4991_Therya_BrazilianMammals

THERYA, 2023, Vol. 14(3):329-341 DOI:10.12933/therya-23-4991 ISSN 2007-3364

Influence of forest type on the diversity, abundance, and naïve occupancy of the mammal assemblage in the southeastern Brazilian Atlantic Forest

Alejandra Soto-Werschitz1, 2*, Salvador Mandujano3, and Marcelo Passamani1

1 Programa de Pós-Graduação em Ecologia Aplicada, Instituto de Ciências Naturais, Departamento de Ecologia e Conservação, Laboratório de Ecologia e Conservação de Mamíferos, Universidade Federal de Lavras. C. P 37.200-900, Lavras. Minas Gerais, Brasil. Email: alewerschitz@gmail.com (AS-W); mpassamani@ufla.br (MP).

2 Universidad de Los Andes, Departamento de Biología. C. P. 5101, Mérida. Mérida, Venezuela.

3 Red de Biología y Conservación de Vertebrados, Instituto de Ecología. A.C., Antigua Carretera a Coatepec 351, El Haya C. P. 91070, Xalapa. Veracruz, México. Email: salvador.mandujano@inecol.mx (SM).

* Corresponding author: https://orcid.org/orcid0000-0001-5416-632X

The Brazilian Atlantic Forest has the highest rate of native vegetation destruction, which is one of the principal drivers of mammal extinctions. Therefore, reducing information gaps regarding diversity patterns, abundance, and habitat use is crucial to understand mammal persistence in fragmented landscapes. Our objective was to establish the γ diversity and to assess the extent to which the α, and β diversity, the relative abundance, and naïve occupation of medium and large-sized mammal communities differ between seasonal and ombrophilous forests. Between January 2019 and March 2020, we placed 22 camera traps in the Atlantic Forest of Minas Gerais. We calculate Hill's numbers using iNEXT.4steps package, the β-diversity with the Betapart package, as well as the relative abundance index (RAI), and naïve occupancy (PAO). We used Kruskal-Wallis and Mann-Whitney statistical tests to compare the RAIs between different species and forest types. Finally, we calculated the correlation between the RAIs and PAOs. We found 32 species, principally from the orders Carnivora and Artiodactyla. The alpha diversity and evenness profiles were not different between the two forest types (seasonal q0 = 0.91, q1 = 0.99, q2 = 1, J = 0.83; ombrophilous q0 = 0.96, q1 = 0.99, q2 = 1, J = 0.85). The beta diversity was low (βJAC = 0.37) which was mostly associated with species turnover (βJTU = 0.34), while nestedness was almost non-existent (βJNE = 0.02). The RAIs varied among mammalian species (H = 115.24, P = 0.000), with the highest values for Didelphis aurita (RAI = 4.55 ± 7.66) and Cuniculus paca (RAI = 2.35 ± 3.73) and the minor values for Speothos venaticus (RAI = 0.04 ± 0.24) and Galictis cuja (RAI = 0.06 ± 1.19). The RAIs of species was not significantly different between forests (U = 453.5; Z = 0.37; P = 0.70), and only Leopardus wiedii showed significant differences between forests (U = 84.5; P = 0.01). Most of the mammalian species had restricted occupancy to a few localities (< 50 %). The species Eira barbara and Didelphis aurita had the highest PAOs in both forests (> 50 %), and the species Tayassu pecari, Tamandua tetradactyla, and Speothos venaticus, the lowest values (5 %). We found a correlation of 75 % between the average RAI and naïve occupancy. The γ diversity was representative and consistent with the species found in the Atlantic Forest, and the relative abundance and naïve occupancy reflected the rarity of most species in the area. Additionally, the only difference between the two forests corresponds to species turnover. Therefore, we must conserve native remnants of both forests to ensure the existence of native mammals, mainly the most threatened species, to prevent more dramatic scenarios of local extinction in Minas Gerais.

La Mata Atlántica Brasileña presenta la mayor tasa de destrucción de vegetación nativa, causa principal de la extinción de mamíferos. Por lo tanto, reducir las lagunas de información relacionadas con los patrones de diversidad α, β y γ, abundancia, ocupación y uso de hábitat es una prioridad, para comprender la persistencia de mamíferos en paisajes fragmentados. Nuestro objetivo fue establecer la diversidad γ, y evaluar en qué medida la diversidad α y β, la abundancia relativa y la ocupación naïve de las comunidades de mamíferos medianos y grandes, contrastan entre bosques estacionales y ombrófilos. Entre enero (2019) y marzo (2020), colocamos 22 cámaras trampa en el Bosque Atlántico de Minas Gerais. Calculamos los números de Hill utilizando el paquete iNEXT.4steps, la β-diversidad con el paquete Betapart, el índice de abundancia relativa (RAI) y la ocupación naïve (PAO). Utilizamos pruebas de Kruskal-Wallis y Mann-Whitney para comparar los RAI entre especies y bosques. Finalmente, calculamos la correlación entre los RAIs y PAOs. Registramos 32 especies, principalmente de los órdenes Carnivora y Artiodactyla. La diversidad alfa y los perfiles de uniformidad no difirieron entre bosques (estacional q0 = 0.91, q1 = 0.99, q2 = 1, J = 0.83; ombrófilo q0 = 0.96, q1 = 0.99, q2 = 1, J = 0.85). La beta-diversidad fue baja (βJAC = 0.37), correspondiente al recambio de especies (βJTU = 0.34), y a una anidación casi inexistente (βJNE = 0.02). Los RAIs difirieron entre especies (H = 115.24, P = 0.000), los valores más altos fueron para Didelphis aurita (RAI = 4.55 ± 7.66) y Cuniculus paca (RAI = 2.35 ± 3.73), y los valores más bajos para Speothos venaticus (RAI = 0.04 ± 0.24) y Galictis cuja (RAI = 0.06 ± 1.19). Los RAIs no difirieron entre bosques (U = 453.5; Z = 0,37; P = 0.70), y exclusivamente Leopardus wiedii mostró diferencias significativas entre bosques (U = 84.5; P = 0.01). La mayoría de las especies presentaron una ocupación restringida (< 50 %). Eira barbara y Didelphis aurita presentaron los valores más altos en ambos bosques (> 50 %), mientras que Tayassu pecari, Tamandua tetradactyla y Speothos venaticus, los valores más bajos (5 %). Observamos una correlación del 75 % entre los RAIs y PAOs. La diversidad γ fue representativa de la riqueza de especies reportada para la Mata Atlántica. Los RAIs y PAOs reflejaron la rareza de la mayoría de las especies. Adicionalmente, la única diferencia entre los dos bosques correspondió a la rotación de especies. Por lo tanto, debemos conservar los remanentes nativos de ambos bosques para asegurar la existencia de los mamíferos nativos, principalmente las especies más amenazadas, para evitar escenarios más dramáticos de extinción local en la Mata Atlántica de Minas Gerais.

