Stretching for more than 10,000 ha in the Metropolitan Area of São Paulo, southeastern Brazil, Serra da Cantareira comprises the largest native urban rainforest in the World, harboring a rich and diverse Atlantic Forest avifauna. Despite its closeness to major urban areas, few bird surveys have been conducted there. In this article we present an updated compilation of all bird species recorded for Serra da Cantareira, including personal records from the authors. A total of 326 species have been recorded for Serra da Cantareira since 1901; of these, nine have not been sighted there for the last two decades. The number of bird species endemic to the Atlantic Forest is high (80), and seven of its species are globally threatened. According to multivariate analyses the species diversity at Serra da Cantareira is similar to other regions of the Atlantic Forest, such as Carlos Botelho and Intervales state parks, where the vegetation is also ombrophilous dense forest. We discuss local changes in the avifaunal composition over the last decades and suggest the incorporation of large forest remnants to the Cantareira State Park to mitigate the impact of the northern section of Rodoanel Mário Covas, a highway (SP-21) that will soon be operational and will negatively impact the biodiversity of Serra da Cantareira.

Atlantic Forest, bird conservation, hierarchical cluster analysis, principal coordinate analysis, Serra da Cantareira.

The Atlantic Forest (AF) is the second largest rainforest in South America and has a rich and diverse avifauna (900 species, 24% of which are endemic to the AF; Lima 2013). New bird species are still being discovered in this forest, even near large urban areas (Buzzetti et al. 2013), demonstrating that the local avifauna is not completely known, and reinforcing the importance of bird surveys there. Despite the high levels of species richness and endemism, only ~12% of the AF forest remains, and as a result, a large number of its component taxa are at imminent risk of extinction (Marini and Garcia 2005, Ribeiro et al. 2009). In addition to the importance of the AF for bird conservation, its forest remnants provide several ecosystem services for human populations, such as guaranteeing water supply near major urban areas (Dean 1995).

After 1893, due to a water collapse in the city São Paulo, the government started to expropriate farms at Serra da Cantareira for reforestation, to ensure the recovery of the streams and river that supply the city with water (Vilar 2007). Considered the central piece in the São Paulo City Green Belt Biosphere Reserve by UNESCO and an Important Bird Area by BirdLife International (IBA SP03), Serra da Cantareira still provides water and other ecosystem services to the city, and houses a significant portion of the AF biodiversity, including several threatened and endemic species (Bencke et al. 2006, Whately and Cunha 2007, Ayres 2008). Serra da Cantareira stretches for more than 10,000 ha in the Metropolitan Area of the city of São Paulo, the most populated and urbanized region in South America, and currently comprises the largest native urban rainforest in the world (Ayres 2008). The main impacts to the biodiversity of Serra da Cantareira are urban expansion and large infrastructure projects in nearby towns (Bencke et al. 2006, Fig. 1).

Figure 1. 

Serra da Cantareira locations surveyed in this study. In the urban slope are Cantareira State Park admnistrative zones and their trails: AC = Águas Claras; PG = Pedra Grande; Eng = Engordador; Cab = Cabuçu; EL = Estrada do Lenhador; Alberto Löfgren State Park (ALSP) and A2. In the inner slope are A1, A3, A4, A5 and SC = Sítio Capuavinha private reserve.

Although Serra da Cantareira is continuous with one of the largest cities in the world and near several research institutions and universities, few ornithological studies have been performed there. Museum collectors were the first to survey these mountains in the end of the XIX century (see Pinto 1938, 1944, Paynter and Traylor 1991). Graham (1992) published the only comprehensive bird inventory of the region, more than 20 years ago. Other than museum specimens and Graham’s article, there have been few ornithological studies and bird surveys. They range from comparison of communities in different vegetation types (Graham 1991, Antunes et al. 2009; frugivory (Fonseca and Antunes 2007, Ikuta and Martins 2013); new species records (Antunes et al. 2008a), plumage anomaly (Cavarzere and Tonetti 2015); nest description (V.R. Tonetti et al. unpublished data), studies on species ecology (Tonetti and Pizo 2016), technical reports (Antunes and Eston 2009, 2012, DERSA 2010) and records provided by birdwatchers (WikiAves 2016, Xeno-Canto 2016).

In this article, we present an updated compilation of all bird species recorded for Serra da Cantareira, including our personal records. We compare the avifauna between Serra da Cantareira and other regions of AF using multivariate analysis, discuss changes in avifaunal composition during the last decades, and provide additional information on some threatened and/or uncommon species. Lastly, we focus on specific localities and measures for bird conservation in our study area.

