On the use of 10-minute point counts and 10-species lists for surveying birds in lowland Atlantic Forests in southeastern Brazil

On the use of 10-minute point counts and 10-species lists for surveying birds in lowland Atlantic Forests in southeastern Brazil

Vagner Cavarzere^I,II,III; Thiago Vernaschi Vieira da Costa^I,II; Luís Fábio Silveira^II

^IPós-Graduação, Departamento de Zoologia, Instituto de Biociências, Universidade de São Paulo. Rua do Matão, travessa 14, 101, CEP 05508-900, São Paulo, SP, Brasil.

^IISeção de Aves, Museu de Zoologia da Universidade de São Paulo. Avenida Nazaré, 481, CEP 04218-970, São Paulo, SP, Brasil.

^IIICorresponding author e-mail: [e-mail]

ABSTRACT

Due to rapid and continuous deforestation, recent bird surveys in the Atlantic Forest are following rapid assessment programs to accumulate significant amounts of data during short periods of time. During this study, two surveying methods were used to evaluate which technique rapidly accumulated most species (> 90% of the estimated empirical value) at lowland Atlantic Forests in the state of São Paulo, southeastern Brazil. Birds were counted during the 2008-2010 breeding seasons using 10-minute point counts and 10-species lists. Overall, point counting detected as many species as lists (79 vs. 83, respectively), and 88 points (14.7 h) detected 90% of the estimated species richness. Forty-one lists were insufficient to detect 90% of all species. However, lists accumulated species faster in a shorter time period, probably due to the nature of the point count method in which species detected while moving between points are not considered. Rapid assessment programs in these forests will rapidly detect more species using 10-species lists. Both methods shared 63% of all forest species, but this may be due to spatial and temporal mismatch between samplings of each method.

Key-Words: Bertioga, Bird species richness; MacKinnon lists; Rapid assessment programs; Surveying methods.

RESUMO

Devido ao rápido e contínuo desmatamento, inventários avifaunísticos realizados na Mata Atlântica têm seguido protocolos de levantamentos rápidos para acumular uma grande quantidade de informação em períodos de tempo relativamente curtos. Para averiguar qual metodologia deve mais rapidamente acumular o maior número de espécies (cerca de 90% das espécies estimadas a partir de dados empíricos), foram realizados pontos de escuta de 10 minutos e listas de 10 espécies em florestas de baixada da Mata Atlântica de São Paulo durante as estações reprodutivas entre 2008 e 2010. De maneira geral, pontos detectaram tantas espécies quanto as listas (79 versus 83, respectivamente), sendo que 88 pontos (14,7 h) detectaram 90% da riqueza estimada. Quarenta e uma listas foram insuficientes para o registro de 90% da riqueza estimada, porém acumularam mais rapidamente o número de espécies e num período mais curto de tempo, provavelmente devido à natureza da metodologia de pontos de escuta, na qual espécies registradas durante a locomoção entre pontos não são consideradas. Levantamentos rápidos nessas florestas irão rapidamente detectar mais espécies com a utilização das listas de 10 espécies. As metodologias compartilharam 63% das espécies registradas, mas este resultado pode refletir diferenças nas amostragens definidas em localidades e intervalo de tempo distintos durante a realização das amostragens.

Palavras-Chave: Levantamentos rápidos da avifauna; Listas de MacKinnon; Metodologias de censo; Bertioga; Riqueza de aves.

INTRODUCTION

In light of increasing destruction of forests and wide gaps in the understanding of tropical bird communities, several researchers have recently applied a rapid assessment approach to maximize data collection with limited funds, time, and personnel (Parker & Bailey, 1991; Poulsen et al., 1997). Some auditory-visual rapid assessments employ standard techniques for surveying birds, such as point counts (Poulsen & Krabbe, 1998) or line transects (Karr, 1971). Although those methods generate quantifiable data with well-established sampling protocols, they present several disadvantages for rapid assessments, such as: they are often difficult to apply under tropical field conditions; require highly qualified observers (Poulsen et al., 1997); and tend to underestimate the richness and abundance of some groups (e.g., nocturnal species). Nonetheless, point counts remain widely used in surveying birds in Neotropical forests.

MacKinnon & Phillips (1993) suggested a quantitative approach to analyzing auditory-visual survey data that accounts for differences in effort, observer qualifications and weather (Poulsen et al., 1997). In this method, observations are grouped into consecutive lists of 20 species, and a species accumulation curve is generated from the addition of those species not recorded on any previous list to the total species number, which is then plotted as a function of list number. It is crucial to include even observations that cannot be positively identified at first (Poulsen et al., 1997). Because the method relates species richness to the number of observations rather than to time, area or walking speed, this method allows for comparison of data obtained by different observers or under varying field conditions (Herzog et al., 2002).

