti et al. 2010; Oliveira-
Santos et al. 2012; Massara et al. 2016; Penido et al. 2017; Dias et al. 2018, 2019; Nagy-Reis et al.
2019), northern tiger cat as nocturnal-crepuscular (Penido et al. 2017; Marinho et al. 2018b; Dias et
al. 2019), and jaguarundi as diurnal (Giordano 2016; Massara et al. 2016; Dias et al. 2019). In the
case of the lesser grison, although the low number of records prevented any analysis, all five records
were obtained between 06:15 and 07:15 h, suggesting a daytime activity as already reported in the
literature (Kasper et al. 2013).
Potential competitor species tend to develop mechanisms to alleviate competition, especially in
the case of morphologically similar and closely related species (Schoener 1974). In the case of
carnivorous mammals, besides the exploitation competition, the risk of aggression (interference
competition) and intraguild predation can induce submissive species to be active at hours with a lower
probability of finding a dominant competitor (Polis et al. 1989). Thus, temporal avoidance is often
the most important mechanism of coexistence (Bianchi et al. 2016; Carothers and Jaksic 1984). Our
results provide partial support for the hypothesis that temporal segregation represents a mechanism
that facilitates the coexistence of mesocarnivores in a semiarid region. Although most pairs of species
exhibited a high or moderate activity overlap, almost all mesocarnivores segregated at least their
activity peaks throughout the circadian cycle, suggesting a partial avoidance that may decrease
competition as well as the risk of intraguild predation (Carothers and Jaksic 1984), especially in
relation to the larger species in the area (i.e. the ocelot). For example, the ocelot has a relatively
82
continuous higher activity between 18:00 p.m. and 2:00 a.m., followed by a decrease until
approximately 5:00 a.m., while the other species exhibited a higher proportion of activity in the
second part of the night, with two of them (crab-eating fox and northern tiger cat) displaying a larger
peak near sunrise hours similar to diurnal jaguarundi. The studies have pointed out a high activity
overlap between ocelot and omnivorous mesocarnivore species (Bianchi et al. 2016; Massara et al.
2016), and a low to moderate temporal overlap between ocelot and small sympatric felid species (Di
Bitetti et al. 2010; Massara et al. 2016; Nagy-Reis et al. 2019).
Regarding generalist or omnivorous species, we found some evidence of temporal segregation
between the striped hog-nosed skunk and the crab-eating fox, but no segregation between the crab-
eating raccoon and the other two species. In the case of the crab-eating raccoon, its larger body size
in relation to the other mesocarnivores prevents it from being attacked or predated (Oliveira and
Pereira 2014). In addition, its preference for environments near water bodies (Cheida et al. 2013), as
well as the consumption of aquatic and semi-aquatic prey, may decrease competition with other
mesocarnivores through spatial and dietary segregation, respectively, at least when there are enough
water bodies. However, considering the scarcity of water bodies in the Caatinga, especially in the dry
season, it is possible that during most of the year there is a considerable trophic niche overlap among
all omnivorous mesocarnivores. In an environment with greater resource availability such as the
Pantanal, temporal segregation was more important for generalist mesocarnivores, including crab-
eating fox and crab-eating raccoon (Bianchi et al. 2016). Further studies are necessary for a better
understanding of the ecology of the crab-eating raccoon in a semiarid and seasonal environment like
the Brazilian Caatinga.
The three felid species, which due to their ecological similarities can be strong competitors,
partially segregated their temporal activity. While the diurnal jaguarundi exhibited a daily activity
pattern very different from the other two felids, the ocelot and the northern tiger cat separated their
peaks of higher activity despite overlapping much of their daily activity. This pattern is similar to that
found in other areas of the Caatinga where larger predators occur (Penido et al. 2017; Dias et al.
83
2019), but it is somewhat different from the pattern reported for a very closely related species, the
southern tiger cat Leopardus gutullus Hensel, 1872 (Massara et al. 2016; Nagy-reis et al. 2019). In
the mesic Atlantic Forest of southern and southeastern Brazil, the southern tiger cat exhibited a
nocturnal behavior in the absence of the ocelot and the puma (Oliveira-Santos et al. 2012) compared
to a cathemeral or mostly diurnal activity where larger felids are present with the overlap between
them ranging from moderate to low (Oliveira-Santos et al. 2012; Massara et al. 2016; Nagy-Reis et
al. 2019). Unlike what happens in these more humid environments, where temporal segregation seems
to be stronger (Massara et al. 2016; Nagy-Reis et al. 2019), in the Brazilian semiarid, differentiation
in spatial dimensions may be important to promote mesocarnivore coexistence (Dias et al. 2019). In
this sense, future studies should investigate whether spatial segregation is acting in main or
complementary way toward coexistence of this carnivore guild.
We found a high overlap in temporal activity patterns between northern tiger cat and crab-eating
fox, which contrasts with the temporal segregation found in an area of Caatinga where large felids
coexist with such mesocarnivores (Penido et al. 2017). Their coexistence can be facilitated by the fact
that these two species partially differ in their diet, with crab-eating fox being omnivorous and northern
tiger cat strictly carnivorous. The crab-eating fox, for instance, can be found near human habitations
looking for food resources, while the northern tiger cat normally avoids such areas (Marinho et al.
2018b). These observations also suggest that an alternative mechanism of coexistence can be in the
space use. Spatial segregation could alleviate possible agonistic interactions, despite the chances for
intraguild predation are considered to be low (Oliveira and Pereira 2014). It should be notice that the
temporal overlap between these two mesocarnivores may be forced by the avoidance of the peak
activity of ocelot and hottest part of the day.
In addition to interspecific interactions, the environmental conditions may also be critical in
shaping the species activity. In semiarid environments such as the Caatinga, where there are high
solar radiation and high diurnal temperatures, both predators and prey are expected to be active during
nocturnal and twilight periods to avoid overheating and water loss (Terrien et al. 2011; Penido et al.
84
2017). Thus, this should limit the temporal window within the circadian cycle available for species
to adjust their activity in response to competition or predation risk. The considerable overlap in
activity of even ecologically close species such as the ocelot and the northern tiger cat suggests this
limitation. Thus, the differentiation in activity peaks may be a more efficient mechanism of
coexistence.
In contrast to what one could expect, we did not detect important variations in the daily activity
patterns of mesocarnivores between dry and wet periods. Since the Brazilian Caatinga experiences a
marked seasonal variation in precipitation and resource availability along the year (Andrade et al.
2017), our results indicate that mesocarnivores are well adapted to such variable conditions. Possible
behavioral mechanisms explain such resilience include diet and spatial shifts. For instance, during
months of greater heat and water stress, the animals may be moving to more mesic locations such as
mountains or to sites with water sources (Carmignotto and Astúa 2017).
Predators tend to manage the pressure for intraguild segregation, but they also need to
synchronize their activity with their main prey to reduce energy expenditure on food hunting (Foster
et al. 2013; Monterroso et al. 2013). According to our results, for most mesocarnivores, the activity
pattern seems to reflect a balance between partial temporal segregation with competitors/predators
and partial temporal overlap with as many potential prey as possible, since few species such as
jaguarundi synchronized their activity with specific potential prey. All of the mesocarnivore species
have a high or moderate activity overlap with at least one type of prey, although the peaks of activity
were generally different. This may allow a sequential exploration of the resources while decreasing
the chances of agnostic encounters (Monterroso et al. 2013). Indeed, the other mesocarnivores that
exhibited high overlap and synchrony in their activity with potential prey are the omnivorous striped
hog-nosed skunk and crab-eating raccoon, which should only occasionally feed on small mammals
(Olmos 1993; Dias and Bochiglieri 2015).
In the case of the jaguarundi, the significant overlap in the distribution of activity with the rocky
cavy and tinamous suggests that these must be important preys in the region, while the other wild
85
felid species should exploit these preys in a complementary way. In other Caatinga areas, the rocky
cavy has a more nocturnal activity which is more overlapped by the activity of the northern tiger cat
(Penido et al. 2017; Dias et al. 2019) and the ocelot (Dias et al. 2018) than by the jaguarundi. Dias et
al. (2018) also found a high and significant overlap of the ocelot activity with the nine-banded
armadillo, a species apparently rare in our area, while the yellow armadillo is abundant but has a
markedly distinct activity compared to the ocelot.
However, it is important to highlight that the temporal overlap alone does not define the
vulnerability of the prey or the preference of the predator, since there must also be spatial overlap,
not investigated here. In addition, some prey can be captured while resting or taking refuge in their
dens (Emsens et al. 2013), such as suggested to crab-eating fox and hog-nosed skunk in Caatinga,
which seem to prey on lizards removing them from their dens (Olmos 1993). Therefore, studies on
spatial interactions and diet of mesocarnivores in this semiarid region are important to elucidate
predator-prey relationships and the level of feeding overlap between mesocarnivores.
Finally, our results suggest that the mesocarnivore species have generally a higher proportion of
nighttime activity, and that there is a separation in their activity peaks rather than a stronger temporal
segregation which remains throughout the seasons, possibly as a trade-off between avoiding the
aggressive encounters with competitors/predators and hotter periods of the circadian cycle. Although
mesocarnivores presented a high overlap with at least one prey, species such as the crab-eating fox,
the northern tiger cat, and the ocelot did not exhibit a strong synchrony of their activity with any of
the preys evaluated, suggesting a more generalist behavior that could contribute to mediate intraguild
interactions. Our results contribute to the understanding of the ecology of mesocarnivores, intraguild
interactions and predator-prey relationships in semiarid environments in a scenario of potential
disturbance caused by the eradication of top predators. It is important to keep in mind that high
overlap of daily activity between species does not necessarily determine a high potential of encounter
if these species segregate spatially, so further studies should seek to understand the role of the spatial
86
and trophic dimensions of the ecological niche in intraguild interactions, and ideally to compare the
ecology of species in areas with different degrees of carnivore guild integrity.
Acknowledgments
We are grateful to João B. de Lima (Seu João) and Joana Darc for the field assistance and
hospitality, and to Eugenia C. Schmidt, Juan C. V. Mena, Felipe Marinho, Raul dos Santos, Maria L.
Falcão, V. Paixão and T. Oliveira for field or analysis assistance. We would like to thank two
anonymous reviewers who helped to improve the work. This study was partially supported by
Restaurante Camarões and The Mohamed bin Zayed Species Conservation Fund (#172516360).
PHM, PS and CF would like to thank University of Aveiro and FCT/MEC for the financial support
to CESAM RU (UID/AMB/50017) through national funds and co-financed by the FEDER, within
the PT2020 Partnership Agreement. EMV (308040/2017-1) and CRF (305304/2013-5; 306812/2017-
7) were supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and
PHM was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil
(CAPES; financing code 001) and Santander Universities (Santander Mundi scholarship).
References
Andrade EM, Aquino DN, Chaves LCG, Lopes FB (2017) Water as capital and its uses in the
Caatinga. In: Silva JC, Leal I, Tabarelli M (eds) Caatinga: the largest tropical dry forest region in
South America. Springer, Cham, pp 281–302.
Antongiovanni M, Venticinque EM, Fonseca CR (2018) Fragmentation patterns of the Caatinga
drylands. Land Ecol 33:1353–1367. https://doi.org/10.1007/s10980-018-0672-6
Astete S, Marinho-Filho J, Kajin M, Penido G, Zimbres B, Sollmann R, Jácomo ATA, Tôrres NM,
Silveira L (2017a). Forced neighbours: Coexistence between jaguars and pumas in a harsh
environment. J Arid Environ 146:27–34. https://doi.org/10.1016/j.jaridenv.2017.07.005
87
Astete S, Marinho-Filho J, Machado RB, Zimbres B, Jácomo ATA, Sollmann R, Tôrres NM, Silveira
L (2017b) Living in extreme environments: modeling habitat suitability for jaguars, pumas, and
their prey in a semiarid habitat. J Mammal 98:464–474.
http://dx.doi.org/10.1093/jmammal/gyw184.
Azevedo FC, Lemos FG, Almeida LB, Campos CB, Beisiegel BDM, Paula RC, Crawshaw Junior
PG, Ferraz KMP, Barros M, De Oliveira TG (2013). Avaliação do risco de extinção da onça-parda
Puma concolor (Linnaeus, 1771) no Brasil. Biodiv Brasil 3:107–121.
Bennie JJ, Duffy JP, Inger R, Gaston KJ (2014) Biogeography of time partitioning in mammals. P
Natl A Sci 111:13727–13732. https://doi.org/10.1073/pnas.1216063110
Bianchi RC, Olifiers N, Gompper ME, Mourão G (2016) Niche partitioning among pesocarnivores
in a Brazilian Wetland. PLoS ONE 11:e0162893. https://doi.org/10.1371/journal.pone.0162893
Bu H, Wang F, McShea WJ, Lu Z, Wang D, Li S (2016) Spatial co-occurrence and activity patterns
of mesocarnivores in the temperate forests of Southwest China. PLoS ONE 11:e0164271.
https://doi.org/10.1371/journal.pone.0164271
Cabús R (2015) Tropsolar 5.0. Grupo de Pesquisa em Iluminação (GRILU).
http://www.ctec.ufal.br/grupopesquisa/gril. Accessed 15 April 2019
Carmignotto AP, Astúa D (2017). Mammals of the Caatinga: diversity, ecology, biogeography, and
conservation. In: Silva JC, Leal I, Tabarelli M (eds) Caatinga: the largest tropical dry forest region
in South America. Springer, Cham, pp 211–254.
Carothers JH, Jaksic FM (1984) Time as a niche difference: the role of interference competition.
Oikos 42:403–406. https://doi.org/10.2307/3544413
Cavalcanti GN, Alfaro-Alvarado LD, Rodrigues FHG (2014) Home range and activity of Conepatus
semistriatus (Carnivora, Mephitidae) in Emas National Park, Brazil. Anim Biol 64:151–162.
https://doi.org/10.1163/15707563-00002436
Cheida, CC, Rodrigues FHG, Beisiegel BM (2013) Avaliação do risco de extinção do Guaxinim
Procyon cancrivorus (Cuvier, 1798) no Brasil. Biodiv Brasil 3:283–90.
88
Crooks KR, Soulé ME (1999) Mesopredator release and avifaunal extinctions in a fragmented system.
Nature 400:563–566. https://doi.org/10.1038/23028
De Angelo C, Paviolo A, Di Blanco Y, Di Bitetti M, Chiappe A (2008) Guía de huellas de los
mamíferos de Misiones y otras áreas del subtrópico de Argentina. Ediciones del Subtrópico,
Tucumán
Di Bitetti MS, De Angelo CD, Di Blanco YE, Paviolo A (2010) Niche partitioning and species
coexistence in a Neotropical felid assemblage. Acta Oecol 36:403e412.
https://doi.org/10.1016/j.actao.2010.04.001
Dias DM (2017) Spatiotemporal ecology of the striped hog-nosed skunk Conepatus semistriatus
(Carnivora, Mephitidae) in a seasonally dry forest of northeastern Brazil. Anim Biol 2:119–131.
https://doi.org/10.1163/15707563-00002525
Dias DM, Bocchiglieri A (2015) Dieta de carnívoros (Mammalia, Carnivora) em um remanescente
de Caatinga, Nordeste do Brasil. Bioikos 29:13–19.
Dias DM, Bocchiglieri A (2016) Trophic and spatio-temporal niche of the crab-eating fox, Cerdocyon
thous (Linnaeus, 1766) (Carnivora: Canidae), in a remnant of the Caatinga in northeastern Brazil.
Mammalia 80:281–291. https://doi.org/10.1515/mammalia-2014-0108
Dias DM, Campos CB, Rodrigues FHG (2018) Behavioural ecology in a predator-prey system.
Mamm Biol 92:30–36. https://doi.org/10.1016/j.mambio.2018.04.005
Dias DM, Massara RL, Campos CB, Rodrigues FHG (2019) Feline predator-prey relationships in a
semi-arid biome in Brazil. J Zool 307:282-291. https://doi.org/10.1111/jzo.12647
Emsens WJ, Hirsch BT, Kays R, Jansen PA (2013) Prey refuges as predator hotspots: ocelot
(Leopardus pardalis) attraction to agouti (Dasyprocta punctata) dens. Acta Theriol 59:257–262.
https://doi.org/10.1007/s13364-013-0159-4
Feijó A, Langguth A (2013). Mamíferos de médio e grande porte do Nordeste do Brasil: distribuição
e taxonomia, com descrição de novas espécies. Rev Nord Biol 22:3–225.
89
Fonseca CR, Antongiovanni M, Matsumoto M, Bernard E, Venticinque EM (2017) Conservation
opportunities in the Caatinga. In: Silva JC, Leal I, Tabarelli M (eds) Caatinga: the largest tropical
dry forest region in South America. Springer, Cham, pp 429–443.
Foster VC, Sarmento P, Sollmann R, Tôrres N, Jácomo ATA, Negrões N, Fonseca C, Silveira L
(2013) Jaguar and puma activity patterns and predator-prey interactions in four Brazilian biomes.
Biotropica 45:373–379. https://doi.org/10.1111/btp.12021
Gaynor KM, Hojnowski CE, Cater NH, Bashares JS (2018) The influence of human disturbance on
wildlife nocturnality. Science 360:1232–123. https://doi.org/10.1126/science.aar7121
Giordano AJ (2016) Ecology and status of the jaguarundi Puma yagouaroundi: a synthesis of existing
knowledge. Mammal Rev 46:30–43. https://doi.org/10.1111/mam.12051
Gómez H, Wallace RB, Ayala G, Tejada R (2005) Dry season activity periods of some Amazonian
mammals. Stud Neotrop Fauna Environ 40:91–95. https://doi.org/10.1080/01650520500129638
Goulart FVB, Cáceres NC, Graipel ME, Tortato MA, Ghizoni Jr IR, Oliveiras-Santos LGR (2009)
Habitat selection by large mammals in a southern Brazilian Atlantic Forest. Mamm Biol 74:182–
190. https://doi.org/10.1016/j.mambio.2009.02.006
Kasper CB, Leuchtenberger C, Bornholdt R, Pontes ARM, Beisiegel BM (2013) Avaliação do risco
de extinção do furão Galictis cuja (Molina, 1782) no Brasil. Biodiv Brasil 3:203-210.
Linkie M, Ridout MS (2011) Assessing tiger-prey interactions in Sumatran rainforests. J Zool
284:224–229. https://doi.org/10.1111/j.1469-7998.2011.00801.x
Lucherini M, Reppucci JI, Walker RS, Villalba ML, Wurstten A, Gallardo G, Iriarte A, Villalobos R,
Perovic P (2009) Activity pattern segregation of carnivores in the high Andes. J Mammal 90:1404–
1409. https://doi.org/10.1644/09-MAMM-A-002R.1
Marinho PH, Bezerra D, Antongiovanni M, Fonseca CR, Venticinque EM (2018a). Mamíferos de
médio e grande porte da Caatinga do Rio Grande do Norte, nordeste do Brasil. Mastozool Neotrop
25:345–362. https://doi.org/10.31687/saremMN.18.25.2.0.15
90
Marinho PH, Bezerra D, Antongiovanni M, Fonseca CR, Venticinque EM (2018b). Activity patterns
of the threatened northern tiger cat Leopardus tigrinus and its potential prey in a Brazilian dry
tropical forest. Mamm Biol 89:30–36. https://doi.org/10.1016/j.mambio.2017.12.004
Massara RL, Paschoal AMO, Bailey LL, Doherty PFJr, Chiarello AG (2016). Ecological interactions
between ocelots and sympatric mesocarnivores in protected areas of the Atlantic Forest,
southeastern Brazil. J Mammal 97:1634–1644. https://doi.org/10.1093/jmammal/gyw129
Meredith M, Ridout M (2018) Overview of the overlap package. R Project. 1–9. https://cran.r
project.org/web/packages/overlap/vignettes/overlap.pdf. Accessed 04 February 2019
MMA (2016) Áreas Prioritárias para conservação, uso sustentável e repartição de benefícios da
biodiversidade do bioma Caatinga – 2ª atualização - Portaria N°223, de 21 de junho de 2016.
Ministério do Meio Ambiente. http://www.mma.gov.br/areas-protegidas/cadastro-nacional-de-
ucs/item/10724-resultados-da-2%C2%B0-atualiza%C3%A7%C3%A3o-das-%C3%A1reas-
priorit%C3%A1rias.html Accessed 20 December 2018
Monterroso P, Alves PC, Ferreras P (2013) Catch me if you can: diel activity patterns of mammalian
prey and predators. Ethology 119:1044–1056. http://dx.doi.org/10.1111/eth.12156
Monterroso P, Aves PC, Ferreras P (2014) Plasticity in circadian activity patterns of mesocarnivores
daily activity coexistence. Behav Ecol Sociobiol 68:1403–1417.
http://dx.doi.org/10.1007/s00265-014-1748-1
Morato RG, Beisiegel BM, Ramalho EE, Campos CB, Boulhosa RLP (2013) Avaliação do risco de
extinção da onça-pintada Panthera onca (Linnaeus, 1758) no Brasil. Biodiv Brasil 3:122–132.
Moreno RS, Kays RW, Samudio RJr (2006) Competitive release in diets of ocelot (Leopardus
pardalis) and puma (Puma concolor) after jaguar (Panthera onca) decline. J Mamm 87:808–816.
https://doi.org/10.1644/05-MAMM-A-360R2.1
Nagy-Reis MB, Iwakami VHS, Estevo CA, Setz EZF (2019) Temporal and dietary segregation in a
Neotropical small-felid assemblage and its relation to prey activity. Mamm Biol 95:1–8.
https://doi.org/10.1016/j.mambio.2018.12.005
91
Nouvellet P, Rasmussen GSA, MacDonald DW, Courchamp F (2012) Noisy clocks and silent
sunrises: measurement methods of daily activity pattern. J Zool 286:179–184.
https://doi.org/10.1111/j.1469-7998.2011.00864.x
Oliveira TG, Pereira JA (2013) Intraguild predation and interspecific killing as structuring forces of
carnivoran communities in South America. J Mammal Evol 21:427–436.
https://doi.org/10.1007/s10914-013-9251-4
Oliveira-Santos LGR, Graipel ME, Tortato MA, Zucco CA, Cáceres NC, Goulart FVB (2012)
Abundance changes and activity flexibility of the northern tiger cat, Leopardus tigrinus
(Carnivora: Felidae), appear to reflect avoidance of conflict. Zoologia 29:115–120.
https://doi.org/10.1590/S1984-46702012000200003
Olmos F (1993) Notes on the food habitats of Brazilian “Caatinga” carnivores. Mammalia 57:126–
130.
Paglia AP, Fonseca GAB, Rylands AB, Herrmann G, Aguiar LMS, Chiarello AG, Leite YLR, Costa
LP, Siciliano S, Kierulff MCM, Mendes SL, Tavares VC, Mittermeier RA, Patton JL (2012)
Annotated Checklist of Brazilian Mammals. 2nd Edition. Occasional Papers in Conservation
Biology 6:1–76.
