The effectiveness of acoustic indices for forest monitoring in Atlantic rainforest fragments
Introduction
Biodiversity monitoring is necessary to detect changes in ecosystem patterns (Noss, 1990, Hooper et al., 2012, Gardner, 2012, Siddig et al., 2016), which has become more pressing issue due to global climate change (Thuiller, 2007, Heller and Zavaleta, 2009, Urban, 2015, Ceballos et al., 2015). This monitoring supports the decision-making processes in public policies that aim for management actions to achieve goals established in international agreements on environmental conservation directives (UNEP, 2017). In this context, bird monitoring is considered as a suitable option (Koskimies, 1989, Bibby, 1999, Carignan and Villard, 2002, Krause and Farina, 2016), because birds are ecologically well studied (Birdlife, 2016) and respond to environmental changes over many spatial scales (Bregman et al., 2014). Changes in bird species composition and population dynamics might represent changes in environmental quality, as they respond to multiple environmental variables, such as natural resources supply, inter- and intra-specific interactions, habitat quality, and landscape configuration (Larsen et al., 2010, Carrara et al., 2015).
To establish an appropriate protocol for bird monitoring, it is necessary to consider characteristics such as team experience, equipment, permits, field effort, transport, and species detection by each method (Bispo et al., 2016). In tropical regions, the most used ornithological survey methods are audiovisual ones, such as point-counts, line transect, and complete list of species, which become more efficient when combined with sound recording (Larsen et al., 2016). Cost constraints, however, may be an obstacle to the implementation of such methods, which require logistic investments and specialists in fieldwork and statistics. Alternatively, there is an emerging method of bird monitoring, namely acoustic monitoring methods, using autonomous recording units (ARUs) (Brandes, 2008). The main advantage of ARU sampling is to simultaneously collect data at several points (Digby et al., 2013, Gasc et al., 2013). This automated technique also allows data collection in remote areas (Hutto and Stutzman, 2009), with minimum disturbance to wildlife (Acevedo and Villanueva-Rivera, 2006). Additionally, ARUs can be executed with a minimum of training, staff, and operative costs (Sueur et al., 2012).
By using ARUs the researcher can follow two approaches. The first is trying to identify the recorded species, using automated recognition software (Mammides et al., 2017). This approach, however, is relatively time-consuming in data processing (Hutto and Stutzman, 2009), and is not an appropriate approach for long-term community monitoring, because an automated recognizer is more restricted to the recognition of one or few species of interest (Alquezar and Machado, 2015, Mammides et al., 2017). The second approach, in turn, allows the researcher to determine acoustic indices from the ARU data.
Acoustic indices can be used for the determination of bird species diversity, based on the assumption that higher species diversity translates into higher acoustic complexity (Sueur et al., 2014). Usually, the studies using ARUs tried to compare species detection using this automated method with the data obtained when directed by researchers in the field (Shonfield and Bayne, 2017) and, so far, they have yielded controversial results. For instance, Boelman et al. (2007) recorded correlation between the bioacoustic index (BI) with direct ornithological survey methods in Hawaiian sub montane ecosystems. Pieretti et al. (2011) also found correlation between the acoustic complexity index (ACI) with avifauna counts in a beech mountain forest. Mammides et al. (2017), however, found low and sometimes controversial correlations between acoustic indices with the number of bird species in small fragments and large protected reserves of tropical rainforest.
As in these studies the authors aimed to compare the species’ detection methods, they concomitantly conducted direct surveys with the ARU sampling. Thus, disturbance caused by the researcher can result in greater occurrence of alertness calls by some bird species, which make them more conspicuous, and in a bias toward point-counts survey, as verified by Digby et al. (2013). Furthermore, without the visual cues used during the point-counts survey (Leach et al., 2016), the researcher can miss species that vocalize just occasionally (Alquezar and Machado, 2015). It is possible, therefore, that in more biodiverse habitats, such as the subtropical forest studied by Mammides et al. (2017), the researchers’ presence in the field caused higher disturbances in the bird vocal activity than in less complex habitats, such as the ones studied by Boelmam et al. (2007) and Pieretti et al. (2011).
The potential of the use of ARUs for long-term biodiversity monitoring in remote locations, however, is remarkable, as reviewed by Shonfield and Bayne, (2017), who claimed that this new methodology is more than simply a potential substitute for a human observer in the field. However, as highlighted by Mammides et al. (2017), the acoustic indices, as well as the autonomous recording units, should be tested in more environments to reveal their potential and limitations before they can be widely applied. Thus, in this study we aimed to compare the effectiveness of sampling birds by the traditional point-count survey with six acoustic indices calculated from ARU data in a tropical rainforest fragment. We predicted stronger correlations between the number of bird species determined by the point-count survey with the acoustic complexity index (ACI) and the bioacoustic index (BI), because both indices estimate the diversity of an acoustic community based on sound intensity variation, as verified in previous studies (Farina et al., 2011, Pieretti et al., 2011, Gasc et al., 2013). We also aimed to quantify the bias caused by human presence during data collection. Additionally, as a secondary result of the study, we compared the cost benefit between the point-count survey and the ARU method.
