Avifauna of the region of the Volta Grande Hydroelectric Power Plant in Southeast Brazil

The Volta Grande region in the Brazilian state of Minas Gerais is classified as of “High Biological Importance” for the conservation of birds in the state. Despite this, the avifauna of the region has yet to be adequately sampled, as evidenced by the few papers on the region in ornithological literature. The results of an avifaunal inventory conducted between March 2013 and January 2014 in five sites located throughout the region of the Volta Grande Hydroelectric Power Plant in Southeast Brazil are presented. Point-counts, ad libitum observations and mist-nets recorded 224 species for the region, with insectivorous and omnivorous being the predominant feeding habits. Species that are “independent” of forest habitats comprised 52% of the recorded species. Three species are endemics of the Cerrado, 83 are migratory and 32 are under some degree of threat of extinction. These findings demonstrate that the region of the Volta Grande Hydroelectric Power Plant is important for bird conservation. Additional inventories, adequate management plans and public policies aimed the preserving species are needed for more effective conservation of the biodiversity of the studied region.


INTRODUCTION
The states of Minas Gerais and São Paulo in the Southeast Region of Brazil have 785 (Drummond et al., 2005) and 789 (Silveira et al., 2009) recorded bird species, respectively. The greatest threat to these species in both states is the degradation of natural environments and the consequent fragmentation and loss of habitats (Drummond et al., 2005;Bressan et al., 2009).
The Cerrado is the second largest biome of Brazil, encompassing approximately 200 million ha located mainly in the center of the country but with small areas in northeastern Paraguay and eastern Bolivia (Silva, 1995a;Ratter et al., 1997;Pinheiro & Monteiro, 2010). It is characterized by a mosaic of phytophysiognomies ranging from grasslands to forest formations (Silva, 1995a;MMA, 2007).
The diversity of the vegetation mosaic of the Cerrado is further enhanced by the influence of adjacent biomes, such as the Atlantic Forest, Amazon rainforest, Caatinga and Chaco (Ratter et al., 1997;Myers et al., 2000;Silva & Bates, 2002). This great species diversity accompanied by drastic loss of original vegetation over a relatively short period of time due to anthropic actions has led to the Cerrado being classified as one of 34 global terrestrial hotspots (Ratter et al., 1997;Myers et al., 2000;Myers, 2003). The Cerrado is also considered the largest, richest and probably most endangered tropical savanna in the world (Silva & Bates, 2002). The avifauna of the Cerrado comprises 837 species (Silva, 1995b;Stotz et al., 1996), of which 30 are endemic and 34 endangered according to the Brazilian red list (Silva, 1995b;Zimmer et al., 2001;ICMBio, 2018).
The Volta Grande Hydroelectric Power Plant (hereafter Volta Grande HPP) is located in the region of the lower Grande River at the boundary between the states of Minas Gerais and São Paulo in Southeast Brazil. This region is classified as of "High Biological Importance" for the conservation of birds in Minas Gerais (Drummond et al., 2005). Despite its importance, the avifauna of the region has yet to be adequately sampled (Silva, 1995b), as evidenced by the low number of papers currently available in the ornithological literature (Silveira, 1998;Andrade & Marini, 2002;MMA, 2005;Marçal-Júnior et al., 2009;Bessa et al., 2011;IEF, 2011;Malacco et al., 2013).
The objective of this study was to inventory the avifauna of the region of Volta Grande HPP, with the aim of increasing knowledge of its bird community. Specifically, the bird species of the community were characterized with regard to: (i) frequency of occurrence, (ii) trophic guild, (iii) dependence on forest habitat, (iv) endemism, (v) migratory condition and (vi) conservation status.

Study area
Fieldwork was conducted in five sites distributed among four municipalities around the Volta Grande HPP reservoir (20°01′27″S, 48°12′45″W) ( Table 1). The reservoir is located on the Grande River at the boundary between the states of Minas Gerais and São Paulo in Southeast Brazil. The reservoir is 46 years old, has an average elevation of 524 m and encompasses 19,800 ha (Fig. 1). The predominant morphoclimatic domain of the study region is Cerrado (Neotropical savanna). The region has a tropical climate with a dry winter (April to September) and a rainy summer (October to March) (Moreira et al., 2008). Mean annual rainfall ranges 1400-1600 mm, with the rainiest months being December and January, and the driest months June and July (Fujaco & Leite, 2016). The mean annual temperature ranges from 22°C to 24°C (Filardi et al., 2007).
