Abstract: Brazil has a high diversity of birds and presents the largest number of threatened bird species in the neotropical region. Even so, there are gaps in the bird knowledge, especially in ecotonal montane regions. Given this panorama, this study aimed to analyse the bird community distribution (richness, composition, and beta diversity between phytophysiognomies) of an ecotonal montane landscape of southeastern Brazil, with the purpose of detecting substitution patterns of bird species on a meso-scale. Using bird data performed during the years 1998 to 2015 in 46 sampling points, we found high bird richness in montane phytophysiognomies along an ecotone between Cerrado and Atlantic Forest hotspots. The composition present species of both domains, with high turnover component. We highlight the field environments and candeais are considered homogeneous and threathened, which would directly affect birds. The present study contributes to future conservation strategies, as it demonstrates ecotonal regions as transition zones and reinforces the need to consider as particular ecological units. These ecotonal regions are key locations for understanding ecological patterns in response to environmental changes or phytophysiognomies. Knowing how partitioning of the composition occurs within an environmental mosaic is essential to understand the limits and distributions of the species and conserve them.
Keywords: biodiversity, beta diversity, composition, conservation.
Ecologia
Mesoscale bird distribution pattern in montane phytophysiognomies along an ecotone between two hotspots

Recepción: 03 Diciembre 2020
Aprobación: 18 Mayo 2021
Brazil has the greatest diversity of birds in the world (Marini & Garcia, 2005; Piacentini et al., 2015; Morelli, Benedetti, Hanson, & Fuller, 2021) with approximately 57% of the species recorded throughout South America (Marini & Garcia, 2005), and 10% of this total are endemic species. This suggests that the Brazilian territory y is a priority for conservation investments (Sick, 1997). In addition, Brazil is the country with the largest number of threatened bird species in the neotropical region (Collar et al., 1992; Piacentini et al., 2015).
The bird community distribution is heterogeneous among biomes (Morelli et al., 2021). Therefore, knowledge about its distribution among Brazilian vegetation physiognomies (Sick, 1997; Gonzaga, Carvalhaes, & Buzzetti, 2007; Vasconcelos, 2008b) and in ecotonal regions is incipient. Ecotonal regions usually have their own characteristics and high ecological complexity resulting from a mixture of adjacent formations. There is ecological tension in biotas which produce high biodiversity because they enable species substitution at different scales (i.e., small mammals in Machado, Gregorin, & Mouallen, 2013; and plants in Machado, Fontes, Santos, Garcia & Farrapo, 2016).
The mountain landscapes of southeastern Brazil are within this ecologically tense context and present highly endemic areas in tropical regions for both flora and fauna (Eiten, 1992; Giulietti, Pirani, & Harley, 1997; Sick, 1997; Stattersfield, Crosby, Long, Wege, & BirdLife International, 1998; Safford, 1999; Silva & Bates, 2002; Gonçalves, Myers, Vilela, & Oliveira, 2007; Thom et al., 2020; Moura, Machado, Mariano, Leite, & Fontes, 2021). To birds in mountains, there are exclusive species directly associated with the vegetation, presenting half of the local species pool. Similar results have been found in research on various mountain ranges such as the Peruvian Andes (Lloyd & Maridem, 2008).
The ecotonal region between the Atlantic Forest-Cerrado of Minas Gerais State stands out for the high occurrence of areas covered by montane fields which are considered the most threatened environments (Stotz, Fitzpatrick, Parker, & Moskovits, 1996; Vasconcelos & Rodrigues, 2010; Moura et al., 2021). And this bird diversity linked to high altitude areas are among the most endangered species (Machado, Fonseca, Machado, Aguiar, & Lins, 1998; Lopes et al., 2009; BirdLife International, 2011). The region’s landscape also has areas with other montane phytophysiognomies in addition to the phytophysiognomies of the Cerrado domain, such as Semi-deciduous forests, Cloud Forests and the Candeais. Limited knowledge about the floristic composition and biogeography of the Cloud Forest (Bertoncello, Yamamoto, Meireles, & Sheperd, 2011; Pompeu et al., 2018) and Candeais makes it difficult to implement an effective management plan which focuses on its conservation (Scolforo, Oliveira, Davide, & Camolesi, 2002), and consequently the fauna which use it which is considered threatened (Moura et al., 2021).