Palomares F, Caro TM (1999) Interspecific killing among mammalian carnivores. Am. Naturalist
153:492–508. https://doi.org/10.1086/303189
Penido G, Astete S, Jácomo ATA, Sollmann R, Tôrres N, Silveira L, Marinho Filho J (2017)
Mesocarnivore activity patterns in the semiarid Caatinga: limited by the harsh environment or
affected by interspecifc interactions? J Mammal 98:1732–1740.
https://doi.org/10.1093/jmammal/gyx119
Pita R, Mira A, Beja P (2011) Circadian activity rhythms in relation to season, sex and interspecific
interactions in two Mediterranean voles. Anim Behav 81:1023–1030.
https://doi.org/10.1016/j.anbehav.2011.02.007
92
Polis GA, Myers CA, Holt RD (1989) The ecology and evolution of intraguild predation: potential
competitors that eat each other. Annu Rev Ecol Syst 20:297–330.
https://doi.org/10.1146/annurev.es.20.110189.001501
Porfirio G, Foster VC, Fonseca C, Sarmento P (2016) Activity patterns of ocelots and their potential
prey in the Brazilian Pantanal. Mamm Biol 81:511–517.
https://doi.org/10.1016/j.mambio.2016.06.006
Projeto Caatinga Potiguar (2015) Projeto Caatinga Potiguar – Cartograma. WCS-Brazil/UFRN.
http://programs.wcs.org/brazil/Lugares-naturais/Projeto-Caatinga.aspx. Accessed 02 November
2017
Prugh LR, Golden CD (2014) Does moonlight increase predation risk? Meta-analysis reveals
divergent responses of nocturnal mammals to lunar cycles. J Anim Ecol 83:504–514.
https://doi.org/10.1111/1365-2656.12148
Prugh LR, Stoner CJ, Epps CW, Bean WT, Ripple WJ, Laliberte AS, Bashares JS (2009) The rise of
the mesopredator. Bioscience 59:779–791. https://doi.org/10.1525/bio.2009.59.9.9
R Development Core Team (2012) R: a Language and Environment for Statistical Computing. R
Foundation for Statistical Computing. http://www.R-project.org/. Accessed 30 April 2018
Ridout MS, Linkie M (2009) Estimating overlap of daily activity patterns from camera trap data. J
Agr Biol Envir St 14:322-337. https://doi.org/10.1198/jabes.2009.08038
Ritchie EG, Johnson CN (2009) Predator interactions, mesopredator release and biodiversity
conservation. Ecol Lett 12:982–998. https://doi.org/10.1111/j.1461-0248.2009.01347.x
Roemer GW, Gompper ME, Valkenburgh BV (2009) The ecological role of the mammalian
mesocarnivore. BioScience 59:165–173. https://doi.org/10.1525/bio.2009.59.2.9
Schoener TW (1974) Resource partitioning in ecological communities. Science 185:27-39.
https://doi.org/10.1126/science.185.4145.27
93
Silva JMC, Barbosa LCF, Leal I, Tabarelli M (2017) The Caatinga: understanding the challenges. In:
Silva JC, Leal I, Tabarelli M (eds) Caatinga: the largest tropical dry forest region in South America.
Springer, Cham, pp 3–19.
Stoner K, Timm RM (2011) Seasonally dry tropical forest mammals: adaptations and seasonal
patterns. In: Dirzo R, Young HS, Mooney HA, Ceballos G (eds) Seasonally Dry tropical forests:
ecology and conservation. Island Press, Washington, pp 85-106.
Terrien J, Perret M, Aujard F (2011) Behavioral thermoregulation in mammals: a review. Front.
Biosci 16:1428–1444. https://doi.org/10.2741/3797
Tomasella J, Silva RMPV, Barbosa AA, Rodriguez DA, Oliveira MS, Sestini MF (2018)
Desertification trends in the Northeast of Brazil over the period 2000–2016. Int J Appl Earth Obs
73:197–206. https://doi.org/10.1016/j.jag.2018.06.012
Vanak AT, Fortin D, Thaker M, Ogden M, Owen C, Greatwood S, Slotow R (2013) Moving to stay
in place: behavioral mechanisms for coexistence of African large carnivores. Ecology 94:2619–
2631. https://doi.org/10.1890/13-0217.1
Velloso AL, Sampaio EVSB, Pareyn FGC (2002) Ecorregiões propostas para o Bioma Caatinga.
Associação Plantas do Nordeste, Instituto de Conservação Ambiental, The Nature Conservancy
do Brasil, Recife.Valeix M, Chamaillé-Jammes S, Fritz H (2007) Interference competition and
temporal niche shifts – elephants and herbivore communities at waterholes. Oecologia 153:739–
748. https://doi.org/10.1007/s00442-007-0764-5
Wang E (2002) Diets of Ocelots (Leopardus pardalis), Margays (L. wiedii), and Oncillas (L. tigrinus)
in the Atlantic Rainforest in Southeast Brazil. Stud. Neotrop. Fauna E 37:207–212.
https://doi.org/10.1076/snfe.37.3.207.8564
Wang Y, Allen ML, Wilmers CC (2015) Mesopredator spatial and temporal responses to large
predators and human development in the Santa Cruz Mountains of California. Biol
Conserv.190:23–33. https://doi.org/10.1016/j.biocon.2015.05.007
94
Woodroffe R, Ginsberg JR (1998) Edge effects and the extinction of populations inside protected
areas. Science 280:2126–2128. https://doi.org/10.1126/science.280.5372.2126
Ximenez A (1982) Notas sobre felidos neotropicales viii observaciones sobre el contenido estamacal
y el comportamiento alimentar de diversas espécies de felinos. Rev Nord de Biol 5:89–91.
95
Supplementary Material 1
2
Table S1 Data from three camera-trapping campaigns carried out to collect mesocarnivore activity data in a Caatinga dry forest, northeastern Brazil. 3
Survey period Cameras/station Stations (N) Effort (camera-days)
May - June 2016 2 24 1,114
January - October 2017 1 43 9,255
October 2018 - February 2019 1 37 3,607
Total 50 13,976
4
96
Table S2 Temporal niche overlap coefficient (∆), with its respective 95% confidence intervals (between parentheses), among mesocarnivores during wet 5
(upper diagonal) and dry season (lower diagonal) in a Caatinga dry forest, northeastern Brazil. The ∆ values with asterisk indicate significant difference in 6
activity distribution (p<0.05) according to Mardia-Watson-Wheller test. 7
Mesocarnivores Striped hog-nosed skunk
Crab-eating fox
Crab-eating raccoon
Ocelot Northern tiger cat
Jaguarundi
Striped hog-nosed skunk - 0.70*
(0.65-0.76)
0.79
(0.65-0.91)
0.73*
(0.63-0.81)
0.71*
(0.64-0.78)
0.17*
(0.06-0.27)
Crab-eating fox 0.72*
(0.66-0.77) -
0.75
(0.61-0.87)
0.77*
(0.69-0.85)
0.92
(0.88-0.96)
0.44*
(0.32-0.55)
Crab-eating raccoon 0.87*
(0.67-1.00) 0.81
(0.61-0.97) -
0.69 (0.53-0.83)
0.75 (0.62-0.88)
0.28* (0.15-0.41)
Ocelot 0.80
(0.71-0.88)
0.78*
(0.70-0.85)
0.87
(0.66-1.00) -
0.80*
(0.71-0.89)
0.37*
(0.26-0.48)
Northern tiger cat 0.71*
(0.64-0.78)
0.91
(0.87-0.95)
0.79
(0.58-0.95)
0.75*
(0.65-0.84) -
0.45*
(0.36-0.54)
Jaguarundi 0.23*
(0.12-0.34)
0.50*
(0.38-0.61)
0.34*
(0.14-0.55)
0.31*
(0.19-0.44)
0.52*
(0.39-0.64) -
8
97
Table S3 Description of the activity patterns of mosocarnivore’s potential prey based on records obtained in a Caatinga dry forest, northeastern Brazil). 9
Statistically significant values of uniformity Rayleigh test (based in all records) and of Mardia-Watson-Wheller test (based on the comparison of dry and wet 10
season records) are presented in bold. ∆ represents the overlap coefficient values with their respective 95% confidence intervals (95% CI). The few records of 11
nine-banded armadillo and tinamou prevented some analysis. 12
Potential prey (N)
Total
Rayleigh Z
(p)
Activity
(%day/night/twilight)
N
(dry/wet)
∆dry-wet (95%
CI)
MWWdry-wet
(p)
Mammals
Spix's cavy 362 120.4
(<0.001) Nocturnal-crepuscular
(1.9 / 66.9 / 31.2) 139 / 223
0.94 (0.86-0.99)
1.45 (0.48)
Punaré rat 142 78.0
(<0.001)
Nocturnal
(0.0 / 93.7 / 6.3) 86 / 56
0.92
(0.82-0.99)
0.25
(0.88)
Rock cavy 74 3.2
(0.04)
Cathemeral (48.5 / 25.0 / 26.5)
43 / 31 0.79
(0.64-0.91) 0.40
(0.82)
White-eared opossum 151 82.2
(<0.001)
Nocturnal
(0.0 / 96.0 / 4.0) 94 / 57
0.94
(0.84-1.00)
0.51
(0.77)
Yellow armadillo 448 98.9
(<0.001)
Mostly diurnal
(69.2 / 19.9 / 10.9) 232 / 216
0.78
(0.71-0.86)
20.64
(<0.001)
Nine-banded armadillo 7 - Unclassified
(0.0 / 100 / 0.0) 1 / 6 - -
Birds
Dove 1015 466.1
(<0.001) Diurnal
(79.0 / 0.0 / 21.0) 306 / 709
0.90 (0.84-0.95)
13.51 (<0.01)
Tinamou 16 6.37
(<0.001)
Crepuscular
(37.5 / 0.0 / 62.5) 6 / 10 - -
Reptiles
Lizard 179 134.3
(<0.001)
Diurnal
(100.0 / 0.0 / 0.0) 95 / 84
0.88
(0.79-0.96) 3.18 (0.20)
13
98
Table S4 Temporal niche overlap coefficient (∆), with its respective 95% confidence intervals (between parentheses), among mesocarnivores and potential 14
prey during dry and wet season in a Caatinga dry forest, northeastern Brazil. The ∆ values with asterisk indicate significant difference (p <0.05) in activity 15
distribution according to Mardia-Watson-Wheller test. 16
Mesocarnivores
Potential prey
Spix's cavy Punaré rat Rock cavy White-eared
opossum
Yellow
armadillo Dove Lizard
Dry season
Striped hog-nosed skunk 0.71*
(0.62-0.79)
0.89
(0.79-0.98)
0.37*
(0.25-0.50)
0.96
(0.89-1.00) -
0.07*
(0.01-0.09)
0.01*
(<0.01-0.03)
Crab-eating fox 0.80*
(0.72-0.86)
0.71*
(0.62-0.78)
0.53*
(0.41-0.66)
0.69*
(0.62-0.76) -
0.29*
(0.25-0.33)
0.08*
(0.04-0.12)
Crab-eating raccoon 0.78
(0.56-0.93)
0.84
(0.62-1.00)
0.47*
(0.28-0.66)
0.86
(0.66-1.00) -
0.15*
(<0.01-0.31)
0.04*
(<0.01-0.16)
Ocelot 0.71*
(0.60-0.81)
0.73*
(0.60-0.84)
0.48*
(0.34-0.62)
0.80*
(0.69-0.90)
0.26*
(0.19-0.33)
0.12*
(0.05-0.19)
0.04*
(<0.01-0.09)
Northern tiger cat 0.78*
(0.70-0.85) 0.72*
(0.62-0.80) 0.57*
(0.44-0.69) 0.68*
(0.60-0.76) -
0.30* (0.25-0.36)
0.12* (0.08-0.17)
Jaguarundi 0.39*
(0.25-0.52)
0.25*
(0.13-0.37)
0.68
(0.53-0.82)
0.22*
(0.12-0.33) -
0.73*
(0.60-0.83)
0.43*
(0.29-0.58) Wet season
Striped hog-nosed skunk 0.81*
(0.72-0.90 )
0.94
(0.84-1.00)
0.32*
(0.18-0.48)
0.87
(0.79-0.92) -
0.04*
(<.01-0.08)
<0.01*
(<0.01-0.03)
Crab-eating fox 0.83*
(0.76-0.90 ) 0.68*
(0.60-0.75) 0.54*
(0.39-0.68) 0.68*
(0.60-0.76) -
0.27* (0.23-0.31)
0.09* (0.06-0.13)
Crab-eating raccoon 0.81
(0.67-0.92)
0.80
(0.65-0.93)
0.44*
(0.26-0.61)
0.75*
(0.60-0.88) -
0.11*
(0.02-0.20)
0.05*
(<0.01-0.13)
Ocelot 0.73*
(0.64-0.82)
0.72*
(0.59-0.84)
0.54*
(0.39-0.69)
0.78
(0.67-0.88)
0.50*
(0.41-0.60)
0.22*
(0.14-0.32)
0.16*
(0.08-0.24)
Northern tiger cat 0.82* 0.68* 0.59* 0.68* - 0.27* 0.15*
99
(0.75-0.88) (0.60-0.77) (0.45-0.74) (0.59-0.76) (0.22-0.32) (0.10-0.21)
Jaguarundi 0.29*
(0.21-0.38)
0.16*
(0.08-0.26)
0.74
(0.60-0.87)
0.16*
(0.07-0.25) -
0.67*
(0.56-0.77)
0.42*
(0.30-0.54)
17
100
18
101
Fig. S1 Density estimates of daily activity patterns and extension of overlap between mesocarnivores and their potential prey in a Caatinga dry 19
forest, northeastern Brazil. Overlap is represented by the shaded grey area. The dashed vertical lines represent the average legal time of sunrise 20
[5:20 h] and sunset [17:18 h]) during the study period. Significant differences in MWW test are indicated by an asterisk after overlap coefficient 21
values (∆). The time of the records is shown as ticks in the bottom of the figures. We considered yellow-armadillo, which is a medium-sized 22
mammal (5.4 kg, Paglia et al. 2012), a potential prey only for larger mesocarnivore in the area, the ocelot. 23
24
102
Capítulo 3 25
26
CO-OCCURRENCE PATTERNS BETWEEN OCELOT AND SYMPATRIC 27
MESOCARNIVORES IN A BRAZILIAN DRY FOREST 28
29
30
31
32
33
Este capítulo está escrito de acordo com as regras da revista Mammalian Biology 34
(www.journals.elsevier.com/mammalian-biology). 35
103
Running title: Co-occurrence between ocelot and mesocarnivores 36
37
Co-occurrence patterns between ocelot and sympatric mesocarnivores in a Brazilian 38
dry forest 39
40
Paulo Henrique Marinhoa*, Carlos Roberto Fonsecaa, and Eduardo Martins Venticinquea 41
42
a Departamento de Ecologia, Centro de Biociências, Universidade Federal do Rio Grande 43
do Norte, Natal, Rio Grande do Norte, Brazil 44
45
*Corresponding author: Paulo Henrique Marinho, Programa de Pós-graduação em 46
Ecologia, Departamento de Ecologia, Centro de Biociências, Universidade Federal do 47
Rio Grande do Norte (UFRN), Campus Universitário UFRN, Lagoa Nova, Natal, RN, 48
59078-970, Brasil 49
51
52
53
54
55
56
57
58
⸭ 59
104
Abstract 60
Understanding the mechanisms for maintaining the diversity of mesocarnivores is 61
important for the management and conservation of altered communities. Related species 62
compete strongly through exploitation or interference competition through aggressive 63
encounters; thus they can develop mechanisms of coexistence to avoid competitive 64
exclusion. Subordinate mesocarnivores are generally pressured to occupy suboptimal 65
habitats that minimize intraguild interference competition with competitively superior 66
species, which tend to become more intense whether a dominant mesopredator benefits 67
from the absence of top predators. We aimed to investigate the effects of dominant 68
ocelot’s presence on the habitat use and detection of subordinate mesocarnivores in an 69
area of Caatinga dry forest where top predators are functionally absent. We used camera 70
trapping data associated with occupancy models of single and two species, also 71
considering the potential effect of seasonality, prey availability, and anthropogenic 72
pressures on species' habitat use. Our results suggest that most submissive 73
mesocarnivores occur regardless of the presence of ocelot, while the jaguarundi uses 74
more habitats shared with this dominant mesopredator, probably due to their strong 75
temporal segregation. The absence of spatial avoidance among mesocarnivores may be 76
related to a probable ocelot’s low density in the area, reducing the effects of intraguild 77
suppression. Our results allow us to better understand how mesocarnivores interact 78
spatially in semiarid environments, where top predators have become extinct or are in 79
decline due to chronic human disturbance. 80
Keywords: Caatinga; Intraguild competition; Mesopredator; Semiarid; Two-species 81
occupancy models. 82
83
105
Introduction 84
Mammalian carnivores structure biological communities and regulate the functioning 85
of ecosystems through control of prey populations and intraguild interactions (Terborgh 86
et al., 1999; Miller et al., 2001; Ripple et al., 2014). They directly or indirectly affect 87
various taxa and ecological processes through trophic cascades and thus, representing a 88
key group for the maintenance of healthy and resilient ecosystems, so that the current and 89
accentuated decline in top predators has negative consequences for the entire environment 90
(Woodroffe, 2000; Miller et al., 2001; Beschta and Ripple, 2009; Ripple et al. 2014). One 91
of the consequences of the local extinction or top predator populations’ depletion is the 92
mesopredator release from top-down control (Crooks and Soulé, 1999; Ritchie and 93
Johnson, 2009). In this new situation, mesocarnivores (mammalian carnivores < 15 kg; 94
Roemer et al., 2009) tend to become more abundant and generate greater pressure on their 95
prey (Crooks and Soulé, 1999; Jiménez et al., 2019). Consequently, understanding the 96
mechanisms for maintaining the diversity of mesocarnivores is essential for the 97
management and conservation of altered communities (Bu et al. 2016). 98
Many mammalian carnivores, especially more related species, compete strongly 99
through exploitation, which reduces limiting resources for competitors, or interference 100
competition, through direct aggressive encounters (Case and Gilpin, 1974; Schoener, 101
1974a). Interspecific killing and intraguild predation of a subordinate (usually smaller or 102
solitary predator) by a dominant (larger or social) predator are additional pressures that 103
regulate coexistence in carnivore guild (Palomares and Caro, 1999; Oliveira and Pereira, 104
2014). Two morphologically or ecologically similar species must develop coexistence 105
mechanisms to avoid competitive exclusion (Hutchinson, 1959; Macarthur and Levins, 106
1964; Diamond, 1978), such as diverge in their diet or segregate temporally and/or 107
spatially (Schoener, 1974a,b). Where spatial segregation is the most efficient strategy for 108
106
reducing competition, subordinate species can be pressured to occupy suboptimal habitats 109
to minimize competition and the risks of agonistic encounters with competitively superior 110
species (Harrison et al., 1989). Therefore, habitat (space) can be an important dimension 111
in promoting the coexistence of closely related species (Schoener, 1974a; Harrison et al., 112
1989). 113
In neotropical ecosystems, the ocelot Leopardus pardalis Linnaeus, 1758 (8-11 kg) 114
acts as the main competitor and potential intraguild predator of the mesocarnivore 115
community (Oliveira and Pereira, 2014). In locations where top predators are absent, the 116
ocelot can expand its realized ecological niche and emerge as a surrogate top predator 117
(Terborgh et al., 1999; Moreno et al., 2006; Prugh et al., 2009). However, other studies 118
have found a positive relationship between the presence of ocelot and top predators, 119
probably due to the higher habitat quality that benefits both (Massara et al., 2015, 2018). 120
Different theoretical and empirical studies have indicated the negative effect of ocelot on 121
the abundance, distribution and activity patterns of smaller carnivores, especially felids 122
(Oliveira et al., 2010; Di Bitetti et al., 2010; Oliveira-Santos et al., 2012; Oliveira and 123
Pereira, 2014; Massara et al., 2016). This might occur mainly via the potential for 124
intraguild predation and killing (Polis et al., 1989; Oliveira et al., 2010) since most 125
neotropical mesocarnivores are 2.0–5.4 times smaller than the dominant ocelot (Donadio 126
and Buskirk 2006). Indeed, different studies have documented the predation of 127
mesocarnivores by ocelot (Moreno et al., 2006; Bianchi et al., 2010; Bolze et al., 2019). 128
However, studies have also demonstrated the absence of ocelot’s effects on some species, 129
especially those with different diets (Massara et al., 2016). 130
In the seasonally dry tropical forests of northeastern Brazil (Caatinga), most of the 131
places where puma Puma concolor Linnaeus, 1771 and jaguar Panthera onca Linnaeus, 132
1758 are locally extinct, which represent a large part of this semiarid ecoregion (Paula et 133
107
al., 2013; Azevedo et al., 2013), the ocelot remains the largest predator acting in the 134
control of potential prey and competitors’ suppression. This study aimed to investigate 135
the effects of ocelot’s presence on the occupancy and detection of sympatric 136
mesocarnivores in a not legally protected arera of the Caatinga. In this region top 137
predators have not been recorded for approximately 10 years (Marinho et al., 2018a). We 138
predicted that the subordinate mecocarnivores would occupy less places most occupied 139
by ocelot to avoid agonistic encounters and explotative competition, especially those that 140
strongly overlap their nighttime activity with this dominant mesopredator (Marinho et al., 141
2020). Considering that competition and intraguild interactions may vary seasonally with 142
resource availability and habitat structure (Schmitt and Holbrook, 1986; Vanak et al., 143
2013), we evaluated potential changes in the relationships between ocelot and 144
mesocanivores considering the well-marked dry and wet seasons of the Caatinga. The 145
occurrence of carnivores is closely related to prey availability (Santos et al., 2019), as 146
well as impacted by anthropogenic activities (Cruz et al., 2018; Marinho et al., 2018b). 147
In addition, we assessed the role of resource availability (e.g. prey and water), 148
anthropogenic threats (hunters and their dogs), and disturbances (livestock) on habitat use 149
of species, considering that these factors may modulate interspecific interactions and 150
species distributions (Wang et al., 2015; Nagy-Reis et al., 2017; Dias et al., 2019a; Santos 151
et al., 2019). 152
At least seven species of mesocarnivores occur in our study area (Marinho et al., 153
2020). Three of these mesocarnivores have a generalist diet: crab-eating fox Cerdocyon 154
thous Linnaeus, 1766 (adult body mass = 5.7-6.5 kg), crab-eating raccoon Procyon 155
cancrivorus (G. [Baron] Cuvier, 1798) (5.4-8.8 kg ), and striped hog-nosed skunk 156
Conepatus amazonicus Lichtenstein, 1838 (2.4 kg); the other four are mostly strict 157
carnivorous: lesser grison Galictis cuja (Molina, 1782) (1-3 kg), northern tiger cat 158
108
Leopardus tigrinus Thomas, 1904 (1.5-3 kg), jaguarundi Herpailurus yagouaroundi (É. 159
Geoffroy Saint-Hilaire, 1803) (3- 6 kg), and ocelot Leopardus pardalis (8-11 kg) (Paglia 160
et al., 2012; Oliveira and Pereira 2014). Northern tiger cat and jaguarundi are threatened 161
with extinction in Brazil (MMA, 2014) and the former is categorized as vulnerable by the 162
IUCN (IUCN, 2020). 163
Our group has studied this area since 2014, and despite the large sampling effort 164
employed in it (more than 16,800 camera-days), we have never recorded the presence of 165
puma in the area, not even through tracks and feces. In fact, the most convincing reports 166
of its presence in the region are over 20 years old (Pichorim et al., 2014; Marinho et al., 167
2018a), and for the jaguar, probably extinct during the beginning of the region's 168
colonization, there are not even old reports. 169
170
Material and methods 171
Study area 172
Our study was carried out in an area composed of private properties in Lajes 173
municipality, in Rio Grande do Norte state, northeastern Brazil (5º44'-5º51'S latitude, 174
36º11'-36º06'W longitude, Fig. 1). This region, which lies between Feiticeiro and Bonfim 175
mountains, is considered a priority area for the conservation of the Caatinga (Fonseca et 176
al., 2017); however, it does not have any legal protection. The main economic activity in 177
the region is cattle raising (cows, goats, and sheep), in addition to cutting firewood from 178
exotic and native species for charcoal production. The region is strongly impacted by the 179
illegal hunting of mammals and birds (Marinho et al., 2018a). The vegetation varies from 180
thorny shrub formations in the lower areas to arboreal habitats (~ 5 m canopy height) in 181
the mountains, typical phytophysiognomies of the Caatinga dry tropical forest. The 182
climate is semiarid, with high temperatures (average of 27 ° C and maximum of 33 ° C), 183
109
and low rainfall (350 mm) that usually concentrate between February and May (Szilagyi, 184
2007; Tomasella et al., 2018). 185
186
187
Fig. 1 Study area and distribution of the 43 camera trap stations used to investigate spatial co-188
occurrence patterns between the ocelot and sympatric mesocarnivore in a Caatinga dry forest area 189
in northeastern Brazil. 190
191
Camera trapping 192
We obtained records of mesocarnivores through camera trapping carried out from 193
January to October 2017 and from October 2018 to February 2019. These periods covered 194
both the wet (January to June) and the dry season (July to December) (adapted from 195
Tomasella et al., 2017). We distributed 43 sampling points over an area of approximately 196
70 km² (as estimated by minimum convex polygon). At each point, we installed a camera 197
trap (Bushnell® Trophy Cam ™ HD) on trails used by people and livestock, dirt roads, 198
and temporary stream, which are preferred places for the movement of carnivores 199
(Goulart et al., 2009). These cameras remained active 24 hours a day and were 200
programmed to take two or three photos per shot at minimum intervals of 5 minutes. 201
110
Seeking to follow a minimum distance of 1 km between cameras, our points were away 202
from each other at an average of 929.5 m (SD = 262.1 m). We did not use any type of bait 203
to attract animals. We considered consecutive records of the same species at the same 204