The original vegetation of the region was mostly deforested or flooded during the construction of the Volta Grande HPP reservoir (from 1970 to 1974) (Pires et al., 2016). Four out of the five study sites contain revegetated riparian forest patches that were the result of a reforestation program undertaken by the Minas Gerais state Figure 1. Grande River Basin located at the border between the states of Minas Gerais and São Paulo in Southeast Brazil. The five study sites are shown as: 1 and 2 located in Minas Gerais, and 3, 4 and 5 located in São Paulo.  , 2020; v.60: e20206016 2/15 energy company (CEMIG). One of the sites has fragments of native vegetation. Together the five sites encompass a total of 423.8 ha.

Sampling
Field surveys were conducted during four months of each climatic season: April, May, June and July 2013 (dry season), and October, November, December 2013 and January 2014 (rainy season). The five study sites were sampled over the course of five consecutive days each month such that each site was sampled once per month. Quantitative data were collected using the point-count method (Vielliard et al., 2010), with six points in each site (three within riparian forest and three in surrounding areas) being sampled per campaign. Points were separated by 350 m to avoid overlapping of territory of some species (Ralph et al., 1995;Vielliard et al., 2010). Due to this distance it wasn't necessary to establish a limiting radius for the detection of bird individuals. Point-counts of a given site were performed in the mornings (between 06:00 and 09:00 h), with points being sampled in a randomly determined order. The number of individuals of all bird species seen and/or heard during 20 min at each point was recorded (six points × 20 min × eight monthly visits = 960 min for each site).
Qualitative bird surveys were conducted through ad libitum observations during and after quantitative sampling between 06:00 to 11:00 and 13:00 to 18:00 h, for a sampling effort of 10 h of sampling/day (10 h sampling × eight monthly visits = 80 h sampling per site; 400 h total for all five sites).
Adequacy of sampling effort was evaluated by constructing a species accumulation curve (Santos, 2003). Bird richness was estimated by the Chao 1 index using EstimateS 9.1 software (Colwell, 2009), which is based on the number of singletons and doubletons in each point count (Hortal et al., 2006).
Birds were identified visually with binoculars, by photography (Canon PowerShot SX50 HS) or audibly. A Sony ICD-PX312 recorder was used to document vocalizations. Photographs and recordings that were archived on WikiAves (2019) are identified in Table 2.

Data analysis
Frequency of Occurrence (FO) was determined for each bird species based on the proportion of effort units (days) during which a given species was detected (adapted from Costa & Rodrigues (2012)) as: C (common) -species recorded between 75-100% of sampling days; RC (relatively common) -species recorded between 50-74% of sampling days; UC (uncommon) -species recorded between 25-49% of sampling days; R (rare)species recorded between 6-24% of sampling days and O (occasional) -species recorded in less than 5% of the sampling days (effort was one day per sampling site and the presence of each species was only recorded once per sampling day).
Bird species were grouped into 25 trophic guilds considering not only food items, but also the environment and vertical strata most utilized during foraging (adapted from Willis (1979) and Antunes (2005) Bird species were classified into three categories of forest habitat dependence following Silva (1995a): (1) independent -species that forage and breed mainly in open vegetation; (2) semi-dependent -species that forage and breed in both forest habitats and open areas; and (3) dependent -species that forage and breed mainly in forest habitats. Endemism for the Cerrado was based on Silva (1995a) while endemism for riparian forest habitats followed Silva & Vielliard (2000).
The species were also classified according to their migratory condition according Chesser (1994), Sick (1997) and Somenzari et al. (2018) as: septentrional migrants (SM) -species from the northern hemisphere; and austral migrants (AM) -species that breed in temperate continental South America and migrate north, towards Amazonia, during the austral winter. Also included in this category are those resident species whose southernmost populations perform small-scale seasonal migrations.