The bird distribution on mountain landscapes associated to ecotonal regions are not yet fully resolved and described. For instance, there is little understanding of separated situations [only mountain (Santillán et al., 2020; Thom et al., 2020) or only ecotonal (Gonçalves, Santos, Cerqueira, Juen, & Bispo, 2017; Sementili-Cardoso, Vianna, Gerotti, & Donatelli, 2019)]. Given this panorama, this study aimed to present and analyse the bird richness, composition (by beta diversity), and structure between phytophysiognomies of an ecotonal montane landscape of southeastern Brazil in South America on a meso-scale and in a wide sampling over a decade of ornithological observations and records. Here we hypothesize that the richness is high due to the high number of phytophysiognomies, heterogeneity and complexity of these environments. The structure will have high amplitude. And composition (beta diversity) will have a high component of species substitution due to the specificity of bird diversity with each phytophysiognomy.
The study area is situated in Carrancas city, South Minas Gerais State, Southeastern Brazil (21° 29' 29.45” S/44° 38’ 42.47” W – 1097 m). The landscape corresponds to an ecotonal region between two hotspots, namely the Cerrado and Atlantic Forest (Myers, Mittermeier, Mittermeier, Fonseca, & Kent, 2000), and is composed by montane fields, Cerrado Stricto sensu, riparian forests, montane semi-deciduous forests, cloud forests, anthropic areas (pastures, agricultural areas, Eucalyptus forests, hydroelectric dam lake), and Candeais (forests dominated by Eremanthus erythropappus (DC.) Macleish). In addition, montane field areas are predominant in the landscape. The climate is mostly CWA type according to the Köppen classification; however it evolves into the CWB type for mountain tops in the areas with the highest elevation (Alvares, Stape, Sentelhas, Gonçalves, & Sparovek, 2013).
We highlight that the high lands of the study area are considered a ‘Hotspot’ (Drummond, Martins, Greco, & Vieira, 2009), regionally called ‘Chapada das Perdizes’, where the landscape is composed of montane phytophysiognomies (Cloud forests, Candeais, upper-montane semi-deciduous forests, and montane fields), with elevations ranging from 1000 to 1600 m. This region also houses the largest remnant of continuous forest in the south of Minas Gerais State, known as ‘Mata Triste’ (Oliveira-Filho, Carvalho, Fontes, Van Den Berg, & Carvalho, 2004). In addition, it also contains some of the Capivari River headwaters, a tributary of the Grande River. In turn, the Grande River joins the Paranaíba River, forming the Paraná River, which is the main lotic system of the second largest basin in South America (Pereira, Oliveira-Filho, & Lemos Filho, 2006). The region is strategic for conservation purposes, as it connects two large mountain ranges from two different biodiversity hotspots: the Espinhaço Complex (Cerrado) and the Mantiqueira Montain Range (Atlantic Forest).
Observations and data collections were performed during the years 1998 to 2015 in 46 sampling points which represent all the phytophysiognomies of the study area (Figure 1, Table 1). Each year a sampling was carried out in the hot and humid season, and another in the cold and dry season to reach resident and migratory birds. The semi-deciduous forest has high altitudinal variation, but we decided to separate the ‘Mata Triste’ and ‘Semi-deciduous Forest’ samples for this article to achieve proximity to the bird sampling points. The nomenclature used to identify bird species followed Piacentini et al. (2015).


We initially quantified the richness and bird families of each phytophysiognomy for data analysis in order to assess the specificities of each one. We also evaluated the pattern in the phytophysiognomies, quantifying the species number occurring in one or more phytophysiognomies, and considering all possible combinations for each category. We then obtained a Jaccard dissimilarity matrix from the data for the presence and absence of species in phytophysiognomies in order to make comparisons between phytophysiognomies. We subsequently performed a Principal Coordinate Analysis (PCoA) based on this matrix (Ter Braak, 1995), with the objective of ordering phytophysiognomies and observing possible aggregations and gradients.