point after 1 hour to be independent (Goulart et al. 2009). 205
Covariates 206
We used the number of records of small mammals detected in the camera traps, 207
divided by the sampling effort in camera-days times 100 (the relative abundance index or 208
RAI), expecting a positive effect of this covariate on the species’ habitat use, especially 209
for strictly carnivorous species like wild cats (Santos et al., 2019). We used data of 210
potential prey like rock cavy Kerodon rupestris (Wied-Neuwied, 1820), Spix's cavy 211
Galea spixii (Wagler et al., 1831), and punaré rat Thrichomys laurentius Thomas, 1904. 212
One of the main threats to neotropical mammals is hunting (Ferreguetti et al., 2019), 213
especially in unprotected environments such as our study area. In addition to the direct 214
persecution of carnivores to prevent or in retaliation for livestock or poultry predation, 215
hunting also reduces their prey (Peters et al., 2017; Ferreguetti et al., 2019). Thus, we 216
used the relative abundance index of hunters and dogs recorded during sampling with 217
camera trapping, which are widely used in this activity in the region, as a covariate with 218
a potential negative effect on mesocarnivore’s occurrence. Another anthropic disturbance 219
present in semiarid environments is cattle, which degrades vegetation and displaces native 220
species (Marinho et al., 2016; Pudyatmoko, 2017). We expected that RAI of cattle (cows, 221
sheep, and goats recorded during camera trapping) would negatively affect the presence 222
of mesocarnivores, mainly for species such as ocelot that prefer less degraded habitats 223
and avoid human activities (Cruz et al., 2018). Finally, considering the variations in our 224
effort between points due to operational problems and logistical limitations, we used the 225
sampling effort in camera-days as a detection covariate, expecting a positive relationship 226
111
between them (Massara et al., 2016). We assumed that our sampling points vary little in 227
relation to habitat structure and integrity, so we did not model any habitat variables. 228
Based on a Pearson’s correlation, we found no indication of collinearity between the used 229
covariates (e.g. |r |< 0.6) (Table S1). 230
Occupancy and co-occurrence analysis 231
MacKenzie et al. (2002, 2006) developed hierarchical models based on maximum 232
likelihood capable to estimate the occupancy of a species in an area (ψ, biological 233
parameter) considering the possibility of false absences when estimating the probability 234
of detection (p, observational parameter) and to correct the occupancy estimate. From 235
these models, Richmond et al. (2010) developed a new approach that allows the modeling 236
of conditional two-species occupancy incorporating predictor covariates. This 237
parameterization makes it possible to estimate the occupancy probability of a subordinate 238
species B conditioned to the presence of a dominant species A (ψA x ψAB or ψAb) 239
(Richmond et al., 2010). In addition, it allows modeling if the detection of the subordinate 240
species changes in the absence (pB) or presence (rBA or rBa) of the dominant species 241
(Richmond et al., 2010). This approach also allows estimating the species interaction 242
factor (SIF) (MacKenzie et al., 2004; Richmond et al., 2010). SIF values < 1 suggest that 243
the species occur together less than expected by chance, suggesting spatial avoidance, 244
while SIF values >1 indicate that they co-occur more than expected by chance, suggesting 245
aggregation, and finally, a SIF value =1 suggests that the two species occur independently 246
(MacKenzie et al., 2004; Richmond et al., 2010). Considering that the home range of the 247
species studied in general covers more than one sample point, we interpreted occupancy 248
as the probability of habitat use (MacKenzie et al., 2006; Nagy-Reis et al., 2017). 249
We created the detection (1) and non-detection (0) histories for each mesocarnivore 250
species with sufficient data for modeling using 10-day occasions to maximize the 251
112
detection estimates. We used 12 occasions in 2017 wet season (March to June, discarding 252
February data for a better balance between seasons), 11 occasions in the beginning of 253
2017 dry season (July to October) and 12 occasions at the end of 2018 dry season (October 254
2018 to January 2019) to balance the occasion number between seasons and sampling 255
years. To consider possible variations in the species' occupancy or habitat use over time 256
we incorporated potential seasonal variation into the modeling, what is an important 257
premise of this type of modeling (MacKenzie et al., 2002). For that, we used three groups 258
that represented the three sample periods throughout the wet and dry throughout the 259
seasons. As our main interest was the co-occurrence patterns rather than investigating 260
dynamic process such as colonization and extinction (e.g., multi-season modeling), we 261
chose to use single-season models considering the surveys in the dry period and in the 262
two wet periods as groups, which also decrease the number of parameters and may 263
improve the model convergence (Gutiérrez-González et al., 2017; Rich et al., 2017). 264
We opted for a two-step approach (Nagy-Reis et al., 2017). We used single-season 265
single-species models and then single-season two-species models with conditional 266
occupancy. First, we modeled the covariate effects on the occupancy or habitat use (prey 267
abundance, hunter pressure, and livestock abundance) and detection (sample effort) of the 268
species individually, besides considering the survey periods as three groups (one wet and 269
two dry season periods) for both parameters. We modeled all possible combinations 270
(Doherty et al. 2012), but we limited the maximum number of five covariates and their 271
additive effects per model to avoid lack of models’ convergence. 272
In the second step, we used the covariates from the top-ranked models (ΔAICc ≤ 2) 273
in the first step for each species in the investigation of conditional occupancy. We 274
considered ocelot as the dominant species capable of affecting habitat use and the 275
detection of subordinate mesocarnivores (Oliveira et al., 2010). We evaluated whether 276
113
the subordinate species’ occupancy is affected by the presence of the dominant species 277
(ψBA≠ψBa) or not (ψBA=ψBa), as well as whether the detection of the other mesocarnivores 278
is affected by ocelot presence (pB≠rBA=rBa) or detection (pB≠rBA≠rBa). In this second step, 279
we also used all possible combinations of additive effects, not keeping the variables 280
ranked in the first step as fixed effects. All modeling was performed using the free 281
program Mark (White and Burnham, 1999). 282
The ranking of the candidate models was performed using the Akaike Information 283
Criterion corrected for small samples (AICc) (Burnham and Anderson, 2002). Models 284
with ΔAICc ≤ 2 were considered adjusted (Burnham and Anderson, 2002). Akaike's 285
weight (w) was used to interpret the relative importance of each model and its covariates. 286
The balanced number of models at each step allowed us to use the cumulative AICc 287
weights (w+) as an indicator of relevance for each covariate evaluated (Burnham and 288
Anderson, 2002; Massara et al., 2016). In addition, continuous covariates with regression 289
coefficient (ß) that did not include zero were considered significantly important for 290
predicting habitat use and species detection. We used the model averaging to estimate the 291
overall occupancy and detection of species (general and per season when it was 292
important). We used the goodness-of-fit (GOF) test incorporated in program PRESENCE 293
(Hines, 2006) to correct possible cases of overdispersion and lack-of-fit (e.g. 𝑐̂ > 1 and P 294
< 0.05) of the most parameterized models (global) of each species in the single-season 295
single-species approach, through the Quasi Akaike Information Criterion (QAICc) 296
(Burnham and Anderson, 2002; MacKenzie and Bailey 2004). For species with 𝑐̂ values 297
> 1 we applied the QAIC, while for 𝑐̂ values < 1 we kept the value of 1 unchanged as 298
suggested by Cooch and White (2017). 299
300
114
Results 301
Data collection 302
With a total effort of 12,862 camera-days (9,255 camera-days in January - October 303
2017 and 3,607 camera-days in October 2018 - February 2019), we obtained 2,035 304
records of the seven mesocarnivore species known for our study area (Table 1). We also 305
obtained 492 records of potential prey such as rock cavy (74), Spix's cavy (286), and 306
punaré rat (132). Regarding human activities, we obtained records of cows (2,054), goats 307
(170), sheep (107), in addition to hunters (113), and probable hunting dogs (89). For crab-308
eating raccoon and lesser grison, the low number of records and recaptures did not allow 309
further analysis. 310
311
Table 1 Number of records and naïve occupancy of mesocarnivores, their potential prey, and 312
anthropogenic activities (divided by seasons/sampling period and in general) in a Caatinga dry 313
forest area, northeastern Brazil. 314
Nº of records Naïve occupancy
Mesocarnivores General Wet
2017
Dry
2017
Dry
2018 General
Wet
2017
Dry
2017
Dry
2018
Striped hog-nosed
skunk 266 91 119 56 0.65 0.74 0.80 0.57
Crab-eating fox 1031 313 433 285 0.69 0.65 0.73 0.73
Crab-eating raccoon 29 19 10 0 0.14 0.23 0.17 0.00
Jaguarundi 75 38 26 11 0.32 0.44 0.44 0.19
Northern tiger cat 494 198 171 125 0.79 0.84 0.90 0.76
Ocelot 137 46 52 39 0.42 0.51 0.46 0.38
Lesser grison 4 3 0 1 0.03 0.07 0.00 0.03
315
Occupancy patterns 316
For all analyzed species the GOF test suggested some level of overdispersion or 317
underdispersion in data, as follows: jaguarundi (𝑐̂ = 0.51, P = 0.49), striped hog-nosed 318
skunk (𝑐̂ = 1.59, P = 0.10), crab-eating fox (c-hat = 1.75, P < 0.01), ocelot (𝑐̂ = 2.12, P = 319
0.06), and northern tiger cat (𝑐̂ = 4.72, P < 0.01). For the jaguarundi, the global model 320
115
used was simpler (ψ [season + prey], p [season]), since more parameterized models did 321
not reach convergence, probably due to the low detection rate of this species (p < 0.10). 322
Thus, for jaguarundi eight additive models were built with all possible combinations, 323
while for the other species 63 additive models were run in each analysis, considering all 324
possible combinations of up to five covariates per model (including season) (Table 2). 325
We found little evidence of the effect of continuous covariates on habitat use and 326
detection of mesocarnivores (Table 2; Table 3). Although factors such as hunter and 327
livestock, in the case of striped hog-nosed skunk, and prey for crab-eating fox, presented 328
a high cumulative weight (w+> 0.50), the coefficient (ß) of all covariates overlapped with 329
zero values (Table 3), suggesting that there is no clear effect of these covariates on the 330
evaluated parameters. On the other hand, the results suggest that the detection of the 331
striped hog-nosed skunk and the crab-eating fox varied between the sampled seasons, the 332
same occurred for the habitat use of the jaguarundi (Table 2; Table 3; Fig. 2; Table S2). 333
For the northern tiger cat and the ocelot the top ranked models had no effect of the 334
covariates (e.g. ψ[.], p[.]) (Table 2), suggesting no effect of the evaluated factors (Table 335
3). Based on the model averaging, the detection probability varied widely between 336
species, from 0.07 (jaguarundi) to 0.45 (for crab-eating fox) (Fig. 1A; Table S2), 337
meanwhile, in general, the species presented a moderate to high estimated occupancy (e.g. 338
ψ = 0.50 - 0.99) over the seasons and sampled years (Fig. 1B; Table S2). 339
340
Table 2 Models used to assess the patterns of occupancy and detection of mesocarnivores in a 341
Caatinga dry forest area, northeastern Brazil. Only the models considered adjusted (ΔQ/AICc ≤ 342
2) are presented. Akaike’s weight (w). Number of parameters (K). 343
Species
Model Q/AICc ΔQ/AICc w Likelihood K QDeviance -2log(L)
Striped hog-nosed skunk
ψ(prey+hunt+livest), p(season) 623.107 0.000 0.105 1.000 7 608.116 966.904
116
ψ(hunt+livest), p(season) 623.308 0.201 0.095 0.904 6 610.571 970.808
ψ(season+hunt), p(.) 624.215 1.108 0.060 0.575 5 613.693 975.772
ψ(season+hunt+livest), p(season) 624.858 1.751 0.044 0.417 8 607.572 966.040
ψ(prey+hunt+livest), p(season) 625.007 1.900 0.040 0.387 8 607.721 966.277
Crab-eating fox
ψ(prey), p(season) 735.148 0.000 0.213 1.000 5 724.627 1268.097
ψ(prey), p(season+effort) 736.546 1.398 0.106 0.497 6 723.809 1266.666
ψ(prey+livest), p(season) 736.770 1.622 0.095 0.445 6 724.033 1267.058
ψ(.), p(season) 736.840 1.691 0.091 0.429 4 728.495 1274.866
Jaguarundi*
ψ(season+prey), p(.) 431.988 0.000 0.407 1.000 5 421.466 421.466
ψ(season), p(.) 433.054 1.066 0.239 0.587 4 424.709 424.709
Northern tiger cat
ψ(.), p(.) 276.076 0.000 0.204 1.000 2 271.974 1283.717
ψ(livest), p(.) 277.952 1.876 0.080 0.391 3 271.747 1282.645
Ocelot
ψ(.), p(.) 306.099 0.000 0.112 1.000 2 301.997 634.195
ψ(livest), p(.) 306.298 0.199 0.101 0.905 3 300.093 630.195
ψ(hunt), p(.) 307.300 1.201 0.061 0.549 3 301.095 632.299
ψ(hunt+livest), p(.) 307.392 1.293 0.058 0.524 4 299.048 628.000
ψ(.), p(effort) 307.609 1.510 0.052 0.470 3 301.404 632.948
ψ(livest), p(effort) 308.083 1.984 0.041 0.371 4 299.738 629.450
344
Table 3 Cumulative weight (w+) and coefficient estimates (β) with their respective standard errors 345
(SE) and 95% confidence intervals (CI) of the covariates used to assess the occupancy and 346
detection patterns of mesocarnivores in a Caatinga dry forest area, northeastern Brazil. Estimates 347
were obtained using single-season single-species models. Duplicate values for beta estimates are 348
related to the modeling of the three separate sampling periods (one wet season and two dry 349
seasons). 350
Species Covariates w+ β SE (β) Lower CI (β) Upper CI (β)
Striped hog-nosed
skunk
Seasonp* 0.623 -0.378/0.362 0.325/0.308 -1.016/-0.242 0.258/0.966
Effortp 0.253 -0.003 0.005 -0.012 0.006
Seasonψ 0.456 0.874/2.754 0.756/1.704 -0.608/-0.586 2.357/6.094
Preyψ 0.373 -0.078 0.056 -0.188 0.031
Huntψ 0.711 -0.652 0.363 -1.363 0.060
Livestψ 0.526 0.365 0.193 -0.013 0.744
Crab-eating fox
Seasonp* 0.985 -0.677/0.098 0.243/0.228 -1.154/-0.349 -0.200/0.546
Effortp 0.325 -0.004 0.004 -0.012 0.004
Seasonψ 0.097 -0.228/-0.171 0.756/0.753 -1.710/-1.647 1.253/1.305
Preyψ 0.700 -0.068 0.041 -0.148 0.012
117
Huntψ 0.268 -0.023 0.051 -0.124 0.078
Livestψ 0.299 -0.010 0.012 -0.034 0.014
Jaguarundi
Seasonp 0.202 0.739/0.743 0.433/0.430 -0.109/-0.100 1.587/1.585
Seasonψ* 0.724 2.217/2.733 1.085/1.374 0.089/0.041 4.344/5.427
Preyψ 0.565 0.131 0.103 -0.072 0.333
Northern tiger cat
Seasonp 0.186 0.220/0.475 0.377/0.408 -0.506/-0.325 0.946/1.274
Effortp 0.255 -0.001 0.007 -0.014 0.012
Seasonψ 0.132 0.709/-0.442 1.793/1.499 -2.805/-3.382 4.224/2.498
Preyψ 0.262 0.026 0.112 -0.194 0.245
Huntψ 0.262 0.045 0.170 -0.288 0.378
Livestψ 0.276 0.019 0.049 -0.077 0.115
Ocelot
Seasonp 0.266 -0.618/-0.018 0.497/0.450 -1.592/-0.892 0.357/0.870
Effortp 0.314 -0.006 0.008 -0.021 0.009
Seasonψ 0.138 0.632/0.649 0.847/0.831 -1.028/-0.980 2.292/2.278
Preyψ 0.255 0.009 0.047 -0.083 0.101
Huntψ 0.351 0.066 0.088 -0.106 0.239
Livestψ 0.461 -0.022 0.018 -0.058 0.013
351
352
118
353
Fig. 2. Estimates of detection probability (A) and occupancy probability (B) of mesocarnivores 354
in a Caatinga dry forest area, northeastern Brazil, obtained through single-season single-species 355
models. The general estimates and for each season/sampling period are shown. 356
357
Co-occurrence patterns 358
We found no evidence of spatial avoidance among any of the four assessed potential 359
competitors pairwise (e.g., SIF = 1; Table 4; Table 5). In fact, our results suggest that the 360
jaguarundi used more the sites where ocelot was present (Table 4; Table 5). Regarding 361
119
detection, three subordinate mesocarnivores (striped hog-nosed skunk, crab-eating fox 362
and northern tiger cat) were more detected in locations where ocelot was present (Table 363
4; Table 5). The ocelot, in turn, was more detected on sites most used by crab-eating fox 364
and jaguarundi (Table 4; Table 5). 365
366
367
120
Table 4 Co-occurrence occupancy models used to evaluate the role of a dominant competitor (ocelot, A or a) on the habitat use and detect ion of sympatric 368
mesocarnivores (B or b) in a Caatinga dry forest area, northeastern Brazil. Only the models considered adjusted (AICc ≤ 2) are presented. 369
Model AICc ΔAICc w Likelihood K Deviance -2log(L)
Ocelot – Striped hog-nosed skunk
ψA, ψBA=ψBa, pA=rA, pB≠rBA=rBa(season) 1620.07 0.00 0.14 1.00 7 1605.08 1605.08
ψA, ψBA=ψBa, pA=rA, pB≠rBA=rBa 1620.11 0.04 0.13 0.98 5 1609.59 1609.59
ψA, ψBA=ψBa, pA≠rA, pB≠rBA=rBa(season) 1620.35 0.28 0.12 0.87 8 1603.07 1603.07
ψA, ψBA=ψBa, pA≠rA, pB≠rBA=rBa 1620.45 0.38 0.11 0.83 6 1607.71 1607.71
ψA, ψBA≠ψBa, pA=rA, pB≠rBA=rBa 1621.52 1.45 0.07 0.48 6 1608.79 1608.79
ψA, ψBA≠ψBa, pA=rA, pB≠rBA=rBa(season) 1621.74 1.67 0.06 0.43 8 1604.46 1604.46
ψA, ψBA≠ψBa, pA≠rA, pB≠rBA=rBa 1621.91 1.83 0.05 0.40 7 1606.91 1606.91
ψA, ψBA≠ψBa, pA≠rA, pB≠rBA=rBa(season) 1621.99 1.92 0.05 0.38 9 1602.37 1602.37
Ocelot – Crab-eating fox
ψA, ψBA=ψBa, pA≠rA, pB≠rBA=rBa(season) 1857.33 0.00 0.52 1.00 8 1840.04 1840.04
Ocelot – Jaguarundi
ψA, ψBA≠ψBa(season), pA≠rA, pB=rBA=rBa 1056.14 0.00 0.37 1.00 8 1038.85 1038.85
ψA, ψBA≠ψBa(season), pA≠rA, pB≠rBA≠rBa 1058.00 1.86 0.15 0.39 9 1038.38 1038.38
Ocelot – Northern tiger cat
ψA, ψBA=ψBa, pA=rA, pB≠rBA=rBa 1917.71 0.00 0.25 1.00 5 1907.18 1907.18
ψA, ψBA=ψBa, pA≠rA, pB≠rBA=rBa 1918.13 0.43 0.20 0.81 6 1905.40 1905.40
ψA, ψBA=ψBa, pA=rA, pB≠rBA≠rBa 1919.27 1.56 0.11 0.46 6 1906.53 1906.53
ψA, ψBA≠ψBa, pA≠rA, pB≠rBA=rBa 1919.40 1.69 0.11 0.43 7 1904.41 1904.41
370
371
121
Table 5 Occupancy probability (ψ), detection probability (p and r), and species interaction factor (SIF) estimated from co-occurrence occupancy models used 372
to evaluate the role of a dominant competitor (ocelot, A or a) on the habitat use and detection of sympatric mesocarnivores (B or b) in a Caatinga dry forest 373
area, northeastern Brazil. According to the ranking of the models, general values or estimates for each season/sample period are shown. 374
Species pairwise ψA(SE) ψBA(SE) ψBa(SE) pA(SE) rA(SE) pB(SE) rBA(SE) rBa(SE) SIF(SE)
Ocelot-Striped hog-nosed sunk 0.60(0.06) 0.88(0.05) 0.92(0.07) 0.19(0.06) 0.14(0.01) 0.06(0.03) 0.08(0.04)
0.06(0.03)
0.25(0.05) 0.30(0.04)
0.26(0.04)
0.29(0.03) 0.26(0.04)
0.25(0.04)
1(0.00)
Ocelot-Crab-eating fox 0.66(0.05) 0.82(0.05) 0.83(0.11) 0.24(0.05) 0.12(0.01)
0.06(0.02)
0.12(0.03) 0.09(0.03)
0.43(0.05)
0.64(0.04) 0.56(0.05)
0.43(0.04)
0.64(0.03) 0.56(0.03)
1(0.00)
Ocelot-Jaguarundi 0.55(0.06)
0.98(0.03)
0.97(0.03)
0.77(0.14)
0.74(0.29)
0.68(0.26)
0.29(0.31)
0.54(0.09) 0.14(0.02) 0.06(0.02) 0.07(0.02) 0.06(0.01)
1.16(0.16)
1.20(0.15)
1.65(0.21)
Ocelot-Northern tiger cat 0.58(0.07) 0.85(0.06) 0.88(0.09) 0.13(0.05) 0.17(0.02) 0.22(0.04) 0.37(0.04) 0.39(0.03) 1(0.00)
375
376
377
122
Discussion 378
In the absence of pumas and jaguars in neotropical environments, ocelot should 379
emerge as a secondary top predator (Prugh et al., 2009; Ritchie and Johnson, 2009). 380
However, in general, the habitat use of submissive mesocarnivores was not affected by 381
the presence of ocelot in the seasonally dry tropical forest area that we studied. The results 382
described here, together with previous work in the same area (e.g., Marinho et al., 2020), 383
indicate that this dominant mesopredator does not affect the mesocarnivore community 384
to the point of significantly changing its behavior or distribution to minimize interference 385
competition. However, the absence of control areas in the present study limits our ability 386
to extrapolate these results. 387
Species like the striped hog-nosed skunk, crab-eating fox and northern tiger cat seem 388
to use the habitat regardless of the presence of ocelot. Meanwhile, we found a suggestion 389
of spatial aggregation between ocelot and jaguarundi. In this case, it is likely that this 390
apparent aggregation is related to places with greater prey availability preferred by 391
jaguarundi, and that it is allowed by the clear temporal segregation between a diurnal and 392
a mainly nocturnal species, respectively (Dias et al., 2019b; Marinho et al., 2020). 393
Jaguarundi was the species that prey abundance was most explanatory, although we have 394
not found a clear effect of this covariate on the occurrence of mesocarnivores, which may 395
be related to the low detection of this wild cat in our study. In the Caatinga, jaguarundi 396
feeds mainly small mammals (Olmos, 1993; Dias and Bocchiglieri, 2015) and 397
synchronizes its activity with rock cavy (Marinho et al., 2020), the least represented prey 398
species in our records, probably because it inhabits mainly rocky outcrops, environments 399
that are spatially constrained and poorly recorded in our study. 400
It is well established in the literature that competitors need to differentiate one or 401
more dimensions of their ecological niche in order to reduce exploitation and interference 402
123
competition and allow coexistence (Case and Gilpin, 1974; Schoener, 1974a, 1974b). 403
Mainly in the case of mammalian carnivores, it means reducing the chances of 404
interspecific killing and intraguild predation (Polis et al., 1989; Palomares and Caro, 405
1999; Oliveira and Pereira, 2014). However, a recent study investigated patterns of co-406
occurrence among carnivores on a worldwide scale and suggested that pairs of more 407
ecologically similar species (e.g., body size, diet and activity pattern) are more likely to 408
share the same habitat than segregate or occur independently (Davis et al., 2018). 409
Spatial interactions between carnivores can vary with prey and water availability, 410
habitat type, ecosystem productivity, and the abundance of competitors (Rich et al., 2017; 411
Davis et al., 2018). It is true that competition relationships can be more severe in stressful 412
environments such as seasonally dry tropical forests (Stoner and Timm, 2011). However, 413
there is evidence that in environments with low productivity the coexistence of predators 414
may be facilitated because higher predators may not reach sufficient density to suppress 415
smaller species (Ritchie and Johnson, 2009). Studies on intraguild interaction of 416
carnivores like big cats in semiarid environments have found more evidence of spatial 417
aggregation than segregation. Astete et al. (2017) suggest that the jaguar and puma are 418
forced to live closely together to escape warmer habitats and human pressure in the 419
Caatinga and that the lower density of jaguar in this environment would facilitate this 420
coexistence. Similarly, in the dry tropical forest of northern Mexico, pumas and jaguars 421
overlap much of their habitat, which was related to the lower density and worse adaptation 422
of the jaguar to semiarid conditions compared to puma (Gutiérrez-González and López-423
González, 2017). Human persecution is another factor that decreases the density of wild 424
