ArticlePDF Available

Methods to measure biological sounds and assess their drivers in a tropical forest

Authors:

Abstract and Figures

The study of soundscapes and biological sounds is becoming the focus of increasing scientific attention. Studying biological sounds involves the deployment of acoustic sensors to record sounds and the identification of animal species and other sources of sound in audio recordings. In addition, data extracted from audio recordings may be pooled together with ecological and human activity data to investigate the drivers of biological sounds. We provide a detailed method description of our study on biological sounds in a tropical forest and their drivers along a gradient of disturbance in Southeast Cameroon. We designed and implemented a research protocol to: - make large scale audio recordings and identify animal species detected; - collect ground-truth data on mammal and bird species; - collect climate, habitat, and human activity data and describe remoteness and accessibility.
Content may be subject to copyright.
MethodsX 9 (2022) 101619
Contents lists available at ScienceDirect
MethodsX
j o u r n a l h o m e p a g e: w w w . e l s e v i e r . c o m / l o c a t e / m e x
Method Article
Methods to measure biological sounds and assess
their drivers in a tropical forest
Johan Diepstraten
a , , Jacques Keumo Kuenbou
b
, Jacob Willie
c
a
Animal Behaviour and Cognition, Department of Biology, Faculty of Science, Utrecht University, The Netherlands
b
Association de la protection des grands singes, Cameroon
c
Centre for Research and Conservation, Royal Zoological Society of Antwerp, Belgium
a b s t r a c t
The study of soundscapes and biological sounds is becoming the focus of increasing scientific attention. Studying
biological sounds involves the deployment of acoustic sensors to record sounds and the identification of animal
species and other sources of sound in audio recordings. In addition, data extracted from audio recordings may
be pooled together with ecological and human activity data to investigate the drivers of biological sounds. We
provide a detailed method description of our study on biological sounds in a tropical forest and their drivers
along a gradient of disturbance in Southeast Cameroon. We designed and implemented a research protocol to:
make large scale audio recordings and identify animal species detected;
collect ground-truth data on mammal and bird species;
collect climate, habitat, and human activity data and describe remoteness and accessibility.
©2022 The Author(s). Published by Elsevier B.V.
This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )
a r t i c l e i n f o
Method name: Passive acoustic monitoring and analysis
Keywo rds: Ecoacoustics, Audio recordings, Human listeners, Audiomoth, Transects, Mammal surveys, Bird surveys, Human
disturbance
Article history: Received 23 September 2021; Accepted 9 January 2022; Available online 14 January 2022
DOI of original article: 10.1016/j.gecco.2021.e01819
Corresponding author.
E-mail address: johandiepstraten@gmail.com (J. Diepstraten).
https://doi.org/10.1016/j.mex.2022.101619
2215-0161/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license
( http://creativecommons.org/licenses/by/4.0/ )
2 J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619
Specifications table
Subject Area: Environmental Science
More specific subject area: Ecoacoustics
Method name: Passive acoustic monitoring and analysis
Name and reference of original
method:
N.A.
Resource availability: AudioMoth: https://www.openacousticdevices.info/
iNext: https://chao.shinyapps.io/iNEXTOnline/
Methods
Study area
This study was conducted in the northern part of the Dja Faunal Reserve’s buffer zone in
Cameroon. Data were obtained in three study sites (Ngouleminanga, La Palestine, and La Belgique)
that differ in land-use type and conservation management ( Table 1 ; Fig. 1 ). Since the overall level
of disturbance in a site depends on these two factors, these three sites are expected to represent a
gradient of disturbance. Logging was absent throughout the study area.
Ngouleminanga is a forest site and eponymous village located 24 km north of the Dja Faunal
Reserve. The village is easily accessible by motorised vehicles and supports an estimated population
of 130 inhabitants [8] . Ngouleminanga is used as a community forest (CF), indicating that local
communities can derive their livelihoods in a sustainable manner from this forest [4] . Furthermore,
no active conservation management is present in Ngouleminanga. La Palestine is a forest site adjacent
to the villages Malen V, Doumo-Pierre, and Mimpala. Toge th er, these villages have an estimated
population size of 350 inhabitants [8] . Malen V is easily accessible by motorised vehicles, as opposed
to Doumo-Pierre and Mimpala. La Palestine is located 17 km north of the Dja Faunal Reserve. The
entire site is considered to be a CF. However, active conservation management in this area has been
conducted by APGS since 2001. APGS focusses on the protection of wildlife through conservation-
applied research and support to the local community. The organisation promotes educational
programmes within the local community to increase their awareness of the local natural environment.
La Belgique is a research site founded and managed by APGS and located 2 km north of the Dja Faunal
Reserve. The site is only accessible by foot and the nearest villages (Doumo-Pierre and Mimpala)
house approximately 180 inhabitants [8] . The forest is officially unprotected and is classified as a
forest management unit (FMU), meaning that it is property of the state and can be used for timber
production purposes [1] . However, APGS has signed an agreement with local villages prohibiting
human activities, such as hunting, within the research site. Accordingly, the presence of APGS has
proven to counteract the negative effects of hunting and deforestation in the area [18,20] .
