Research PaperDo acoustic indices reflect the characteristics of bird communities in the savannas of Central Brazil?
Introduction
Biodiversity loss caused by humans is a major and challenging problem globally (Pimm, Raven, Peterson, Sekercioglu, & Ehrlich, 2006), and threats to species and ecosystems are set to continue in the future (Pereira et al., 2010). Deforestation and habitat fragmentation caused by expansion of croplands and pastures impose serious threats to species and native ecosystems, especially in South America, where the most globally significant changes in cropland expansion occurred between 1960 and 1990 (Ramankkutty, Foley, & Olejniczak, 2002). Most of cropland expansion has occurred in the Brazilian Cerrado (woodland savanna), a region, which has lost half of its original area (more than 1 million km2) in recent decades (Brasil, 2009, Françoso et al., 2015, Klink and Machado, 2005). How biodiversity responds to habitat loss and fragmentation is one of the key topics in ecology and conservation biology (Sala et al., 2000).
Considering the rapid deforestation rate in the Brazilian Cerrado, (Brasil, 2009, Klink and Machado, 2005) it is important to develop and apply methods that can be effective for rapid biodiversity assessment. The field of Ecoacoustics, an emerging discipline that investigates spatial and temporal variation of the sounds associated with landscape structures (Sueur & Farina, 2015), may provide the answers to the ecological and conservation issues. Bioacoustics has been traditionally used in behavioural studies, and only recently has been applied in conservation biology (Ritts, Gage, Picard, Dundas, & Dundas, 2016; Sueur, Pavoine, Hamerlynck, & Duvail, 2008; Towsey, Parsons, & Sueur, 2014). The field of ecoacoustics, an emerging discipline that investigates spatial and temporal variation of the sounds associated with population, community or landscape structures (Sueur & Farina, 2015), may provide the answers to ecological and conservation issues.
Ecoacoustics may provide an excellent tool in conservation biology because it is non-invasive, it can synthesize a wide range of species and many locations can be surveyed simultaneously using automated recorders. In addition, the recordings can be sent to experts for thorough analysis, and data can be collected without human interference. Unlike the in situ observations made by researchers, recordings can provide voucher material for biodiversity inventories, creating a permanent record of a field study. The availability of analytical tools for data processing has also increased in recent years, and semi-automatic or automatic signal extraction and even identification is now possible, although there are still obstacles in automatic identification of soundscape entities. Acoustic indices for ecological studies and biodiversity monitoring are one of the automatic approaches for data analysis (Kasten, Gage, Fox, & Joo, 2012; Pieretti, Farina, & Morri, 2011; Pijanowski et al., 2011; Sueur, Farina, Gasc, Pieretti, & Pavoine, 2014), and acoustic indices can potentially be used as a surrogate for observational data when quick assessments are necessary. Another advantage of the use of acoustic indexes is to overcome the problem of analysing hours and hours of records obtained by passive recorder units.
In order to be useful, however, an acoustic index should have some congruence with the patterns of biodiversity or species diversity of the taxonomic group under investigation. If the acoustic index can be associated with species diversity or species activity, it can provide a valid tool to rapidly assess biodiversity using a soundscape approach.
The soundscape, according to Dumyahn and Pijanowski (2011), represents the entire acoustic environment of a particular landscape. In this paper, we used acoustic data to (a) characterize species richness and composition of birds associated with the Cerrado in Brazil and (b) to compare the behaviour of two indices, an acoustic diversity index (ADI) (Villanueva-Rivera, Pijanowski, Doucette, & Pekin, 2011) and the normalized difference soundscape index (NDSI) (Kasten et al., 2012) in two different Cerrado habitat types: the cerrado stricto sensu (in lower case to indicate the phytophysiognomy) and the gallery forest (evergreen forest-like vegetation that exists along rivers and streams).
Gallery forest is a narrow strip of forest along the rivers and streams with three vegetation layers and a closed canopy, reaching 20–25 m in height. On the other hand, the cerrado stricto sensu is a woodland savanna-like vegetation with sparse trees reaching 6–10 m high, with unconnected canopy coverage from 50 to 70%, allowing sunlight to directly reach the soil (Eiten, 1972, Ribeiro and Walter, 1998).
We tested predictions from three hypotheses concerning the potential associations between biodiversity data and the acoustic indices. First, we predicted that acoustic indices would reflect the characteristics of bird communities associated with the cerrado stricto sensu (s.s.) and gallery forest, the two habitats with highest species richness in the Cerrado biome. The basis for this prediction was that the values of the acoustic diversity index [ADI; is a measurement of the degree of acoustic complexity found at a site (see Villanueva-Rivera & Pijanowski (2014) and the Methods for details)] should be higher in gallery forest, which has greater structural complexity compared with the cerrado. Second, we predicted that the biophony/technophony ratio (i.e. the ratio of sounds produced by wildlife and by human activities respectively) should be lower in the cerrado than in gallery forest due to the higher attenuation of anthropophonic sound by the complex vegetation in gallery forest. We used the normalized difference soundscape index (NDSI) to indicate the amount of biophony relative to technophony in the two habitat types (see Gage & Axel (2014) and the Methods for more details). Finally, we predicted that the spatial arrangement of technophony noise sources such as roads, houses and buildings or airports, should influence the values of the NDSI, such that lower values would be observed near such noise sources. Thus, the NDSI would be a valid proxy for the impact of technophony on biodiversity regardless of the type of habitat (cerrado or gallery forest), and areas located close to technophony sources will have lower values of NDSI.
