Introduction
The Restinga Antwren Formicivora littoralis is a threatened species endemic to restinga vegetation within a highly restricted area of coastal eastern Brazil. Formicivora littoralis was described by Gonzaga and Pacheco (Reference Gonzaga and Pacheco1990) as a subspecies of Serra Antwren F. serrana. Because of the Restinga Antwren's limited distribution, and its morphological and ecological differentiation, it has been treated as a separate species since Collar et al. (Reference Collar, Gonzaga, Krabbe, Madroño Nieto, Naranjo, Parker and Wege1992). A recent study shows that the taxon is clearly in the process of differentiation, although no diagnostic morphological or vocalization characters were found (Firme Reference Firme2008).
The Restinga Antwren's known distribution extends 80 km along the coast of Rio de Janeiro state, from Saquarema to Cabo Frio municipalities (Gonzaga and Pacheco Reference Gonzaga and Pacheco1990). It is a habitat specialist restricted to restinga scrub with abundant cacti and terrestrial bromeliads, on white sand soil of sandbars and sand dunes, and coastal and island hillsides (Zimmer and Isler Reference Zimmer, Isler, del Hoyo, Elliot and Christie2003). Its habitat is under great threat as it coincides with a vacation destination and prime beach real estate. There are 21 restinga remnants in the state of Rio de Janeiro, totalling 105,285 ha (Rocha et al. Reference Rocha, Bergallo, Van Sluys, Alves and Jamel2007). There are 14 causes of degradation to restingas in Rio de Janeiro, with the greatest threat to the Restinga Antwren's habitat being village construction, salt ponds and invasions of natural reserves (Alves et al. Reference Alves, Pacheco, Gonzaga, Cavalcanti, Raposo, Yamashita, Maciel, Castanheira, Bergallo, Rocha, Alves and Van Sluys2000, Reis and Gonzaga Reference Reis, Gonzaga, Esteves and Lacerda2000, Rocha et al. 2003, Reference Rocha, Bergallo, Van Sluys, Alves and Jamel2007). The habitat of the species is steadily shrinking and most remnants lack legal protection. Under current human pressure, most of its unique habitat is likely to be lost from Rio de Janeiro state within the next few years (Rocha et al. Reference Rocha, Bergallo, Van Sluys, Alves and Jamel2007).
In 1994, IUCN listed the Restinga Antwren as ‘Endangered’ and in 2004 uplisted the species to ‘Critically Endangered’ (BirdLife International 2008). The listing was warranted because of its small known and severely fragmented range, together with a declining habitat linked to beachfront development. Information about the Restinga Antwren, however, is still very limited. The only pertinent literature published since Collar et al. (Reference Collar, Gonzaga, Krabbe, Madroño Nieto, Naranjo, Parker and Wege1992) is Cerqueira (Reference Cerqueira1995) on the potential distribution of the species, Tobias and Williams (Reference Tobias and Williams1996) on behaviour, threats and conservation, and Vecchi and Alves (Reference Vecchi and Alves2008) on the inland expansion of its range. None of these studies gathered the necessary data to evaluate the species’s conservation status according to the IUCN criteria.
Following the recommendations of Collar et al. (Reference Collar, Gonzaga, Krabbe, Madroño Nieto, Naranjo, Parker and Wege1992), we conducted field surveys of the Restinga Antwren to better ascertain and monitor its status. The information was used to re-assess the species’s conservation status based on three IUCN criteria: Extent of Occurrence, Area of Occupancy, and number of mature individuals (IUCN 2001).
Methods
Bird surveys: geographic range
From October 2005 through June 2006, we surveyed the previously known range of the Restinga Antwren, from Saquarema to Cabo Frio municipalities (Gonzaga and Pacheco Reference Gonzaga and Pacheco1990), as well as adjacent areas with potential habitat for the species (Figure 1). The surveys were conducted from Maricá municipality to Armação dos Búzios municipality, all in the state of Rio de Janeiro (Table 1). We used playback to determine the presence and absence of the Restinga Antwren, stopping every 2 km along the motorways RJ-102 and RJ-140. At each point, we broadcast a five-minute vocalization of Formicivora serrana (from Gonzaga and Castiglioni Reference Gonzaga and Castiglioni2001) and waited another five minutes, totalling 10 minutes of sampling. Gonzaga and Pacheco (Reference Gonzaga and Pacheco1990) found that the Restinga Antwren responds well to Formicivora serrana's playback. We also tested and confirmed that in previous, exploratory surveys. We conducted playback to improve the chances of detecting the birds, as suggested by Johnson et al. (Reference Johnson, Brown, Haight and Simpson1981) and Vale et al. (Reference Vale, Bell, Alves and Pimm2007).