Keywords: Biodiversity; Brazil; conservation; habitat use; mammals; relative abundance index.

Introduction

The Atlantic Forest is widely recognized as a biodiversity hotspot (Myers et al. 2000), yet it is also one of the most severely defaunated subregions on the planet, primarily due to anthropogenic drivers (Bogoni et al. 2020; Galetti et al. 2021). It harbors one of the world's most depleted mammal faunas, with an average historical loss of 62 % (Bogoni et al. 2020), primarily resulting from habitat loss, landscape fragmentation (Ribeiro et al. 2009; Haddad et al. 2015; Bogoni et al. 2018; Püttker et al. 2020), agricultural expansion, wildlife trafficking, urbanization, industrial development (Campanili and Schäffer 2010) and lack of protected areas (Bogoni et al. 2020).

Brazil is home to 770 mammal species, of which 384 are found in the Atlantic Forest (Abreu et al. 2021). Among these, 262 are terrestrial, and 109 are endemic to the region (Figueiredo et al. 2021). The state of Minas Gerais contains roughly 70 % of all mammal species found in the Brazilian Atlantic Forest, including nearly 60 endemic species (Campanili and Schäffer 2010). Unfortunately, Minas Gerais also harbors 45 species threatened with extinction, according to the List of Endangered Fauna Species in the State of Minas Gerais (COPAM 2010).

Mammals in the region are highly sensitive to changes in landscape structure (Regolin et al. 2020), floristic composition (Galetti et al. 2009), alterations in species abundance (Fahrig et al. 2019), changes in community composition (Beca et al. 2017), and the size of assemblages (Bogoni et al. 2020a). This vulnerability has resulted in the local extinctions of several mammal species (Bogoni et al. 2020), as well as alterations in ecological processes (Oliveira et al. 2020) and ecosystem services (Bogoni et al. 2020). The degree of threat faced by mammal species in Minas Gerais justifies the urgent need for conservation action in the region.

Insufficient data on population sizes, distribution, and occurrence of many threatened mammal species in Minas Gerais included in the Brazilian Red List pose challenges to understanding their conservation status (Corrêa et al. 2021). Particularly for rare species, data gaps occur mainly in the interior of forests, highlighting the need for increasing sampling efforts in the Atlantic Forest to identify priority areas for conservation (Corrêa et al. 2021; Figuereido et al. 2021). Among various monitoring methods, camera trapping has proven highly effective for medium to large mammals in neotropical forests (Srbek-Araujo and Chiarello 2005; Rovero and Spitale 2016), providing standardized data at regional or global scales (Wearn et al. 2019). This method can be especially useful for obtaining information about uncommon, rare, or cryptic species, supporting decision-making for biodiversity conservation in hotspots such as the Atlantic Forest (Trolliet et al. 2014).

Species richness, abundance, and distribution patterns are influenced by a variety of environmental and landscape factors, which can vary depending on the spatial scale (Bogoni et al. 2017). While forest type may have a greater impact on species response at smaller scales, larger-scale patterns remain poorly understood (Bogoni et al. 2016). Therefore, it is crucial to gather additional data (De Barros et al. 2021) on species abundance (Galetti et al. 2009; Bogoni et al. 2020), diversity, habitat use, species occurrence (Beselga et al. 2010; Vilas et al. 2022), and assemblage organization to better comprehend the ecological requirements of mammals across various habitat types (Regolin et al. 2020; De Barros et al. 2021; Vilas et al. 2022).

It is imperative to understand how native vegetation contributes to the persistence of many species in fragmented agricultural landscapes (Ikin et al. 2014). This includes identifying the α and β-diversity components to gain insight into the processes that drive differences in species assemblages (Regolin et al. 2020). Such knowledge can serve as a baseline for making comparisons between current, past, and future species richness, and for inferences about species distributions and differences across locations (Cooke et al. 2019; Steinbeiser et al. 2019). Additionally, it can contribute to the management of natural areas (Corrêa et al. 2021) and land-use decisions (Wearn et al. 2017), to expand and strengthen public policies and good agricultural practices. This, in turn, can support the implementation of Brazil's Vegetation Protection Law aimed at restoring native vegetation in the Atlantic Forest (Brancalion et al. 2016).