Data collection
Field work was conducted between February and May 2020. During this time, acoustic
measurements were performed for the detection and identification of individual vocalising species
and the determination of biological sounds. Eighteen AudioMoth bioacoustics sensors were used
throughout the study area (6 per site). Six transects of 1 km each were opened in each site and
one sensor was deployed in the middle of every transect, at the 500-m mark [13] . To create enough
space between the transects within each site, a cascading design was used. In this design, the distance
Tabl e 1
Overview of how the three study sites represent a disturbance gradient.
Site Population size (#) Land-use type Conservation management Disturbance
Ngouleminanga 130 Community forest Absent High
La Palestine 176 Community forest Present Medium
La Belgique 182 Forest management unit Present Low
J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619 3
Fig. 1. Location of the three study sites and the adjacent villages in the northern periphery of the Dja Faunal Reserve. Six 1-km transects were opened in each site using a cascading
design. The transects were cut with a constant compass bearing of 140 °, 180 °, and 45 °in Ngouleminanga, La Palestine, and La Belgique, respectively.
4 J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619
Fig. 2. Audio sampling design. A) An AudioMoth bioacoustics sensor wrapped in a zip lock bag, placed in a protective, labelled
case with foam in the back to keep the sensor in place and a small hole in the front at the location of the microphone. B) A
sensor deployed in the forest, attached to a tree at a height of 2 m, orientated at 90 °.
between the start and the end of the adjacent transects was also 1 km. This resulted in a
2-km
distance between sensors. The transects were cut with a randomly chosen, constant compass bearing
of 140 °, 180 °, and 45 °in Ngouleminanga, La Palestine, and La Belgique, respectively ( Fig. 1 ).
Different setups were tested to protect the sensors against rain and animals. The setup where all
sensors were kept in zip lock bags within a protective case resulted in the clearest audio quality
while also protecting the case. Furthermore, at the location of the sensors’ microphone, a small hole
was made in the case to ensure optimal audio quality of the recordings ( Fig. 2 A). For consistency, all
sensors were attached to a small tree at a height of 2 m, oriented at 90 °( Fig. 2 B).
Bats, which produce ultrasonic sounds, were not included in this study. Therefore, only sounds
within a human audible range were recorded. For proper recording, the sampling rate of the sensors
should be over twice as high as the highest frequency of interest, or Nyquist frequency [3] . Therefore,
the sampling rate of all sensors was set to record at 48 kHz. Since the tested environment did not
show high levels of noise, recordings were made at 30.6 dB [13] . All sensors were set to record the
first minute of every hour, resulting in 24 min of sound recordings per transect per day.
In total, 20 485 min of sound were obtained from the three sites. 5949, 7712, and 6824 min were
recorded in Ngouleminanga, La Palestine, and La Belgique, respectively ( Table 2 ). All recordings made
during rainy periods were excluded from the listening process and the analyses. This was the case for
1895 audio files. To expedite the listening process, all recordings made during the night were screened
by JD beforehand. Only night recordings that contained vocalisations other than those of insects,
amphibians or western tree hyraxes ( Dendrohyrax dorsalis ), which were all easily recognisable after
some training, were played to the local expert listeners for identification. All remaining recordings
were played to two local villagers who could identify the audible species. These listeners consistently
identified the same sounds as being the same species. The listening process took two months. On
average, 340 audio recordings were listened to each day. For each recording, these local experts were
asked to independently write down the names of all the species they heard in their local language,
Badjué. The English and scientific translation of many of the local names were already known, if
not, the local experts were asked to pinpoint the species in local identification guides [2,14] . When
the local experts did not unanimously agree on the identification of an audible species, they were
asked to reach a consensus through discussion or replaying a recording as many times as necessary.
J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619 5
Tabl e 2
Recording schedule and number of recordings per sensor.
Site Sensor # Recordings Start date End date
La Belgique BT1 1244 26-2-2020 26-4-2020
La Belgique BT2 1214 26-2-2020 25-4-2020
La Belgique BT3 1234 25-2-2020 25-4-2020
La Belgique BT4 1250 27-2-2020 27-4-2020
La Belgique BT5 1072 25-2-2020 25-4-2020
La Belgique BT6 810 25-2-20 8-4-2020
La Palestine PT1 1299 5-3-2020 5-5-2020
La Palestine PT2 13 04 5-3-2020 5-5-2020
La Palestine PT3 1202 1-3-2020 25-4-2020
La Palestine PT4 13 00 3-3-2020 3-5-2020
La Palestine PT5 13 00 3-3-2020 3-5-2020
La Palestine PT6 13 07 29-2-2020 29-4-2020
Ngouleminanga NT1 1322 8-3-2020 8-5-2020
Ngouleminanga NT2 1091 8-3-2020 30-4-2020
Ngouleminanga NT3 1030 9-3-2020 8-5-2020
Ngouleminanga NT4 1261 9-3-2020 9-5-2020
Ngouleminanga NT5 654 10-3-2020 8-4-2020
Ngouleminanga NT6 591 10-3-2020 4-4-2020
To avoid bias, they were kept uninformed about the site in which each recording was made. Since
this study used mammals and birds to evaluate the impact of ecological and anthropogenic factors
on vocalisation patterns, vocalisations from these taxa were identified by species. Vocalisations from
amphibians and insects were identified by class. Unidentifiable animal sounds were noted down as
‘Animal unknown’ or, if the local experts were sure that the sound was produced by a bird, ‘Bird
unknown’.