– Restinga Antwren extent of occurrence and regional occurrence (presence and absence data over 166 surveyed sites) in eastern Rio de Janeiro state, Brazil. Local abundance survey sites are indicated with stars. New record is in bold (Tucuns). Municipalities: ABU = Armação dos Búzios, ACA = Arraial do Cabo, ARA = Araruama, CFR = Cabo Frio, IGU = Iguaba Grande, MAR = Maricá, SAQ = Saquarema, SPE = São Pedro da Aldeia.
Regional abundance of the Restinga Antwren (presence and absence data) in the most important restinga fragments within its geographic range in eastern Rio de Janeiro state, Brazil.
Additionally, from 2005 to 2007, we conducted further surveys of vegetation fragments where the species could potentially occur, in the following municipalities: Maricá, Saquarema, Arraial do Cabo, Araruama, Iguaba Grande, São Pedro da Aldeia, Armação dos Búzios and Cabo Frio.
Bird surveys: regional occurrence
We estimated the regional occurrence as the probability of finding the species within its geographic range (Vale et al. Reference Vale, Bell, Alves and Pimm2007). We divided the number of points where the species was present within its geographic range by the total points surveyed within its geographic range. It is important to note that although presence data are definite, absence data are not. If the species responded to playback, we were sure it was present at that point but if it did not respond, it could either be absent or present but not responsive. Therefore, our estimate of the probability of finding a species in a given site is likely underestimated.
To double check the presence and absence data, we conducted an a posteriori visit to 30 presence and 30 absence points. Those points were drawn at random from the pool of all the points that had been surveyed. False-negatives (points recorded as absences where later the bird was detected) were considered as presence point on the final analysis. Points where the species was previously present but was not detected on the second survey were considered as presence.
Bird surveys: local abundance
The fieldwork for local abundance ran from November 2005 to February 2006. The surveys took place in Praia do Peró (22°50’41”S, 41°59’14”W) and Praia das Conchas (22°52’09”S, 41°59’14”W) in Cabo Frio municipality (Figure 1). These sites have dense restinga vegetation on sand dunes. In addition, Praia das Conchas also has restinga vegetation on hills. In the two sites, we used pre-existing trails that run through the middle of the restinga fragments.
We used the point-count with playback method to estimate the Restinga Antwren local abundance. We flagged points at 200 m intervals, within a 2-km transect, enabling a total of 10 points at each site. We visited each site eight times (four mornings and four afternoons), totalling 16 surveys. During each visit, the same observer did the survey in the morning and again in the afternoon. Each survey started at 07h00 and took three hours to complete. At each survey point, we broadcast the same pre-set vocalization used for the regional occurrence survey, recording all individuals heard or seen within a 100 m radius from the observer. We standardized the time spent at each point by the length of the playback routine (five minutes of playback plus five minutes of silence).
Data analysis: local population density
We used local abundance to estimate population density for each surveyed site, following the procedure in Vale et al. (Reference Vale, Bell, Alves and Pimm2007). The total number of individuals at a site (n) was defined as the sum of the maximum number of individuals recorded at each point surveyed for that site. We did not consider the sex when determining the total number of individuals because both male and female Restinga Antwrens were responsive to playback. We calculated the population density (D) per site as the total number of individuals (n) divided by the area surveyed. The total area surveyed was defined as the area of the 100-m radius circle around the 10 points surveyed in each site.
Data analysis: available habitat, extent of occurrence, area of occupancy and global population size
We used the polygon and area tools of Google Earth Professional to map the available restinga habitat of the Restinga Antwren. Google Earth images in the target area have a 1m-resolution that enables visual identification of the restinga fragments. We mapped all the restinga fragments within the species’s geographic range (Figure 1).