RESULTS
A total of 224 bird species distributed among 24 orders and 58 families was recorded (Table 2). Passeriformes was the most representative order with 113 species, corresponding to 50% of the total recorded. The most representative family was Tyrannidae, with 35 Table 2. Bird species recorded in the five study sites located in the region of the reservoir of the Volta Grande Hydroelectric Power Plant in Southeast Brazil, and their frequency of occurrence (FO), trophic guild (TG), forest habitat dependence (FHD), threat status, and migratory condition. Reference numbers for the online database www.wikiaves.com.br (WA) are also provided. Taxonomy and nomenclature follow Piacentini et al. (2015).  species (16%), followed by Thraupidae (23 species; 10%). Among non-Passeriformes, the most representative families were Ardeidae, Psittacidae and Trochilidae, each one with nine species (4%).
The species richness curve produced by the Chao 1 index estimated (mean ± SD) 232 ± 8.7 species for the study area (Fig. 2). Thus, the observed richness represented 97% of the estimated richness.
According to the FO estimates, only 16% (n = 36) of the recorded species were considered common in the study area (Table 2). Among the common species, only five (2%) had an FO of 100% by being present in all samples ( Table 2). The majority of the species (50%, n = 113) were classified as occasional or rare. Thirty-two species (14%) were recorded only once in the study area (single records) ( Table 2).
Fifty-two percent of the bird species recorded (n = 117) were forest independent, while 19% (n = 42) were strictly dependent on forest habitats for breeding or foraging. Three recorded species were endemic to the Cerrado: Herpsilochmus longirostris, Clibanornis rectirostris, and Antilophia galeata. These three species were also considered endemic of the Central Brazil riparian forests (Silva & Vielliard, 2000).
Eighty-five migratory species were recorded, corresponding to 38% of the total sampled. Only four species came from the Northern Hemisphere (Table 2), while the remaining migratory species came from the southernmost regions of South America. Thirty-three of the recorded species appeared in at least one of the lists of endangered, near threatened and data-deficient species considered in this study (Table 2). In study sites located in the state of Minas Gerais three species were considered threatened in the state: Crax fasciolata, Mycteria americana, and Ara ararauna (COPAM, 2010). C. fasciola-ta also appears on the IUCN red list, in the "vulnerable" category (IUCN, 2020). Two other species present in the Minas Gerais red list were registered only in areas located in the state of São Paulo: Platalea ajaja and Sporophila angolensis.
According to the red list of threatened species of the state of São Paulo, 24 species recorded in study sites located in this state were under some degree of threat of extinction (Table 2). Amazona aestiva was classified as Near Threatened in the Brazilian national list (ICMBio, 2018) ( Table 2).

DISCUSSION
The region of the lower Grande River was found to possess high bird species richness, with various threatened (almost exclusively locally, present only in State lists; with the exception of C. fasciolata, present in the global list) and some endemic species. The area was also found to be a feeding and resting area for migratory birds, confirming its status as an area of "High Biological Importance" for the conservation of birds in the states of Minas Gerais (Drummond et al., 2005). The species accumulation curve indicated that the sampling effort satisfactorily detected a significant portion of the region's avifauna. In fact, at 50% of the total sampling effort, 89% (n = 200) of the species had already been recorded. It is important to emphasize that most of the sampling sites were reforested areas, reaffirming their importance to bird communities.
Studies addressing the bird community of non-urban areas of the region of the lower Grande River and Paranaíba River (Triângulo Mineiro region) are extremely scarce in the scientific literature. The few that do exist have found 403 bird species for Serra da Canastra National Park (2,000,000 ha, and 90 km from the study area) (Silveira, 1998;Andrade & Marini, 2002;MMA, 2005;Bessa et al., 2011); 231 species for RPPN Panga Ecological Reserve (410 ha, and 130 km from the study area) (Marçal-Júnior et al., 2009); 202 species for a private reserve located in the municipality of Uberlândia (640 ha, and about 130 km from the study area) (Malacco et al., 2013), 221 species for Pau Furado State Park (2,000 ha, and 140 km from the study area) (IEF, 2011), and 328 species for the 17 forest remnants in northwestern São Paulo state (Bispo et al., 2011).