To assess the contribution of different animal classes to the soundscape, vocalisation abundance
was determined for each vocalising animal. All 1-minute recordings (from all transects) were pulled
together. The number of recordings in which an animal class was present was divided by the total
number of recordings in order to obtain the vocalisation rate for each class.
To assess how biological sounds vary among sites with differing levels of disturbance, abundance
and diversity of vocalisations were compared across study sites. For vocalisation abundance,
differences in vocalisation rates per sensor per day across study sites were evaluated for each species.
To ensure reliable analysis, only data recorded in all sites on the same day, at the same time, and
by sets of sensors with similar spatial designs were used. Sets of sensors were considered to have a
similar spatial design in all study sites when the geographical distance between the used sensors
was the same (the spatial configuration of the sensors providing data for analysis was consistent
across sites). This resulted in data from a total of 30 days. On each of these days, in all sites, one
or more sensors with the same spatial design recorded sounds without background noise between
6am–3pm and 9pm–10pm. This totalled 10 min of sound recordings per day. Therefore, calculated
vocalisation rates ranged between 0 and 10 for each vocalising species. To compare diversity of
vocalising species across sites, sound recordings were only used from times where all acoustic sensors
in all three study sites had recorded without background noise. This resulted in a total of 1512 1-
minute sound recordings in each site, spread over 23 days. For each vocalising species, the number
of sound recordings in which the species was present was determined per site. With these numbers,
rarefaction curves were plotted with iNEXT [6] . These curves were extrapolated to larger sample sizes
to estimate asymptotic species richness and compare diversity across sites.
To assess the drivers of biological sounds, data on anthropogenic and ecological factors were
collected during field surveys. All field surveys were conducted between 8AM and 1PM. In each
transect, habitat description and surveys of human activities, mammals (both direct and indirect
observations), great ape nests, and birds were conducted. During the surveys, a researcher walked
along the transect accompanied by one or more local guides who were able to detect and identify
signs. Additionally, data on precipitation, temperature, and humidity were obtained.
6 J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619
For all transects, the vegetation type was described at every marked 50-m interval. For this
description, vegetation consisted of 8 different habitat types (mature forest, old secondary forest,
young secondary forest, light gap, riparian forest, swamp, old plantation site, and current plantation)
based on previous vegetation classifications used in the area [9,24,25] .
All signs of human activity were recorded within a 2-m range perpendicular to the transect. Both
items left by humans, such as cartridges or rubbish, and human constructions, such as trails or traps,
were considered human activity. For each observation, the type of human activity, location along the
transect (m), and vegetation type were noted. Human activity surveys were conducted twice for every
transect, with one month in between surveys. Signs were removed from the transect after counting.
Signs that were impossible to remove, such as human trails, were marked during the first survey to
avoid recounting during the second survey.
All signs of animal activity present within 2-m on either side of the transect were recorded. The
local guide indicated the type of animal sign (e.g., footprint, dung, feeding remain) and the local
name of the corresponding animal species. For each observation, the location along the transect
(m), perpendicular distance from the transect (m), vegetation type, canopy openness, understorey
openness, and horizontal visibility (m) were recorded. Canopy openness (open, average, or closed),
understorey openness (open, average, or closed), and horizontal visibility (m) were visually estimated.
The openness of the canopy and understorey were always classified as open, average or closed.
Indirect mammal surveys were conducted twice for each transect, with one month in between
surveys. There was no overlap in observations between the surveys, because rainfall washed away
all signs counted during the first survey.
All encountered nests of central chimpanzees ( Pan troglodytes troglodytes ) and western lowland
gorillas ( Gorilla gorilla gorilla ) were recorded. The local guide identified the local names of the
plants used to construct the nests. Furthermore, the nests were counted and, for each nest, age,
location along the transect (m), perpendicular distance from the transect (m), and diameter (cm)
were recorded. To reliably distinguish between central chimpanzee and western lowland gorilla
nests, several criteria were used. First, fresh nests could be distinguished based on the presence
of characteristics like footprints, urine, hairs, and feces [27] . Furthermore, western lowland gorillas
commonly sleep in nests on the ground, whereas central chimpanzees tend to build their nests
in trees [11,22] . Consequently, nests lacking clear signs were distinguished based on their height.