Our study aimed at estimating gamma diversity and comparing alpha and beta diversity, relative abundance, and naïve occupation of medium and large-sized mammals between seasonal and ombrophilous forests. Our specific objectives were: (a) analyzing the mammalian gamma diversity, abundance, and naïve occupation in the Atlantic Forest; (b) comparing mammal alpha diversity, abundance, and naïve occupation between seasonal and ombrophilous forests; and (c) assessing beta-diversity, species turnover and species nesting between the forests. We hypothesized that forest type would affect mammalian assemblages, resulting in significant differences between α-diversity, abundance, and occupancy, mainly due to differences in species' habitat use. Furthermore, we expected a high beta-diversity (>70 %) explained by the nestedness of species between forests.

Materials and methods

Mammal sampling. This study was conducted in 22 fragments located in the southeastern Brazilian Atlantic Forest, spanning 15 municipalities in Minas Gerais and one in Rio de Janeiro (-46° 0’ W to -43° 0’ W, -23° 0’ S to -21° 0’ S; Figure 1; Supplementary material 1), covering an area of approximately 22,049 km2. Fragment selection was based on satellite images from Google Earth Pro software and updated maps of remaining forest fragments provided by the SOS Mata Atlântica Institute (2013/2014). The selection criteria included the presence of both seasonal and ombrophilous forest, with a minimum distance of 2.85 km between fragments.

To survey medium- and large-sized mammals (≥1 kg), we utilized one camera trap per sampling point (Bushnell® HD Bushnell Outdoor Products, California, USA) in each of the 22 fragments. The camera trap in each fragment operated continuously for an average of 130 days between January 2019 and March 2020, capturing images 24 hours per day. To ensure independence between pictures and avoid pseudo-replication, we implemented a 24-hour interval between pictures of the same species (Porfirio et al. 2014).

The study area consisted of a highly fragmented landscape surrounded by agricultural matrices and villages. The regional relief is rugged and occurs at elevations ranging from 887 to 2,087 masl. The lower elevations are predominantly characterized by a humid temperate climate with dry winters and hot summers (Cwa type), while the mountainous areas have dry winters and rainy and moderately hot summers (Cwb type), according to the classification by the Instituto Brasileiro de Geografia e Estatística (IBGE 2012).

We organize the data using the Wild.ID Program 0.9.31 (Conservation International 2018). We identified the species at the lowest taxonomic level possible, grouping the records of Mazama americana and Mazama gouazoubira, as Mazama sp., due to the difficulty of differentiating between these two species.

Gamma, alpha, and beta diversity. We assessed the gamma diversity (Whitaker 1972) of the southeastern Brazilian Atlantic Forest using the TEAM library program 1.7.R for Windows (Rovero and Spitale 2016). To evaluate the sampling effectiveness and compare it with the richness of native species in other Atlantic Forest studies, we excluded domestic species. Accumulation curves were generated with the Vegan package (Oksanen et al. 2013) and used the non-parametric Jackknife 1 estimator from the Biodiversity R package (Kindt and Coe 2005), which is known for its precision in reducing bias of estimated values (Burnham and Overton 1979).

To compare the alpha diversity of seasonal and ombrophilous forests, we used the R package iNEXT.4steps online for rarefaction and extrapolation of species diversity with Hill numbers based on abundance data (Chao et al. 2020). We performed fifty bootstraps to calculate 95 % confidence intervals. Hill numbers included species richness (q = 0), Shannon diversity (q = 1), and Simpson diversity (q = 2; Chao et al. 2014; Hsieh et al. 2016). We followed a four-step procedure to assess each forest: (a) sample completeness profile, (b) size-based rarefaction/extrapolation, (c) asymptotic and empirical diversity profiles, (d) non-asymptotic coverage-based rarefaction and extrapolation analysis, and (e) evenness profile among species abundances (Pielou J), derived from the slopes of the diversity profile (Chao et al. 2020).

To compare species composition between seasonal and ombrophilous forests, we used the Partitioning Beta Diversity into Turnover and Nestedness Components package (betapart package; Baselga and Orme 2012; Baselga et al. 2021) in R v4.1.0 (R Core Team 2021). We computed the beta.JTU value of turnover, measured as the turnover fraction of Simpson's dissimilarity, the beta.JNE of the nesting component, measured as the resulting fraction of Simpson's dissimilarity nesting, and the beta.JAC value of overall beta diversity. We computed using a resampling procedure, taking 100 random samples. We performed this analysis by a) grouping native and non-native species and b) considering only native species.

Species abundance. We calculated the relative abundance index (RAI) for each species as follows: RAI ij = Nj/days j * 100 nights traps, where “Nj” is the number of independent records of the species, “i” is each species, and “j” is each landscape (Mandujano and Pérez 2019), throughout the study area and at each forest. We performed all analyses with the RAI.1 package (Mandujano and Pérez 2019). Finally, we used the Kruskal-Wallis and Mann-Whitney statistical tests to compare the RAIs between species and forests. To determine the magnitude of the differences obtained with the Kruskal Wallis test, we computed the epsilon squared (Tomczak and Tomczak 2014). In all the comparative analyses carried out, we used a significance level of p < 0.05.