Finally, we partitioned the dissimilarity matrix into the Turnover and Nestedness components (Baselga, 2010) and obtained a dendrogram corresponding to each component using UPGMA as a connection method (Gotelli & Ellison, 2016). The partitioning was carried out with the objective of evaluating which component is more significant in differentiating the communities in the phytophysiognomic study set, and if the ecological patterns are different in different perspectives. All analyzes were performed in the R version 3.3.1 (2016) using its default and the ‘vegan’ (Oksanen et al., 2017) and ‘betapart’ (Baselga, Orme, Villeger, Bortoli, & Leprieur, 2013) packages.
We found 310 bird species (Supplementary material) allocated in 60 families. The most represented families were Tyrannidae (N = 43), Throchilidae (N = 17), Tamnophilidae, Psittacidae and Picidae (N = 9).
The richness and families was higher for anthropic environments, followed by the semi-deciduous forest, while the lakes and Candeais showed the lowest richness and number of families, with the other phytophysiognomies varying between these extremes (Figure 2).

The ‘physiognomy number’ refers to physiognomy combinations. In this case, 11 species occur in all nine physiognomies, as along with 28 only occurring in one of the nine physiognomies, among other combinations. Thus, the distribution is relatively heterogeneous with approximately 100 species widely distributed and another approximately 200 species restricted to a few phytophysiognomies (Figure 3).

The first two axes of the PCoA together explained 69.4% of the data variation. The PCoA demonstrated a well-defined separation between forest environments: riparian forests, semi-deciduous forests (including ‘Mata Triste’), and cloud forests; non-forest: non-forest phytophysiognomies from Cerrado (as ‘Dirt Field’ and ‘Rupestrian Fields’), mountain fields, anthropic environments; and lake environments. This was a trend for axis 1, while only the lake environment differed from the other environments for axis 2. The dendrograms showed a pattern similar to the PCoA, with separation of forest and non-forest communities, presenting higher values for turnover in relation to nestedness (Figures 4 and 5).


The richness found in Chapada das Perdizes represents 16.05% of the 1919 records for the Brazilian territory (Piacentini et al., 2015), constituting a very high number if we consider the phytophysiognomies which occurred in the two global Hotspot domains were sampled: the Cerrado and Atlantic Forest (Myers et al., 2000). The high representativeness of the Tyrannidae and Tamnophilidae families was already expected, as they are the most represented in Brazil (Sick, 1997; Piacentini et al., 2015; and similar composition to Sementili-Cardoso et al., 2019), and also in studies previously conducted in Southern Minas Gerais State, which found similar results (Lopes, 2006; Lombardi, Vasconcelos, & D’angelo-Neto, 2007; Moura, Correa, Braga, & Gregorin, 2010; Moura, Corrêa, & Machado, 2015; Moura, Machado, Mariano, Souza, & Fontes, 2020).
In relation to the medium and long term studies in the Atlantic and Cerrado domains, the present study presents a high diversity with 310 species when compared to other high altitude (montane) regions of Southeastern Brazil. Vasconcelos & Rodrigues (2010) found 231 species in a survey compiled for mountains (only non-forest phytophysiognomies) in the states of Bahia, Minas Gerais, São Paulo, Rio de Janeiro and Espírito Santo; Rodrigues et al. (2011) found 151 species for the Serra do Cipó National Park; and Vasconcelos and D’Angelo-Neto (2009) found 206 species for Serra da Mantiqueira. This biodiversity is a research result of many phytophysiognomies at high altitude (similar to Machado et al., 2013 with small mammals) and has a long sampling time.