However, central chimpanzee ground nesting occurs at a low rate in the area [19] . Therefore, nest
groups containing at least one nest in a tree at > 2-m height were attributed to central chimpanzees,
whereas nest groups that were built on the ground or in a tree at < 2-m height were attributed to
western lowland gorillas [21] . For western lowland gorilla nests, the type of nest was described by the
composition of plants used for construction (herbaceous or mix). In turn, for central chimpanzee nests,
the type of nest was described by its position in the tree (on the side or on the top). Additionally,
the height of the nest was estimated, the height and circumference of the tree were estimated,
and the tree was checked for fruits. Finally, vegetation type, canopy openness, understorey openness,
horizontal visibility (m), and coordinates were noted for each nest. Great ape surveys were conducted
twice for each transect, with one month in between surveys. All nests found during the first survey
were marked to avoid overlap with already recorded nests during the second survey.
To avoid disturbing animals before they were observed, observers walked along each transect at a
speed of approximately 1 km/h. Any direct observation of mammals along the transect was recorded.
The local guide indicated the local name of the species and the number of animals seen. Additionally,
the location along the transect (m), the distance between the observer and the animal (m), the angle
of the observation, vegetation type, canopy openness, understorey openness, and horizontal visibility
(m) were noted.
Birds were surveyed using two methods: point counts in fixed stations and direct observations. At
each 500-m interval along the transect, a bird point count station was located. This resulted in three
count stations per transect (0 m, 500 m, and 10 0 0 m). At every station, the observers waited two
minutes to get acquainted with the environment and to neutralise any possible disturbance caused
by the arrival of the observers. Thereafter, an initial observation direction was randomly chosen.
After two minutes of observing that direction, the observers rotated 90 °in a clockwise direction.
This resulted in 8 min of effective observation for each station [23] . Additionally, for every count
J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619 7
station, the vegetation type, canopy openness, understorey openness, and horizontal visibility (m)
were recorded. The local name and number of individuals of each bird species seen or heard in one
direction was noted. If a bird species was seen or heard in multiple observing directions at the same
count station, it was only counted once because it could be the same individual. Furthermore, when
walking the transect between count stations, direct observations of birds were recorded in the same
manner as direct mammal surveys. Finally, since time of day and weather condition can affect bird
behaviour, bird surveys were not undertaken on rainy or windy days [10] .
Data on rainfall, temperature, and humidity were collected in La Belgique and used to describe all
sites. Rainfall was measured in daily precipitation (mm), whereas temperature ( °C) and humidity (RH)
were noted down hourly.
To assess differences between study sites based on data obtained during field surveys, the
encounter rate (observations/km) was used. Thus, for every transect, the mean number of observations
for each type of field survey data was calculated. To further investigate the influence of the habitat
structure, the total length of swamps and terra firma forests (mature forest, old secondary forest,
young secondary forest and light gaps together) in the transects was calculated [24] . Thereafter, the
total amount of human, mammal, and bird signs in swamps and terra firma forests was determined.
Human activity was calculated overall and broken down into hunting signs and other human signs.
The encounter rates of mammal and bird signs were used as index of species abundance. As done
in a previous study [8] , mammals were grouped in seven defined assemblages of species: elephants,
carnivores, even-toed ungulates, pangolins, old world monkeys, great apes, and rodents. Consequently,
analyses were performed on mammals as a whole, on the defined taxonomical mammal guilds, and
at the species level. Since not all mammals are known to vocalise, another analysis was performed
for mammal species that were identified in the acoustic recordings [28] . For elephants, carnivores,
even-toed ungulates, pangolins, and rodents, indirect mammal survey data were used. Great ape
abundance was estimated by nest counts. To avoid possible bias that may have arisen from grouping
individual great ape nests into nest sites, the number of individual nests was used over the number
of nest sites. Furthermore, direct observations were used for old world monkey abundance estimates.
This approach is consistent with methodologies used in previous studies in the area, enabling the
guilds and species to be compared [8,19,29] . For birds, the encounter rate was also used as an index
of abundance; analyses were performed for all birds together and for each identified bird species
separately. To evaluate additional anthropogenic factors that might affect biological sounds, ArcGIS
was used to measure the shortest straight-line distance (m) between the sound recorders and the
closest village and trail. The distance between the recorder and the closest village served as a proxy
for the remoteness of the recording location, whereas distance to the nearest trail was used as a
measure of accessibility.
To assess how anthropogenic and ecological factors affect biological sounds, obtained values of
mammal abundance, bird abundance, human activity, geographical factors, and climatic measurements
were used as predictor variables. Furthermore, two dependent variables were calculated. Since we did
not have the same number of recordings for each site, these variables were calculated per sensor
per day. The first dependent variable is the proportion of files that contained vocal activity. This
proportion, used as a proxy for abundance of vocalisations, was calculated by dividing the number
of recordings with vocal activity by the total number of recordings per sensor per day. The second
dependent variable is the number of species identified per sensor per day. This variable was used as a
proxy for bioacoustic diversity. These dependent variables were calculated for all bird species together
and all mammal species together. It is important to note that peak acoustic activity in tropical forests
soundscapes occurs during dawn and dusk [5] . Troughout the year, sunrise in the research area always
occurs after 5am [26] . Therefore, this study defined a day as a 24-hour time span starting at 6am. This
way, every day started with dawn and ended with a full night. All obtained data were organized
accordingly. Data obtained during field surveys on transects were attributed to the corresponding
sensors. Additionally, the percentage of swamp and terra firma forest was calculated per transect,
thus per sensor. Note that only observations of dependent variables for which values for all predictor
variables were available were used. This resulted in a total of 847 observations.