Species distribution in the study area. We calculated the naïve occupancy, also known as the Percent of Area Occupied (PAO; MacKenzie and Kendall 2002; sensu Kéry and Royle 2015), throughout the study area and at each type of forest. It was calculated as the number of camera trap sites occupied by each species divided by the total number of sites surveyed. The maximum occupancy value is 1, which indicates that a species occupies 100 % of the sites surveyed. Subsequently, to determine the influence of spatial distribution on species abundance, we calculated the correlation between the PAO and the RAI for each species (Mandujano and Pérez 2019). We performed all analyses using the R program 4.2.1 (R Core Team 2021).

Results

Estimation of mammal gamma, alpha and beta diversity. Based on 2,856 camera traps per day, we recorded 589 independent camera trap events of mammalian species in the entire region. The mammal richness in the study area was 32 species (Figures 2-3), distributed among 26 genera, 9 orders, and 17 families (Table 1). Native mammal species richness was 26, distributed among 22 genera, 8 orders, and 15 families (Table 1). The accumulation curve almost reached the asymptote, and according to the expected richness (Jackknife 1 = 29 ± 2 species) of mammalian species, the sampling effort was deemed sufficient. Therefore, we obtained a representative percentage (90 %) of the native species in the area.

The orders Carnivora and Artiodactyla were the most represented, with five and four families, respectively. The order Rodentia was represented by two families, while Cingulata, Didelphimorphia, Lagomorpha, Perissodactyla, Pilosa, and Primates were each represented by one family. Most of the recorded species were native to the Atlantic Forest (81 %), while 16 % were domestic and 3 % exotic (Table 1). Additionally, we recorded one new species in southeastern Minas Gerais, the bush dog Speothos venaticus (Soto-Werschitz et al. 2023).

There are some differences in the classification of threatened species at the global (International Union for the Conservation of Nature IUCN 2021), national (Instituto Chico Mendes de Conservação da Biodiversidade ICMBio 2018), and state levels (COPAM 2010). Among the species observed on the List of Endangered Species of Fauna in Minas Gerais state, the species Leopardus pardalis and L. guttulus, Pecari tajacu, Puma concolor, and Tamandua tetradactyla are classified as vulnerable; L. wiedii is classified as threatened and Sapajus nigritus, Speothos venaticus, and Tayassu pecari are classified as critically endangered (COPAM 2010; Table 1). The other species recorded in the study area are classified as Least Concern (IUCN 2021; ICMBio 2018) or are not included in the state list (COPAM 2010).

We collected a representative number of species in seasonal and ombrophilous forests (24 and 25 respectively), resulting in a sample completeness profile (seasonal q0 = 0.91, q1 = 0.99, q2 = 1; ombrophilous forest q0 = 0.96, q1 = 0.99, q2 = 1; Figures 4a, b, c; Supplementary material 2). Size-based rarefaction and extrapolation analysis and the asymptotic empirical diversity profiles suggest that our asymptotic diversity estimates for these forests are reliable to infer true diversities (q = 1 and q = 2; Figures 4b, c; Supplementary material 2). We calculated diversity and evenness measures up to a standardized coverage value of 99.8 % (Cmax). The alpha diversity (q = 0, q = 1, q = 2, and evenness profiles among species abundances) were similar between the two forests (Figures 4a, b, c, d, and e; Supplementary material 2). We observed that the profile curve decreased sharply in both vegetations, and the values of Pielou J index indicated an equal evenness among species abundances in each forest (seasonal J = 0.83; ombrophilous J = 0.85; Figure 4-e; Supplementary material 2).

We found that the overall beta diversity between sites was βJAC = 0.37, with a turnover of βJTU = 0.34 and species nestedness of βJNE = 0.02 (Figure 5a). When considering only the native species, the total beta diversity between sites was βJAC = 0.28, with a turnover of βJTU = 0.25 and species nestedness of βJNE = 0.03 (Figure 5b). In each forest, we observed a small number of species exclusive to it. Specifically, we found six species unique to the seasonal forest (Dasypus septemcinctus, Didelphis albiventris, Speothos venaticus, Tamandua tetradactyla, Equus ferus, and Felis silvestris), and seven species were unique to the ombrophilous forest (Euphractus sexcinctus, Galictis cuja, Herpailurus yagouaroundi, Leopardus wiedii, Tayassu pecari, Bos taurus, and Equus africanus). However, the overall beta diversity is primarily driven by turnover, while nestedness is almost absent.

Relative abundance. The Relative Abundance Indices (RAIs) between species were significantly different (H = 115.24, d.f. = 30, P = 0.000), but the effect size was not strong (Epsilon-squared ε2 = 0.16). Four species, Didelphis aurita (RAI = 4.55 ± 7.66), Cuniculus paca (RAI = 2.35 ± 3.73), Pecari tajacu (RAI = 1.65 ± 7.1), Eira barbara (RAI = 1.65 ± 6.03), had the highest RAIs, (Figure 6a; Supplementary material 3). In contrast, the species with the lowest abundance indices were Speothos venaticus (RAI = 0.04 ± 0.24) and Galictis cuja (RAI = 0.06 ± 1.19; Figure 6a; Supplementary material 3). Among the non-native species, Canis lupus familiaris (RAI = 1.08 ± 2.55) had the highest relative abundance, while the Sus scrofa (RAI = 0.15 ± 0.56), and Equus ferus (RAI = 0.04 ± 0.18), Equus africanus (RAI = 0.04 ± 0.62), and Bos taurus (RAI = 0.06 ± 0.27), the lowest values (Figure 6a; Supplementary material 3).