Point counts have been tested in Neotropical systems and some authors concluded that 5- to 10-min counts are ideal for detecting a significant number of species, including endemic and threatened species (Develey, 2004; Esquivel & Peris, 2008). In addition, Herzog et al. (2002) suggested the use of 10-species lists in tropical regions as a solution between 5-, 10- and 20-species lists. Ribon (2010) further suggested that species lists are better than any other method for surveying birds, but the author did not give empirical examples or comparisons for his statement. The list method of avifaunal assessment has been increasingly adopted for tropical bird studies worldwide, from Indonesia (MacKinnon & Phillips, 1993; Trainor, 2002a,b), to mainland Africa (Fjeldså, 1999), Madagascar (O’dea et al. 2004), and South America (Poulsen et al., 1997; O’dea et al., 2004; Herzog & Kessler, 2006; Herzog, 2008). It has also been promoted as a potentially useful technique in a manual on bird census methods (Bibby et al., 2000). In this study, we surveyed forest birds using 10-min point counts and 10-species lists in lowland Atlantic Forests in southeastern Brazil to test the hypothesis that both methods equally detect most bird species (> 90% of the estimated species richness) during short periods of time, making them suitable for rapid assessment programs.

MATERIAL AND METHODS

Study area

We carried out bird counts in the coastline municipalities of Bertioga and Santos (23º57’S, 46º19’W and 23º51’S, 46º08’W, respectively) in the state of São Paulo, southeastern Brazil (Fig. 1). Bertioga has about 480 km² of Atlantic Forest, 85% of which constitutes areas of environmental protection (Maia et al., 2008). According to Köppen’s classification, the climate of the region is "Af", humid or super humid tropical, with rains distributed throughout the year (Nascimento & Pereira, 1988). Climatologic data monitored between 1941 and 1970 indicates that the mean annual temperature is 24.8ºC, with lowest and highest monthly means of 20.7ºC in July and 28.3ºC in February, respectively. Bertioga is one of the most humid regions in Brazil, with mean annual rainfall of more than 3,200 mm, with the lowest mean rainfall in July (111 mm) and the highest in February (410 mm; Martins et al., 2008). Two locations were surveyed in the municipality of Santos, which abuts Bertioga to the west. Forests are continuous and the climate is quite similar.

Bird survey

We cleared an existing trail ranging from 0 to 400 m in Bertioga (bordered by the Rio Guaratuba). This steep trail is narrow (< 1 m wide) and is still used to access the Boraceia Biological Station in the municipality of Salesópolis at the 800 m high plateaus of Serra do Mar (Cavarzere et al., 2010). Lower elevations had signs of human disturbance, such as selective logging (especially the palm tree Euterpe edulis) and hunting trails, while stations above 300 m seemed undisturbed with scarce understory. VC used point counts at this Bertioga location for two breeding seasons (August 31-September 1, October 24-November 8, November 25-28, 2008; November 14-29, 2009), beginning 15 min before sunrise. This elevational gradient included five 100 m elevational bands; to each band was assigned three point counts 200 m apart, which were visited for six nonconsecutive days. The same individuals were not recorded again if it was certain that they had been previously detected. As most species were registered by vocalization, successive recordings were made of simultaneously singing/calling individuals, and/or of vocalizations widely separated in space.

TVVC surveyed birds with 10-species lists during one reproductive season in Bertioga and Santos, including the vicinity of the Rio Itatinga, on the 16^th, 18^th, 20^th and 22^nd of December 2010. We followed Herzog et al.’s (2002) recommendations for using n-species lists, adopting their modifications of the method as follows: 10-species lists were chosen as an intermediate solution among 5-, 10- and 20-species lists regarding estimation of species richness, sample size and bird community pool. The observer took descriptions or sound recordings of any bird not immediately identified, but that were seen or heard sufficiently well for identification. These individuals were subsequently identified using standard reference work. More than one individual was considered if a species was sexually dimorphic (e.g., antbirds).

Records were restricted to a 50 m limited radius of detection for both methods, thus data were recorded from a standardized survey area (path length of lists ca. 12 km). Bird records consisted of individuals heard and/or seen with the help of 8 × 40 binoculars. Observers had been very familiar with Atlantic Forest birds for more than five years and we believe that differences in identification did not introduce major detection bias to our analyses. There was no bird counting on rainy or windy days. Over a total of 63 non-continuous days of fieldwork, 90 point counts were conducted (15 h) and 41 10-species lists were accumulated (~ 16 h). Recordings of individuals (with a Sennheiser ME-66 directional microphone and a Zoom H4N digital recorder) were deposited at the Seção de Aves of the Museu de Zoologia da Universidade de São Paulo (MZUSP).