Serra da Cantareira ranges from 750 to 1,250 m asl. The vegetation there is predominantly montane ombrophilous dense forest. The climate is mesothermal and humid, with rainy summers and dry winters (CWA Köppen), and annual temperature and rainfall averaging 20 °C and 1,500 mm, respectively (Bencke et al. 2006). The slope facing the urban area comprises the most preserved area, with continuous forests, which is mostly protected by Cantareira (7,900 ha; hereafter CSP) and Alberto Löfgren state parks (187 ha; ALSP). CSP is divided into four administrative zones: Águas Claras, Cabuçu, Engordador and Pedra Grande (Fig. 1). Patches in the inner slope have variable sizes (~1 to 100 ha) and are at different stages of regeneration. In this study we considered bird records from CSP, ALSP, neighboring patches in the inner slope, including the 20-ha private reserve Sítio Capuavinha (Lemos 2014) and records that indicated only “Serra da Cantareira” as a location, without any further detail.

We obtained species records over the last decade during systematic and unsystematic bird surveys. In unsystematic surveys, samplings consisted of recording all aurally or visually detected birds at unlimited distance, while walking on trails (Ribon 2010). Surveys were performed during the morning in different seasons and were mostly concentrated on the available trails of the four CSP administrative zones and a dirt road named “estrada do lenhador”, at the park’s boundaries (Fig. 1). Birds were recorded with the aid of binoculars and song recordings and photographs were occasionally taken during surveys.

VRT performed a systematic survey using point-counts. In this study, 100 fixed-points were established in the four CSP administrative zones and visits lasted 10 minutes during the morning. Each point was visited six times on different days between May and December 2014, resulting in a total of 600 samples, which include only birds detected within a 50-m radius (Vielliard et al. 2010, Tonetti and Pizo 2016). MAR, along with other researchers (see names in acknowledgements), performed another systematic study using line transects (80 h of sampling effort) and mist-nets (2,600 net-hours) in five different areas (A1 to A5; Fig. 1). These surveys were conducted during the morning and late afternoon from August 12^th 2009 to April 29^th 2010. Some netted specimens were collected and deposited in the MZUSP collection (details in DERSA 2010).

In order to compile our data, we searched for museum specimens collected from “Cantareira State Park”, “Parque Estadual da Cantareira”, “Cantareira Mountains”, “Serra da Cantareira”, “Cantareira”, “Alberto Löfgren State Park”, “Parque Estadual Alberto Löfgren”, “Horto Florestal”, “Cuca, Horto Florestal” and “Parque Florestal”. We surveyed the collections of the following museums: Museum of Zoology of the University of São Paulo (MZUSP, Brazil), Museum of Zoology of the Campinas University (MUNICAMP, Brazil), and the Field Museum of Natural History (FMNH, USA). The localities cited above correspond to CSP, ALSP or not specified sites at Serra da Cantareira (Paynter and Traylor 1991). Using the same terms as outlined above, we searched for indexed articles potentially containing bird records in the citation databases Web of Science, Google Scholar and SciELO. Additionally, we used Google to search for non-indexed articles and “gray” literature, such as technical reports and theses. We also compiled records from the WikiAves (www.wikiaves.com; WA) and Xeno-canto (www.xeno-canto.org; XC) databases. All searches were performed until June 1^st 2016. We discarded erroneous or doubtful records – i.e. misidentified pictures and song records, as well as species pending confirmation of their occurrence in our study area according to the literature. We followed the nomenclature of the CBRO (Brazilian Ornithological Records Committee) (Piacentini et al. 2015).

We used Mountford’s dissimilarity index (Wolda 1981) to perform a hierarchical cluster analysis (HCA) and a Principal Coordinate Analysis (PCoA) to compare the bird communities from Serra da Cantareira with other AF regions. After calculating the Mountdord’s index using presence/absence data with the vegdist function in the R package vegan (Oksanen et al. 2016), we performed the HCA using the Ward’s minimum variance criterion as the objective function, the results of which define which clusters merge at each step (Ward Jr 1963). The HCA and the PCoA were performed using the hclust and ordiplot functions (Oksanen et al. 2016, R Core Team 2016). We also calculated the Jaccard dissimilarity index using the function vegdist (Oksanen et al. 2016). The Jaccard coefficient ranges from 0 (identical pairwise) to 1 (totally different pairwise) and it is a robust index to ascertain similarity among groups (Krebs 1999).