Section snippets
Study site
The fieldwork was carried out in Serra do Teimoso Natural Reserve, an Atlantic rainforest fragment (UTM 24L 442733 E/8324297 S) in Southern Bahia, Brazil. The RST has a total of 200 hectares of hillside forest ranging from 250 to 850 meters above sea level. The surrounding landscape includes pasture areas, cacao under a shaded system, small crops, and forests in natural regeneration. The average annual rainfall is 1323 mm, relative air humidity in this area is around 84%, and the average annual
Point-counts survey
At the 12-point stations of the RST, 97 bird species were identified from 823 registers (Appendix A in Supplementary). The most frequent species recorded was Thripophaga macroura with relative abundance Pi = 0.066 (Table 1); the lowest abundance was Pi = 0.001 for a single-record species, such as Campylorhamphus trochilirostris and Automolus leucophthalmus. Some of the most frequently recorded bird species in the RST are classified as Vulnerable or Endangered according to the Brazilian Red List
Discussion
As far as we know, this is the first study comparing the composition of bird community by using point-counts survey and acoustic indices in the Brazilian Atlantic forest. Using the point-counts survey, we described 97 species in the bird community and determined the occurrence of 10 threatened bird species in the relatively small studied Atlantic rainforest fragment. We determined correlations between the number of bird species with five of the six acoustic indices calculated from using
Conclusions
By quantifying the bias caused by human presence during data collection we verified that among the six acoustic indices tested, just the acoustic evenness index (AEI) showed moderate correlation with the number of bird species and is effective in forest monitoring in Atlantic rainforest. Thus, important changes in this acoustic index during monitoring can be used to trigger local actions and prevent severe environmental impacts in threatened Atlantic forest fragments. Further studies must be
Acknowledgements
We thank Ricardo Machado and Maíra Benchimol de Souza, and the anonymous reviewer of the Ecological Indicators for their pertinent and valuable comments, which helped to improve the original manuscript. This work was supported by the Brazilian National Council for Scientific and Technological Development (CNPq Process#130666/2015-7) and by the Brazilian Federal Agency for Support and Evaluation of Graduate Education. Sérgio Nogueira-Filho was funded by the Brazilian National Council for
References (46)
- et al.
Global patterns and predictors of bird species responses to forest fragmentation: implications for ecosystem function and conservation
Biol. Conserv.
(2014) - et al.
Impact of landsccape composition and configuration on forest specialist and generalist bird species in the fragmented Lacondona rainforest, Mexico
Biol. Conserv.
(2015) - et al.
Connecting soundscape to landscape: which acoustic index best describes landscape configuration?
Ecol. Ind.
(2015) - et al.
Biodiversity management in the face of climate change: a review of 22 years of recommendations
Biol. Conserv.
(2009) - et al.
The remote environmental assessment laboratory's acoustic library: an archive for studying soundscape ecology
Ecol. Inf.
(2012) - et al.
Using ecoacoustic methods to survey the impacts of climate change on biodiversity
Biol. Conserv.
(2016) - et al.
Do acoustic indices correlate with bird diversity? Insights from two biodiverse regions in Yunnan Province, south China
Ecol. Ind.
(2017) - et al.
A new methodology to infer the singing activity of an avian community: the acoustic complexity index (ACI)
Ecol. Ind.
(2011) - et al.
How do ecologists select and use indicator species to monitor ecological change? insights from 14 years of publication in Ecological Indicators
Ecol. Ind.
(2016) - et al.
Comparisons between autonomous acoustic recordings and avian point counts in open woodland savanna
Wilson J. Ornithol.
(2015)
Using automated digital recording systems as effective tools for the monitoring of birds and amphibians
Wildlife Soc. Bull.
BioEstat: Aplicações Estatísticas nas áreas das Ciências Biológicas e Médicas, Version 5.0
Making the most of birds as environmental indicators
Ostrich
Protocolo para monitoramento de comunidades de aves em Unidades de Conservação Federais
Biodiversidade Brasileira
Multi-trophic invasion resistance in Hawaii: bioacoustics, field surveys, and airborne remote sensing
Ecol. Appl.
Automated sound recording and analysis techniques for bird survey and conservation
Bird Conserv. Int.
Selecting indicator species to monitor ecological integrity: a review
Environ. Monit. Assess.
Accelerated modern human-induced species losses: entering the sixth mass extinction
Sci. Adv.
A practical comparison of manual and autonomous methods for acoustic monitoring
Methods Ecol. Evol.
Avian soundscapes and cognitive landscapes: theory, application and ecological perspectives
Landscape Ecol.
Monitoring Forest Biodiversity: Improving Conservation Through Ecologically Responsible Management
Cited by (49)
Tree diversity and liana infestation predict acoustic diversity in logged tropical forests
2024, Biological ConservationPoor performance of acoustic indices as proxies for bird diversity in a fragmented Amazonian landscape
2023, Ecological InformaticsQuantifying the Soundscape: How filters change acoustic indices
2023, Ecological IndicatorsEffectiveness of acoustic indices as indicators of vertebrate biodiversity
2023, Ecological Indicators