cats in these areas (Astete et al., 2017; Gutiérrez-González and López-González, 2017). 425
In fact, the density of ocelot in the Caatinga is one of the lowest in its entire 426
distribution (Penido et al., 2016), which may facilitate coexistence with other smaller 427
124
mesocarnivores, especially in areas where intense hunting can reduce the abundance of 428
predators. Other studies have found the absence of effect of ocelot on the distribution or 429
habitat use of mesocarnivores in the Atlantic Forest (Massara et al., 2016; Nagy-Reis et 430
al., 2017). Meanwhile, other authors have reported an apparent avoidance from places 431
where this dominant mesopredator is present by the southern tiger cat in the Atlantic 432
Forest (Cruz et al., 2018) and northern tiger cat in a large and well preserved Caatinga 433
area (Dias et al., 2019b). In most of these areas there is an opposite pattern to that 434
presented here, with northern tiger cat being less abundant than ocelot. Thus, these 435
differences in species interaction may be related to variations in the conditions and 436
resources of the study areas as well as in ocelot density throughout its wide distribution. 437
Temporal segregation is an important mechanism for promoting interspecific 438
coexistence and avoiding agonistic interactions (Carothers and Jaksić, 1984), which can 439
be used especially by mesocarnivores that overlap much of their diet and morphology to 440
co-occur (Bianchi et al., 2016; Di Bitetti et al., 2010; Massara et al., 2016). However, 441
previous studies in the same area reported that species segregate only its peak activity 442
(Marinho et al., 2020), which is when there is a greater risk of encounters (Rafiq et al., 443
2020), suggesting that other mechanisms should mediate this coexistence in a 444
complementary way. 445
The possible inclusion of larger prey in the ocelot diet in the absence of top predators 446
(Moreno et al., 2006) can lessen the pressure on smaller prey shared with other 447
mesocarnivores, and thus minimize competition for exploitation in an environment with 448
scarce resources. This mechanism has already been suggested to explain the co-449
occurrence of African golden cats Caracal aurata and sympatric mesocarnivores in the 450
absence of the leopard Panthera pardus in the forests of Uganda (Mills et al., 2019). 451
125
It is difficult to explain behaviorally how the ocelot presence increases the detection 452
of striped hog-nosed skunk, crab-eating fox, and northern tiger cat since we expected the 453
opposite to occur as a mechanism to avoid aggressive encounters. Other studies have 454
found this type of relationship between subordinate and dominant species (Cruz et al., 455
2015; Nagy-Reis et al., 2017; Astete et al., 2017; Mills et al., 2019). Associating camera 456
trapping with scat sampling data, Nagy-Reis et al. (2017) suggest that the use of scent 457
marks to avoid aggressive encounters may increase the detection of the southern tiger cat 458
(Leopardus guttulus) in the presence of ocelot. Another possible explanation would be 459
the greater availability of prey in the places where the dominant species occurs (Mills et 460
al., 2019). However, we have found no evidence on the effect of prey considered. 461
The detection probability or the photographic rate might be correlated with the local 462
species abundance (Carbone et al., 2001; Royle and Nichols, 2003). Therefore, it is 463
possible that the sites occupied by ocelot represent more favorable refuges and food 464
resources (although we have not been able to detect variations here) for other 465
mesocarnivores, with those being more abundant in these sites, and consequently more 466
detected. This goes for the opposite case, with ocelot being more detected in the presence 467
of jaguarundi and crab-eating fox, which in turn can guarantee coexistence through the 468
temporal and food niche, respectively. However, studies on species density are essential 469
to advance this issue, especially considering that the theory suggests that ocelot 470
negatively affect the abundance of small neotropical cats (Oliveira et al., 2010). 471
Habitat preferences and prey availability may be more important than interspecific 472
interactions in the mesocarnivore distribution and spatial structure of this guild (Cruz et 473
al., 2015; Nagy-Reis et al., 2017; Santos et al., 2019). A study on a regional scale 474
suggested that the occurrence of the northern tiger cat in the Caatinga may be higher in 475
more forested and far from human settlements environments (Marinho et al., 2018b). In 476
126
our case, mesocarnivores used much or practically the entire studied landscape, with little 477
variation between seasons and no effect of the investigated covariates. It is likely that on 478
larger scales that better cover the heterogeneity of Caatinga environments, clearer patterns 479
of habitat use will be found. On the other hand, mesocarnivores are more tolerant of 480
human presence than apex predators (Roemer et al., 2009), and the absence of 481
anthropogenic activities’ effects (hunters and cattle) on the use of the species' habitat 482
suggests some level of tolerance. However, we do not know how much hunting of these 483
species and the degradation of habitats due to grazing, so common in this semiarid region 484
(Alves et al., 2016; Melo, 2015), affect other populations and behavioral parameters. 485
The spatial structure of the studied mesocarnivore guild seems to have remained 486
stable over time. The same is not true for species detection. Semiarid and highly seasonal 487
environments put pressure on mammals to develop physiological, behavioral, and 488
ecological strategies to deal with the scarcity and fluctuation of resources (Stoner and 489
Timm, 2011). However, we found no evidence of change in habitat use for most species 490
between seasons. The only suggestion of a change in habitat use was for jaguarundi, 491
which seems to have shrunken its distribution in the drier season (2018). However, the 492
uncertainty linked to these estimates due to the low detection of this species decreases its 493
reliability. Regarding detection, striped hog-nosed skunk and mainly crab-eating fox were 494
more detected in the dry seasons. Other studies in the Caatinga have already shown that 495
the activity of these species increases during the dry season (Dias and Bocchiglieri, 2015; 496
Dias, 2017), when the species need to move more to forage and obtain food, consequently 497
increasing their home range (Stoner and Timm, 2011). Thus, future sampling of these 498
species in the Caatinga focused on the dry period can optimize data collection and save 499
financial resources. 500
127
Our results suggest that most submissive mesocarnivores occur regardless of the 501
ocelot’s presence, while the jaguarundi uses more habitats shared with this dominant 502
mesopredator, probably due to the strong temporal segregation between these two 503
species. As the temporal segregation within the studied guild is low and generally 504
restricted to the species’ activity peaks (Marinho et al., 2020), the absence of spatial 505
avoidance among mesocarnivores may be related to a probable low density of ocelot in 506
the area, reducing the effects of intraguild suppression. However, it is necessary to 507
investigate how much the species' diets overlap in order to understand the 508
mesocarnivores’ coexistence mechanisms in this seasonal environment and with scarce 509
resources, where top predators are increasingly rare due to intense human interference. 510
Our results allow us to better understand how mesocarnivores interact spatially in 511
semiarid environments and where top predators have become extinct or are in decline due 512
to chronic human disturbance. 513
514
Acknowledgments 515
We are very grateful to João B. de Lima (seu João) and Joana Darc for the field 516
assistance and hospitality, and to Eugenia C. Schmidt, Juan C. V. Mena, Felipe Marinho, 517
Raul dos Santos, Maria L. Falcão, V. Paixão, and T. Oliveira for field or analysis 518
assistance. We would like to thank Claudia Campos, Mauro Pichorim, Fabiana Rocha, 519
and Rodrigo Massara for the valuable suggestions that helped to improve the manuscript. 520
This study was partially supported by Restaurante Camarões and The Mohamed bin 521
Zayed Species Conservation Fund (#172516360). EMV (#308040/2017-1) and CRF 522
(#305304/2013-5; #306812/2017-7) were supported by Conselho Nacional de 523
Desenvolvimento Científico e Tecnológico (CNPq) and PHM was supported by 524
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES; 525
128
financing code 001), and by Santander Universities (Santander Mundi scholarship) during 526
internship in Portugal. 527
528
References 529
Alves RRRN, Feijó A, Barboza RRD, Souto WMD, Fernandes-Ferreira H, Cordeiro-530
Estrela P, Langguth A (2016) Game mammals of the Caatinga biome. Ethnobio 531
Conserv 5:1-51. https://doi.org/10.15451/ec2016-7-5.5-1-51 532
Astete S, Marinho-Filho J, Kajin M, Penido G, Zimbres B, Sollmann R, Jácomo ATA, 533
Tôrres NM, Silveira L (2017) Forced neighbours: Coexistence between jaguars and 534
pumas in a harsh environment. J Arid Environ 146:27-34. 535
https://doi.org/10.1016/j.jaridenv.2017.07.005 536
Azevedo FC, Lemos FG, Almeida LB, Campos CB, Beisiegel BM, Paula RC, Crawshaw 537
PGJr, Ferraz KMMB, Oliveira TG (2013). Avaliação do risco de extincão da onca-538
parda Puma concolor (Linnaeus, 1771) no Brasil. Biodiv Bras 3:107-121. 539
Beschta RL, William JR (2009) Large predators and trophic cascades in terrestrial 540
ecosystems of the western United States. Biol Conserv 142:2401-2414. 541
https://doi.org/10.1016/j.biocon.2009.06.015 542
Bianchi RdC, Mendes SL, Marco Júnior PD (2010) Food habits of the ocelot, Leopardus 543
pardalis, in two areas in southeast Brazil. Stud Neotrop Fauna E 45:111-119. 544
https://doi.org/10.1080/01650521.2010.514791 545
Bianchi RdC, Olifiers N, Gompper ME, Mourão G (2016) Niche Partitioning among 546
mesocarnivores in a Brazilian wetland. Plos One 11:e0162893. 547
https://doi.org/10.1371/journal.pone.0162893 548
129
Bolze GJ, Tirelli FP, Horn PE, Almeida LR, Figueiró HV, Kessler A, Eizirik E, Pereira 549
MJR (2019) Registro de interação antagonista entre Leopardus pardalis e Cerdocyon 550
thous no limite sul da Mata Atlântica. Bol Soc Bras Mastozool 85: 110-113. 551
Bu H, Wang F, McShea WJ, Lu Z, Wang D, Li S (2016) Spatial Co-Occurrence and 552
Activity Patterns of Mesocarnivores in the Temperate Forests of Southwest China. 553
Plos One 11:e0164271. https://doi.org/10.1371/journal.pone.0164271 554
Burnham KP, Anderson D (2002) Model selection and multi-model inference: A practical 555
information–theoretic approach. Springer Science, New York 556
Carbone C, Christie S, Conforti K et al (2001) The use of photographic rates to estimate 557
densities of tigers and other cryptic mammals. Anim Conserv 4:75-79. 558
https://doi.org/10.1017/S1367943001001081 559
Carothers JH, Jaksić FM (1984) Time as a niche difference: The role of interference 560
competition. Oikos 42:403-406. http://www.jstor.org/stable/3544413 561
Case TJ, Gilpin ME (1974) Interference competition and niche theory. P Natl Acad Sci 562
USA 71:3073-3077. https://doi.org/10.1073/pnas.71.8.3073 563
Cooch EG, White GC (2017) Program MARK: A gentle introduction, seventeenth ed. 564
Colorado State University, Fort Collins, 565
Crooks KR, Soulé ME (1999) Mesopredator release and avifaunal extinctions in a 566
fragmented system. Nature 400:563-566. https://doi.org/10.1038/23028 567
Cruz J, Sarmento P, White PCL (2015) Influence of exotic forest plantations on 568
occupancy and co-occurrence patterns in a Mediterranean carnivore guild. J Mammal 569
96:854-865. https://doi.org/10.1093/jmammal/gyv109 570
Cruz P, Iezzi ME, De Angelo C, Varela D, Di Bitetti MS, Paviolo A (2018) Effects of 571
human impacts on habitat use, activity patterns and ecological relationships among 572
130
medium and small felids of the Atlantic Forest. Plos One 13:e0200806. 573
https://doi.org/10.1371/journal.pone.0200806 574
Davis CL, Rich LN, Farris ZJ et al (2010) Ecological correlates of the spatial co-575
occurrence of sympatric mammalian carnivores worldwide. Ecol Lett 21:1401-1412. 576
https://doi.org/10.1111/ele.13124 577
Di Bitetti MS, De Angelo CD, Di Blanco YE, Paviolo A (2010) Niche partitioning and 578
species coexistence in a Neotropical felid assemblage. Acta Oecol 36:403e412. 579
https://doi.org/10.1016/j.actao.2010.04.001 580
Diamond JM (1978) Niche Shifts and the Rediscovery of Interspecific Competition. Am 581
Sci 66:322-331. 582
Dias DM (2017) Spatiotemporal ecology of the striped hog-nosed skunk Conepatus 583
semistriatus (Carnivora, Mephitidae) in a seasonally dry forest of northeastern Brazil. 584
Anim Biol 2:119-131. https://doi.org/10.1163/15707563-00002525 585
Dias DM, Bocchiglieri A (2015) Trophic and spatio-temporal niche of the crabeating fox, 586
Cerdocyon thous (Linnaeus, 1766) (Carnivora: Canidae), in a remnant of the Caatinga 587
in northeastern Brazil. Mammalia 80:281-291. https://doi.org/10.1515/mammalia-588
2014-0108 589
Dias DM, Massara RL, Campos CB, Rodrigues FHG (2019a) Feline predator-prey 590
relationships in a semi-arid biome in Brazil. J Zool 307:282-291. 591
https://doi.org/10.1111/jzo.12647 592
Dias DM, Massara RL, Campos CB, Rodrigues FHG (2019b) Human activities influence 593
the occupancy probability of mammalian carnivores in the Brazilian Caatinga. 594
Biotropica 51:253-265. https://doi.org/10.1111/btp.12628 595
Doherty PF, White GC, Burnham KP (2012) Comparison of model building and selection 596
strategies. J Ornithol 152:S317-S323. https://doi.org/10.1007/s10336-010-0598-5 597
131
Donadio E, Buskirk SW (2006) Diet, morphology, and interspecific killing in carnivora. 598
Am Nat 167:524-36. https://doi.org/10.1086/501033 599
Ferreguetti AC, Rocha CFD, Bergallo HG (2019) Poaching in non–volant mammals in 600
the Neotropical region: the importance of a metric to assess its impacts. Anim Biodiv 601
Conserv 42.2:203-211. https://doi.org/10.32800/abc.2019.42.0203 602
Fonseca CR, Antongiovanni M, Matsumoto M, Bernard E, Venticinque EM (2017) 603
Conservation opportunities in the Caatinga. In: Silva JC, Leal I, Tabarelli M (eds) 604
Caatinga: the largest tropical dry forest region in South America. Springer, Cham, pp 605
429–443 606
Goulart FVB, Cáceres NC, Graipel ME, Tortato MA, Ghizoni Jr IR, Oliveiras-Santos 607
LGR (2009) Habitat selection by large mammals in a southern Brazilian Atlantic 608
Forest. Mamm biol 74:182–190. https://doi.org/10.1016/j.mambio.2009.02.006 609
Gutiérrez-González CE, López-González CA (2017). Jaguar interactions with pumas and 610
prey at the northern edge of jaguars’ range. PeerJ 5:e2886. 611
https://doi.org/10.7717/peerj.2886 612
Harrison DJ, Bissonette JA, Sherburne JA (1989) Spatial Relationships between Coyotes 613
and Red Foxes in Eastern Maine. J Wildlife Manage 53: 181-185. 614
https://doi.org/10.2307/3801327 615
Hines JE (2006) PRESENCE - Software to estimate patch occupancy and related 616
parameters. USGS-PWRC. www.mbr-pwrc.usgs.gov/software/presence.html 617
Accessed 20 April 2018 618
Hutchinson GE (1959) Homage to Santa Rosalia or why are there so many kinds of 619
animals? Am Nat 93:145-159. https://www.jstor.org/stable/2458768 620
IUCN (2020) The IUCN Red List of Threatened Species, version 2020-1. 621
https://www.iucnredlist.org. Accessed 19 March 2020 622
132
Jiménez J, Nuñez-Arjona JC, Mougeot F et al (2019) Restoring apex predators can reduce 623
mesopredator abundances. Biol Conserv 238:108234. 624
https://doi.org/10.1016/j.biocon.2019.108234 625
MacArthur R, Levins R (1964) Competition, habitat selection, and character 626
displacement in a patchy environment. P Natl Acad Sci USA 51:1207–1210. 627
https://doi.org/10.1073/pnas.51.6.1207 628
MacKenzie DI, Bailey LL (2004) Assessing the fit of site-occupancy models. J Agr Biol 629
Envir St 9:300-318. https://doi.org/10.1198/108571104X3361 630
MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm CA (2002) 631
Estimating site occupancy rates when detection probabilities are less than one. Ecology 632
83:2248-2255. https://doi.org/10.1890/0012-633
9658(2002)083[2248:ESORWD]2.0.CO;2 634
MacKenzie DI, Nichols JD, Royle JA, Pollock KH, Bailey LL, Hines JE (2006) 635
Occupancy estimation and modeling: inferring patterns and dynamics of species 636
occurrence. Elsevier/Academic Press, Burlington 637
Marinho FP, Mazzochini GG, Manhães AP, Weisser WW, Ganade G (2016) Effects of 638
past and present land use on vegetation cover and regeneration in a tropical dryland 639
forest. J Arid Environ 132:26–33. http://dx.doi.org/10.1016/j.jaridenv.2016.04.006 640
Marinho PH, Bezerra D, Antongiovanni M, Fonseca CR, Venticinque EM (2018a) 641
Mamíferos de médio e grande porte da Caatinga do Rio Grande do Norte, nordeste do 642
Brasil. Mastozool Neotrop 25:345-362. 643
https://doi.org/10.31687/saremMN.18.25.2.0.15 644
Marinho PH, Fonseca CR, Sarmento P, Fonseca C, Venticinque EM (2020) Temporal 645
niche overlap among mesocarnivores in a Caatinga dry forest. Eur J Wildl Res 66:34. 646
https://doi.org/10.1007/s10344-020-1371-6 647
133
Massara R, Paschoal AMO, Bailey L, Doherty PF Jr, Hirsch A, Chiarello AG (2018) 648
Factors influencing ocelot occupancy in Brazilian Atlantic Forest reserves. Biotropica 649
50:125–34. https://doi.org/10.1111/btp.12481 650
Massara RL, Paschoal AMO, Bailey LL, Doherty PFJr, Chiarello AG (2016) Ecological 651
interactions between ocelots and sympatric mesocarnivores in protected areas of the 652
Atlantic Forest, southeastern Brazil. J Mammal 97:1634–1644. 653
https://doi.org/10.1093/jmammal/gyw129 654
Massara RL, Paschoal AMO, Doherty PF Jr, Hirsch A, Chiarello AG (2015) Ocelot 655
population status in protected Brazilian Atlantic Forest. Plos One 10:e0141333. 656
https://doi.org/10.1371/journal.pone.0141333 657
Melo FPL (2017) The socio-ecology of the Caatinga: understanding how natural resource 658
use shapes an ecosystem. In: Silva JMC, Leal IR, Tabarelli M (eds) Caatinga: The 659
largest tropical dry-forest region in South America. Springer, Cham, pp 369-382 660
Miller B, Dugelby B, Foreman D, Martinez del Rio C, Noss R, Phillips M, Reading R, 661
Soulé ME, Terborgh J, Willcox L (2001) The importance of large carnivores to healthy 662
ecosystems. Endanger Species Updat 18:202-210. 663
Mills DR, San EDL, Robinson H, Isoke S, Slotow R, Hunter L (2019) Competition and 664
specialization in an African forest carnivore community. Ecol Evol 9:10092-10108. 665
https://doi.org/10.1002/ece3.5391 666
MMA (2014) Lista Nacional Oficial de Espécies da Fauna Ameaçadas de 667
Extinção. Ministério do Meio Ambiente. 668
https://www.icmbio.gov.br/portal/images/stories/biodiversidade/fauna-669
brasileira/portarias/PORTARIA_N_445_DE_17_DE_DEZEMBRO_DE_2014670
.pdf. Accessed 16 June 2019 671
134
Moreno RS, Kays RW, Samudio R (2006) Competitive release in diets of ocelot 672
(Leopardus pardalis) and puma (Puma concolor) after jaguar (Panthera onca) decline. 673
J Mammal 87:808–816. https://doi.org/10.1644/05-mamm-a-360r2.1 674
Nagy-Reis MB, Nichols JD, Chiarello AG, Ribeiro MC, Setz EZF (2017) Landscape use 675
and co-occurrence patterns of Neotropical spotted cats. Plos One 12: e0168441. 676
https://doi.org/10.1371/journal.pone.0168441 677
Oliveira TG, Pereira JA (2014) Intraguild predation and interspecific killing as structuring 678
forces of carnivoran communities in South America. J Mamm Evol 21:427–436. 679
https://doi.org/10.1007/s10914-013-9251-4 680
Oliveira TG, Tortato MA, Silveira L et al (2010) Ocelot ecology and its effect on the 681
small-felid guild in the lowland neotropics. In: Macdonald DW, Loveridge A (eds) 682
Biology and conservation of wild felids. Oxford University Press, Oxford, pp 563-584 683
Oliveira-Santos LGR, Graipel ME, Tortato MA, Zucco CA, Cáceres NC, Goulart FVB 684
(2012) Abundance changes and activity flexibility of the oncilla, Leopardus tigrinus 685
(Carnivora: Felidae), appear to reflect avoidance of conflict. Zoologia 29:115–120. 686
https://doi.org/10.1590/S1984-46702012000200003 687
Olmos F (1993) Notes on the food habitats of Brazilian “Caatinga” carnivores. Mammalia 688
57:126–130. 689
Paglia AP, Fonseca GAB, Rylands AB et al (2012) Annotated Checklist of Brazilian 690
Mammals. 2nd Edition. Occasional Papers in Conservation Biology 6:1–76. 691
Palomares F, Caro TM (1999) Interspecific killing among mammalian carnivores. The 692
Am Nat 153:492-508. https://doi.org/10.1086/303189 693
Paula RC, Campos CB, Oliveira TG (2012) Red list assessment for the jaguar in the 694
Caatinga Biome. Cat News 7:19–24. 695
135
Penido G, Astete S, Furtado MM, Jácomo ATA, Sollmann R, Torres N, Silveira L, 696
Marinho-Filho J (2016) Density of Ocelots in a semiarid environment in northeastern 697
Brazil. Biota Neotropica 16:e20160168. https://doi.org/10.1590/1676-0611-bn-2016-698
0168 699
Peters FB, Mazim FD, Favarini MO, Soares JBG, Oliveira TG. (2016) Caça preventiva 700
ou retaliativa de felinos por humanos no extremo sul do Brasil. In: Castanõ-Uribe C, 701
Lasso CA, Hoogesteijn R, Diaz-Pulido A, Payán E (eds) Conflictos entre felinos y 702
Humanos em América Latina. Instituto de Investigación de Recursos Biológicos 703
Alexander von Humboldt, Bogotá, pp. 311-325. 704
Pichorim M, França Câmara TP, Oliveira-Júnior TM, Valdenor-de-Oliveira D, Galvão-705
do-Nascimento EP, Mobley JÁ (2014) A population of Blue-winged Macaw Primolius 706
maracana in northeastern Brazil: recommendations for a local conservation action 707
plan. Trop Conserv Sci 7:488–507. https://doi.org/10.1177/194008291400700309 708
Polis GA, Myers CA, Holt RD (1989) The ecology and evolution of intraguild predation: 709
potential competitors that eat each other. Annu Rev Ecol Syst 20:297-330. 710
https://doi.org/10.1146/annurev.es.20.110189.001501 711
Prugh LR, Stoner CJ, Epps CW, Bean WT, Ripple WJ, Laliberte AS, Bashares JS (2009) 712
The rise of the mesopredator. Bioscience 59:779-791. 713
https://doi.org/10.1525/bio.2009.59.9.9 714
Pudyatmoko S (2017) Free-ranging livestock influence species richness, occupancy, and 715
daily behaviour of wild mammalian species in Baluran National Park, Indonesia. 716
Mamm Biol 86:33–41. http://dx.doi.org/10.1016/j.mambio.2017.04.001 717
Rafiq K, Jordan NR, Wilson AM, McNutt JW, Hayward MW, Meloro C, Wich SA, 718
Golabek KA (2019) Spatio‐temporal factors impacting encounter occurrences between 719
136
leopards and other large African predators. J Zool 310:191-720
200. https://doi.org/10.1111/jzo.12746 721
Rich LN, Miller DAW, Robinson HS, McNutt JW, Kelly MJ (2017) Carnivore 722
distributions in Botswana are shaped by resource availability and intraguild species. J 723
Zool 303:90–98. https://doi.org/10.1111/jzo.12470 724
Richmond OMW, Hines JE, Beissinger SR (2010) Two-species occupancy models: a new 725
parameterization applied to co-occurrence of secretive rails. Ecol App 20:2036–2046. 726
https://doi.org/10.1890/09-0470.1 727
Ripple WJ, Estes JA, Beschta RL, Wilmers CC, Ritchie EG, Hebblewhite M, Berger J, 728
Elmhagen B, Letnic M, Nelson MP, Schmitz OJ, Smith DW, Wallach AD, Wirsing 729