We compared bird communities between Serra da Cantareira and the following AF regions that also have dense ombrophilous forest: Boraceia Biological Station (Cavarzere et al. 2010), Carlos Botelho State Park (Antunes 2013), Intervales State Park (Antunes et al. 2008b), Morro Grande Reserve (Develey and Martensen 2006), Municipality of Ilha Comprida (Gussoni 2010), Municipality of Ubatuba (Simpson et al. 2012), and in Serra da Mantiqueira, where the vegetation is predominantly mixed ombrophilous forest: Campos do Jordão (composed by the Mananciais de Campos do Jordão and Campos do Jordão State Parks; Willis and Oniki 1981) and Itatiaia National Park (Barreto et al. 2013), as well as regions of semi-deciduous forest: Barreiro Rico Ecological Station (Antunes 2005), Caetetus Ecological Station (Tabanez et al. 2005, Cavarzere et al. 2009), Mata dos Godoy State Park (Anjos 2001), Morro do Diabo State Park (Uezu and Metzger 2016), Porto Ferreira (composed by Porto Ferreira and Vassununga State Parks, Uezu and Gaban-Lima 2003) and Rio Claro Farm (Donatelli et al. 2004). In addition to the references cited above, to complement lists of birds from those regions, we also used records from WikiAves and Xeno-Canto databases (WikiAves 2016, Xeno-Canto 2016). Most of those regions are or incorporate large conservation areas (reserves) where bird richness is high. Additionally, whenever possible, we also compiled records from patches surrounding the parks and reserves, to better characterize the avifauna, in the same manner that we did for Serra da Cantareira. These regions are ~230 km distant from our study area (50 km the nearest and 600 km the farthest; Fig. 2). For comparisons, we ignored seabirds records for Boracéia Biological Station and municipalities of Ilha Comprida and Ubatuba (e.g. the Magnificent Frigatebird Fregata magnificens [Mathews, 1914]) as well as exotic species, such as the House Sparrow Passer domesticus (Linnaeus, 1758), and birds from other regions in Brazil that escaped or were intentionally released from captivity (e.g. the Yellow-rumped Cacique Cacicus cela [Linnaeus, 1758]).

Figure 2. 

Dendogram with Atlantic Forest regions based on their bird community using Hierarchical Cluster Analysis (Mountford’s distance coefficient; a) and the Principal Coordinate Analysis of these areas (b). 1 = Morro do Diabo State Park; 2 = Mata dos Godoy State Park; 3 = Caetetus Ecological Station; 4 = Rio Claro Farm; 5 = Intervales State Park; 6 = Barreiro Rico Ecological Station; 7 = Carlos Botelho State Park; 8 = Ilha Comprida; 9 = Morro Grande Reserve; 10 = Porto Ferreira (Porto Ferreira and Vassununga State Parks); 11 = Serra da Cantareira; 12 = Boracéia Biological Station; 13 = Campos do Jordão (composed by Mananciais de Campos do Jordão and Campos do Jordão State Parks); 14 = Ubatuba; 15 = Itatiaia National Park.

Overall, 326 species were found at Serra da Cantareira, including seven exotic species, e.g. the Rock Pigeon Columba livia (Gmelin, 1789), or species that may have escaped from captivity, such as the Red-cowled Cardinal Paroaria dominicana (Linnaeus, 1758). For nine species with previous records for Serra da Cantareira there have been no further records for the last two decades, and 55 species records are not adequately documented (through either museum specimen, photograph or song record; Appendix 1). Additionally, nine questionable literature records of bird species, for instance the Vinaceous Parrot Amazona vinacea (Kuhl, 1820), were not included in this report (Appendix 2). Non-passerines accounted for 138 species in 32 families and passerines are represented by 188 species in 34 families (Piacentini et al. 2015). Tyrannidae was the most representative family (35 species), followed by Thraupidae (33) and Furnariidae (18). AF endemics accounted for 24.5% of the records according to Lima’s (2013) classification and 227 are forest birds (Parker et al. 1996). Seven species recorded in our study area are globally threatened (BirdLife 2016); five are threatened in Brazil (MMA 2014) and 17 in the state of São Paulo (São Paulo 2014). The ornithological knowledge of Serra da Cantareira has increased significantly since Graham’s study. Only 71 birds were reported before 1992, all represented by museums specimens (Fig. 3). Records exclusively from the authors of this study contributed to add four species (Appendix 1). Species in the urban slope accounted for most records (Table 1).

Figure 3. 