Analyses

When analyzing point counts each point was considered a sample, whereas the number of 10-species lists was considered a sample for the list method. We used the Sørensen Similarity Index to evaluate similarities of species richness between point counts and lists (Chao et al., 2005). Herzog et al. (2002) suggested using the Chao1 estimator to compare the estimated species richness to the observed species richness. This should be done until the observed species richness is > 90% of the respective Chao1 estimate. This nonparametric estimator, as well as 50-times randomized accumulation curves, was produced with EstimateS 8.2 (Colwell, 2009). A plateau in species accumulation was defined as the point at which the rate of species accumulation over a 10-sample interval fell below 0.10 (O’dea et al., 2004). Atlantic Forest endemic species follow Parker et al. (1996), except for Florisuga fusca, Thalurania glaucopis, Baryphthengus ruficapillus, Trogon surrucura and Automolus leucophthalmus, which may not be endemics according to Cavarzere et al. (2011).

RESULTS AND DISCUSSION

Point counting recorded 992 individuals of 79 species (71% of all observed species) including 28 exclusive records, whereas lists recorded 306 individuals of 83 (74%) species with 32 species in addition to those recorded during point counting; the accumulated number of species was 111 (Appendix). The similarity index between methods was 0,63 with 51 shared species, indicating that as many as 37% of all species were absent from either method. These differences could be due to spatial and temporal mismatch between the times when each survey was carried out, the high turnover expected among samples of diverse communities, or the similar nature of detection of methods (visual and aural). Figure 2 shows the change in estimation of total species richness calculated by the Chao1 estimator as sample size increases. For point counts the predicted value was 88 species, in which case 88 point counts (14.7 h) were enough to detect > 90% of all species. On the other hand, lists recorded only 83% of the estimated number of species (99), suggesting that a greater number of lists (> 41) were still needed to record most species in those areas. Compared to point counts, 10-species lists recorded a lesser percentage of the predicted species richness, but recorded more species than point counts with only 22% of the total number of samples. Species-rich environments represent difficult case studies because applying rapid assessment surveys require more samples or sampling effort (time). However, we are confident that our results accurately describe empirical patterns, as estimations remain effective, even with as many as 23 list samples in the survey effort (MacLeod et al., 2011).

The estimate of total species richness did not stabilize within the available sampling effort because we continued to encounter new species throughout the sampling period. As such, it is not possible to state with confidence the magnitude of species richness for the area based on empirical data, even after extensive sampling (O’dea et al., 2004). Although the rate of species accumulation did not drop below 0.10, 80 point counts obtained a rather similar rate (~ 0.12) as 40 lists with roughly the same number of hours spent on both methods (Fig. 3). The Chao1 list estimate approached the rate of 0.10 more rapidly (less samples) than the estimate for the point count data, indicating a faster accumulation of species richness. This evidences that the use of lists may maximize species detection under a fixed time period. O’dea et al. (2004) have also demonstrated that list accumulation curves tend to approach the recorded species richness of the area more closely while point counts probably need many samples, as the species recorded while moving between points are not considered.

Because sampling efforts differ between methods, we produced species curves to compare the accumulation of species richness as a function of the number of surveying days. Within four days point counts detected 45 species, whereas 83 species were accumulated from lists (Fig. 4). Point counts present the advantage of detecting species that would have gone unnoticed if an observer walked past through a particular site without staying for at least 10 minutes. However, the fact that species detected during movement between points are not considered by this method seems to strongly hinder the rapid assessment of the avifauna. Also, the list method may better record rare and vagrant species because it allows continuous recording and active searching (Fjeldså, 1999). It also allows a greater proportion of the available time to be spent in the field collecting bird survey data, and more species are recorded using the list technique due to its flexibility in continuously recording data and actively searching out new species (O’dea et al. 2004). Given that the duration of field expeditions is usually defined as the amount of surveying days, our results again favor the use of species lists if the goal is to maximize the number of species detected in a fixed amount of time.