We multiplied the available habitat by the mean population density (D) to estimate the species’s maximum global population size. This is the maximum global population because it assumes that the entire available habitat is occupied by the species. We calculate the minimum global population size by multiplying the maximum global population by the probability of finding the species in a given site (regional occurrence) (Vale et al. Reference Vale, Bell, Alves and Pimm2007). This is a minimum population size because the probability of finding the species in a given site is likely underestimated. Therefore, the species’s global population size should lie somewhere between the estimated maximum and minimum population size.
In order to estimate the Area of Occupancy (sensu IUCN 2001), we divided the entire known geographic range of the species into a grid with 2 x 2 km-cells (Standards and Petitions Working Group 2006). The grids were created using the Hawth's Tools in ArcMap 9.2. We counted the total number of grid cells with presence points of Restinga Antwren and multiplied it by the area of the grid cell to obtain the Area of Occupancy. In addition, we estimated the Extent of Occurrence (sensu IUCN 2001) by using an ‘imaginary boundary which encompasses all the sites of presence occurrences.’
Results
Geographic range limits, extent of occurrence and regional occurrence
In 2006, we surveyed a total of 166 points in the region where the Restinga Antwren is known to occur and adjacent areas where we suspected the species might be present (Figure 1). We confirmed the species’s presence from Saquarema to Cabo Frio municipalities, which is in agreement with Gonzaga and Pacheco (Reference Gonzaga and Pacheco1990). It is important to note that Gonzaga and Pacheco's point at Praia de Jaconé (42°39’59”W, 22°56’28”S) is actually within Maricá municipality, not Saquarema municipality as reported by these authors. We also found a new location where the species was present in a 16-ha fragment of restinga at Praia de Tucuns, Armação dos Búzios municipality (41°55’53” W, 22°48’00” S) (Figure 1, Table 1). This is the easternmost point for the species, 5 km from the previously known easternmost limit (Praia Peró, Cabo Frio municipality, Gonzaga and Pacheco Reference Gonzaga and Pacheco1990), expanding the species’s distribution to Armação dos Búzios municipality.
The point at Praia de Jaconé is on the border between Maricá and Saquarema municipalities, and marks the westernmost limit of the species’s distribution. The species was absent between Praia de Tucuns and Praia do Peró, where there is a large fragment of semi-deciduous forest, unsuitable for the Restinga Antwren. In 2007, we conducted further surveys between Praia do Peró and Praia de Tucuns (at Praia de Caravelas and Serra das Emerências), as well as in Armação dos Búzios municipality, but these did not produce any additional records. We conclude, therefore, that the population in Praia de Tucuns represents a peripheral and isolated population of the Restinga Antwren.
We propose a geographic range of the Restinga Antwren with a continuous distribution from Maricá to Cabo Frio municipalities and an isolated population in Armação dos Búzios municipality (Figure 1). This range represents an Extent of Occurrence (sensu IUCN 2001) of 233.5 km2 (that excludes the area within Lagoa de Araruama).
Of the 166 points surveyed, 104 were within restinga vegetation in the species’s geographic range. The Restinga Antwren was present at 65% of these 104 points. The species was very common on coastal restinga between Saquarema and Cabo Frio municipalities (Figure 1, Table 1), often withstanding substantial habitat disturbance or persisting in small fragments of restinga. Praia de Tucuns, with an intermediate abundance, represents an expansion of the species range, where the species is present only where restinga vegetation is also present at the beachfront. The hills of Armação dos Búzios were the only site with restinga vegetation inside the species’s geographic range where the Restinga Antwren was absent.
On the a posteriori revisit to 60 survey points, we detected three presence points previously considered as absence and one absence previously recorded as presence. These four errors translate into a 93% accuracy of the survey. We did not detect any visible habitat changes in these sites and therefore we assume that it is a detectability error.
Local abundance
The bird abundance and density were similar for both sites surveyed. We recorded a maximum number of 29 individuals at Praia do Peró, which translates into a population density of 92.3 individuals km−2 (around 46 pairs). At Praia das Conchas, we recorded a maximum number of 28 individuals and a density of 89.1 individuals km−2 (around 44.5 pairs). Most commonly, we observed solitary individuals (either male or female) or pairs. On rare occasions, we observed groups of four individuals (male, female and two young). Both male and female Restinga Antwrens responded well to playback, the birds closely approaching the observer. Although both males and females responded to the playback, males seem to exhibit stronger territorial behaviour. There was a significant difference between the numbers of males and females observed per survey (paired t-test, t = 18.6, df = 15, P < 0.001), with 58% of the 235 visual contacts being males.