About composition, a total of eight taxa among the records are threatened, including the species Urubitinga coronata, Spizaetus tyrannus, Amazona vinacea, Geositta poeciloptera, Culicivora caudacuta, Alectrurus tricolor, Phibalura flavirostris, Anthus nattereri and Coryphaspiza melanotis (Fundação Biodiversitas, 2008; International Union for Conservation of Nature [IUCN], 2020; and similar to Sementili-Cardoso et al., 2019). Furthermore, the species Malacoptila striata, Aratinga auricapillus, Sarcoramphus papa, Platalea ajaja and Mycteria americana are in the almost threatened category (Lopes et al., 2017), thus showing the importance of this region for the conservation of the Brazilian bird community (Moura et al., 2021), and also highlighting the urgency of creating protected areas for wildlife conservation in the studied region (as proposed by Zambaldi, Louzada, Carvalho, & Scolforo, 2011). Mainly by the fragments of wide territorial extension that can be considered reference environments of vital importance for the conservation of the species of birds (Torezan, Calsavara, Bochio, & Anjos, 2021)
In view of the ecotonal characteristics of the two Cerrado and Atlantic hotspot domains, three typical Cerrado species of birds were recorded, namely Synallaxis spixi, Saltatricula atricollis and Antilophia galeata, and 19 typical species from the Atlantic Forest, including Aramides saracura, Florisuga fusca, Thalurania glaucopis, Baryphthengus ruficapillus, Malacoptila striata, Campephilus robustus, Pyrrhura frontalis, Pyriglena leucoptera, Conopophaga lineata, Ilicura militaris, Chiroxiphia caudata, Mionectes rufiventris, Todirostrum poliocephalum, Myiornis auricularis, Hemitriccus nidipendulus, Knipolegus nigerrimus, Hemithraupis ruficapilla, Tachyphonus coronatus and Sporophila ardesiaca (Silva, 1995; D’Angelo-Neto, Venturin, Oliveira-Filho, & Costa, 1998; Silva & Santos, 2005; Lopes et al., 2017).
The forest vegetation types presented a similar number of species with some of the greatest richness for the study area due to greater heterogeneity and complexity environmental (Willrich, Lima, & Dos Anjos, 2019) which provides more niches (Johnson, 1975; Terborgh, 1985; Santillán et al., 2020). Two situations deserve to be highlighted, namely the case of montane fields and the anthropic environments. Montane fields showed richness in line with other studies in different natural altitude fields in wildlife conservation areas (Conservation Units – Brasil, 2000) in Southeastern Brazil (e.g., Vasconcelos, 2008b), with 108 species for Cadeia do Espinhaço, and Rodrigues et al. (2011), with 151 species for Serra do Cipó). The high richness of anthropic environments suggests that the heterogeneous conditions caused by human actions in natural environments supplies a large variety of resources to avifauna (Willrich et al., 2019).
The PCO analysis also separated the communities into forest and non-forest environments generated by the aforementioned environmental complexity and heterogeneity (sensu August 1983) of the studied region. The beta diversity analysis indicates high turnover of bird species along the sampled environments (similar turnover results to De Deus, Schuchmann, Arieira, Oliveira Tissiani, & Marques, 2020; Gomez, Ponciano, Londoño, & Robinson, 2020), demonstrating the specificity of each phytophysiognomy or environment (Castaño-Villa, Ramos-Valencia, & Fontúrbel, 2014; Gomez et al., 2020). This pattern of beta diversity of the birds in Chapada das Perdizes is mainly driven by the local dynamics of phytophysiognomy. These findings indicate that the maintenance of several phytophysiognomies at meso-scale will guarantee a high turnover of species and is the key to the maintenance of a diverse biota (Roos, Giehl, & Hernández, 2021, Adorno, Barros, Ribeiro, Silva, & Hasui, 2021). In addition, flight capacity was not a factor which favored similarity for the ability to migrate between areas, therefore once again we emphasize the need for preservation (as highlighted by Zambaldi et al., 2011 and Moura et al., 2021), and demonstrating that each area can have a unique diversity which is difficult to find in other locations in the south of Minas Gerais or in the southeastern of Brazil.
Another important factor to be mentioned is the proximity of the turnover values of forest environments. This similarity (also expressed in the PCoA and in the richness graph) is an expression of the forest similarity for different areas resulting from soil characteristics and consequently of vegetation (Oliveira-Filho et al., 2004). The altitudinal variation influences the appearance of highly humid areas called cloud forests, which have connections with riparian forests and with large forest fragments such as the ‘Mata Triste’ and other semi-deciduous forests. This interconnection by ecological corridors favors an analogous composition of the avifauna community (Correa, Louzada, & Moura, 2012).