The relative abundance of species did not show significant differences between the two forests (U = 453.5; Z = 0.37; P = 0.70). However, the species Leopardus wiedii showed significant differences in relative abundance between forests (U = 84.5; P = 0.01; Figure 6b). In the seasonal forest, the species Didelphis aurita (RAI = 4.70 ± 8.28), Puma concolor (RAI = 2.40 ± 7.55), and Sylvilagus brasiliensis (RAI = 2.38 ± 8.91) showed the highest abundance indices. On the other hand, Speothos venaticus (RAI = 0.08 ± 0.03) and Sus scrofa (RAI = 0.08 ± 0.22) had the lowest abundance indices (Figure 6b; Supplementary material 3). In the ombrophilous forest, the species Didelphis aurita (RAI = 4.28 ± 6.97), Cuniculus paca (RAI = 3.26 ± 5.3), and Leopardus guttulus (RAI = 2.23 ± 2.16) showed the highest abundance indices. Finally, the species Cabassous unicinctus (RAI = 0.08 ± 0.22) and Sapajus nigritus (RAI = 0.08 ± 0.22) showed the lowest indices of abundance (Figure 6-b; Supplementary material 3).

Distribution of species. The naïve occupancy of mammal species in the study area did not reach totality (naïve < 1; detection rate, min = 0, max = 0.68). Most of the mammalian species had restricted occupancy to a few localities and values lower than 50 % (Supplementary Material 3). The highest values were for the species Eira barbara (68 %), Didelphis aurita (55 %), Cuniculus paca (45 %), Pecari tajacu and Puma concolor (41 %), Leopardus guttulus and Leopardus pardalis (36 %). Instead, the species Tayassu pecari, Tamandua tetradactyla, and Speothos venaticus presented the lowest values (5 %). Regarding the non-native species, Canis lupus familiaris had the highest naïve occupancy values (32 %), whereas Sus scrofa (9 %), Bos taurus, Equus ferus, Equus africanus, and Felis silvestris had the lowest values (5 %; Supplementary Material 3).

In the seasonal forest, the naïve occupancy of the 24 mammal species ranged from 7 % to 57 %. The species with the highest values of naïve occupancy were Eira barbara (57 %), Didelphis aurita (50 %), Puma concolor, and Cuniculus paca (43 %; Supplementary Material 3). In contrast, in the ombrophilous forest, the naïve occupancy of the 25 mammal species ranged from 12 % to 88 %. The species with the highest values were Eira barbara (88 %), Didelphis aurita, Pecari tajacu, and Leopardus guttulus (62 %; Supplementary Material 3). Finally, we found a correlation of 75 % between the average RAI and naïve occupancy (r2 = 0.75). The species Didelphis aurita, Cuniculus paca, Pecari tajacu and Eira barbara showed the high occupancy and RAIs (Figure 7; Supplementary Material 3).

Discussion

The gamma diversity was 32 mammalian species, which accounts for approximately 12 % of the terrestrial non-flying mammal species recorded in the Brazilian Atlantic Forest (Figueiredo et al. 2021). The richness of species found in our study area was similar to the reported in other localities within the Atlantic Forest, which varies from 17 to 39 species (Bogoni et al. 2016, 2017, 2018; Souza et al. 2019; Ríos et al. 2021). Therefore, our results suggest that the study area is preserving a representative fraction of the original richness of medium to large-sized in the Atlantic Forest. Furthermore, our study demonstrates that camera trapping is a suitable method to record rare, uncommon, and cryptic species (Figueiredo et al. 2021; Trolliet et al. 2014; Wearn et al. 2019), such as Speothos venaticus.

The order Carnivora exhibited the highest species richness, which is consistent with previous studies in the Atlantic Forest (Santos et al. 2016; Bogoni et al. 2018; Souza et al. 2019). Mesocarnivores accounted for the majority of the diversity (87 %), supporting the findings of Bogoni et al. (2018, 2020), Regolin et al. (2017), and the mesopredator release hypotheses (Crooks and Soulé 1999), which explains how the extirpation of apex predators benefits medium-sized carnivores. The detection of the puma, the only large mammal species in the area, is particularly noteworthy in the Atlantic Forest of Minas Gerais due to its crucial role in ecosystem structure, services, and functioning (Botelho et al. 2018).

Contrary to our expectations, we found no evidence for differences in alpha diversity or evenness profiles between the two forests, and the β-diversity values (<37 %), corresponded almost entirely to species turnover. Therefore, the expected degree of species homogenization in the Atlantic Forest in Minas Gerais, was not found in the sampling area. The observed β-diversity could be due not only to differences between vegetation types but also likely to a mixture of factors such as body size (Bogoni et al. 2017), activity areas (Botelho et al. 2018), and habitat loss and fragmentation (Püttker et al. 2020).

Of the registered mammals, 35 % corresponded to species on the List of Endangered Species of Fauna in Minas Gerais, which explains their restricted distribution (naïve <0.5), and the low relative abundances (RAI <1). Our results confirm that variation in species abundance and occupancy of mammals in the Atlantic Forest (Souza et al. 2019), but in our case the species with the highest RAI were Didelphis aurita and Cuniculus paca. Additionally, our results confirm the high levels of defaunation present in the Atlantic Forest (Bogoni et al. 2016, 2018; Ríos et al. 2021), corresponding to 46 % (Galetti et al. 2021).

The only mammal species with the highest abundance indices and naïve occupancy in the seasonal and ombrophilous forest was Didelphis aurita. This didelphid exhibits strong plasticity in habitat, resource use, and higher matrix tolerance (Bogoni et al. 2016), which probably explains its presence in landscapes with agricultural matrices such as those in this study. Furthermore, we must consider that species with small home ranges have a higher probability of detection (Santos et al. 2021). Therefore, it is likely that we overestimated their abundance by photographing the same individuals several times at some localities, because the cameras were placed in areas that covered their home range.