8 J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619
Statistical analysis
Rstudio (version 4.0.2.) was used for all statistical analyses. To determine the structure of biological
sounds, normality of all processed data, obtained during the identification of sound recordings,
was tested per study site using the Shapiro–Wilk’s test. Normally distributed data were tested for
homogeneity of variances using a Bartlett test. If data for one of the compared study sites followed a
non-normal distribution, a Kruskal–Wallis one-way analysis of variance test was always used because
this test does not assume equal variances. As post-hoc analysis, to determine which study sites differ
significantly, Dunn’s multiple comparison test with Benjamini–Hochberg correction was performed
[16] . If data for all study sites were normally distributed and showed equal variances, a one-way
ANOVA test was used with Tukey Honest Significant Differences post-hoc analysis. For normally
distributed data with unequal variances between the study sites, a Welch ANOVA test with Games-
Howell post-hoc analysis was performed.
Sound data in this study follow a hierarchical pattern, where moment of recording is nested in
the transects, which are nested in the different study sites. However, preliminary multi-level analyses
did not result in a need to treat data from different sites and times differently when assessing the
drivers of biological sounds. Therefore, to evaluate the effect of anthropogenic and ecological factors
on biological sounds, generalised estimating equations (GEE) were used. GEE are an extension of
generalised linear models that allow for the analysis of repeated measurements where observations in
separate clusters are independent [12] . To assess multicollinearity among the predictor variables, the
variance inflation factor (VIF) of each variable was calculated. Variables with VIF > 5 were excluded
from the analyses [15] . Additionally, correlation analyses for all pairs of quantitative variables were
run. Only temperature and humidity were strongly correlated. Since temperature was measured
with more precision, humidity was excluded from the analyses. Poisson models for all dependent
variables were used to assess dispersion. Since the number of recordings per sensor per day was
not always equal, an offset variable (calculated as log of the total amount of recordings per sensor
per day) was added to the models. Models with a dispersion statistic of 0.8 < σp
< 1.2 were
considered normally dispersed [17] . Bioacoustic diversity of all wildlife together and birds separately
was normally dispersed, whereas total abundance of vocalisations and bird vocalisation abundance
was underdispersed. Mammal bioacoustic diversity was also underdispersed and mammal vocalisation
abundance was overdispersed. For models that showed under- and overdispersion, binomial GEE
analyses were run without an offset variable. Howe ver, binomial GEE models require the dependent
variable to be a proportion, but mammal bioacoustic diversity was represented as a count value.
Therefore, these values were divided by the total number of vocalising mammal species identified
throughout the study period to obtain a proportion of bioacoustic diversity. For the GEE analyses, the
“exchangeable” correlation structure was used and waves were added to maintain the chronological
order of the repeated measurements. The fitted GEE models were compared to similar models to
which weights were added to account for the different amounts of recordings that were available per
sensor per day. The models were compared using the QIC programme to select the GEE model that
best fits the dataset [7] . Models without weights proved to fit the dataset better. Therefore, results
from these models were saved.
Acknowledgements
We want to sincerely thank Maxwell Ndju’u, Julia van Plateringen, and all the local guides and
experts (Assim, Casimir, Florent, Marco, and Martial) who helped during the field surveys and the
identification of animals from sound recordings. This research would not have been possible without
funding from the Antwerp Zoo Centre for Research and Conservation (core funded by the Flemish
Government) and the contributions of the Association de la protection des grands singes and Utrecht
University. Finally, thanks are also extended to Stichting FONA and Stichting het Kronendak for
providing personal financial support to Johan Diepstraten.
J. Diepstraten, J.K. Kuenbo u and J. Willie / MethodsX 9 (2022) 101619 9
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
References
[1] P. Auz el , F. Feteke , T. Fomete , S. Nguiffo , R. Djeukam , Social and environmental costs of illegal logging in a forest
management unit in Eastern Cameroon, J. Sustain. Forestry 19 (1–3) (2004) 153–180 .
[2] N. Borrow , R. Demey , Birds of Weste rn Africa. Series ‘Helm Identification Guides, Gardners Books, 2002 .
[3] E. Browning , R. Gibb , P. Glover-Kapfer , K.E. Jones , Passive acoustic monitoring in ecology and conservation, WWF Conserv.
Technol. Ser. 1 (2) (2017) 1–75 .
[4] D. Bruggeman , P. Meyfroidt , E.F. Lambin , Production forests as a
conservation tool: effectiveness of Cameroon’s land use
zoning policy, Land Policy 42 (2015) 151–164 .
[5] Z. Burivalova , M. Towsey , T. Boucher , A. Truskinger , C. Apelis , P. Roe , E.T. Game ,Using soundscapes to detect variable
degrees of human influence on tropical forests in Papua New Guinea, Conserv. Biol. 32 (1) (2018) 205–215 .