The distance between lake environments and other phytophysiognomies in PCoA analysis is due to the presence of species with narrow phenotypic flexibility and highly specific to aquatic environments (e.g., ducks such as Amazonetta brasiliensis, Cairina moschata, and herons such as Nycticorax nycticorax, among others). This specificity is closely linked to fish-eating habits (Paszkowski & Tonn, 2001), which is only possible in this environment.
From a conservationist point of view, cloud forests are a refugee which has yet to be explored, as their occurrence is restricted to high altitude regions above sea level (Carvalho, Fontes, & Oliveira-Filho, 2000; Bertoncello et al., 2011; Pompeu et al., 2018). There are very few locations in Southeast Brazil which present this characteristic, being more commonly found in Serra da Mantiqueira, Ibitipoca and cities of Aiuruoca, Baependí and Itamonte. Biogeographic studies of birds of cloud forests in Brazil are non-existent, and this is the first report which reinforces the high diversity for these environments.
The montane fields are a threatened phytophysiognomy from the expansion of Brachiaria sp. exotic grass (as mentioned by Klink & Machado, 2005). The composition of birds found in the montane fields was highly specific with the occurrence of endangered species such as C. caudacuta, A. nattereri and C. melanotis. Taxa such as the abovementioned are closely associated with the fields, and are among the most threatened birds in the Cerrado domain (Machado et al., 1998; Lopes et al., 2009). This group has a high affinity with the environment which is the result of an evolutionarily-shaped interaction (as mentioned by Santillán et al., 2020 when mentions about specifity in evolutionary history). Other studies in the mountain fields from the sampling area demonstrate specificity of other taxonomic groups with open environmens, such as rodents (Oxymycterus delator), marsupials (Monodelphis domestica) (Machado et al., 2013), and bats (Desmodus rotundus and Histiotus velatus) (Moras, Bernard, & Gregorin, 2013), among others. Therefore, the loss of grassland environments will lead to a co-extinction, without considering the loss in functional diversity and its respective ecosystem services.
Even though there are forests monodominated by Eremanthus (‘Candeais’) in the studied regions, there is no mention of the bird community in these forests in studies previously conducted (D'Angelo-Neto et al., 1998; Ribon, 2000; Vasconcelos et al., 2002;Vasconcelos, D’angelo-Neto S. & Nemesio, 2005; Lopes, 2006; Lombardi et al., 2007; Vasconcelos, 2008a; Moura & Corrêa, 2012; Moura et al., 2015) probably because they do not perceive this phytophysiognomy as a differentiated unit (similar to Willrich et al., 2019), as in the case of ‘Paratudal’ (Tabebuia aurea) forests in the ‘Pantanal’, and of the ‘Caxeitais’ (Tabebuia cassinoides) on the Brazilian coast, among others, thus highlighting the importance of these data for the ecology and conservation of the bird community in these forests and even for the phytophysiognomy itself, mainly as natural occurrence are homogeneous and in only high altitudinal elevation, so threathened as other montane phytophysiognomies.
We conclude that the montane phytophysiognomies along an ecotone between Cerrado and Atlantic Forest hotspots present high species richness. The composition present species of both domains, with high turnover component. We highlight the field environments and candeais are considered homogeneous and threathened, which would directly affect birds. The present study contributes to future conservation strategies, as it demonstrates ecotonal regions as transition zones (mixed composition form both domains) and reinforces the need to consider as particular ecological units. These ecotonal regions are key locations for understanding ecological patterns in response to environmental changes or phytophysiognomies. Knowing how partitioning of the composition occurs within an environmental mosaic is essential to understand the limits and distributions of the species and conserve them.
The authors thank the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) – Financing code 001, who supported this work by granting the doctoral scholarship to Aloysio Souza de Moura. For the funding from the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).

epilefsama@hotmail.com