Surprisingly, the relative abundance and occupancy values of non-native species were not high in the sampled areas. The specie Canis lupus familiaris had a moderate abundance and a restricted distribution (RAI = 1.8; 32 % naïve occupancy), while Sus scrofa had one of the lowest values (RAI = 0.15; 9 % naïve occupancy). It is possible that they have not yet extended their population to the areas we sampled, or that we underestimated their abundance. Therefore, we should not ignore their presence as these species can cause increased disturbances to native fauna, predation, competition for resources, and transmission of diseases (Quintela et al. 2020). The wild boar continues to expand their distribution in the forest remnants (Galetti et al. 2021; Rosa et al. 2016). Therefore, we need planning management strategies to control their presence (De Assis Morais et al. 2020; Rosa et al. 2016).

Our study highlights the rarity of most species in the Atlantic Forest and provides valuable insights into the habitat use of mammals. Notably, while both forests exhibit similar species richness, their compositions differ. Therefore, we emphasize the importance of conserving all remaining native forest remnants, to favor particularly rare and threatened species (35 %) that are susceptible to local extinctions in the Minas Gerais Atlantic Forest (Ferreira et al. 2020).

In conclusion, in order to prevent further extinctions and mitigate the risk of dramatic scenarios, it is crucial to understand the alpha and beta diversity of forest fragments in different landscapes (Bogoni et al. 2016). We need to prioritize restoration programs and establish functional connectivity between fragments to ensure species dispersal between the two forests and across landscapes (Beselga 2010; Bogoni et al. 2018).

Acknowledgments

This project received partial funding from the Biodiversity Conservation in Hyperfragmented Tropical Forest Project (Hyper-Frag) supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq-Brazil). A. Soto-Werschitz received a Ph.D. scholarship from the PAEC-OAS-GCUB Program: Organization of American States (OEA), International Cooperation Group of Brazilian Universities (GCUB), Alliance Program for Education and Training (PAEC) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES-Brazil). We also express our gratitude to all the reviewers of this article for their valuable contributions.

Literature cited

Abreu, E. F., et al. 2021-2. Lista de Mamíferos do Brasil. Dataset Open Access. Zenodo. Comitê de Taxonomia da Sociedade Brasileira de Mastozoologia (CT-BMz). https://doi.org/10.5281/zenodo.5802047.

de Assis Morais, T., et al. 2020. The influence of population-control methods and seasonality on the activity pattern of wild boars (Sus scrofa) in high-altitude forests. Mammalian Biology 100:101-106.

De Barros, R.A., et al. 2021. The value of a small urban green area to the medium and large-sized mammals conservation. Research, Society and Development 10:e11710817043.

Beca, G., et al. 2017. High mammal species turnover in forest patches immersed in biofuel plantations. Biological Conservation 210:352-359.

Baselga, A., 2010. Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography 19:134-143.

Baselga, A., and C. D. L. Orme. 2012. Betapart: an R package for the study of beta diversity. Methods in Ecology and Evolution 3:808–812.

Baselga, A., et al. 2021. Betapart: partitioning beta diversity into turnover and nestedness components. R package ver. 1.5.2.

Bogoni, J. A., et al. 2016. Landscape features lead to shifts in communities of medium- to large-bodied mammals in subtropical Atlantic Forest. Journal of Mammalogy 97:713-725.

Bogoni, J. A., et al. 2017. What would be the diversity patterns of medium- to large-bodied mammals if the fragmented Atlantic Forest was a large metacommunity? Biological Conservation 211:85-94.

Bogoni, J. A., et al. 2018. Wish you were here: How defaunated is the Atlantic Forest biome of its medium- to large-bodied mammal fauna? Plos One 13:e0204515.

Bogoni, J. A., C.A. Peres, and K. M. Ferraz. 2020. Extent, intensity and drivers of mammal defaunation: a continental-scale analysis across the Neotropics. Scientific Reports 10:14750.

Bogoni, J. A., C.A. Peres, and K. M. Ferraz. 2020a. Effects of mammal defaunation on natural ecosystem services and human well being throughout the entire Neotropical realm. Ecosystem Services 45:101173.

Bothelo, A. L. M., L. H. M. Borges, and B. Mcfarland. 2018. Abundance and composition of the medium to large-sized mammals in a private area of a REDD+ project in Acre, Brazil. Biota Neotropica 18:e20170487.

Brancalion, P. H., et al. 2016. A critical analysis of the Native Vegetation Protection Law of Brazil (2012): updates and ongoing initiatives. Natureza and Conservação14:1-15.

Burnham, K. P., and W. S. Overton. 1979. Robust estimation of population size when capture probabilities vary among animals. Ecology 60:927-936.

Campanili, M., and W. B. Schäffer (eds.). 2010. Mata Atlântica: patrimônio nacional dos brasileiros. Ministério do Meio Ambiente. Secretaria de Biodiversidade e Florestas. Núcleo Mata Atlântica e Pampa. Brasília, Brazil.

Chao, A., C. H. Chiu, and L. Jost. 2014. Unifying species diversity, phylogenetic diversity, functional diversity, and related similarity and differentiation measures through Hill Numbers. Annual Review of Ecology, Evolution, and Systematics 45:297-324.

Chao, A., et al. 2020. Quantifying sample completeness and comparing diversities among assemblages. Ecological Research 35:292-314.

Conservation International. 2018. Wild.ID 0.9.31. TEAM Network Supercomputer Center. San Diego U.S.A.