AJ (2014) Status and ecological effects of the world’s largest carnivores status and 730
ecological effects of the world’s largest carnivores. Science 343:1241484. 731
https://doi.org/10.1126/science.1241484 732
Ritchie EG, Johnson CN (2009) Predator interactions, mesopredator release and 733
biodiversity conservation. Ecol Lett 12:982–998. https://doi.org/10.1111/j.1461-734
0248.2009.01347.x 735
Roemer GW, Gompper ME, Valkenburgh BV (2009) The ecological role of the 736
mammalian mesocarnivore. BioScience 59:165-173. 737
https://doi.org/10.1525/bio.2009.59.2.9 738
Royle A, Nichols JD (2003) Estimating abundance from repeated presence absence data 739
or point counts. Ecology 84:777-790. https://doi.org/10.1890/0012-740
9658(2003)084[0777:EAFRPA]2.0.CO;2 741
Santos F, Carbone C, Wearn OR, Rowcliffe JM, Espinosa S, Lima MGM, Ahumada JA, 742
Gonçalves ALS, Trevelin LC, Alvarez-Loayza P, Spironello R, Jansen PA, Juen L, 743
Peres CA. (2019) Prey availability and temporal partitioning modulate felid 744
137
coexistence in Neotropical forests. Plos One 14:e0213671. 745
https://doi.org/10.1371/journal.pone.0213671 746
Schmitt RJ, Holbrook SJ (1986) Seasonally fluctuating resources and temporal variability 747
of interspecific competition. Oecologia 69:1-11. https://doi.org/10.1007/BF00399030 748
Schoener TW (1974a) The compression hypothesis and temporal resource partitioning. 749
Proc Natl Acad Sci USA 71:4169-4172. https://doi.org/10.1073/pnas.71.10.4169 750
Schoener TW (1974b) Resource partitioning in ecological communities. Scince 185: 27-751
39. https://doi.org/10.1126/science.185.4145.27 752
Stoner K, Timm RM (2011) Seasonally dry tropical forest mammals: adaptations and 753
seasonal patterns. In: Dirzo R, Young HS, Mooney HA, Ceballos G (eds) Seasonally 754
dry tropical forests: Ecology and conservation. Island Press, Washington, pp 85-105 755
Szilagyi G (2007) Diagnóstico ambiental do processo de desertificação do município de 756
Lajes/RN. Dissertation, Universidade Federal do Rio Grande do Norte 757
Terborgh J, Estes JA, Paquet P, Ralls K, Boyd-Heigher D, Miller BJ, Noss RF (1999) 758
The role of top carnivores in regulating terrestrial ecosystems. In: Terborgh J, Soulé 759
M (eds) Continental Conservation: Scientific foundations of regional reserve 760
networks. Island Press, Washington, pp 39-64 761
Tomasella J, Silva RMPV, Barbosa AA, Rodriguez DA, Oliveira MS, Sestini MF (2018) 762
Desertification trends in the Northeast of Brazil over the period 2000–2016. Int J Appl 763
Earth Obs 73:197-206. https://doi.org/10.1016/j.jag.2018.06.012 764
Vanak AT, Fortin D, Thaker M, Ogden M, Owen C, Greatwood S, Slotow R (2013) 765
Moving to stay in place: behavioral mechanisms for coexistence of African large 766
carnivores. Ecology 94:2619–2631. https://doi.org/10.1890/13-0217.1 767
138
Wang Y, Allen ML, Wilmers CC (2015) Mesopredator spatial and temporal responses to 768
large predators and human development in the Santa Cruz Mountains of California. 769
Biol Conserv 190:23–33. http://dx.doi.org/10.1016/j.biocon.2015.05.007 770
White GC, Burnham KP 1999 Program mark: survival estimation from populations of 771
marked animals. Bird Study 46:120-139. https://doi.org/10.1080/00063659909477239 772
Woodroffe R (2000) Predators and people: using human densities to interpret declines of 773
large carnivores. Anim Conserv 3:165-173. https://doi.org/10.1111/j.1469-774
1795.2000.tb00241.x 775
776
Supplementary Material 777
778
Table S1. Correlation coefficients and range of covariates used to investigate patterns of habitat 779
use and detection of mesocarnivores in an area of dry tropical forest in northeastern Brazil. 780
Period Pearson's correlation coefficients Covariate range
Covarite Effort Prey Livestock Hunter Mean Min. Max.
Wet - 2017
Effort 1.000 -0.298 0.276 -0.121 86.490 23.000 110.000
Prey 1.000 -0.010 -0.113 4.225 0.000 48.000
Livestock 1.000 0.114 19.547 0.000 123.809
Hunter 1.000 1.102 0.000 6.780
Dry - 2017
Effort 1.000 -0.247 0.228 -0.120 96.950 0.000 110.000
Prey 1.000 -0.022 -0.014 3.229 0.000 71.739
Livestock 1.000 -0.035 17.433 0.000 114.414
Hunter 1.000 2.597 0.000 29.333
Dry - 2018
Effort 1.000 0.076 0.144 -0.147 78.910 0.000 112.000
Prey 1.000 0.029 0.123 5.654 0.000 38.462
Livestock 1.000 0.128 2.529 0.000 22.727
Hunter 1.000 2.426 0.000 40.385
781
782
783
139
Table S2 Estimates of probability of detection (A) and occupancy (B) of mesocarnivores in an 784
area of dry tropical forest in northeastern Brazil obtained through single-season single-species 785
models. The general estimates and for each season/sampling period are shown. Estimates in bold 786
(varying by season or general) were more adjusted according to the ranking of the models. 787
Species Parameter Estimate SE Lower CI Upper CI
Striped hog-nosed
skunk
p(general) 0.217 0.020 0.179 0.260
p(wet2017) 0.175 0.042 0.108 0.272
p(dry2017) 0.240 0.034 0.181 0.312
p(dry2018) 0.205 0.038 0.140 0.291
ψ(general) 0.992 0.018 0.604 1.000
ψ(wet2017) 0.858 0.135 0.406 0.982
ψ(dry2017) 0.916 0.105 0.429 0.994
ψ (dry2018) 0.778 0.208 0.249 0.974
Crab-eating fox
p(general) 0.449 0.023 0.404 0.494
p(wet2017) 0.336 0.042 0.260 0.422
p(dry2017) 0.518 0.039 0.442 0.592
p(dry2018) 0.494 0.043 0.410 0.578
ψ(general) 0.724 0.060 0.594 0.825
ψ(wet2017) 0.721 0.066 0.576 0.831
ψ(dry2017) 0.723 0.064 0.582 0.831
ψ(dry2018) 0.724 0.065 0.581 0.832
Jaguarundi
p(general) 0.073 0.016 0.047 0.110
p(wet2017) 0.073 0.018 0.045 0.116
p(dry2017) 0.072 0.018 0.044 0.116
p(dry2018) 0.069 0.023 0.036 0.129
ψ(general) 0.672 0.140 0.372 0.877
ψ(wet2017) 0.761 0.176 0.322 0.955
ψ(dry2017) 0.815 0.190 0.272 0.981
ψ(dry2018) 0.421 0.231 0.102 0.823
Northern tiger cat
p(general) 0.314 0.034 0.251 0.385
p(wet2017) 0.306 0.044 0.228 0.397
p(dry2017) 0.315 0.039 0.243 0.396
p(dry2018) 0.326 0.050 0.237 0.430
ψ(general) 0.850 0.081 0.618 0.952
ψ(wet2017) 0.854 0.097 0.561 0.964
ψ(dry2017) 0.862 0.089 0.591 0.964
ψ(dry2018) 0.844 0.100 0.549 0.960
Ocelot
p(general) 0.181 0.030 0.130 0.247
p(wet2017) 0.167 0.042 0.100 0.267
p(dry2017) 0.192 0.040 0.126 0.282
p(dry2018) 0.190 0.042 0.120 0.287
ψ(general) 0.509 0.083 0.352 0.664
ψ(wet2017) 0.521 0.112 0.312 0.724
140
ψ(dry2017) 0.513 0.095 0.334 0.689
ψ(dry2018) 0.497 0.102 0.308 0.687
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
141
Capítulo 4 811
812
MULTI-SPECIES OCCUPANCY MODELLING REVEALS MAMMALS’ 813
PREFERENCE FOR FORESTED HABITATS IN AN OVERGRAZED 814
SEMIARID LANDSCAPE 815
816
817
818
819
820
821
Este capítulo foi escrito de acordo com as regras da revista Journal of Arid Environments 822
(https://www.journals.elsevier.com/journal-of-arid-environments). 823
142
Multi-species occupancy modelling reveals mammals’ preference for forested 824
habitats in an overgrazed semiarid landscape 825
826
Paulo Henrique Marinhoa,*, Camile Lugarinib, Gabriel Penidoc, Carlos Roberto 827
Fonsecaa, Eduardo Martins Venticinquea 828
829
a Departamento de Ecologia, Cendro de Biociências, Universidade Federal do Rio 830
Grande do Norte, Campus Universitário, Lagoa Nova, Natal, 59078-970, Rio Grande do 831
Norte, Brazil 832
b Centro Nacional de Pesquisa e Conservação das Aves Silvestres (CEMAVE), 833
ICMBio, BR 230 - KM 10 Floresta Nacional da Restinga de Cabedelo, Cabedelo, 834
58108-012, Paraíba, Brazil 835
c Laboratório de Ecomorfologia e Macroevolução LEMA, Departamento de Ecologia, 836
Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves 9500, Campus do 837
Vale, Porto Alegre, 91501-970 UFGRS, Rio Grnde do Sul, Brazil 838
839
*Corresponding author. 840
E-mail address: [email protected] (P. H. Marinho) 841
842
843
844
845
846
143
Highlights 847
Generalist and threatened terrestrial mammal species can benefit from forested 848
environments in disturbed semiarid landscapes under intense drought. 849
More productive and structurally complex semiarid environments can benefit the 850
persistence of mammals by providing better resources and refuges. 851
The expansion of protected areas in perturbed semiarid landscapes should consider 852
forested habitats, as well as the recovery of degraded vegetation. 853
854
Abstract 855
Arid and semiarid environments challenge species to adapt to extreme conditions. The 856
association between low productivity and high rates of human occupation has led to high 857
levels of degradation in these regions. In this scenario, medium to large-sized mammals 858
are especially vulnerable. We investigated the occupancy patterns of medium to large-859
sized mammals in an overgrazed landscape in Brazilian Caatinga. We collected camera 860
trapping data at the end of an extreme drought period and used Bayesian hierarchical 861
multi-species occupancy models to investigate the relative effects of anthropogenic 862
disturbance and environmental predictors on species-specific and community occupancy. 863
We obtained 566 records from 12 medium to large-sized wild mammal species. Among 864
the environmental and anthropogenic predictors evaluated, forest cover influenced 865
significantly and positively the occupancy rate of Cerdocyon thous, Dasypus 866
novemcintus, Leopardus tigrinus, Mazama gouazoubira, and Herpailurus yagouaroundi 867
as well as the community level occupancy. C. thous and Euphractus sexcinctus were more 868
detected on wider trails, which affected the community level as well. More forested 869
habitats can provide better resources and shelters, being an important predictor of 870
mammal’s occurrence in a disturbed semiarid landscape with scarce resources, benefiting 871
144
both generalist and endangered species. On the other hand, anthropogenic factors did not 872
affect mammals’ occurrence, suggesting some level of adaptation, especially for the most 873
recorded species, since we obtained rare records of the most sensitive ones. Our results 874
must serve as a baseline for future mammals’ population monitoring in semiarid regions, 875
as well as for the expansion of protected areas and degraded vegetation restoration in 876
Caatinga. 877
Keywords: Bayesian occupancy models; Caatinga; Dry tropical forest; Extreme drought; 878
Terrestrial mammals. 879
880
Introduction 881
Drylands cover approximately 41% of Earth's surface (Niemeijer et al., 2005) and 882
are among the most threatened and least known environments in the world (Sunderland 883
et al., 2015; Banda et al., 2016). The association between low productivity and high rates 884
of human occupation has led to high levels of degradation (Niemeijer et al., 2005; Davies 885
et al., 2012). Extensive livestock farming, for example, is one of the most widespread 886
human activities in arid and semiarid regions, where is 50% of all cattle in the world 887
(Niemeijer et al., 2005). In this scenario, where natural resources play a key role in the 888
local economy and combating poverty (Davies et al., 2012), the growing impacts of 889
human activities on wildlife and which semiarid landscapes’ attributes guarantee the 890
persistence of the species still poorly understood (Drouilly et al., 2018). 891
Arid and semiarid environments challenge mammals to adapt to extreme conditions 892
(Stoner and Timm, 2010; Astete et al., 2016), making it harder for animals to resist the 893
intense human pressures in these places. Many mammals, for example, persist in semiarid 894
conditions under low population densities and with increased nocturnal activity (Bennie 895
et al., 2014; Jędrzejewski et al., 2017; Santini et al., 2018), usually occupying 896
145
environments with forest structure to take refuge and obtain better resources (Stoner and 897
Timm, 2010). In this scenario, medium to large-sized mammals (> 1 kg) are especially 898
vulnerable due to their generally higher food and habitat requirements (Stoner and Timm, 899
2010; Astete et al., 2017), besides they are heavily hunted (Benítez-López et al., 2017). 900
At the regional level, for instance, jaguar (Panthera onca) populations are less resistant 901
to human persecution in the driest regions of Venezuela, since poor environmental quality 902
makes the recovery of impacted populations difficult (Jędrzejewski et al., 2017). 903
The seasonally dry tropical forest of northeastern Brazil (Caatinga) is one of the most 904
diverse and populous semiarid environments in the world, harboring 28 million 905
inhabitants (Silva et al., 2017). In this ecoregion, large and iconic Neotropical mammals 906
such as top predators (Panthera onca and Puma concolor) and ungulates (Tayassu pecari 907
and Pecari tajacu) still persist (Carmignotto and Astúa, 2017), however, they restricted 908
to the small number of protected and less impacted sites of this ecosystem. Indeed, less 909
than 2% of the Caatinga’s territory is fully protected (Fonseca et al., 2017), and the 910
remaining original vegetation that covers 50% of the region (Antongiovanni et al., 2018) 911
is under high levels of chronic anthropogenic disturbance (Ribeiro et al., 2015; 912
Antongiovanni et al., in press) such as the overgrazing promoted by 19 million goats and 913
sheep present in this semiarid region (Silva et al., 2017). Unlike strictly forest 914
environments, the Caatinga is covered by a mosaic of shrub formations and patches of 915
dry forests and woody vegetation that are distributed according to environmental and 916
anthropic factors (Velloso et al., 2002; Silva et al., 2017), since chronic anthropogenic 917
disturbance has converted more accessible forested environments into open shrub 918
formations (Ribeiro et al., 2015). In addition, between 2012 and 2017, the region 919
experienced the most intense drought in the last decades, which tends to become more 920
frequent with climate change (Brito et al., 2018). 921
146
It is necessary, however, to better understand how medium to large-sized mammals 922
persist in semiarid landscapes under intense anthropogenic disturbance and long periods 923
of extreme resource scarcity (Drouilly et al., 2018), especially bearing in mind that both 924
factors tend to become more common. This knowledge is essential for the conservation 925
of these environments considering the ecological roles played by mammals in ecosystem 926
functioning and forest regeneration (Magioli et al., 2020), as well as being urgent for the 927
compatibility of human activities that rely on natural resources with biodiversity 928
conservation. 929
In this study, we investigated the occupancy of medium to large-sized mammals in a 930
Brazilian Caatinga landscape marked by intense and historical anthropogenic 931
disturbances, mainly through extensive cattle raising. Specifically, with data collected 932
during the sixth year of a long drought period, we investigated the relative effects of 933
attributes related to anthropogenic disturbance and environmental predictors on mammal 934
occupancy using Bayesian hierarchical multi-species modelling. 935
Site occupancy is an ecological parameter widely used for investigating habitat use 936
and monitoring the status of animal populations through hierarchical models (Ferreira et 937
al., 2017; Drouilly et al., 2018). This approach does not require specimen 938
individualization, while it considers imperfect detection and allows modeling of predictor 939
covariate effects from multiple sampling (Mackenzie et al., 2002). These hierarchical 940
models have been extended to a multi-species approach that allows for further 941
understanding of community occurrence and richness patterns including those rare and 942
low-deteced species, thereby improving species-specific estimates from whole 943
community data (Dorazio and Royle, 2005; Royle and Dorazio, 2008; Zipkin et al., 2010). 944
We expected a lower mammal’s occurrence in locations closer to human residences 945
and to roads due to increased hunting and habitat disturbances (Benítez-López et al., 946
147
2017). Cattle, in turn, could displace wild species spatially or temporally, through 947
environmental degradation, competition for space and resources, and through an 948
increased probability of interaction with human ranchers (Pudyatmoko, 2017; Gaynor et 949
al., 2018). Finally, we expected that more arboreal and productive environments, 950
generally related to drainage (riparian habitats) and on the humid slopes of hills and 951
mountains, will benefit the persistence of the species, as these places may represent better 952
resources and refuges (Stoner and Timm, 2010), as well as being less accessible to human 953
activities (Lopes et al., 2017). 954
955
Methods 956
Study area 957
This study was conducted in a landscape between Curaçá and Juazeiro 958
municipalities, in the north of the Bahia state, northeastern Brazil (Fig. 1). Recently, two 959
protected areas have been established in the region: the Blue Macaw Environmental 960
Protection Area (89,996 ha) and Wildlife Refuge (29,986 ha). Both reserves were created 961
to protect the remaining habitats that historically housed the last population of the blue 962
macaw (Cyanopsitta spixii), currently extinct in the wild, and where its reintroduction is 963
planned for the coming years. The climate of the region is semiarid, with high 964
temperatures (21.1 − 30.1 °C, annual average: 24.9 °C) and one of the lowest annual 965
average rainfall of the Caatinga (452 mm) (Freitas et al., 2005). This region was one of 966
the most affected by the intense drought that occurred between 2012 and 2017 (Brito et 967
al., 2018). 968
Regarding vegetation in our study area, in the flatter, low-lying places, there is 969
mainly shrubby vegetation, while in higher altitude forested formations predominate, with 970
denser and arboreal vegetation. Along rivers and streams, gallery forests predominate, 971
148
where large trees of Tabebuia aurea can reach up to 20 m height (Freitas et al., 2005). 972
The exploitation and degradation of native vegetation in the region are historic and 973
intense, either for domestic use by the various existing communities or as food for the 974
abundant herds of extensively raised cattle and spectially goats and sheep, which in the 975
region reach the largest abundances in Brazil (IBGE, 2016). 976
977
978
Fig. 1. Study area where medium to large-sized mammals were surveyed in an overgrazed 979
semiarid landscape in Caatinga dry forest, northeastern Brazil. The 60 camera trap stations and 980
the polygons of the newly created protected areas are shown (Environmental Protection Area and 981
Wildlife Refuge of the Blue Macaw, Área de Proteção Ambiental - APA and Refúgio de Vida 982
Silvestre - RVS da Ararinha Azul, respectively) in northern Bahia, northeathern Brazil. 983
984
Data collection 985
We used automatic camera traps (Bushnell Essential E2 and Bushnell Trophy Cam 986
HD) triggered by heat and motion to study the mammal presence. We followed the 987
Tropical Ecology Assessment and Monitoring Network - TEAM (2011) protocol for 988
149
surveying terrestrial vertebrates in tropical forests. We established 60 sample stations 989
seeking to establish a minimum distance of 1.5 km between nearest. The stations were 990
established considering the possibility of access, seeking to homogeneously cover the 991
main environments (riparian forest, shrub, and arboreous vegetation), as well as covering 992
the altitudinal gradient of the area, as suggested by TEAM (2011). We installed camera 993
traps on human and animal trails (livestock and wild mammals), 30-40 cm from the 994
ground, programmed to shot 2 or 3 pictures, with at least 5 minutes intervals, running 24 995
hours. We concentrated sampling in dry months of the year (October to December) when 996
there is little or no rainfall in the region. We did not use any bait for animals. Due to 997
equipment restrictions, sampling was performed in two blocks of 30 sample stations each 998
time, trying to keep the equipment at least 30 days in each sample station. We have 999
accumulated a sampling effort of 2,250 camera-days, with the 60 camera trap stations 1000
remaining on average 37.5 days in the field (standard deviation: 2.9; range: 17-40 days). 1001
Data analysis 1002
Descriptive analysis 1003
We calculated the capture success or RAI (relative abundance index) as the ratio 1004
between the number of records and sampling effort times 100, for each species (Rovero 1005
et al., 2017). For all analyzes, we used records considered to be independent, those over 1006
60 minutes (1 h) between consecutive detections of the same species at the same station 1007
(i.e. Rovero et al., 2017). The efficiency of our effort to characterize the mammalian 1008
community was ascertained through the species accumulation curve based on 1009
independent records and randomized 1000 times in the free program i-NEXT (Chao et 1010
al., 2016). We obtained the naïve occupancy (observed) as a simple proportion of 1011
sampling sites in which each species was recorded in relation to total sampling sites 1012
surveyed (Zimbres et al., 2018). 1013
150
Occupancy and detection covariates 1014
Occupancy site modeling allows the investigation of which biological factors may 1015
explain the occupancy of a species (i.e., habitat use), while correcting for the probability 1016
of the species being detected (i.e., sampling bias) when present, allowing those 1017
probabilities to be a function of site or sample covariates (Mackenzie et al., 2002). We 1018
listed eight environmental and anthropic covariates that are potentially capable of 1019
explaining the occupancy probabilities of the sampled species (Table 1). Among 1020
anthropic covariates are the distance of nearest residence, nearest road distance, and 1021
relative abundance index (RAI) of goats, sheep, and cows (cattle) (Table 1), while forest 1022
cover, elevation, terrain roughness, and temporary watercourse distance were the 1023
environmental covariates investigated (see Table 1 for details). 1024
Regarding detection, all camera traps were installed on pre-existing trails in the area, 1025
however, these trails varied in width and intensity of use, and this may influence species 1026
detectability (Ferreira et al., 2017). Therefore, we classified as main trails those relatively 1027
wide (i.e. > 0.5 m) and with signs of intense use by animals (wild or domestic), while the 1028
secondary trails (i.e. < 0.5 m) were those less marked and trails that are usually branches 1029
of a main trail or road. We assumed that wider trails could benefit the detection of species 1030
that often move along trails or roads, such as carnivores (Goulart et al., 2009; Harmsen et 1031
al., 2010). 1032
We investigated possible correlations between covariates using Pearson's correlation 1033
coefficient in program R (R Core Team, 2016) and covariates with r > | 0.6 | were 1034
excluded from modeling. Altitude was strongly correlated with forest cover, river 1035
distance, and terrain roughness (Table S1), so it was excluded from modeling. 1036
1037
1038
151
Table 1 1039
Covariates used to model detection and occupancy probabilities of medium to large-sized wild mammals in an overgrazed semiarid 1040
landscape in Caatinga dry forest, northeastern Brazil. 1041
Covariate (code) Description Parameter
of interest
Expected
effect
Mean
(SD; min. – max.)