The advantage of point counts relative to the list method remains for estimating relative abundances. Point counting with spatial and temporal standardization means that their biases are more readily quantifiable and controllable (O’dea et al., 2004). Lists, when standardized, are neither easier nor superior to point counts; what they provide is an efficient means of determining species richness in a species-rich environment, as well as cataloguing the species systematically missed by point counts (Poulsen et al., 1997). Since lists usually do not take the number of individuals into account, their abundance indices are rougher than those obtained by point counts. As lists do not standardize area or time surveyed, they can hinder comparisons of abundances among surveys. Thus, the complementarity of methods should maximize bird surveys if their strengths and weaknesses are applied together, i.e., rapid detections with lists and abundance estimate with point counts (O’dea et al., 2004; MacLeod et al., 2011). Lists have been shown to produce consistent abundance indices for the majority of species in a tropical forest bird community in Bolivia within a short time period, especially by comparing abundance estimates for species of conservation importance. Although its efficiency in generating consistent abundance indices suggests that it has excellent potential as a rapid assessment and monitoring tool, it will be particularly important to study how relative abundance indices generated by the list method compare to density estimates generated by more traditional methodologies (MacLeod et al., 2011).

Major advantages of lists are their simplicity, their potential to collect large amounts of data, and that they are highly time-efficient compared with point and transect data. This is mainly in view of the logistical realities in the tropical forests and for securing a broad coverage of the study area. Furthermore, lists do not appear to be more biased than other observational approaches, judging from the very close correlation with data obtained by a more time-consuming method (Fjeldså, 1999). Whereas the purpose of a survey is simply to assess the overall species richness of an area, lists appear to be the most effective tool. It seems to be a promising technique, as it is a means of accumulating species richness at a faster pace. Complete surveys continue to require the use of a combination of methods in which the nature of detection differs, such as visual-auditory versus mist nets (Silveira et al., 2010; Somenzari et al., 2011). Due to the fact that few publications use lists for surveying birds in Neotropical regions, we wish to encourage and extend its application. Its wide use will help elucidate its advantages and flaws while generating reliable information on species distribution. Formal comparisons of abundances generated by points and lists in Neotropical forests still remain to be studied.

ACKNOWLEDGEMENTS

Gabriel P. Moraes helped during point counting in Bertioga. We extend many thanks to Andrés Calonge-Méndez for inviting TVVC to survey this Bertioga-Santos region. IdeaWild supported us with field equipment. VC benefitted from a Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) scholarship and LFS receives a grant from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, 313818/2009-6). Anonymous reviewers made important contributions to this manuscript.

Aceito em: 25.07.2012

Publicado em: 28.09.2012

APPENDIX

Species recorded with 10-min point counts (PC) and 10-species lists (SL) in the municipalities of Bertioga and Santos, state of São Paulo, southeastern Brazil. The Atlantic Forest column indicates endemic species.