Cooke, R. S. C., F. Eingenbrod, and A. E. Bates. 2019. Projected losses of global mammal and bird ecological strategies. Nature Communications 10:1-8.

Corrȇa, T. C. V., et al. 2021. Medium and large-sized mammals in Private Natural Heritage Reserves in the Quadrilátero Ferrífero of Minas Gerais, Brazil. Neotropical Biology and Conservation 16:383-396.

Crooks, K. R., and M. E. Soulé. 1999. Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400:563-566.

Faharig, L., et al. 2019. Is habitat fragmentation bad for biodiversity? Biological Conservation 230:179-186.

Ferreira, A. S., et al. 2020. Multi-scale mammal responses to agroforestry landscapes in the Brazilian Atlantic Forest: the conservation value of forest and traditional shade plantations. Agroforestry Systems 94:2331-2341.

Figuereido, M. D. S. L., et al. 2021. Tetrapod Diversity in the Atlantic Forest: Maps and Gaps. Pp. 185–204, in The Atlantic Forest (Marques, M. C. M., and C. E. V. Grelle, eds.). Springer International Publishing. Cham, Germany.

Galetti, M., et al. 2009. Priority areas for the conservation of Atlantic Forest large mammals. Biological Conservation 142:1229-1241.

Galetti, M., et al. 2021. Causes and Consequences of Large-Scale Defaunation in the Atlantic Forest. Pp. 297–324, in The Atlantic Forest (Marques, M. C. M., and C. E. V. Grelle, eds.). Springer International Publishing. Cham, Germany.

Haddad, N. M., et al. 2015. Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances 1:e1500052.

Hsieh, T. C., K. H. Ma, and A. Chao. 2016. iNEXT: an R package for rarefaction and extrapolation of species diversity (Hill numbers). Methods in Ecology and Evolution 7:1451-1456.

Ikin, k., et al. 2014. Multi-scale associations between vegetation cover and woodland bird communities across a large agricultural region. Plos One 9: e97029.

Instituto Brasileiro de Geografia e Estatística (eds.). 2012. Manual técnico da vegetação brasileir. Instituto Brasileiro de Geografia e Estatística-IBGE. Rio de Janeiro, Brazil.

Instituto Chico Mendes de Conservação da biodiversidade (ICMBio, eds.). 2018. Livro Vermelho da Fauna Brasileira Ameaçada de Extinção. Brasilia, Brazil.

International Union for Conservation of Nature and Natural Resources (IUCN). 2021. IUCN Red List categories. IUCN.

Invasive Species Specialist Group (ISSG). 2015. The Global Invasive Species Database. University of Auckland, New Zealand.

Kéry, M. and J. A. Royle. 2015. Applied Hierarchical Modeling in Ecology: Analysis of Distribution, Abundance and Species Richness in R and BUGS. Academic Press. Londres, U.K.

Kindt, R., and R. Coe. 2005. Tree diversity analysis: a manual and software for common statistical methods for ecological and biodiversity studies. World Agroforestry Centre (ICRAF). Nairobi, Kenya.

Lista das Espécies da Fauna Ameaçadas de Extinção no Estado de Minas Gerais, COPAM. 2010. Deliberação Normativa Conselho Estadual de Política Ambiental Nº 147. Diário do Executivo. Minas Gerais, Brazil.

Mackenzie, D. I. and W. L. Kendall. 2002. How should detection probability be incorporated into estimates of relative abundance? Ecology 83:2387-2393.

Mandujano, S. and L. A. Pérez-Solano. 2019. Fototrampeo en R: organización y análisis e datos. Instituto de Ecología A. C. Xalapa, México.

De Matos, T. P. V., et al. 2021. Protected areas and forest fragmentation: sustainability index for prioritizing fragments for landscape restoration. Geology, Ecology, and Landscapes 5:19-31.

Myers, N., et al. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853-858.

Oliveira, R. F., A. R. de Morais, and L. C. Terribile. 2020. Effects of landscape and patch attributes on the functional diversity of medium and large-sized mammals in the Brazilian Cerrado. Mammal Research 65:301-308.

Oksanen, J., et al. 2013. Package 'vegan'. R Package.

Porfirio, G., et al. 2014. Medium to large size mammals of southern Serra do Amolar, Mato Grosso do Sul, Brazilian Pantanal. Check List 10:473-482.

Püttker, T., et al. 2020. Indirect effects of habitat loss via habitat fragmentation: A cross-taxa analysis of forest-dependent species. Biological Conservation 241:108368.

Quintela, F. M., C. A. da Rosa, and A. Feijó. 2020. Updated and annotated checklist of recent mammals from Brazil. Anais da Academia Brasileira de Ciências 92:e20191004.

R Core Team 4.1.0. 2021. R: A language and environment for statistical computing. R for Statistical Computing. Vienna, Austria.

Ríos, E., et al. 2022. Spatial predictors and species’ traits: evaluating what really matters for medium‐sized and large mammals in the Atlantic Forest, Brazil. Mammal Review 52:236-251.

Regolin, A. L., et al. 2020. Spatial heterogeneity and habitat configuration overcome habitat composition influences on alpha and beta mammal diversity. Biotropica 52:969-980.

Ribeiro, M. C., et al. 2009. The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biological Conservation 142:1141-1153.

Rovero, F., and D. Spitale. 2016. Presence/absence and species inventory. Pp. 43-67 in Camera Trapping for Wildlife Research (Rovero, F., and F. Zimmermann, eds.). Pelagic Publishing, U.K.