Available habitat, area of occupancy and global population size
The Restinga Antwren was present in restinga fragments covering an area of 42.3 km2, or 84% of the total restinga within the species’s geographic range. This percentage is likely an overestimate because it assumes that the species is present in the entire fragment, even if that fragment was not entirely surveyed.
The non-restinga areas were mostly human-modified landscapes (urban areas, roads, salt ponds and farms), with some natural habitats (bare sand dunes and semi-deciduous forest). The combination of the species’s available habitat (42.3 km2) and its mean population density (90.7 individuals km−2) translates into a global population size of 3,837 individuals. If we factor in the species’s regional abundance (65.4% of points surveyed), we have a global population size of 2,509 individuals. The true population size should fall somewhere in between these two estimates. The species was present in 37 of the 2 x 2 km cells within its geographic range, translating into an estimated Area of Occupancy (sensu IUCN 2001) of 148 km2.
Discussion
Local abundance
The playback method has been used by different authors for population estimates of secretive species or species that occur in nearly impenetrable habitats (Hutto et al. Reference Hutto, Pletschet and Hendricks1986, Bibby et al. Reference Bibby, Burgess, Hill and Mustoe2000, Vale et al. Reference Vale, Bell, Alves and Pimm2007). The Restinga Antwren falls into both categories – its usually high local abundance can be completely overlooked by an observer who does not use playback. This method adequately estimates population abundance and greatly expedites fieldwork. For a discussion of the implications of the use of playback for estimating population abundance, refer to Vale et al. (Reference Vale, Bell, Alves and Pimm2007).
Our estimate of population density of about 45 pairs km−2 is within the range of what has been found for similar species. Using the same method as in the present study, Vale et al. (Reference Vale, Bell, Alves and Pimm2007) found a population density of 38 pairs km−2 for the Rio Branco Antbird Cercomacra carbonaria, also a typical antbird of scrubby vegetation. Using spot-map methods, Robinson and Terborgh (Reference Robinson and Terborgh1997) found an average density of 30 pairs km−2 for understorey insectivorous birds in shrubby stands of Tessaria integrifolia on sandy soils. For the two typical antbirds with similar body size to Restinga Antwren, Robinson and Terborgh (Reference Robinson and Terborgh1997) recorded 19 pairs km−2 (Blackish Antbird, Cercomacra nigrescens) and 56.5 pairs km−2 (Black-throated Antbird, Myrmeciza atrothorax). It seems that the Restinga Antwren has a population density somewhat at the higher end of the spectrum for the family. The Restinga Antwren, therefore, seems to have a relatively high local abundance. According to Parker et al. (Reference Parker, Stotz, Fitzpatrick, Stoltz, Fitzpatrick, Parker and Moskovits1996), the Relative Abundance classification for the Restinga Antwren is ‘Fairly Common’. Our abundance estimates and field observations, however, indicate that the species should be classified as ‘Common’ which means that the species, ‘occurs throughout its range in moderate to large numbers and are found easily during brief periods of fieldwork.’ It is important to highlight that the species is only common locally. Regionally and globally, the Restinga Antwren is extremely rare, occurring only in the sparsely distributed restinga fragments within a very limited geographic range.
Conservation status
The Restinga Antwren is listed as ‘Critically Endangered’ in IUCN's Red List of threatened species because its Area of Occupancy is estimated at < 10 km2, and estimates indicating a severely fragmented habitat, and a continuing decline in the Extent of Occurrence, Area of Occupancy, habitat and number of mature individuals. The listing was based on the best available knowledge to date (Birdlife International 2008).