Source
Trail type (trail) Main trails are wider (e.g.> 0.5 m)
and apparently more used by animals and people than secondary
trails (e.g. <0.5 m)
Detection + / - 1 (principal trail),
0 (secondary trail)
Classification in the
field
Forest cover (forest) Percentage of vegetation with a predominance of continuous
canopy (forested steppe-savannah
and seasonal semideciduous and
deciduous forest) within 1 km of camera (buffer)
Occupancy + 31.9% (34.2; 0 – 99.3)
MAPBIOMAS
Elevation (elevat) Elevation average within 1 km of
camera (buffer)
Occupancy + 459 m
(69.4; 367 – 619)
SRTM
Terrain roughness (rough) Altitude standard deviation within 1
km of camera (buffer)
Occupancy + 13.4
(19.4; 1.6 – 76.5)
DPI/INPE
River or stream distance (river) Euclidean distance to the nearest waterway, usually temporary
Occupancy + 1319.2 m (1363.7; 0 – 7106)
Measured on Google Earth Pro
Livestock abundance (livest) Relative abundance index (RAI) of
cows, sheep and goats (together)
Occupancy - 104.4 records/100
camera-days
(112.9; 0 – 469.4)
Camera trapping
data
Proximity to a road (road) Euclidean distance to the nearest
road
Occupancy - 818.3 m
(742.6; 0 – 3029)
Measured on
Google Earth Pro
Proximity of a residence (house) Euclidean distance to the nearest human residence
Occupancy - 2476 m (2085.2; 270 – 8112)
Measured on Google Earth Pro
1042
152
Modelling framework 1043
We applied the model developed by Zipkin et al. (2009), with modifications, to 1044
determine site-specific occupancy (𝒛) for each species (𝒊 = 𝟏, 𝟐, … , 𝑵) at every site (𝒋 =1045
𝟏, 𝟐, … , 𝑱). In this form, 𝒛(𝒊, 𝒋) = 𝟏 if species i occurred at site j, and 𝒛(𝒊, 𝒋) = 𝟎, if not. 1046
Thus, we can specify the model for occurrence as a Bernoulli random variable 1047
𝒛(𝒊, 𝒋)~𝑩𝒆𝒓𝒏(𝝍𝒊,𝒋), where 𝝍𝒊,𝒋 is the probability of that species i occurs at site j. 1048
However, we do not observe directly this state variable 𝒛(𝒊, 𝒋), instead, we gather 1049
information on observational data, 𝒙(𝒊, 𝒋, 𝒌), for species i, at site j during sample k. This 1050
observational data is also considered as a Bernoulli random variable if species i is present 1051
𝒛(𝒊, 𝒋) = 𝟏 or, otherwise, 𝒙(𝒊, 𝒋, 𝒌) = 𝟎 if 𝒛(𝒊, 𝒋) = 𝟎. This is a logical statement because 1052
it is not possible to observe species i at site j during any sample k if that species does not 1053
occur at site j in the first place. Hence, the model of the observational process is specified 1054
as: 𝒙(𝒊, 𝒋, 𝒌)~𝑩𝒆𝒓𝒏(𝜽𝒊,𝒋,𝒌 ∗ 𝒛(𝒊, 𝒋)) where 𝜽𝒊,𝒋,𝒌 represents the probability of detection for 1055
species i at site j in the kth sampling period, conditional on the species being present (i.e. 1056
𝒛(𝒊, 𝒋) = 𝟏). 1057
The effects of a site (environmental and anthropogenic) and survey (trail) covariates 1058
on both occupancy and detection probabilities, 𝝍 and 𝜽, respectively, can be incorporated 1059
in the linear model in the logit scale: 𝒍𝒐𝒈𝒊𝒕(𝝍𝒊,𝒋) = 𝒖𝒊 + 𝜶𝒋 and 𝒍𝒐𝒈𝒊𝒕(𝜽𝒊,𝒋) = 𝒗𝒊 + 𝜷𝒋, 1060
where 𝒖𝒊 and 𝒗𝒊 are different species effects and 𝜶𝒋 and 𝜷𝒋 are habitat effects on 1061
occupancy and detection respectively. Hence the model for the occurrence of species i at 1062
site j is: 1063
𝒍𝒐𝒈𝒊𝒕(𝝍𝒊,𝒋) = 𝒖𝒊 + 𝜶𝟏𝒊𝒇𝒐𝒓𝒆𝒔𝒕𝒋 + 𝜶𝟐𝒊𝒓𝒐𝒖𝒈𝒉𝒋 + 𝜶𝟑𝒊𝒉𝒐𝒖𝒔𝒆𝒋 + 𝜶𝟒𝒊𝒍𝒊𝒗𝒆𝒔𝒕𝒋1064
+ 𝜶𝟓𝒊𝒓𝒊𝒗𝒆𝒓𝒋 + 𝜶𝟔𝒊𝒓𝒐𝒂𝒅𝒋 1065
Considering closure of the community, i.e. there was no increases or decreases in the 1066
species numbers during the study, the detection model, then, is specified as: 1067
153
𝒍𝒐𝒈𝒊𝒕(𝜽𝒊,𝒋,𝒌) = 𝒗𝒊 + 𝜷𝟏𝒊𝒕𝒓𝒂𝒊𝒍𝒋 1068
We then considered a community-level response to all variables by assuming that 1069
occupancy and detection on a species level (𝒖𝒊, 𝒗𝒊) are random effects regulated by a 1070
“hyper-parameter” (Zipkin et al. 2009, 2010), for example, 𝜶𝟏𝒊~𝑵(𝝁𝜶𝟏, 𝝈𝜶𝟏), where 1071
𝝁𝜶𝟏 is the mean effect for parameter 𝜶1𝒊 (forest cover) across all species, and 𝝈𝜶1 is the 1072
standard deviation. In this manner, the mean response, and its variance, for each 1073
occupancy and detection covariates, 𝜶𝒋 and 𝜷𝒋, for all species, is the hyper-parameter for 1074
the community distribution. In other words, if the estimative of a covariate hyper-1075
parameter is deemed significant (i.e. its posterior distribution does not overlap 0), it would 1076
suggest that the mean occupancy and/or detection probabilities for every species – the 1077
community – is a response to that specific covariate. Although all recorded species were 1078
considered in community modeling and to estimate the occupancy and detection 1079
probabilities, in relation to species-specific responses to covariates, we considered only 1080
species with more than five records in the interpretation of species-level results. 1081
All model parameters and analysis were estimated using a Bayesian framework, 1082
through the package R2jags within the R3.4.4 (R Core Team, 2016), which implemented 1083
a Markov Chain Monte Carlo (MCMC) to estimate the posterior distribution of the 1084
variables. We used vague priors for hyper-parameters (see code in supplementary 1085
material) and random initial values for all variables. We run 3 chains with 15,000 1086
iterations, discarding the first 3,000 as burn-ins. Convergence was checked visually and 1087
through the Gelman-Rubin statistic (Rhat, where values �̂� < 𝟏. 𝟏 suggests convergence; 1088
Kéry et al., 2010). We also assessed model fitness using a Bayesian P-value approach 1089
(Zipkin et al., 2010), described in the supplementary material. 1090
1091
154
Results 1092
General results 1093
We obtained a total of 3082 records of wild and domestic mammals; of these, 566 1094
records were from 12 medium to large-sized wild mammal species (Table 2, Fig. S1), and 1095
other 69 records were from small mammals (Galea spixii [33 records], Kerodon rupestris 1096
[25], Thrichomys sp. [7], and non-identified species [4]). Among the recorded species, 1097
three are considered threatened globally, nationally or regionally (Table 2). The lack of 1098
stabilization of the species accumulation curve suggests that more species occur in the 1099
area (Fig. S2). In addition to wild species, we obtained 2,447 records of domestic 1100
mammals (goats [1745], sheep [512], cows [108], donkeys [59], horses [10], dogs [12], 1101
and cat [1]), which is more than 300% higher than records of medium to large-sized wild 1102
mammals, besides six people records. 1103
Cerdocyon thous was by far the most recorded species, representing just over 50% 1104
of medium to large-sized wild mammal records, followed by Dasypus novemcinctus, 1105
Euphractus sexcinctus and Leopardus tigrinus (Table 2). Following Zimbres et al. (2018), 1106
of the remaining species, three of them were rare (< 30 records) and the other five species 1107
were recorded very rarely (< 10 records) (Table 2).1108
155
Table 2 1109
Checklist of medium to large-sized wild mammal species detected by camera trapping in an overgrazed semiarid landscape in Caatinga dry forest, 1110
northeastern Brazil. 1111
Taxon Common name Body mass
(Kg) Feeding
guild Conservation status
(Bahia/Brazil/IUCN) Records RAI
PILOSA
Tamandua tetradactyla Southern tamanduá 5.2 In LC/LC/LC 18 0.80
CINGULATA
Dasypus novemcinctus Nine-banded armadillo 3.6 In/On LC/LC/LC 85 3.78
Euphractus sexcinctus Yellow armadillo 5.4 In/On LC/LC/LC 66 2.93
CARNIVORA
Leopardus tigrinus Nothern tiger cat 2.2 Ca VU/EN/VU 40 1.78
Herpailurus yagouaroundi Jaguarundi 4.5 Ca VU/VU/LC 7 0.31
Puma concolor Puma 46 Ca VU/VU/LC 1 0.04
Cerdocyon thous Crab-eating fox 6.5 In/On LC/LC/LC 317 14.09
Conepatus amazonicus Striped hog-nosed skunk 2.4 In/On LC/LC/LC 11 0.49
ARTIODACTYLA
Mazama gouazoubira Gray brocket deer 20 Fr/Hb LC/LC/LC 17 0.76
Pecari tajacu Collared peccary 26 Fr/Hb NT/LC/LC 1 0.04
RODENTIA
Dasyprocta nigriclunis Highland black-rumped agouti
3 Fr/Gr LC/LC/LC 2 0.09
DIDELPHIMORPHIA
Didelphis albiventris White-eared opossum 1.6 Fr/On LC/LC/LC 1 0.04
Conservation status are based on Cassano et al. (2017, Bahia state), MMA (2014, Brazil) and IUCN (2020). Body mass and feeding guild are based on Paglia et al. 1112 (2012) (Fr = frugivore; On = omnivore; In = insectivore; Myr = myrmecophage; Hb = herbivore grazer; Gr = granivore; Ca = carnivore.). RAI (relative abundance 1113 index = detection events per camera trap days x 100). 1114
156
Occupancy and detection patterns 1115
Considering the naïve occupancy rate, C. thous was also the most widely distributed 1116
species in our study area, being registered in 79% of the sampled sites (Table 3), while 1117
Euphractus sexcinctus was detected in 50% of the sites, and the other species occurred in 1118
35% or less of the sites, with four of them being record at a single sample site (2%) (Table 1119
3). Regarding the average occupancy, the species exhibited a dissimilar pattern, varying 1120
from 0.89 for E. sexcinctus to 0.13 for D. nigriclunis (Table 3). Most of the recorded 1121
species showed low average detection (< 0.10; Table 3), while C. thous exhibited the 1122
highest detection rate (0.48; Table 3). 1123
Among the environmental and anthropogenic covariates that we hypothesized to be 1124
able to explain the occupancy of medium to large-sized wild mammals, only forest cover 1125
(forest) influenced significantly and positively the occupancy rate of five out of 12 1126
analyzed species: C. thous, L. tigrinus, M. gouazoubira, D. novemcintus, and H. 1127
yagouaroundi (Table 4; Fig. 2). In addition, community-level occupancy was also 1128
correlated with the percentage of forested habitat on the sampled site (Table 4). For all 1129
other predictors evaluated, besides forest cover for some species, the confidence intervals 1130
of beta coefficients overlapped zero, which suggests the absence of significant effect 1131
(Table 4). Regarding the detection probability, only C. thous and E. sexcinctus presented 1132
a significant variation in their detection between trail types (Table 4; Fig. 3), both species 1133
were more detected in main trails (wider) than on secondary trails, which also influenced 1134
the community-level detection (Table 4). 1135
1136
1137
1138
1139
157
Table 3 1140
Naïve occupancy (observed) and average occupancy (ѱ) and detection (Ɵ), and their average 1141
standard deviations (SD), of medium to large-sized mammal species recorded by camera trapping 1142
in an overgrazed semiarid landscape in Caatinga dry forest, northeastern Brazil. The estimates 1143
come from the model with all the variables investigated. 1144
Species Naïve
occupancy ѱ (SD) Ɵ (SD)
C. thous 0.79 0.81 (0.10) 0.48 (0.04)
E. sexcinctus 0.50 0.89 (0.14) 0.13 (0.02)
D. novemcinctus 0.35 0.41 (0.16) 0.33 (0.06)
L. tigrinus 0.33 0.58 (0.18) 0.12 (0.04)
T. tetradactyla 0.22 0.86 (0.20) 0.05 (0.02)
M. gouazoubira 0.12 0.26 (0.10) 0.14 (0.04)
C. amazonicus 0.12 0.65 (0.32) 0.05 (0.03)
H. yagouaroundi 0.12 0.43 (0.26) 0.04 (0.02)
D. nigriclunis 0.02 - -
D. albiventris 0.02 - -
P. tajacu 0.02 - -
P. concolor 0.02 - -
1145
1146
1147
158
Table 4 1148
Regression coefficients (and 95% credible confidence intervals) for all covariates tested to influence detection probability (Ɵ) and occupancy probability (ѱ) of 1149
medium to large-sized mammal species and entire community in an overgrazed semiarid landscape in Caatinga dry forest, northeastern Brazil. Coefficients that 1150
do not overlap zero considering their 95% Bayesian confidence intervals are shown in bold. Estimates are shown only for species with more than five records, 1151
but all 12 species were considered in the modeling. The estimates come from the model with all the variables investigated. 1152
Detection (Ɵ) Occupancy (ѱ)
Species Trail Forest Rough River Livest Road House
C. thous
0.40
(0.18 − 0.62)
3.03
(0.95 − 6.15)
0.41 (-0.77 − 2.08)
-0.67 (-1.65 − 0.17)
0.15 (-0.52 − 0.87)
0.12 (-0.84 − 1.40)
0.11 (-0.85 − 1.32)
D. novemcinctus
0.25
(-0.10 − 0.61) 1.15
(0.05 − 3.78)
0.56
(-0.21 − 1.76)
-0.37
(-1.22 − 0.39)
0.286
(-0.30 − 1.01)
-0.37
(-1.31 − 0.41)
-0.38
(-1.45 − 0.41) E. sexcinctus
0.61
(0.28 − 0.97)
2.51
(-0.06 − 6.42)
-0.52
(-3.38 − 1.95)
-0.24
(-1.59 − 1.31)
0.21
(-0.89 − 1.46)
0.40
(-1.11 − 2.56)
-0.57
(-2.77 − 1.47)
L. tigrinus
0.13
(-0.26 − 0.51) 3.60
(1.27 − 7.12)
-0.01
(-1.65 − 2.29)
-0.45
(-1.47 − 0.50)
0.36
(-0.41 − 1.38)
0.28
(-0.87 − 1.80)
-0.69
(-2.48 − 0.75) T. tetradactyla
0.32
(-0.10 − 0.76)
2.62
(-0.59 − 7.73)
0.26
(-1.99 − 3.11)
-0.26
(-1.67 − 1.28)
0.23
(-0.97 − 1.48)
0.27
(-1.65 − 2.57)
-0.07
(-2.39 − 2.44)
M. gouazoubira
0.18 (-0.33 − 0.66)
4.79
(1.87 − 11.33)
-0.67 (-3.56 − 1.03)
-0.12 (-1.39 − 1.55)
0.09 (-1.13 − 1.17)
0.30 (-1.19 − 2.14)
0.12 (-1.25 − 1.96)
C. amazonicus
0.16
(-0.37 − 0.67)
2.31
(-1.15 − 8.26)
0.53
(-1.57 − 3.23)
-0.76
(-2.84 − 0.69)
0.16
(-1.03 − 1.42)
0.17
(-1.72 − 2.47)
-0.41
(-2.64 − 1.57)
H. yagouaroundi
0.42 (-0.13 − 1.04)
3.26
(0.69 - 7.31)
-0.53 (-3.87 - 2.12)
-0.58 (-2.26 - 0.79)
0.03 (-1.25 - 1.10)
0.77 (-0.63 − 3.12)
-0.24 (-2.08 − 1.54)
Community
0.305
(<0.01 − 0.60)
3.05
(1.21 − 6.50)
-0.02
(-1.59 − 1.30)
-0.43
(-1.32 − 0.38)
0.17
(-0.52 − 0.87)
0.16
(-0.77 − 1.27)
-0.27
(-1.48 − 0.90)
1153
159
Fig. 2 Relationship between forest cover and estimated occupancy probability (with their 95%
Bayesian confidence intervals) of medium to large-sized wild mammals in an overgrazed semiarid
landscape in Caatinga dry forest, northeastern Brazil. Only significant species-specific responses
to the investigated predictors are presented: (a) Cerdocyon thous, (b) Dasypus novemcinctus, (c)
Leopardus tigrinus, (d) Mazama gouazoubira, and (e) Herpailurus yagouaroundi. Illustrations
by Aldo Chiappe (use permitted).