Rosa, C. A., et al. 2017. Alien terrestrial mammals in Brazil: current status and management. Biological Invasions 19:2101-2123.

Santos, K. K., G. S. M. Pacheco, and M. Passamani. 2016. Medium-sized and large mammals from Quedas do Rio Bonito Ecological Park, Minas Gerais, Brazil. Check List 12:1830.

Santos, F., et al. 2021. Site and species contribution to β-diversity in terrestrial mammal communities: Evidence from multiple Neotropical Forest sites. Science of the Total Environment 789:147946.

Da Silva, P. G., M. I. M. Hernández, and J. Heino. 2018. Disentangling the correlates of species and site contributions to beta diversity in dung beetle assemblages. Diversity and Distributions 24:1674-686.

Souza, Y., et al. 2019. Atlantic mammals: a data set of assemblages of medium- and large-sized mammals of the Atlantic Forest of South America. Ecology 100:e02785.

Soto-Werschitz, A., S. Mandujano and M. Passamani. 2023. First record of the bush dog Speothos venaticus in the Atlantic Forest of Minas Gerais, Brazil. Oryx, 1-3. doi:10.1017/S0030605323000236.

Srbek-Araujo, A. C., and A. G. Chiarello. 2005. Is camera-trapping an efficient method for surveying mammals in Neotropical forests? A case study in south-eastern Brazil. Journal of Tropical Ecology 21:121-125.

Steinbeiser, C. M., et al. 2019. Relative abundance and activity patterns explain method-related differences in mammalian species richness estimates. Journal of Mammalogy 100:192-201.

Tomczak, M., and E. Tomczak. 2014. The need to report effect size estimates revisited. An overview of some recommended measures of effect size. Trends in Sport Sciences 1:19-25.

Trolliet, F., et al. 2014. Use of camera traps for wildlife studies: a review. Biotechnologie, Agronomie, Société et Environnement 18:446-454.

Vilas Boas, A. H., et al. 2022. Survey of medium-and large-sized mammals in Atlantic Forest remnants of Conceição dos Ouros, Minas Gerais, Brazil. Biodiversity Data Journal 10:e82139.

Wearn, O. R., et al. 2017. Mammalian species abundance across a gradient of tropical land-use intensity: A hierarchical multi-species modelling approach. Biological Conservation 212:162-71.

Wearn, O. R. and Glover-Kapfer. 2019. Snap happy: camera traps are an effective sampling tool when compared with alternative methods. Royal Society open science, 6:181-748.

Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon 21:213-251.

Associated editor: Eduardo Mendoza

Submitted: Junio 4, 2023; Reviewed: Junio 24 2023

Accepted: July 4, 2023; Published on line: August 29, 2023

Supplementary material 1

https://www.revistas-conacyt.unam.mx/therya/index.php/THERYA/article/view/4991/4991_Supplementary%20material

Table 1. Mammalian species richness, conservation status categories, and definition as non-native, native, or exotic species in the southeastern Brazilian Atlantic Forest of Minas Gerais (IUCN 2021; ICMBIO 2018; COPAM 2010).

Order	Family	Species	IUCN	ICMBIO	COPAM	Sp type
Artiodactyla	Bovidae	Bos taurus	NA	NA	NA	N-N
	Cervidae	Mazama americana	DD	DD	NA	N
		Mazama gouazoubira	LC	LC	NA	N
	Suidae	Sus scrofa	LC	NA	NA	E
	Tayassuidae	Pecari tajacu	LC	LC	VU	N
		Tayassu pecari	VU	VU	CR	N
Carnivora	Canidae	Canis lupus familiaris	NA	NA	NA	N-N
		Cerdocyon thous	LC	LC	NA	N
		Speothos venaticus	NT	VU	CR	N
	Felidae	Felis silvestris	NA	NA	NA	N-N
		Herpailurus yagouaroundi	LC	VU	NA	N
		Leopardus guttulus	VU	VU	VU	N
		Leopardus pardalis	LC	LC	VU	N
		Leopardus wiedii	NT	VU	EN	N
		Puma concolor	LC	VU	VU	N
	Mephitidae	Conepatus semistriatus	LC	LC	NA	N
	Mustelidae	Eira barbara	LC	LC	NA	N
		Galictis cuja	LC	LC	NA	N
	Procyonidae	Nasua nasua	LC	LC	NA	N
Cingulata	Dasypodidae	Cabassous unicinctus	LC	LC	NA	N
		Dasypus novemcinctus	LC	LC	NA	N
		Dasypus septemcinctus	LC	LC	NA	N
		Euphractus sexcinctus	LC	LC	NA	N
Didelphimorphia	Didelphidae	Didelphis albiventris	LC	LC	NA	N
		Didelphis aurita	LC	LC	NA	N
Lagomorphia	Leporidae	Sylvilagus brasiliensis	EN	LC	NA	N
Perissodactyla	Equidae	Equus asinus	NA	NA	NA	N-N
		Equus caballus	NA	NA	NA	N-N
Pilosa	Myrmecophagidae	Tamandua tetradactyla	LC	LC	VU	N
Primates	Cebidae	Sapajus nigritus	NT	NT	EN	N
Rodentia	Cuniculidae	Cuniculus paca	LC	LC	NA	N
	Sciuridae	Sciurus aestuans	LC	LC	NA	N

Not apply (NA), Insufficient Data-Know (DD), least concern (LC), vulnerable (VU), near threatened (NT), Endangered (EN) and critically endangered (CR). Non-native species (N-N). Native species (N). Exotic (E).