[6] Chao, A., Ma, K.H., and Hsieh, T.C. (2016) iNEXT (iNterpolation and EXTrapolation) Online: Software for Interpolation and
Extrapolation of Species Diversity. Program and User’s Guide published at http://chao.stat.nthu.edu.tw/wordpress/software _
download/ .
[7] J. Cui , QIC program and model selection in
GEE analyses, Stata J. 7 (2) (2007) 209–220 .
[8] S. Dekegel , Wildlife Under pressure: Assessing the Effects of Biophysical and Anthropogenic Factors On Tropical Forest
Mammals and Their Habitats in Different Land-Use Types in the Dja Conservation Complex, Cameroon Master’s thesis,
Universit y of Antwerp, Antwerp, Belgium, 2018 .
[9] J. Dupain , P. Guislain , G.M. Nguenang , K. De Vleeschouwer , L. Van Elsacker , High chimpanzee and gorilla densities in a
non-protected area on the northern periphery of the Dja Faun al Reserve, Cameroon, Oryx 38 (2) (2004) 209–216 .
[10] A. Esquivel , S. Peris , Influence
of time of day, duration and number of counts in point count sampling of birds in an
Atlantic forest of Paraguay, Ornitol. Neotrop. 19 (2008) 229–242 .
[11] J.M. Goodall , Nest building behavior in the free ranging chimpanzee, Ann. N. Y. Acad. Sci. 102 (2) (1962) 455–467 .
[12] U. Halekoh , S. Højsgaard , J. Yan , The R package geepack for generalized estimating equations, J. Stat. Softw. 15 (2) (2006)
1–11 .
[13] A.P. Hill , P. Prince , J.L. Snaddon , C.P. Doncaster , A. Rogers , AudioMoth: a low-cost acoustic device for monitoring biodiversity
and the environment, HardwareX 6 (2019) e0 0 073 .
[14] M. Languy , The Birds of Cameroon. Their Status and Distribution . Series ‘Studies in Afrotropical Zoology, 299, Royal Museum
for Central Africa, Ter vuren , 2019 .
[15] R.M. O’Brien , A caution regarding rules of thumb for variance inflation factors, Qual Quant. 41 (5) (2007) 673–690 .
[16] Ogle, D.H., Wheeler, P., & Dinno, A. (2017). FSA: fisheries stock analysis. R Package Version 0.8, 17, 629.
[17] E.H. Payne , M. Gebregziabher , J.W. Hardin , V. Ramakrishnan , L.E. Egede , An empirical approach to determine a threshold
for assessing overdispersion in
Poisson and negative binomial models for count data, Commun. Stat. - Simul. Comput. 47
(6) (2018) 1722–1738 .
[18] N. Tagg , C.A. Petre , J. Willie , Evaluating the effectiveness of a 10-ye ar s old great ape conservation project in Cameroon, Pan
Africa News 18 (2) (2011) 21–23 .
[19] N. Tagg , J. Willie , C.A. Petre , O. Haggis , Ground night nesting in chimpanzees: new insights from central chimpanzees (Pan
troglodytes troglodytes) in south-east Cameroon, Folia Primatol. 84 (6) (2013) 362–383 .
[20] N. Tagg , J. Willie , J. Duarte , C.A. Petre , J.E. Fa , Conservation
research presence protects: a case study of great ape abundance
in the Dja region, Cameroon, Anim. Conserv. 18 (6) (2015) 4 89–4 98 .
[21] L.R.D. Tédonzong , J. Willie , S.T. Makengveu , L. Lens , N. Tagg , Variation in behavioral traits of two frugivorous mammals
may lead to differential responses to human disturbance, Ecol. Evol. 10 (8) (2020) 3798–3813 .
[22] C.E. Tutin , R.J. Parnell , L.J. White , M. Fernandez , Nest building by lowland gorillas in the Lopé Reserve, Gabon:
environmental influences and implications for censusing, Int. J. Primatol. 16 (1) (1995) 53 .
[23] A.
Wheeldon , H.L. Mossman , M.J. Sullivan , J. Mathenge , S.R. de Kort , Comparison of acoustic and traditional point count
methods to assess bird diversity and composition in the Aberdare National Park, Kenya, Afr. J. Ecol. 57 (2) (2019) 168–176 .
[24] J. Willie , C.A. Petre , N. Tagg , L. Lens , Density of herbaceous plants and distribution of western gorillas in different habitat
types in south-east Cameroon, Afr. J. Ecol. 51 (1) (2013) 111–121 .
[25] J. Willie , N. Tagg , C.A. Petre , Z. Pereboom , L. Lens , Plant selection for nest building
by western lowland gorillas in Cameroon,
Primates 55 (1) (2014) 41–49 .
[26] Worlddata.info. (2020). Sunrise and sunset in Cameroon [Cited on 18 Sep 2020]. Available from URL: https://www.
worlddata.info/africa/cameroon/sunset.php
[27] D. Morgan , C. Sanz , J.R. Onononga , S. Strindberg , Ape abundance and habitat use in the Goualougo Triangle, Republic of
Congo, Int. J. Primatol. 27 (1) (2006) 147–179 .