Number of species reported by different sources during different periods in Serra da Cantareira. Museums: specimens deposited in ornithological collections between 1901 and 1992; Graham: Graham 1992; Authors: species recorded by the authors during unsystematic surveys; WA: WikiAves 2016; XC: Xeno-Canto 2016; Recent literature: Figueiredo and Loo 2000, Develey and Endrigo 2004, Fonseca and Antunes 2007, Antunes et al. 2008a, 2008b, 2009, Antunes and Eston 2009, 2012, Minns et al. 2009, Dersa 2010, Ikuta and Martins 2013, Lemos 2014, Cavarzere and Tonetti 2015. All recent encompasses records obtained in recent literature cited above as well as specimens deposited in museuns after the work performed by Graham (1992) and species recorded by the authors

HCA showed two major groups of areas, separated according to their bird communities. One is in a macro-region where the vegetation is predominantly ombrophilous dense or ombrophilous mixed forest near the coast, and comprises the following areas: Boracéia Biological Station, Campos do Jordão, Serra da Cantareira, Carlos Botelho and Intervales State Parks, Ilha Comprida, Itatiaia National Park, Morro Grande Reserve and Ubatuba. The other group, encompassing more inland areas, has lower rainfall rates with seasonal semi-deciduous forests and includes: Barreiro Rico Biological Station, Caetetus Ecological Station, Mata dos Godoy State Park, Morro do Diabo State Park, Porto Ferreira and Rio Claro Farm (Fig. 2). The bird community of Serra da Cantareira is more similar to a clade composed of Boracéia Biological Station, Carlos Botelho and Intervales state parks and Morro Grande Reserve (Fig. 1a) according to HCA. According to the Jaccard index the bird community of Serra da Cantareira is more similar to Carlos Botelho State Park, while Pontal do Parapanapanema has a more distinct avifauna (Table 2). Although the ombrophilous mixed forests, represented here by Campos do Jordão and Itatiaia National Park, are more related to ombrophilous dense forests according to HCA, they form a separated clade within this major group (composed by areas 13 and 15, Fig. 2a). The PCoA also showed two regions of ombrophilous mixed forests that are more related to ombrophilous dense forests in the first dimension, while in the coordinate 2 they were more distant from other omborphilous dense forest areas (Fig. 2b).

Our results are consistent with other ornithological compilations conducted at Boracéia Biological Station, which is an AF reserve located at Serra do Mar, 75 km distant from Serra da Cantareira, and where the vegetation is also dense mountain rain forest (Fig. 2). Cavarzere et al. (2010) performed transect counts, mist netting and unsystematic surveys at the station, where they recorded 242 bird species. After accounting for museums specimens, literature records, and records provided by birdwatchers, 323 species were reported for that reserve, 30% of which are endemic to the AF (Cavarzere et al. 2010, Lima 2013). This suggests that ornithological reviews provide a more comprehensive knowledge on the actual bird diversity of a given region than localized surveys, even when using different methodologies and intensive sampling effort.

Some records of species for Serra da Cantareira are based on only one source of information (records from the authors, those provided by birdwatchers, reported in the literature or represented by museums specimens; Fig. 3), e.g. the Gray-bellied Hawk Accipiter poliogaster (Temminck, 1824). This bird was only recently spotted by the authors and had not been previously recorded there. Despite its broad-range distribution (north Colombia to south Brazil, and more recently found in Costa Rica and Panama), this species is considered uncommon and is sparsely found throughout its range (Bierregard et al. 2016). VRT first recorded a juvenile at Cabuçu (a CSP administrative zone; Fig. 1) on May 5, 2014. On that day, the birds’ vocalization was recorded.

The high number of species recorded by birdwatchers (76% of all species compiled in this study, ten of which reported only by them; Appendix 1) highlights the importance of public websites where digital vouchers can be deposited, such as WikiAves (http://www.wikiaves.com) and Xeno-Canto (http://www.xeno-canto.org). Among the records provided only by these databases is the Shrike-like Cotinga Laniisoma elegans (Thunberg, 1823) (WA1076770). This frugivorous bird, found mostly in old-growth forests, was photographed in a forest edge at the Engordador administrative CSP zone. Birds recorded only in the literature accounted for 13 species, such as the Rufous-tailed AttilaAttila phoenicurus Pelzeln, 1868, song-recorded at Cabuçu (Minns et al. 2009). The Rufous-tailed Attila breeds in southeast Brazil during the austral winter, and is found in greater numbers in Serra do Mar (Walther 2016), being uncommon in our study area (pers. obs.).