160
Fig. 3 Effect of trail type (main and secondary) on the detection probability of medium to large-
sized mammals in an overgrazed semiarid landscape in Caatinga dry forest, northeastern Brazil.
Only significant species-specific responses to the trail type are presented: (a) Cerdocyon thous,
(b) Euphractus sexcinctus. The bars represent the 95% Bayesian confidence intervals.
Discussion
Mammals’ occupancy and detection patterns
The proportion of forested habitat was important to explain the medium to large-
sized mammals’ occupancy in a landscape of the Caatinga dry forest historically impacted
by intense grazing and dominated by shrub and more sparse vegetation. Tallest and
densest vegetation is an important predictor of terrestrial mammal’s occurrence in
different tropical ecosystems (Desbiez et al., 2009; Vynne et al., 2011; Nagy-Reis et al.,
2017; Goulart et al., 2009). Structurally more complex habitats can provide better food
resources and shelters for rest and thermoregulation to mammals, especially in more
modified landscapes and in periods of resource scarcity (Debiez et al., 2009; Vynne et al.,
2011). In seasonally dry tropical forests, forested patches associated with riparian and
higher altitude environments conserve moisture and harbor strictly forest mammals,
especially during periods of greatest resource scarcity (Stone and Timm, 2011), but they
161
also benefit more generalist species, as shown here. In the case of elevated and sloping
areas, these environments may guarantee greater protection against anthropogenic threats
due to less accessibility (Morato et al., 2014; Lopes et al., 2017).
Our results indicate that in a context of disturbance and resource scarcity the entire
community of medium to large-sized mammals can benefit from a greater proportion of
forested habitat, but the effect was especially strong for Cerdocyon thous, Mazama
gouazoubira, Dasypus novemcinctus, Leopardus tigrinus and Herpailurus yagouaroundi,
including from generalist to threatened species. C. thous is a generalist and tolerant of
anthropogenic impacts (Dias and Bocchiglieri, 2016; Dias et al., 2019a), which is
reinforced by its arge amount of records and considerable high occupancy rate even in
non-forested sites found here. However, more forested environments can provide greater
food availability for these species during intense drought periods, when the insect
abundance, its main food item during the dry season (Dias and Bocchiglieri, 2016), falls
significantly (Vasconcellos et al., 2010).
Mazama gouaozubira is known as an ecologically flexible deer, often preferring open
and managed areas, but generally depending on forest patches as a refuge (Ferreguetti et
al., 2015; Rodrigues et al., 2017). However, our results suggest that in highly disturbed
semiarid environments the relationship of this species with more forested habitats appears
to be stronger. In fact, other studies have suggested this (Astete et al., 2016), which may
be related to its dependence of more abundant leaves and fruits on the woody vegetation,
especially in periods of drought and greater resource scarcity, when competition with
livestock may be more intense (Serbent et al., 2011). Similarly, although D. novemcinctus
is relatively abundant and habitat generalist in some environments (Michalski and Peres,
2007; Zimbres et al., 2018), the more facility for thermoregulation and greater
invertebrate availability may favor the occurrence of these species within forested areas
162
(Goulart et al., 2009; Ferreguetti et al., 2016; Rodrigues and Chiarello, 2018). Dense
environments can also guarantee a better capacity for camouflage and escape for prey
species (Goulart et al., 2009). In addition, these places can protect species such as M.
gouazoubira and D. novemcinctus against intense hunting practiced in the Caatinga that
has led some populations to local extinction (Bezerra et al., 2014; Alves et al., 2016;
Marinho et al., 2018a).
Among the species most benefited by the most forested areas area also small wild
cats threatened with extinction. L. tigrinus and H. yagouaroundi are threatened mainly by
habitat loss, human persecution, roadkill, and negative interactions with domestic
carnivores (Oliveira et al., 2013; Giordano, 2016). In the Caatinga, a study in 10 different
landscapes concluded that the occurrence of this species is higher in more forested
habitats and far from human settlements, where there would be better resources and
refuges against anthropic threats (Marinho et al., 2018b). In the case of the less abundant
H. yagouaorundi, its preference for forested habitats is less known in the Caatinga but
corroborates the greater use of deciduous forests with undergrowth in dry environments
(Giordano, 2016). Another factor that has been used to explain the neotropical wild cat’s
abundance and distribution is its potential negative interaction with ocelot (Leopardus
pardalis) (Oliveira et al., 2013; Marinho et al., 2018b; Dias et al., 2019b). The absence
of records from this dominant mesopredator did not allow us to investigate this possible
effect, but on the other hand it suggests an ocelot’s low density in our study area, which
could benefit smaller wild cats.
Puma concolor, Pecari tajacu, and Dasyprocta nigriclunis are among the least
recorded species in our study area and were recorded only in forested sites associated
with mountains, corroborating the relevance of this habitat for the persistence of less
tolerant species. Indeed, these species are among the most affected by habitat disturbance
163
and human persecution in the Caatinga (Alves et al., 2016; Bezerra et al., 2014; Marinho
et al., 2018a, 2019). The extremely low number of records of these species is a cause for
concern and suggests that they may be functionally extinct or shrinking for this if no
action is taken to reverse this scenario (Tilker et al., 2019). This can lead to the loss of
important ecological functions in tropical regions such as large seed dispersion and
population control of prey and mesocarnivores (Magioli et al., 2020).
Despite being one of the most abundant species in Caatinga and other ecosystems
for being omnivorous and tolerant (Freitas et al., 2005; Bezerra et al., 2014; Marinho et
al., 2018a), Didelphis albiventris was only recorded once during our study precisely at a
forested site. Among the species with more than 10 records, Conepatus amazonicus and
Euphractus sexcinctus are habitat generalists or related with more open areas (Dias, 2017;
Ferreguetti et al., 2016) and did not have their distribution explained by any factors
analyzed here. On the other hand, the scansorial Tamandua tetradactyla is usually
associated with forested environments (Desbiez et al., 2009), but it was not associated
with this habitat type.
The distance from rivers and streams was not an important predictor to explain the
mammal occurrence, probably because in many stretches of these environments the
vegetation is restricted to a small and degraded open forest strip because of anthropogenic
pressure like grazing cattle. However, considering other studies and that the forest cover
category used here includes riparian forests, we believe that these places are important
for species’ movement between most degraded and preserved environments in the region
(Goulart et al., 2009; Ferreguetti et al., 2016; Rodrigues and Chiarello, 2018; Zimbres et
al., 2018), functioning as ecological corridors, especially in semiarid environments
(Shuette et al., 2013; Drouilly et al., 2018).
164
In landscapes with high modification or historical and intense anthropogenic
disturbance, it is possible that species pass through an environmental filter and that only
those most tolerant to degradation persist with viable populations, becoming true
“survivors” (Prugh et al., 2008). This may explain the effect’s absence of the investigated
anthropogenic factors on species occupancy in our study area, while these are the main
predictors of abundance and distribution of large and medium-sized mammals in other
tropical landscapes (Nagy-Reis et al., 2017; Tilker et al., 2019; Dias et al., 2019a).
Another possibility is that anthropic factors not investigated here may affect the species’
distribution. In a study in the rainforest of South Asia, Tilker et al. (2019) demonstrated
that dense forests explained the mammal’s distribution in degraded habitats while hunting
pressure measured through village density was more important in conserved
environments. Studies in more preserved areas of the Caatinga, where sensitive species
such as Panthera onca still occur, have found negative anthropic effects on large
mammals’ occupancy (Astete et al., 2017; Dias et al., 2019b). Similarly, an investigation
that included landscapes with different integrity degrees, detected this negative
anthropogenic effect on L. tigrinus in the Caatinga (Marinho et al., 2018b). According to
our results, it is possible that “surviving” species in semiarid environments may be less
impacted using more forested patches and beyond that, taking into account that human
activities occur mainly during daytime (except hunting), greater nocturnality may favor
this coexistence (Gaynor et al., 2018), what should be investigated in future studies.
Additionally, considering the possible impacts of the livestock on vegetation and
fauna of drylands (Tabeni and Ojeda, 2003; Yoshihara et al., 2008; Marinho et al., 2016;
Pudyatmoko, 2017), it is likely that a possible significant effect of livestock disturbance
on native species’ occupancy in Caatinga should be better detected at controlled levels of
grazing pressure, while in our study area it might be widespread in the landscape. So the
165
wild mammal species that occur in this landscape in apparent greater abundance must be
those best able to resist the chronic anthropogenic disturbance using mechanisms as the
higher use of forested and more productive habitats to persist.
In methodological terms, the trail type affected the detection of C. thous and E.
sexcinctus, as well as at the community level detection, with species being better detected
in wider and more intensively used by livestock trails. The positive effect of sampling
mammals on trails is well recognized but varies between species (Goulart et al., 2009;
Harmsen et al., 2010; Ferreira et al., 2017). Carnivores, for example, are best detected on
wide, long-established trails (Harmsen et al., 2010). C. thous and E. sexcinctus are two
species relatively tolerant to disturbance and seem to move better along more open trails,
which must be considered to optimize future sampling.
Conclusions and recommendations
The association between a minimum protocol for survey of tropical mammals
(TEAM, 2011) and Bayesian hierarchical multispecies occupancy modeling proved to be
efficient and must serve as a baseline for future mammals’ studies and population
monitoring programs in ecologically relevant, little known semiarid landscapes like ours.
This type of approach is especially important to obtain robust information on rare species’
ecology, which is essential to improve biodiversity conservation and management in
semiarid environments (Zipkin et al., 2010; Drouilly et al., 2018).
In this study, we assessed the occupancy patterns of medium to large-sized wild
mammals to understand how species persist in a semiarid region of great ecological
relevance historically impacted by anthropogenic disturbance, and which from 2012 to
2017 experienced the most severe drought of the last decades. Our results suggest that
more forested habitats can be an important refuge to terrestrial mammals in disturbed
semiarid landscapes, benefiting both generalist and threatened species. In this context,
166
apparently more abundant species can be considered “survivors” while those more
sensitive are heading towards their functional extinction. Therefore, conservation
strategies for these species and their ecological functions must prioritize the maintenance
of forested habitats and restoration of degraded areas. The knowledge generated here is
especially important in a scenario of intensification of extreme drought events due to
climate change.
The conflict between extensive livestock farming and wildlife conservation is a
concern in semiarid landscapes worldwide (Tabeni and Ojeda, 2003; Kinnaird et al.,
2012), considering that herds are often the only income source for local people. In this
context, Kinnaird et al. (2012) suggest some actions to improve the coexistence between
people and wildlife in these places, which we adapted to our reality, as follows: a)
incentive to ecotourism and access of local residents to the gains with this activity; b)
assistance and incentive for the restoration of degraded areas; c) technical assistance and
credit to improve livestock farming efficiency, prioritizing ecologically more sustainable
techniques; and d) in the case of the landscape studied here, it is important to seek the
ensurance of effective protection of the more dense and forested vegetation remnants in
the mountains and riparian zones. This is essential considering that most of these
remnants are outside the recently created protected areas and that these places can serve
as refuge and source of wild species, besides still generate benefits for communities
through tourism to contemplate its relevant scenic beauty.
Acknowledgments
We are grateful to Ararinha na Natureza Project, carried out by the Centro Nacional
de Pesquisa e Conservação de Aves Silvestres (CEMAVE/ICMBio) and financed by Vale
through the Fundo Brasileiro para a Biodiversidade (Funbio). I would like to thank
167
Cristine Prates, Sueli Damasceno, Damilys Oliveira, Mércia Milena, Tatiane Alves,
Rogério Santos, Paulo da serra, and Leomar Martins (Babá) for their important assistance
during data collection. We thank Virgínia Paixão, Fernanda Lamim, and Barbara Zimbres
for their support in text review or graphics’ elaboration. We would like to thank Claudia
Campos, Mauro Pichorim, Fabiana Rocha, and Rodrigo Massara for the valuable
suggestions that helped to improve the manuscript. EMV (#308040/2017-1) and CRF
(#305304/2013-5; 306812/2017-7) were supported by Conselho Nacional de
Desenvolvimento Científico e Tecnológico (CNPq) and PHM was supported by
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES;
financing code 001).
References
Alves, R.R.R.N., Feijó, A., Barboza, R.R.D., Souto, W.M.D., Fernandes-Ferreira, H.,
Cordeiro-Estrela, P., Langguth, A., 2016. Game mammals of the Caatinga biome.
Ethnobio. Conserv. 5, 1–51. https://doi.org/10.15451/ec2016-7-5.5-1-51
Antongiovanni, M., Venticinque, E.M., Fonseca, C.R., 2018. Fragmentation patterns of
the Caatinga drylands. Landscape Ecol. 33, 1353–1367.
https://doi.org/10.1007/s10980-018-0672-6
Astete, S., Marinho-Filho, J., Machado, R.B., Zimbres, B., Jácomo, A.T.A., Sollmann,
R., Tôrres, N.M., Silveira, L., 2017. Living in extreme environments: modeling habitat
suitability for jaguars, pumas, and their prey in a semiarid habitat. J. Mammal. 98,
464–474. https://doi.org/10.1093/jmammal/gyw184
Astete, S., Marinho-Filho, J., Machado, R.B., Zimbres, B., Jácomo, A.T.A., Sollmann,
R., Tôrres, N.M., Silveira, L., 2017. Living in extreme environments: modeling habitat
168
suitability for jaguars, pumas, and their prey in a semiarid habitat. J. Mammal. 98,
464–474. https://doi.org/10.1093/jmammal/gyw184
Banda, R.K., Delgado-Salinas, A., Dexter, K.G., Linares-Palomino, R., Oliveira-Filho,
A., Prado, D., Pullan, M., Quintana, C., Riina, R., Rodríguez, G.M., Weintritt, J.,
Acevedo-Rodríguez, P., Adarve, J., Álvarez, E., Anairamiz, A.B., Arteaga, J.C.,
Aymard, G., Castaño, A., Ceballos-Mago, N., Cogollo, Á., Cuadros, H., Delgado, F.,
Devia, W., Dueñas, H., Fajardo, L., Fernández, Á., Fernández, M.Á., Franklin, J.,
Freid, E.H., Galetti, L.A., Gonto, R., González-M, R., Graveson, R., Helmer, E.H.,
Idárraga, Á., López, R., Marcano-Vega, H., Martínez, O.G., Maturo, H.M., McDonald,
M., McLaren, K., Melo, O., Mijares, F., Mogni, V., Molina, D., Moreno, N.P., Nassar,
J.M., Neves, D.M., Oakley, L.J., Oatham, M., Olvera-Luna, A.R., Pezzini, F.F.,
Dominguez, O.J.R., Ríos, M.E., Rivera, O., Rodríguez, N., Rojas, A., Särkinen, T.,
Sánchez, R., Smith, M., Vargas, C., Villanueva, B., Pennington, R.T., 2016. Plant
diversity patterns in neotropical dry forests and their conservation implications.
Science 353, 1383-1387. http://dx.doi.org/10.1126/science.aaf5080
Bennie, J.J., Duffy, J.P., Inger, R., Gaston, K.J., 2014. Biogeography of time partitioning
in mammals. P. Natl. Acad. Sci.-Biol. 111, 13727–
13732. https://doi.org/10.1073/pnas.1216063110
Bezerra, A.M.R., Lazar, A., Bonvicino, C. R., & Cunha, A. S. (2014). Subsidies for a
poorly known endemic semiarid biome of Brazil: non-volant mammals of an eastern
region of Caatinga. Zool. Stud. 53, 16. https://doi.org/10.1186/1810-522X-53-16
Carmignotto, A.P., Astúa, D., 2017. Mammals of the Caatinga: diversity, ecology,
biogeography, and conservation, in: Silva, J.M.C., Leal, I.R., Tabarelli, M. (Eds.),
Caatinga: The largest tropical dry forest region in South America. Springer
International Publishing, Cham, pp. 211-254.
169
Cassano, C.R., Almeida-Rocha, J.M., Alvarez, M.R., Bernardo, C.S.S., Bianconi, G.V.,
Campiolo, S., Campos, C.B., Danilewicz, D., Falcão, F.C., García, F.V., Giné, G.A.F.,
Guidorizzi, C.E., Jerusalinsky, L., Pendu, Y.L., Marcondes, C.C., Melo, V.L., Mendes,
F.R., Miretzki, M., Neves, L.G., Oliveira, L.C., Pereira, A.R., Silva, K.F.M., Reis,
M.S., Vleeschouwer. K.M., Vieira, E.M., Ximenes, G.E.I., 2017. Primeira avaliação
do status de conservação dos mamíferos do estado da Bahia, Brasil. Oecol. Aust. 21,
156-170. 10.4257/oeco.2017.2102.06
Chao, A., Ma, K.H., Hsieh, T.C., 2016. iNEXT (iNterpolation and EXTrapolation)
Online: software for interpolation and extrapolation of species diversity.
http://chao.stat.nthu.edu.tw/wordpress/software_download/ (accessed 13 March 2020)
Davies, J., Poulsen, L., Schulte-Herbrüggen, B., Mackinnon, K., Crawhall, N., Henwood,
W.D., Dudley, N., Smith, J., Gudka, M., 2012. Conserving dryland biodiversity. IUCN
Drylands Initiative, Nairobi.
Desbiez, A.L., Bodmer, R.E., Aparecida, S., 2009. Wildlife habitat selection and
sustainable resources management in a Neotropical wetland. Int. J. Biodivers.
Conserv. 1, 11–20.
Dias, D.M., 2017. Spatiotemporal ecology of the striped hog-nosed skunk Conepatus
semistriatus (Carnivora, Mephitidae) in a seasonally dry forest of northeastern Brazil.
Anim. Biol. 2, 119 – 131. https://doi.org/10.1163/15707563-00002525
Dias, D.M., Bocchiglieri, A., 2016. Trophic and spatio-temporal niche of the crabeating
fox, Cerdocyon thous (Linnaeus, 1766) (Carnivora: Canidae), in a remnant of the
Caatinga in northeastern Brazil. Mammalia 80, 281–291.
https://doi.org/10.1515/mammalia-2014-0108
170
Dias, D.M., Massara, R.L., Campos, C.B., Rodrigues, F.H.G., 2019a. Human activities
influence the occupancy probability of mammalian carnivores in the Brazilian
Caatinga. Biotropica 51, 253-265. https://doi.org/10.1111/btp.12628
Dias, D.M., Massara, R.L., Campos, C.B., Rodrigues, F.H.G., 2019b. Feline predator-
prey relationships in a semi-arid biome in Brazil. J. Zool. 307, 282-291.
https://doi.org/10.1111/jzo.12647
Dorazio, R.M., Royle, J.A., 2005. Estimating size and composition of biological
communities by modeling the occurrence of species. J. Am. Stat. Assoc. 100, 389–
398. https://doi.org/10.1198/016214505000000015
Ferreguetti, A.C., Tomas, W.M., Bergallo, H.G., 2015. Density, occupancy, and activity
pattern of two sympatric deer (Mazama) in the Atlantic Forest, Brazil. J. Mamm. 96,
1245–1254. https://doi.org/10.1093/jmammal/gyv132
Ferreguetti, A.C., Tomas, W.M., Bergallo, H.G., 2016. Density and niche segregation of
two armadillo species (Xenarthra: Dasypodidae) in the Vale Natural Reserve, Brazil.
Mamm. Biol. 81, 138–145. http://dx.doi.org/10.1016/j.mambio.2015.10.007
Fonseca, C.R., Antongiovanni, M., Matsumoto, M., Bernard, E., Venticinque, E.M.,
2017. Conservation opportunities in the Caatinga, in: Silva, J.M.C., Leal, I.R.,
Tabarelli, M. (Eds.), Caatinga: The largest tropical dry forest region in South America.
Springer International Publishing, Cham, pp. 429-444.
Freitas, R.R., Rocha, P.L.B., Simões-Lopes, P.C., 2005. Habitat structure and small
mammals abundances in one semiarid landscape in the Brazilian Caatinga. Rev. Bras.
Zool. 22, 119–129. http://dx.doi.org/10.1590/S0101-81752005000100015
Gaynor, K.M., Hojnowski, C.E., Cater, N.H., Bashares, J.S., 2018. The influence of
human disturbance on wildlife nocturnality. Science 360, 1232–123.
https://doi.org/10.1126/science.aar7121
171
Giordano, A.J., 2016. Ecology and status of the jaguarundi Puma yagouaroundi: a
synthesis of existing knowledge. Mammal Rev. 46, 30–43.
https://doi.org/10.1111/mam.12051
Goulart, F.V.B., Cáceres, N.C., Graipel, M.E., Tortato, M.A., Ghizoni Jr, I.R., Oliveiras-
Santos, L.G.R., 2009. Habitat selection by large mammals in a southern Brazilian
Atlantic Forest. Mamm. Biol. 74, 182–190.
https://doi.org/10.1016/j.mambio.2009.02.006
Harmsen, B.J., Foster, R.J., Silver, S.C., Ostro, L.E.T., Doncaster, C.P., 2010. Differential
use of trails by forest mammals and the implications for camera-trap studies: a case
study from Belize. Biotropica 42, 126–133. https://doi.org/10.1111/j.1744-
7429.2009.00544.x
IBGE. 2016. Produção da pecuária municipal 2016, forty-fourth vol. Instituto Brasileiro
de Geografia e Estatística IBGE, Rio de Janeiro.
IUCN, 2020. The IUCN Red List of Threatened Species, version 2020-1.
https://www.iucnredlist.org (accessed19 March 2020).
Jędrzejewski, W., Boede, E.O., Abarca, M., Sánchez-Mercado, A., Ferrer-Paris, J.R.,
Lampo, M., Velasquez, G., Carreño, R., Viloria, A.L., Hoogesteijn, R., Robinson,
H.S., Stachowicz, I., Cerda, H., Weisz, M.M., Barros, T.R., Rivas, G.A., Borges, G.,
Molinari, J., Lew, D., Takiff, H., Schmidt, K., 2017. Predicting carnivore distribution
and extirpation rate based on human impacts and productivity factors; assessment of
the state of jaguar (Panthera onca) in Venezuela. Biol. Conserv. 206, 132–142.
http://dx.doi.org/10.1016/j.biocon.2016.09.027
Kéry, M., 2010. Introduction to WinBUGS for ecologists: a bayesian approach to
regression, ANOVA and related analyses. Academic Press, Cambridge.
172
Kinnaird, M.F., O’brien, T.G., 2012. Effects of private-land use, livestock management,
and human tolerance on diversity, distribution, and abundance of large African
mammals. Conserv. Biol. 26, 1026–1039. http://dx.doi.org/10.1111/j.1523-
1739.2012.01942.x
Lopes, S.F., Ramos, M.B., Almeida, G.R., 2017. The role of mountains as refugia for
biodiversity in Brazilian Caatinga: conservationist implications. Tropical
Conservation Science 10, 1-12. https://doi.org/10.1177/1940082917702651
MacKenzie, D.I., Nichols, J.D., Lachman, G.B., Droege, S., Royle, J.A., Langtimm, C.A.,
2002. Estimating site occupancy rates when detection probabilities are less than one.
Ecology 83, 2248–2255. https://doi.org/10.1890/0012-
9658(2002)083[2248:ESORWD]2.0.CO;2
Magioli, M., Ferraz, K.M.P.M.B., Chiarello, A.G., Galetti, M., Setz, E.Z.F., Paglia, A.P.,
Abrego, N., Ribeiro, M.C., Ovaskainen, O., 2020. Defaunation leads to functional loss
in a tropical biodiversity hotspot. Curr. Biol. http://dx.doi.org/10.2139/ssrn.3550245
Marinho, F.P., Mazzochini, G.G., Manhães, A.P., Weisser, W.W., Ganade, G., 2016.