Two-percent of the species encountered in the present study had 100% FO. Studying rural forest fragments in the state of São Paulo, Almeida et al. (1999) and Donatelli et al. (2004) also found a very low number of species with 100% FO, with 1% and 3%, respectively. The majority of bird species recorded (50%) in our study were classified as occasional and rare, with FOs of less than 25%. A similar result was obtained by Almeida et al. (1999) in two areas located in the Jacaré-Pepira River Basin (rural state of São Paulo), and by Lyra-Neves et al. (2004) in a study in the Gurjau State Reserve, state of Pernambuco. The low percentage of species with 100% FO, and the high per-centage with FOs of less than 25%, can be explained by the presence of wandering species that remain in areas for a few days, species with inconspicuous vocalizations and/or behaviors, accidental species from neighboring habitats and migratory species. It can also be explained by the possibility that the studied sites do not possess the resources and conditions necessary to maintain larger bird populations (Aleixo & Vielliard, 1995;Donatelli et al., 2007).
The guild of edge insectivores is commonly favored by small patches because birds of this guild possesses foraging plasticity, being able to search for food among different strata such as in tangled vegetation and in the middle and canopy levels (Willis, 1979;Borges & Stouffer, 1999;Stouffer et al., 2009). The data obtained in the present study support this assertion because the guild edge insectivores was the trophic guild with the greatest number of species. This result is also likely due to the sampling performed in surrounding open areas, which included edges of the forest fragments. In contrast, insectivorous species foraging in the lower strata, such as predators of small ground arthropods, predators of foliage arthropods and understory omnivores, tend to be more susceptible to habitat degradation due to the consequent structural simplification of the environment, and, thus, are less common in small patches (Willis, 1979;Castaño-Villa et al., 2014;Cid & Caviedes-Vidal, 2014). Structural simplification of the environment and the inability to adapt to the surrounding open habitat has a greater influence on species of these guilds than the availability of arthropods as a food resource (Sekercioglu et al., 2002). Predators of foliage arthropods and understory omnivores were among the guilds with the lowest number of species in the present study. In contrast, the guild of edge omnivores has been shown to be favored by small patches due to the larger proportion of edge environments in these habitats (Willis, 1979;Austen et al., 2001;Anjos et al., 2007), which likely explains why this guild was well represented (second greatest number of species) in the present study.
Species of the guild bamboo or tangle insectivores can be considered specialized microhabitat foragers (Goerck, 1997;Areta & Cockle, 2012;Lebbin, 2013), and thus tend to be rare in small patches since they are competitively excluded by edge insectivores (Willis, 1979;Cockle & Areta, 2013). The only species of the bamboo or tangle insectivores guild recorded in the present study was Poecilotriccus latirostris. This species is the most widely distributed species of its genus in western Brazil, and is characteristic of middle and lower strata of tangled vegetation in riparian forests of Central Brazil (Sick, 1997;Sigrist, 2013).
It is important to highlight the occurrence of only one species of more specialized insectivorous as regards the foraging site: Dendrocolaptes platyrostris. The woodcreepers are trunk and twig insectivores that forage preferentially on senile trees. They are also sensitive to environmental disturbances, and thus are the first to disappear locally when environmental degradation occurs (Laurance et al., 2002;Cleary et al., 2007;Dahal et al., 2015).  Avulsos Zool., 2020;v.60: e20206016 10/15 Large frugivores (e.g., Crypturellus spp., Crax spp. and Penelope spp.) which feed mostly on the ground were poorly represented in the present study. Overall, frugivores are more susceptible to local extinction because their populations are possibly limited by the availability of fruit during periods of scarcity (Antunes, 2005;Cleary et al., 2007;Kennedy et al., 2010) and by hunting (Peres, 2001). Thus, species belonging to this guild tend to disappear from small patches, since they require areas with a wide variety of tree species that produces fruit at different seasons of the year (Willis, 1979;Piratelli et al., 2005;Ferger et al., 2014). Also, large frugivores are cinegetic species, i.e, species that suffer from hunting pressure by humans (Pereira & Schiavetti, 2010), and this impact associated with habitat loss can drive these species to extinction (Simberloff, 1995;Symes et al., 2018). In the study region occurs the Speckled chachalaca (Ortalis remota) in São Paulo state. Deforestation caused mainly by the construction of the Lower Grande reservoir may be responsible for the extinction of this species in Minas Gerais state (Silveira et al., 2017).