Section snippets
Study area
We conducted this study in a protected area named the Environmental Protection Area of Gama e Cabeça do Veado (hereafter referred to as EPA), which is equivalent to the category V Protected Area Category defined by the International Union for Conservation of Nature (Phillips, 2002, Thomas and Middleton, 2003). The EPA covers 25,000 ha and was created in 1986. It is located south of Brasilia city (15° 51′16”–15° 58′ 17”, 47° 59′ 39”–47° 40′ 09” Long W) (Fig. 1), and comprises four reserves: Água
Results
Acoustic identification by listening to recordings and viewing spectrograms revealed the presence of 107 bird species in the 30 sampling locations, where 74 and 47 bird species were identified in the gallery forest the cerrado, respectively (Supplementary Material − Table S3). From this total, 18 birds, mostly with a single call, were not identified at species level. We used two song parameters (duration and frequency) and the spectrogram shape to ensure that they were different from the
Discussion
As far as we know, this is the first study comparing the composition of bird communities by using acoustic indices in South America and in the Brazilian Cerrado. An earlier study in Brazil aimed to define sampling schemes for general passive acoustic monitoring (Pieretti et al., 2015) did not include biodiversity data for comparisons. We believe that the knowledge of biodiversity characteristics allows a better understanding of the behaviour of acoustic indices, which is essential for their
Conclusions
We demonstrate relationships between the characteristics of bird communities associated with different habitats in the Brazilian Cerrado analysed by using acoustic indices. As we initially expected, the acoustic diversity index (ADI) and the normalized difference soundscape index (NDSI) matched differences in species richness and composition that exist between gallery forest and cerrado, as well differences in the spatial arrangement of the sampling locations in relation to landscape
Acknowledgments
The Brazilian National Council for Scientific and Technological Development − CNPq granted RBM and LMSA with a scholarship grant (Science without Border program) and for supporting part of our fieldwork (through PELD and PPBio-DF projects). Digital records were acquired with resources donated by FINEP (Cooperation Agreement #04.12.0110.00) and CAPES (Pro-Equipamentos project associated with the Post-Graduated Master and PhD Courses in Ecology at University of Brasilia). We thank the staff of
References (59)
- et al.
Sonic environment and vegetation structure: a methodological approach for a soundscape analysis of a Mediterranean maqui
Ecological Informatics
(2014) - et al.
Habitat loss and the effectiveness of protected areas in the Cerrado biodiversity hotspot
Natureza & Conservação
(2015) - et al.
Visualization of temporal change in soundscape power of a Michigan lake habitat over a 4-year period
Ecological Informatics
(2014) - et al.
Linking variability in species composition and MODIS NDVI based on beta diversity measurements
Acta Oecologica
(2009) - et al.
Analysis and interpretation of variability in soundscapes along an urban–rural gradient
Landscape and Urban Planning
(2011) - et al.
The remote environmental assessment laboratory's acoustic library: an archive for studying soundscape ecology
Ecological Informatics
(2012) - et al.
Tolerance to noise in 91 bird species from 27 urban gardens of Iberian Peninsula
Landscape and Urban Planning
(2012) - et al.
A new methodology to infer the singing activity of an avian community: the acoustic complexity index (ACI)
Ecological Indicators
(2011) - et al.
Collaborative research praxis to establish baseline ecoacoustics conditions in Gitga’at Territory
Global Ecology and Conservation
(2016) - et al.
Ecology and acoustics at a large scale
Ecological Informatics
(2014)
The use of acoustic indices to determine avian species richness in audio-recordings of the environment
Ecological Informatics
A eficiência do método de amostragem por pontos de escuta na avaliação da riqueza de aves
Revista Brasileira de Ornitologia
Audacity 2.0.6
Sound transmission through vegetation in relation to leag area density, leaf width, and breadth of canopy
The Journal of the Acoustical Society of America
MuMIn: Multi-Model Inference
Relatório técnico de monitoramento do desmatamento no bioma Cerrado, 2002 a 2008: dados revisados
A representatividade das áreas protegidas do Distrito Federal na conservação da avifauna do Cerrado
Ararajuba
Evaluation of methodological protocols using point counts and mist nets: a case study in Southeastern Brazil
Papéis Avulsos De Zoologia
Soundscape conservation
Landscape Ecology
The Cerrado vegetation of Brazil
The Botanical Review
Anthropogenic noise affects song structure in red-winged blackbirds (Agelaius phoeniceus)
Journal of Experimental Biology
Raster v2.3: Geographic data analysis and modeling
Reserva Ecológica do IBGE: ambiente e plantas vasculares
Conservation of the Brazilian Cerrado
Conservation Biology
TuneR: Analysis of music and speech Version 1.2.1
Propagation of environmental noise
Science
Environmental factors affecting bird species diversity
The American Naturalist
A fragmentação do Cerrado e efeitos sobre a avifauna na regiao de Brasília-DF
Cluster analysis extended. Version 2.01
Cited by (66)
The effect of latitude on the efficacy of acoustic indices to predict biodiversity: A meta-analysis
2024, Ecological IndicatorsCicadaNet: Deep learning based automatic cicada chorus filtering for improved long-term bird monitoring
2024, Ecological IndicatorsCOVID-19 shutdown revealed higher acoustic diversity and vocal activity of flagship birds in old-growth than in production forests
2023, Science of the Total EnvironmentPoor performance of acoustic indices as proxies for bird diversity in a fragmented Amazonian landscape
2023, Ecological Informatics