Effects of past and present land use on vegetation cover and regeneration in a tropical
dryland forest. J. Arid Environ. 132, 26–33.
http://dx.doi.org/10.1016/j.jaridenv.2016.04.006
Marinho, P.H., Bezerra, D., Antongiovanni, M., Fonseca, C.R., Venticinque, E.M.,
2018a. Mamíferos de médio e grande porte da Caatinga do Rio Grande do Norte,
nordeste do Brasil. Mastozool. Neotrop. 25, 345–362.
https://doi.org/10.31687/saremMN.18.25.2.0.15
Marinho, P.H., Bezerra, D., Antongiovanni, M., Fonseca, C.R., Venticinque, E.M.,
2018b. Estimating occupancy of the Vulnerable northern tiger cat Leopardus tigrinus
173
in Caatinga drylands. Mamm. Res. 63, 33–42. https://doi.org/10.1007/s13364-017-
0330-4
Marinho, P.H., Fonseca, C.R., Sarmento, P., Fonseca, C., Venticinque, E.M., 2020.
Temporal niche overlap among mesocarnivores in a Caatinga dry forest. Eur. J.
Wildlife Res. 66, 34. https://doi.org/10.1007/s10344-020-1371-6
Marinho, P.H., Silva, M., Lisboa, C.M.C.A., 2019. Presence of the collared peccary
Pecari tajacu (Artiodactyla, Tayassuidae) in the far northeast of its Brazilian
distribution. Neotrop. Biol. Conserv. 14, 499–509.
https://doi.org/10.3897/neotropical.14.e48716
Michalski, F., Peres, C.A., 2007. Disturbance-mediated mammal persistence and
abundance-area relationships in Amazonian forest fragments. Conserv. Biol. 21, 1626-
1640. https://doi.org/10.1111/j.1523-1739.2007.00797.x
MMA, 2014. Lista Nacional Oficial de Espécies da Fauna Ameaçadas de Extinção.
https://www.icmbio.gov.br/portal/images/stories/biodiversidade/fauna-
brasileira/portarias/PORTARIA_N_445_DE_17_DE_DEZEMBRO_DE_2014
Morato, R.G., Ferraz, K.M.P.M.B., Paula, R.C., Campos, C.B., 2014. Identification of
priority conservation areas and potential corridors for jaguars in the Caatinga Biome,
Brazil. Plos One 9, e92950. https://doi.org/10.1371/journal.pone.0092950
Nagy-Reis, M.B., Estevo, C.A., Setz, E.Z.F., Ribeiro, M.C., Chiarello, A.G., Nichols,
J.D., 2017. Relative importance of anthropogenic landscape characteristics on
Neotropical frugivores in a multiple scale perspective. Anim. Conserv. 20, 520–531.
https://doi.org/10.1111/acv.12346
Niemeijer, D., Puigdefabregas, J., White, R., Lal, R., Winslow, M., Ziedler, J., Prince, S.,
Archer, E., King, C., Shapiro, B., Wessels, K., Nielsen, T., Portnov, B., Reshef, I.,
174
Thonell, J., Lachman, E., Mcnab, D., 2005. Dryland systems, in: Hassan, R., Scholes,
R., Neville, A. (Eds.). Ecosystems and human well-being: Current state and trends.
Island Press, Washington, pp. 623–662.
Oliveira, T.G., Tortato, M.A., Almeida, L.B., Campos, C.B., Beisiegel, B.M., 2013.
Avaliação do risco de extinção do gato-do-mato Leopardus tigrinus no Brasil.
Biodivers. Bras. 5, 56-65.
Paglia, A.P., Fonseca, G.A.B., Rylands, A.B., Herrmann, G., Aguiar, L.M.S., Chiarello,
A.G., Leite, Y.L.R., Costa, L.P., Siciliano, S., Kierulff, M.C.M., Mendes, S.L.,
Tavares, V.C., Mittermeier, R.A., Patton, J.L., 2012. Lista anotada dos mamíferos do
Brasil / Annotated checklist of Brazilian mammals, second Ed. Occasional Papers in
Conservation Biology, Conservation International, Arlington.
Prugh, L.R., Hodges, K.E., Sinclair, A.R.E., Brashares, J.S., 2008. Effect of habitat area
and isolation on fragmented animal populations. P. Natl. Acad. Sci.-Biol. 52, 20770-
20775. https://doi.org/10.1073/pnas.0806080105
Pudyatmoko, S., 2017. Free-ranging livestock influence species richness, occupancy, and
daily behaviour of wild mammalian species in Baluran National Park, Indonesia.
Mamm. Biol. 86, 33–41. http://dx.doi.org/10.1016/j.mambio.2017.04.001
R Core Team, 2016. R: A language and environment for statistical computing. R
Foundation for Statistical Computing. https://www.R-project.org/ (accessed 02 May
2016).
Ribeiro, E.M.S., Arroyo-Rodríguez, V., Santos, B.A., Tabarelli, M, Leal, I.R., 2015.
Chronic anthropogenic disturbance drives the biological impoverishment of the
Brazilian Caatinga vegetation. J. Appl. Ecol. 52, 611–620.
http://dx.doi.org/10.1111/1365-2664.12420
175
Rodrigues, T.F., Chiarello, A.G., 2018. Native forests within and outside protected areas
are key for nine-banded armadillo (Dasypus novemcinctus) occupancy in agricultural
landscapes. Agric. Ecosyst. Environ. 266, 133–141.
https://doi.org/10.1016/j.agee.2018.08.001
Rovero, F., Owen, N., Jones, T., Canteri, E., Iemma, A., Tattoni, C., 2017. Camera
trapping surveys of forest mammal communities in the Eastern Arc Mountains reveal
generalized habitat and human disturbance responses. Biodiv. Conserv. 26, 1103–
1119. http://dx.doi.org/10.1007/s10531-016-1288-2
Royle, J.A., Dorazio, R.M., 2008. Hierarchical Modeling and Inference in Ecology.
Academic Press, Boston.
Santini, L., Isaac, N.J.B., Maiorano, L., Ficetola, G.F., Huijbregts, M.A.J., Carbone, C.,
Thuiller, W., 2018. Global drivers of population density in terrestrial vertebrates.
Global Ecol. Biogeogr. 27, 968–979. https://doi.org/10.1111/geb.12758
Schuette, P., Wagner, A.P., Wagner, M.E., Creel, S., 2013. Occupancy patterns and niche
partitioning within a diverse carnivore community exposed to anthropogenic
pressures. Biol. Conserv. 158, 301–312.
http://dx.doi.org/10.1016/j.biocon.2012.08.008
Serbent, M.P., Periago, M.E., Leynaud, G.C., 2011. Mazama gouazoubira (Cervidae) diet
during the dry season in the arid Chaco of Córdoba (Argentina). J. Arid Environ. 75,
87–90. https://doi.org/10.1016/j.jaridenv.2010.09.004
Silva, J.M.C., Leal, I.R., Tabarelli, M. (Eds.), 2017. Caatinga: The largest tropical dry
forest region in South America. Springer International Publishing, Cham.
Stoner, K., Timm, R.M., 2011. Seasonally dry tropical forest mammals: adaptations and
seasonal patterns, in: Dirzo, R., Young, H.S., Mooney, H.A., Ceballos, G. (Eds.),
176
Seasonally dry tropical forests: Ecology and conservation. Island Press, Washington,
pp. 85-105.
Sunderland, T.C.H., Apgaua, D., Baldauf, C., Blackie, R., Colfer, C.J.P., Cunningham,
A.B., Dexter, K., Djoudi, H., Gautier, D., Gumbo, D., Ickowitz, A., Kassa, H.,
Parthasarathy, N., Pennington, R.T., Paumgarten, F., Pulla, S., Sola, P., Tng, D.,
Waeber, P., Wilmé, L., 2015. Global dry forests: a prologue. Int. For. Rev. 17, 1-9.
https://doi.org/10.1505/146554815815834813
Tabeni, S., Ojeda, R.A., 2003. Assessing mammal responses to perturbations in temperate
aridlands of Argentina. J. Arid Environ. 55, 715–726. https://doi.org/10.1016/S0140-
1963(02)00314-2
TEAM Network, 2011. Terrestrial vertebrate protocol implementation manual, v. 3.1.
Conservation International, Arlington.
http://www.teamnetwork.org/files/protocols/terrestrialvertebrate/TEAMTerrestrialVe
rtebrate-PT-EN-3.1.pdf
Tilker, A., Abrams, J.F., Mohamed, A., Nguyen, A., Wong, S.T., Sollmann, R., Niedballa
J, Bhagwat, T., Gray, T.N.E., Rawson, B.M., Guegan, F., Kissing, J., Wegman, M.,
Wilting, A., 2019. Habitat degradation and indiscriminate hunting differentially
impact faunal communities in the Southeast Asian tropical biodiversity hotspot.
Commun. Biol. 2, 396. https://doi.org/10.1038/s42003-019-0640-y
Vasconcellos, A., Andreazze, R., Almeida, A.M., Araujo, H.F.P., Oliveira, E.S., Oliveira,
U., 2010. Seasonality of insects in the semi-arid Caatinga of northeastern Brazil. Ver.
Bras. Entomol. 54, 471–476. https://doi.org/10.1590/S0085-56262010000300019
Velloso, A.L., Sampaio, E.V.S.B., Pareyn, F.G.C. (Eds.), 2002. Ecorregiões propostas
para o bioma Caatinga. Associação Plantas do Nordeste, Instituto de Conservação
Ambiental The Nature Conservancy do Brasil, Recife.
177
Vynne, C., Keim, J.L., Machado, R.B., Marinho-Filho, J., Silveira, L., Groom, M.J.,
Wesser, S.K., 2011. Resource Selection and Its Implications for Wide-Ranging
Mammals of the Brazilian Cerrado. Plos One 6, e28939.
https://doi.org/10.1371/journal.pone.0028939
Yoshihara, Y., Ito, T.Y., Lhagvasuren, B., Takatsuki, S., 2008. A comparison of food
resources used by Mongolian gazelles and sympatric livestock in three areas in
Mongolia. J. Arid Environ. 72, 48–55. https://doi.org/10.1016/j.jaridenv.2007.05.001
Zimbres, B., Peres, C.A., Penido, G., Machado, R.B., 2018. Thresholds of riparian forest
use by terrestrial mammals in a fragmented Amazonian deforestation frontier.
Biodivers. Conserv. 27, 2815–2836. https://doi.org/10.1007/s10531-018-1571-5
Zipkin, E.F., DeWan, A., Royle, J.A., 2009. Impacts of forest fragmentation on species
richness: a hierarchical approach to community modelling. J. Appl. Ecol. 46, 815–822.
https://doi.org/10.1111/j.1365-2664.2009.01664.x
Zipkin, E.F., Royle, J.A., Dawson, D.K., Bates, S., 2010. Multi-species occurrence
models to evaluate the effects of conservation and management actions. Biol. Conserv.
143, 479–484. https://doi.org/10.1016/j.biocon.2009.11.016
178
Supplementary Material
Fig. S1. Medium to large-sized wild mammal species (> 1 kg) recorded in an overgrazed
semiard landscape in Caatinga dry forest, northeastern Brazil. (A) Cerdocyon thous, (B)
Euphractus sexcinctus, (C) Dasypus novemcinctus, (D) Leopardus tigrinus, (E)
Tamandua tetradactyla, (F) Mazama gouazoubira, (G) Conepatus amazonicus, (H)
A
A
B
A
C
A
C
A
E
A
F
A D
A
G
A H
A
I
A
J
A
K
A
L
A
179
Herpailurus yagouaroundi, (I) Dasyprocta nigriclunis, (J) Pecari tajacu, (K) Puma
concolor, and (L) Didelphis albiventris.
Fig. S2. Rarefaction (solid line) and extrapolation curves (dashed line) of medium and large-sized
wild mammal species richness in an overgrazed semiarid landscape in Caatinga dry forest,
northeastern Brazil. The triangle represents the number of recorded species in the area. The 95%
unconditional confidence intervals (gray-shaded regions) were based on 1000 bootstrap
repetitions.
Table S1
Pearson's correlation matrix of the covariates showing the coefficient values (upper diagonal) and
the p values (lower diagonal). High correlations and significant p-values are in bold.
Forest Livest River Road House Rough Elevat
Forest NA -0.06 0.51 0.37 0.44 0.25 0.66
Livest <0.01 NA -0.20 0.12 -0.08 -0.06 -0.17
River <0.01 0.12 NA 0.36 0.42 0.29 0.62
Road <0.01 0.36 0.01 NA 0.20 0.36 0.30
House <0.01 0.52 <0.01 0.12 NA 0.13 0.57
Rough <0.01 0.66 0.03 <0.01 0.31 NA 0.63
Elevat <0.01 0.19 <0.01 0.02 <0.01 <0.01 NA
180
Model code based in Zipkin et al. (2009)
#model code (uses the R2WinBUGS package to run WinBUGS in program R)
sink("modelmulti.txt")
cat("
model{
#Prior distributions on the community level occupancy and detection covariates
psi.mean ~ dunif(0,1)
a <- log(psi.mean) - log(1-psi.mean)
theta.mean ~ dunif(0,1)
b <- log(theta.mean) - log(1-theta.mean)
mu.alpha1 ~ dnorm(0, 0.001)
mu.alpha2 ~ dnorm(0, 0.001)
mu.alpha3 ~ dnorm(0, 0.001)
mu.alpha4 ~ dnorm(0, 0.001)
mu.alpha5 ~ dnorm(0, 0.001)
mu.alpha6 ~ dnorm(0, 0.001)
mu.beta1 ~ dnorm(0, 0.001)
tau1 ~ dgamma(0.1,0.1)
tau2 ~ dgamma(0.1,0.1)
tau.alpha1 ~ dgamma(0.1,0.1)
tau.alpha2 ~ dgamma(0.1,0.1)
tau.alpha3 ~ dgamma(0.1,0.1)
tau.alpha4 ~ dgamma(0.1,0.1)
tau.alpha5 ~ dgamma(0.1,0.1)
tau.alpha6 ~ dgamma(0.1,0.1)
tau.beta1 ~ dgamma(0.1,0.1)
rho ~ dunif(-1,1) # medida de comunidade
var.v <- tau2 /(1.-pow(rho,2))
181
sigma1 <- 1/sqrt(tau1)
sigma2 <- 1/sqrt(tau2)
for (i in 1:species) {
#Prior distributions for the occupancy and detection covariates for each species
u[i] ~ dnorm(a, tau1)
mu.v[i] <- b + (rho*sigma2 /sigma1)*(u[i]-a)
v[i] ~ dnorm(mu.v[i], var.v)
alpha1[i] ~ dnorm(mu.alpha1, tau.alpha1)
alpha2[i] ~ dnorm(mu.alpha2, tau.alpha2)
alpha3[i] ~ dnorm(mu.alpha3, tau.alpha3)
alpha4[i] ~ dnorm(mu.alpha4, tau.alpha4)
alpha5[i] ~ dnorm(mu.alpha5, tau.alpha5)
alpha6[i] ~ dnorm(mu.alpha6, tau.alpha6)
beta1[i] ~ dnorm(mu.beta1, tau.beta1)
#Estimate the occupancy probability (latent Z matrix) for each species at each point
for (j in 1:sites) {
logit(psi[j,i]) <- u[i] + alpha1[i]*forest[j] + alpha2[i]*rough [j] + alpha3[i]*house[j]
+ alpha4[i]*lives[j] + alpha5[i]*river[j] + alpha6[i]*road[j]
Z[j,i] ~ dbin(psi[j,i], 1)
#Estimate the species specific detection probability for every rep at each point where
the
#species occurs (Z=1)
for (k in 1:survey) {
logit(theta[j,k,i]) <- v[i] + beta1[i]*trail[j]
mu.theta[j,k,i] <- theta[j,k,i]*Z[j,i]
X[j,k,i] ~ dbin(mu.theta[j,k,i], 1)
Xnew[j,k,i] ~ dbin(mu.theta[j,k,i], 1)
#Create simulated dataset to calculate the Bayesian p-value
d[j,k,i]<- abs(X[j,k,i] - mu.theta[j,k,i])
dnew[j,k,i]<- abs(Xnew[j,k,i] - mu.theta[j,k,i])
182
d2[j,k,i]<- pow(d[j,k,i],2)
dnew2[j,k,i]<- pow(dnew[j,k,i],2)
}
dsum[j,i]<- sum(d2[j,1:survey,i])
dnewsum[j,i]<- sum(dnew2[j,1:survey,i])
}
}
#Calculate the discrepancy measure, which is then defined as the mean(p.fit >
p.fitnew)
p.fit<-sum(dsum[1:sites,1:species])
p.fitnew<-sum(dnewsum[1:sites,1:species])
}
",fill=TRUE)
sink()
#####
#data
data<-
list("X"=captdata,"survey"=survey,"sites"=sites,"species"=species,"forest"=forest,
"trail"=trail,"house"=house, "rough"=rough,"lives"=lives, "river"=river,
"road"=road)
#inits
zinit <- array(dim = c(sites, species))
for (j in 1:sites) {
for (i in 1:species) {
zinit[j, i] <- max(captdata[j, ,i],na.rm=T)
}
183
}
inits=function(){
list(
Z=zinit,
psi.mean=runif(1, 0, 1),
theta.mean=runif(1, 0, 1),
rho=runif(1, 0, 1),
tau1=runif(1, 0, 1),
tau2=runif(1, 0, 1),
mu.alpha1=runif(1, 0, 1),
mu.alpha2=runif(1, 0, 1),
mu.alpha3=runif(1, 0, 1),
mu.alpha4=runif(1, 0, 1),
mu.alpha5=runif(1, 0, 1),
mu.alpha6=runif(1, 0, 1),
tau.alpha1=runif(1, 0, 1),
tau.alpha2=runif(1, 0, 1),
tau.alpha3=runif(1, 0, 1),
tau.alpha4=runif(1, 0, 1),
tau.alpha5=runif(1, 0, 1),
tau.alpha6=runif(1, 0, 1),
mu.beta1=runif(1, 0, 1),
tau.beta1=runif(1, 0, 1)
)
}
#Parameters
para<-c("u","alpha1","alpha2","alpha3","alpha4","alpha5","alpha6","beta1",
"mu.alpha1","mu.alpha2","mu.alpha3","mu.alpha4","mu.alpha5","mu.alpha6","mu.beta
1",
184
"tau.alpha1","tau.alpha2","tau.alpha3","tau.alpha4","tau.alpha5","tau.alpha6","tau.beta1
",
"v","p.fit","p.fitnew","psi")
#MCMC
nc<-3
nb<-2000
ni<-10000
nt<-3
#gibbs sampler
system.time(outJAGS <-
jags(data=data,inits=inits,parameters.to.save=para,model.file="modelmulti.txt",n.chains
=nc,n.iter=ni,n.burn=nb,n.thin=nt))
185
CONCLUSÕES GERAIS
O conhecimento construído nos últimos anos, tem demonstrado a rica
biodiversidade da Caatinga brasileira e a relevância ecológica dessa floresta tropical
sazonalmente seca nos neotrópicos, embora esforços de conservação ainda estejam muito
aquém da sua importância. Essa tese traz à luz a diversidade de espécies e estratégias de
persistência de mamíferos de médio e grande porte nessa região semiárida.
Especificamete, demonstramos a relevância da mastofauna terrestre do Rio Grande do
Norte, até então desconhecida, através da caracterização das comunidades de 10 áreas
prioritárias para a conservação, que abrigam pelo menos 14 espécies de médio e grande
porte, incluindo felinos ameaçados de extinção como Puma concolor, que parece ser
extremamente raro na região atualmente. Este estudo deve representar um ponto de
partida para novas pesquisas sobre a mastofauna terrestre do estado, além de subsidiar o
planejamento e a execução de ações regionais de conservação e manejo.
Ao direcionarmos um maior esforço amostral de monitoramento de uma das áreas
prioritárias para a conservação mais relevantes do estado, a única reconhecida até aqui
como habitat de sete espécies de mesocarnívoros, e onde predadores de topo como Puma
concolor estão funcionalmente extintos, nós ajudamos a esclarecer os padrões de
coexistência espaço-temporal de mesocarnívoros na Caatinga. Neste ambiente as espécies
sobrepoem grande parte da sua atividade noturna enquanto segregam os picos de maior
atividade, o que parece ser uma estratégia que envolve o balanço entre evitar encontros
agressivos e as elevadas temperaturas durante o dia. Isso é reforçado pela ausência de
segregação espacial das demais espécies de mesocarnívoros com Leopardus pardalis, o
mesopredador dominante. Conforme esperado, a exceção é Herpailurus yagouaroundi,
que ao manter hábitos diurnos nesse ambiente extremo, pode diminuir a competição e as
chances de ataques intraguilda e utilizar ambientes mais frequentados pela espécie
186
dominante, L. pardalis. Os padrões encontrados se mantém relativamente inalterados ao
longo das estações seca e chuvosa, apesar da marcada flutuação de recursos. Investigar o
uso do habitat e os padrões de atividade das espécies em ambientes com diferentes níveis
de integridade da guilda de carnívoros, bem como suas dietas, pode elucidar ainda mais
os mecanismos de manutenção da diversidade desse grupo em ecossistemas semiáridos
Caatinga.
Por fim, observamos que manchas de vegetação com estrutura florestal representam
habitats chave para mamíferos de médio e grande porte em uma paisagem da Caatinga
dominada pela criação extensiva de gado. Apesar da intensa degradação ambiental,
espécies relevantes de mamíferos, tais como Pecari tajacu e Puma concolor, persistem
na área, porém, aparentemente, em baixas abundâncias. Nossos resultados, portanto,
ressaltam características da paisagem que podem beneficiar a persistência desses
mamíferos, como uma maior proporção de vegetação arbórea, fornecendo informações
importantes para o manejo e a expansão de áreas protegidas, bem como para o
monitoramento do estado de conservação das populações.
187
“Todo mundo sabe tratar-se de problema antes de tudo educacional. Mas o
indeferentismo com que vem sendo relegado faz pensar, sem qualquer pessimismo, que
na pisada em que vamos, o sertanejo herdará, em um amanhã bem próximo, um chão
sem rastros de bichos e silencioso de cantos dos pássaros. Paisagem morta e de fauna
sintética já galhofada no dizer matuto: “De bicho de cabelo só vai escapar escova; de
animal de quatro pés, tamborete e bicho de fôlego – o fole...””.
Oswaldo Lamartine (A Caça nos Sertões do Seridó, 1961)
Mata de galeria do riacho Melancia, antigo habitat da extinta ararinha-azul (Cyanopsitta spixii), Curaçá, Bahia.