[28] R.T. Buxton , P.E. Lendrum , K.R. Crooks , G. Wittemyer , Pairing camera traps and acoustic recorders to monitor the ecological
impact of human disturbance, Global Ecology and conservation (2018) .
[29] R.F. Laurance , B.M.
Croes , L. Tchignoumba , S.A. Lahm , A. Alonso , M.E. Lee , C. Ondzeano , Impacts of roads and hunting on
central African rainforest mammals, Conserv. Biol. 20 (4) (2006) 1251–1261 .
... • To assess additional anthropogenic factors, the shortest straight-line distance (m) between each sound recorder and the closest village and trail was measured using ArcGIS to get proxies for remoteness and accessibility, respectively. -See Diepstraten et al. [2] for a detailed description of the experimental design for collecting these data. ...
Article
Full-text available
Obtaining and analysing sound data can be a tedious and lengthy process. We present sound data consisting of 20,485 1-minute sound recordings obtained in three sites within a rainforest landscape in southeast Cameroon. The sites differ in anthropogenic disturbance. We also present meta data corresponding to these recordings with the identification of all animal vocalisations in each 1-minute sound recording. Additionally, we provide a raw database with data on habitat, human activities, remoteness, accessibility, temperature, humidity, rainfall, moon phase, and mammal and bird observations in the area during the recording period. The data were used by Diepstraten & Willie (2021) to investigate the structure and drivers of biological sounds along a disturbance gradient. The data contribute to call libraries of tropical species and can also be used to build classifiers for automatic detection and classification of animal vocalisations.
Article
Full-text available
Studying soundscapes to improve knowledge about an environment with limited visibility is gaining popularity. However, the structure and drivers of biological sounds remain poorly understood although this information can improve understanding of landscape dynamics. To describe biological sounds and provide a better understanding of their drivers, we conducted a study aiming at detecting and identifying individual vocalising animal species, determining the abundance, diversity, and composition of vocalisations, and assessing how ecological and anthropogenic factors affect these sounds. Sounds were recorded from February through May 2020 in southeast Cameroon, using passive acoustic monitoring, in three study sites representing a gradient of disturbance. Local experts listened to 20,485 1-minute files of recorded sounds to identify vocalising species. These identifications, in combination with anthropogenic and ecological data from the field, were used to assess abundance, diversity, and composition of vocalisations. Fifteen vocalising mammal species and 37 vocalising bird species were identified. Overall, insects and, to a lesser extent, birds were the most dominant animal classes in the soundscape. Furthermore, vocalisation abundance and the diversity of vocalising mammals did not vary along the disturbance gradient, whereas vocalisation abundance of birds was lowest in the site where anthropogenic disturbance was least present. Moreover, both mammal and bird vocal activity were negatively impacted by rainfall. Human activity and the amount of swamp habitat reduced the vocal activity of mammals, but these factors did not affect bird vocalisations. For birds, on the other hand, the diversity of vocalisations was positively affected by bird abundance and negatively affected by higher temperatures. Our results indicate that animal classes do not have the same number of acoustically active species and produce differing amounts of sound, thus highlighting that sources of sound differ in their relative contributions to the soundscape and furthering knowledge of soundscape structure. They also show that the effects of anthropogenic and ecological factors on biological sounds vary depending on the sound source, which suggests that soundscape components are affected by soundscape drivers differently. Additionally, our study provides baseline acoustic information on vocalising species in African tropical rainforests. This acoustic information may be used for automated detections of vocalisations in the future. Together, an understanding of soundscape drivers and information on vocalising species can contribute to the development of applications that will monitor soundscapes and detect important changes in landscapes. This will help guide future decision making and landscape conservation planning.
Article
Full-text available
Human activities can lead to a shift in wildlife species' spatial distribution. Understanding the specific effects of human activities on ranging behavior can improve conservation management of wildlife populations in human-dominated landscapes. This study evaluated the effects of forest use by humans on the spatial distribution of mammal species with different behavioral adaptations, using sympatric western lowland gorilla and central chimpanzee as focal species. We collected data on great ape nest locations, ecological and physical variables (habitat distribution, permanent rivers, and topographic data), and anthropogenic variables (distance to trails, villages, and a permanent research site). Here, we show that an-thropogenic variables are important predictors of the distribution of wild animals. In the resource model, the distribution of gorilla nests was predicted by nesting habitat distribution, while chimpanzee nests were predicted first by elevation followed by nesting habitat distribution. In the anthropogenic model, the major predictors of gorilla nesting changed to human features, while the major predictors of chimpanzee nesting remained elevation and the availability of their preferred nesting habitats. Animal behavioral traits (body size, terrestrial/arboreal, level of specialization/ generalization, and competitive inferiority/superiority) may influence the response of mammals to human activities. Our results suggest that chimpanzees may survive in human-encroached areas whenever the availability of their nesting habitat and preferred fruits can support their population, while a certain level of human activities may threaten gorillas. Consequently, the survival of gorillas in human-dominated landscapes is more at risk than that of chimpanzees. Replicating our research in other sites should permit a systematic evaluation of the influence of human activity on the distribution of mammal populations. As wild animals are increasingly exposed to human disturbance, understanding the resulting consequences of shifting species distributions due to human disturbance on animal population abundance and their long-term survival will be of growing conservation importance.