We estimated that the Restinga Antwren has an Extent of Occurrence of 233.5 km2, which, although quite close the cut-off for ‘Critically Endangered’ status, meets the 5,000-km2 cut-off for ‘Endangered’ status (IUCN 2001). We estimated the Area of Occupancy to be 148 km2, which also meets the 500-km2 cut-off for the ‘Endangered’ status. Finally, we estimated a global population size ranging between 2,509 and 3,837 individuals, which meet the 10,000 cut-off for the ‘Vulnerable’ status. We, therefore, recommend a downlisting of the Restinga Antwren from ‘Critically Endangered’ to the ‘Endangered’ threat category of IUCN. The recommended downlisting should be interpreted with caution as it is based on a more accurate assessment of the parameters that determine the species’s status, rather than any suggestion that pressures on the species’s habitat have lessened. Furthermore, the estimated Extent of Occurrence is extremely close to the cut-off size for ‘Critically Endangered’ meaning that the eventual extirpation of F. littoralis from a small area of its range could be sufficient to push it into the ‘Critically Endangered’ category.
Habitat availability and threats
Real estate development constitutes the main threat to the habitat of the Restinga Antwren. Rio de Janeiro is an international tourist destination, especially because of its beautiful coast. The Restinga Antwren occurs in a prime beach area within the state known as Região dos Lagos (Vecchi and Alves Reference Vecchi and Alves2008). This region is under high pressure from urban development. Particularly worrisome to the Restinga Antwren is the coming of a number of mega-investments.
Within the Restinga Antwren's known range, there are two ‘Areas of Environmental Preservation’ (APA, in the Portuguese acronym): APA Pau Brasil with 9,940 ha covering Cabo Frio and Armação dos Búzios municipalities, and APA da Massambaba with 7,630 ha covering Saquarema, Araruama and Arraial do Cabo municipalities. This protected area category, however, has very little restriction on human activities (IBAMA 2001) and does not offer effective protection for the species. Both APA Pau Brasil and APA da Massambaba are highly disturbed by human visitation, logging and inappropriate garbage disposal (Rocha et al. Reference Rocha, Bergallo, Alves and Van Sluys2003). The APA Pau Brasil is extremely threatened by the upcoming construction of the Resort Peró, a mega-investment by a number of international corporations, including Club Méditerranée and Sheraton. The construction of this resort will likely destroy most of the restinga fragment in Praia do Peró (within APA Pau Brasil), an area identified by the present study to harbour an important population of the Restinga Antwren.
Within the APA da Massambaba, there is also the Massambaba State Reserve located in Arraial do Cabo municipality, which is the largest preserved restinga fragment where the species occurs (c. 1,680 ha). Currently, the best restinga fragments and largest populations of the Restinga Antwren are within the APA Pau Brasil and APA Massambaba. Neither, however, provides effective or sufficient protection for the species. We recommend converting the still-preserved areas within both APAs to Biological State Reserves.
The new record at Praia de Tucuns expands the previously known range. The fragment of Restinga at Tucuns beach is small, but also well preserved and hosts an isolated population of the species that needs protection. Unfortunately, this area is also threatened by the recent construction of the Club Breezes Mega Resort (Barros Reference Barros2008). This resort is located about 1 km from where the Restinga Antwren population was found during the present study and is likely to cause highly increased tourism pressure on the area.
Marginal habitats to the Atlantic forest, such as restinga, are often neglected in conservation strategies (Scarano Reference Scarano2002). The restinga vegetation is protected under the Brazilian Federal Law 4.771 of 1965 for its role in the fixation of dunes. There is a need to enforce the conservation actions and environmental laws that already exist. The Restinga Antwren, considered the sole endemic bird of the restinga (Gonzaga and Pacheco Reference Gonzaga and Pacheco1990), could act as flagship species for the conservation of the few restinga remnants in Rio de Janeiro state.
Acknowledgments
We thank Fundação Biodiversitas’ project for endangered species (PEA 056/022004) and Fundação O Boticário de Proteção à Natureza (0730_20062) for their financial support. Without this support we would not have been able to complete this work. We thank our field work team (Flávia G. Chaves and Thiago F. S. Laurindo), Dr. Cyl Farney C. Sá from the Botanical Garden of Rio de Janeiro, Dr. Stuart L. Pimm and Dr. Clinton N. Jenkins from Duke University for all their input and support. We thank Movimento Ambiental Pingo D’Água and Instituto Biomas for the assistance and logistical support. While writing this paper, the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) provided a PhD fellowship to J.C.F.M., Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) provided a research grant to M.A.S.A. (no. 302718/2003-6) and a Post-doc fellowship to M.M.V., Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ) a Post-doc fellowship to M.B.V. Guy Kirwan, Stuart Butchart and an anonymous referee provided valuable comments on the manuscript.