Like large frugivores, nectarivores birds are also dependent on plant species that contribute with food resources throughout the year (Bennett et al., 2014;Rodrigues & Rodrigues, 2015). According to Willis (1979) and Bowen et al. (2009), the few plant diversity and the structural simplification in small patches probably explains the tendency to the nectarivores to disappear from these habitats. On the other hand, there is evidence that in the Neotropical region nectarivores birds are not very sensitive to the habitat fragmentation effects, and may even be benefited by the high abundance of light-dependent plants and the increase of flower production in clearings and forest edges (Vetter et al., 2011). Still according to Vetter et al. (2011), nectarivores often travel long distances to find their food resources and are able to cross open areas. Thus, to better understand this relationship, future studies are needed to evaluate how these trophic guild responds to the consequences of the fragmentation processes and habitat loss.
According to Silva (1995a), 52% of the 759 bird species that reproduce in the Cerrado are dependent on forest habitat. However, species dependent on forest habitat corresponded to only 19% of all the species recorded in the present study (see Table 2). According to Laurance et al. (2002), a reduced number of forest bird species in small, isolated and impacted patches can be due to morphological and behavioral constraints.
Another possible explanation for the reduced number of forest species in isolated fragments is the reluctance of these species to cross open areas, possibly due to the increased predation risk (Watson et al., 2004;Piratelli et al., 2005). These studies reported that the probability that an individual will cross an open field between two forest fragments decreases rapidly as the distance between fragments increases.
The most important action for the conservation of species dependent on forest habitats is to prevent future forest losses, and to conduct habitat restoration. These two activities should be the most effective conservation strategies for ensuring the persistence of forest birds in fragmented landscapes (Mortelliti et al., 2010;Mendoza et al., 2014). In the long-term, conservation of the species dependent on forest habitats in the area of the present study, like C. fasciolata, M. nigrifrons, and C. rectirostris, is unfavorable because the few forest remnants remaining are small, isolated from each other, and among a habitat matrix that is predominantly impermeable to forest species displacements (mostly sugar cane plantations). Furthermore, most of the remnants do not receive effective conservation measures.
On the other hand, most of the birds recorded in the present study were classified as forest independent (52%) or semi-dependent (29%) species (see Table 2). For these species, the matrix may be more permeable, allowing movements among different isolated forest patches (Yabes et al., 2010). Characteristics of the matrix will then favor or not bird movements. In our study area, most of the matrix was composed by sugar cane crops and pasture lands (≈ 90%). This structure will limit movements or modify movements of the birds. In pasture lands, for example, forest-dependent and semi-dependent birds will probably perform fast and straight flights direct to the nearest forest patch, while in sugar cane crops, flights would be slowly because birds could use sugar cane as cover to protection against predators (Biz et al., 2017). Thus, in our study area, due to the structure of the matrix and the isolation of the forest patches, movements across the matrix would be performed mostly by independent bird species, decreasing the genetic flow and persistence of the forest dependent species (Braunisch et al., 2010;Prevedello & Vieira, 2010). The disappearance of forest-dependent birds in the region, like Arremon semitorquatus and Chordeiles pussillus, have been already recorded (Willis & Oniki, 2002;Silveira & Uezu, 2011).
Some authors have indicated that the adverse effects of vegetation degradation and the isolation of the resultant fragments affect more negatively endemic species than species with broad geographic distributions (Aleixo & Vielliard, 1995;Ribon et al., 2003). These authors refer to geographic endemism and not ecological endemism, because data from the literature show that ecologically endemic species of riparian forests often maintain linear territories along riverbanks and, for this reason, do not suffer significant changes due to variation in riparian forest size (Shirley & Smith, 2005;Lees & Peres, 2008).