Although museum records until 1992 did not account for 80% of the species in Serra da Cantareira, some relevant birds were reported only until that year. The most noticeable is the Purple-winged Ground-Dove Claravis geoffroyi (Temminck, 1811) collected in 1937 at a CSP site named “Cuca” (MZUSP 17040; Fig. 1). It was once a fairly common and widely distributed species (occurring from Bahia [Brazil] south through eastern Brazil to northern Argentina and eastern Paraguay). During the last three decades, however, there have beeen only a few non-documented records of this globally “Critically Endangered” AF endemic species (MMA 2014, BirdLife 2016). Claravis geoffroyi feeds on bamboo seeds (Guadua sp.), and most likely travels far in search for flowering events (Sick 1997, Areta et al. 2009). Guadua spp. have synchronous masting events in cycles of ~30 y, attracting birds that eat their seeds, such as the Buffy-fronted Seedeater Sporophila frontalis (Verreaux, 1869) and the Temminck’s Seedeater S. falcirostris (Temminck, 1820), both registered at Serra da Cantareira and globally threatened (Areta et al. 2009, BirdLife 2016). As Guadua sp. bamboo tickets are common in several parts of Serra da Cantareira (Bencke et al. 2006; pers. obs.), we suggest that those searching for the Purple-winged Ground-Dove should try to locate flowering events of that bamboo.

Due to severe forest loss and human intervention at Serra da Cantareira, it is possible that local extinctions have occurred, but they are difficult to document. For example, Cavarzere et al. (submitted) recorded the Squamate Antbird Myrmoderus squamosus (Pelzeln, 1868) after a 190-y time span from the last confirmed record at the Ipanema National Forest, an AF reserve 100 km from our study area. This is a good example of imperfect detection and insufficient sampling effort (Mackenzie et al. 2003). The Spot-billed Toucanet Selenidera maculirostris (Lichtenstein, 1823) is an example of a species recently recorded at Serra da Cantareira after 45 y without records. Six specimens were collected in 1965 (MZUSP 60592 to 60597), and after it, the only confirmed record was provided by MAR and his team in A5 (Fig. 1).

More important than accounting for local bird extinctions per se is identifying functional extinctions and their impact on the ecosystem. Habitat disturbance is one of the main drivers of functional extinctions of sensitive guilds, such as understory insectivorous and large-bodied frugivorous (Martensen et al. 2012, Galetti et al. 2013, Morante-Filho et al. 2015). It is very possible that, of the currently rarest species in our study area – e.g. the Saffron Toucanet Pteroglossus bailloni (Vieillot, 1819), a large-bodied frugivorous, and the Speckle-breasted Antpitta Hylopezus nattereri (Pinto, 1937), an understory insectivorous – were more abundant in the past, given their presence in museums collections. Declines in the populations of these birds have been recorded in other AF regions due to habitat disturbances (Guix et al. 2000, Anjos 2006), and since they participate in key-ecological processes (e.g. seed dispersal), the consequences of their functional extinctions should be better investigated.

In contrast with the decreasing populations of some species in response to habitat disturbances, species that benefit from altered environments are expected to expand their ranges and increase their abundance. Graham (1992), using point-counts (37 hours of sampling effort) between December 11^th 1985 and March 29^th 1986, did not record any individual of the White-eyed Parakeet Psittacara leucophthalmus (Statius Muller, 1776) within the CSP, although this species has been frequently recorded in our study area over the last decade. The White-eyed Parakeet benefits from deforestation and occupies forest edges and areas of open vegetation, and its presence in AF is an indication of environmental degradation (Sick 1997). Increases in the numbers of generalist bird species, together with decreases in the numbers of sensitive species, have also been reported for other large and protected AF remnants after a 30-y time span – at the Caetetus Ecological Station (Cavarzere et al. 2012; 2,800 ha, Fig. 2).

We thank the staff of Cantareira State Park for logistic assistance; the scientific technical committee (COTEC) which authorized VRT to work on the Park (proc. 260108-007.094/2013); Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio) for collecting permits; the Brazilian National Council for Scientific and Technological Development which provided a scholarship to VRT (proc. 130279/2013-7) and for grants and financial support to LFS; people who contribute in the maintenance as well as providing their personal records to WikiAves and Xeno-Canto databases; MAR team: Aline Correa, Fernanda Alves, Fernanda Bocalini, Giuliana Althman, Glaucia Del Rio, Leo Signorini, Rafael Marcondes, Thiago Vernaschi and Vagner Cavarzere. We are also indebted with Tatiana Pongiluppi. Fernanda Alves helped with the English. Maria de Fátima Rodrigues funded part of this research.