• Bibby, C.J.; Burgess, N.D.; Hill, D.A. & Mustoe, s.h. 2000 Bird census techniques, Academic Press, London, 302 p.
• Cavarzere, V.; Moraes, G.P. & Silveira, L.F. 2010. Boracéia Biological Station: an ornithological review. Papéis Avulsos de Zoologia, 50:189-201.
• Cavarzere, V.; Moraes, G.P.; Dalbeto, A.C.; Maciel, F.G. & Donatelli, R.J. 2011. Birds from cerradão woodland, an overlooked forest of the Cerrado Region, Brazil. Papéis Avulsos de Zoologia, 51:259-273.
• Chao, A.; Chazdon, R.L.; Colwell, R.K. & Shen, T.J. 2005. A new statistical approach for assessing similarity of species composition with incidence and abundance data. Ecology Letters, 8:148-159.
• Colwell, R.K. 2009. EstimateS: statistical estimation of species richness and shared species from samples. Version 8.2. Users Guide and application. Available online. <http://viceroy.eeb.uconn.edu/EstimateS>
• Develey, P.F. 2004. Efeitos da fragmentação e do estado de conservação da floresta na diversidade de aves de Mata Atlântica. (Ph.D. Dissertation). University of São Paulo, São Paulo, Brazil.
• Esquivel, A. & Peris, S. 2008. Influence of time of day, duration and number of counts in point count sampling of birds in an Atlantic Forest of Paraguay. Ornitologia Neotropical, 19:229-242.
• Fjeldså, J. 1999. The impact of human forest disturbance on the endemic avifauna of the Udzungwa Mountains, Tanzania. Bird Conservation International, 9:47-62.
• Herzog, S.K. 2008. First ornithological survey of Cordillera Mosetenes, with a latitudinal comparison among lower Yungas bird communities in Bolivia. Revista Boliviana de Ecología y Conservacíon Ambiental, 23:59-71.
• Herzog, S.K. & Kessler, M. 2006. Local vs. regional control on species richness: a new approach to test for competitive exclusion at the community level. Global Ecology and Biogeography, 15:163-172.
• Herzog, S.K.; Kessler, M. & Cahill, T.M. 2002. Estimating species richness of tropical bird communities from rapid assessment data. Auk, 119:749-769.
• Karr, J.R. 1971. Structure of avian communities in selected Panama and Illinois habitats. Ecological Monographs, 41:207-231.
• MacKinnon, S. & Phillipps, K. 1993. A Field Guide to the Birds of Borneo, Sumatra, Java and Bali, Oxford University Press, Oxford, 491 p.
• MacLeod, R.; Herzog, S.K.; MacCormick, A.; Ewing, S.R.; Bryce, R. & Evans, K.L. 2011. Rapid monitoring of species abundance for biodiversity conservation: consistency and reliability of the MacKinnon lists technique. Biological Conservation, 144:1374-1381.
• Maia, V.C.; Magenta, M.A.G. & Martins, S.E. 2008. Ocorrência e caracterização de galhas de insetos em áreas de restinga de Bertioga, São Paulo, Brasil. Biota Neotropica, 8:167-197.
• Martins, S.E.; Rossi, L.; Sampaio, P.S.P. & Magenta, M.A.G. 2008. Caracterização florística de comunidades vegetais de restinga em Bertioga, SP, Brasil. Acta Botanica Brasilica, 22:249-274.
• Nascimento, C.M. & Pereira, M.A.M.G. 1988. Atlas climatológico do Estado de São Paulo: 1977-1986, Fundação Cargill, Campinas. 93 p.
• Odea, N.; Watson, J.E.M. & Whittaker, R.J. 2004. Rapid assessment in conservation research: a critique of avifaunal assessment techniques illustrated by Ecuadorian and Madagascan case study data. Diversity and Distributions, 10:55-63.
• Parker, T.A. & Bailey, B. 1991. A biological assessment of the Alto Madidi region and adjacent areas of northwest Bolivia, May 18-June 15, 1990. RAP Working Papers no. 1, Conservation International, London.
• Parker, T.A.; Stotz, D.F. & Fitzpatrick, J.W. 1996. Ecological and distributional databases. In: Stotz, D.F.; Fitzpatrick, J.W.; Parker, T.A. & Moskovits, D.K. (Eds.). Neotropical birds: ecology and conservation. The University of Chicago Press, Chicago. p. 118-436.
• Poulsen, B.O. & Krabbe, N. 1998. Avifaunal diversity of five high-altitude cloud forests on the Andean western slope of Ecuador: testing a rapid assessment method. Journal of Biogeography, 25:83-93.
• Poulsen, B.O.; Krabbe, N.; Frølander, A.; Hinojosa, B.M. & Quiroga, O.C. 1997. A rapid assessment of Bolivian and Ecuadorian montane avifaunas using 20-species lists: efficiency, biases and data gathered. Bird Conservation International, 7:53-67.
• Ribon, R. 2010. Amostragem de aves pelo método das listas de MacKinnon. In: von Matter, S.; Straube, F.; Accordi, I.; Piacentini, V. & Cândico Jr, J.F. (Eds.), Ornitologia e Conservação: ciência aplicada, técnicas de pesquisa e levantamento. Technical Books, Rio de Janeiro, p. 1-16.
• Silveira, L.F.; Beisiegel, B.M.; Curcio, F.F.; Valdujo, P.H.; Dixo, M.; Verdade, V.K.; Mattox, G.M.T. & Cunningham, P.T.M. 2010. Para que servem os inventários de fauna? Estudos avançados, 24:173-207.
• Somenzari, M.; Silveira, L.F.; Piacentini, V.Q.; Rego, M.A.; Schunck, F. & Cavarzere, V. 2011. Birds of an Amazonia-Cerrado ecotone in southern Pará, Brazil, and the efficiency of associating multiple methods in avifaunal inventories. Revista Brasileira de Ornitologia, 19:260-275.
• Trainor, C.R. 2002a. Status and habitat associations of birds on Lembata Island, Wallacea, Indonesia, with reference to a simple technique for avifaunal survey on small islands. Bird Conservation International, 12:365-381.
• Trainor, C.R. 2002b. The birds of Adonara, Lesser Sundas, Indonesia. Forktail, 28:93-100.

Publication Dates

• Publication in this collection
  21 Sept 2012
• Date of issue
  2012

History

• Received
  25 July 2012
• Accepted
  28 Sept 2012