Of the three Cerrado endemic species recorded in this study, only A. galeata was not recorded in all sampled areas. It is a predominantly frugivores species and forages mainly in the canopy of riparian forests (Marini, 1992). In addition to being considered a rare species in the state of São Paulo (Donatelli et al., 2004) and included in the Near Threatened category in the state , it was also classified as a rare for the study region of the present study according to its FO (14%).
Although most species that occur in the Cerrado depend on forest habitats for reproduction, the majority of migratory species are considered to be independent of these environments. Indeed, according to Silva (1995a), 88% of the septentrional migrants and 67% of the austral Mafia, P.O. & Azevedo, C.S.: Birds from Volta Grande region Pap. Avulsos Zool., 2020; v.60: e20206016 11/15 migrants already recorded in the Cerrado do not depend on forest habitats. The data obtained in the present study corroborate this assertion, since 66% of the migratory species recorded were classified as independent of forest habitat. Other authors have affirmed that migratory species, independent of trophic guild, preferentially exploit open areas for breeding and foraging (Alves, 2007;Robertson et al., 2013). The Blue-tufted Starthroat (Heliomaster furcifer), recorded only once in the present study (a male with post-nuptial plumage), in the municipality of Igarapava, state of São Paulo, appears to be a low-density species, with few records in the literature (Macarrão et al., 2011). The first documented record of these species in São Paulo occurred only in 2010, in a rural environment of the municipality of Brotas (Macarrão et al., 2011). According to these author, H. furcifer appears in São Paulo during his migratory movements, which probably occur during the dry season, suggesting that these species may occur as a winter migrant in the southeastern region of Brazil (Mazzoni & Perillo, 2014). Thus, the record made in the present study, on July 28, 2013, coincides with this period of occurrence described in the literature.
Three of the septentrional migrant species (Tringa solitaria, Hirundo rustica, and Petrochelidon pyrrhonota) were observed only in the rainy season, which corresponds to winter in the Northern Hemisphere. This pattern of occurrence is consistent with that described in the literature (Sick, 1997). The Cliff Swallow (P. pyrrhonota) was observed forming extremely large flocks of more than 150 individuals when foraging through the aerial stratum. The Barn Swallow (H. rustica) was also observed in very numerous groups, but less numerous than the Cliff Swallow. The Solitary Sandpiper (T. solitaria) was recorded only once in November 2013 when two individuals foraged in the midst of swamp vegetation within the sampled area.
The only septentrional migrant recorded during all months of the study, and thus in both in the dry and rainy seasons, was the Osprey (P. haliaetus). This rapinant species, which feeds almost exclusively on fish (Mestre & Bierregaard-Jr., 2009), is most frequently recorded at the end and the beginning of each year in its wintering places, but can be seen during any season in Brazil, with records existing for every month of the year (Sick, 1997).
It is also important to highlight the record of the Greater Rhea (Rhea americana), observed on three occasions only in the municipality of Igarapava, São Paulo: on November 13, 2013, an adult was observed when foraging in the middle of sugarcane monoculture accompanied by eight immature individuals. In January 2014, this family group was recorded again, being possible to notice the development of the immatures. Habitat destruction and fragmentation, hunting, pesticide contamination and few localities with recent records of the species in the state justify their inclusion in the threatened species list of São Paulo as a critically endangered species . Thus, the records made in the present study are important because they provide data about its occurrence (with evidence of reproduction) in a new locality in the state of São Paulo.

CONCLUSION
According to Drummond et al. (2005), the region of the lower Grande River is currently considered as of "High Biological Importance" for bird conservation in the state of Minas Gerais because it has a high number of threatened species. The results of the present study corroborate this statement since a great number of endangered species were recorded in the sampled areas, at least locally (in the State level). The main cause of decline for all threatened species is habitat destruction, mining, dam's constructions and agricultural activities (Drummond et al., 2005;Faria et al., 2008;Bressan et al., 2009). Additional research, with bird inventories, adequate management plans and public policies aimed at conserving habitats, with the creation of conservation units, are needed for more effective conservation of local biodiversity (Drummond et al., 2005).