Article
Full-text available
Environmental sound is a powerful data source for investigating ecosystem health. To capture it, scientists commonly use ruggedized, but expensive acoustic monitoring equipment. In this paper we fully describe the hardware build of a low-cost, small, full-spectrum alternative, called AudioMoth. The credit-card sized device consists of a printed circuit board (PCB), micro-controller and a micro-electro-mechanical systems (MEMS) microphone. This simple to construct device facilitates: (1) deployments in remote locations, with a small size and a simple mechanism that allows it to be retrofitted into numerous low-cost ruggedized enclosures; (2) long-term monitoring, with low-power operation; (3) modular expansion, with easy to access general purpose input and output pins; and (4) acoustic detection, with onboard processing power.
Article
Full-text available
Over the past two decades, the use of camera traps and acoustic monitoring in the investigation of animal ecology have grown rapidly, with each technique enhancing broad-scale wildlife surveying. Camera traps are a cost-effective, noninvasive means of sampling communities of mid-to large-terrestrial species, and acoustic recording devices capture human sounds and sound-producing animals, including species of mammals, birds, anurans, and insects. Rarely are these techniques combined, despite the advantages of merging their respective strengths. Namely, camera traps paired with acoustic recorders can evaluate the abundance, distribution, and behavior of multiple guilds and trophic levels across landscapes while concurrently monitoring multiple human stressors in real time. Moreover, integrating these approaches enhances detection accuracy and strengthens statistical inference at multiple survey scales. We conducted a literature review, and found only 13 studies that combine camera traps and acoustic recorders, 8 of which either compared the ability of each technique to detect species of interest or discussed the advantages of each technique. We outline potential questions that can be addressed by pairing acoustic recorders and camera traps, including enabling the simultaneous assessment of noise pollution and its impacts on mammal and avian communities. Furthermore, we discuss how the analysis of data from each technique face similar challenges; thus, simultaneous innovation offers the ability to apply solutions to both techniques and amplify their respective strengths. Digital technologies and big data are changing nature conservation in increasingly profound ways and integration of camera traps and acoustic recorders will facilitate new, transformative discoveries to meet modern conservation challenges.
Conference Paper
Growing human populations and associated impacts such as forest disruption and overexploitation of wildlife resources, are a rapidly growing threat to African tropical wildlife. Protected areas represent an important global conservation strategy for the preservation of biodiversity. However, inadequate funding and governmental support combined with environmental encroachment occurring in the surrounding landscape, are reducing their conservation capacity. There is growing recognition of the need to extend beyond protected areas and evaluate conservation efforts outside protected areas. This study aims to evaluate the impact of environmental and anthropogenic factors on tropical forest mammals and their habitat in six sites, differing in terms of land-use type and level of active management, in the Dja Conservation Complex, southeast Cameroon. Standardized transects were used to assess species abundances and record vegetation characteristics and signs of human activity. Results show that the study sites differed significantly in anthropogenic activities and mammal abundance. Environmental factors investigated in this study did not influence mammal abundances whereas mammal abundances were influenced by differences in anthropogenic pressures across sites. The results also showed that a lack of effective management in the unprotected sites resulted in lower mammal abundances and that effective management and protection seem to be lacking at forest sites near villages at the northern border of the Dja Biosphere Reserve. This study also provides evidence that the conservation efforts of Project Grand Singes reduced human activity and positively influenced mammal populations. Hence, we encourage the presence and activities of conservation research as secondary conservation activity to protect mammal populations and benefit local communities at the same time.
Article
Afromontane forests, like those in the Aberdare National Park (ANP) in Kenya, sustain unique avifaunal assemblages. There is a growing need for biodiversity inventories for Afromontane forests, especially through the utilization of unskilled observers. Acoustic surveys are a potential aid to this, but more comparisons of this technique with that of traditional point counts are needed. We conducted a systematic survey of the ANP avifauna, assessing whether acoustic and traditional surveys resulted in different species richness scores, and whether this varied with habitat and species characteristics. We also investigated the role of habitat and elevation in driving variation in species richness. The ANP provides habitat types including scrub, moorland, montane, hagenia and bamboo forests. Overall, the surveys yielded 101 identified species. The acoustic method resulted in higher species richness scores compared to the traditional method across all habitats, and the relative performance of the two methods did not vary with habitat type or visibility. Both methods detected different species, suggesting that they should be used together to maximise the range of species recorded. We found that habitat type was the primary driver of variation in species richness, with scrub and montane forest having significantly higher species richness scores than the other habitats.
Article
en Afromontane forests, like those in the Aberdare National Park (ANP) in Kenya, sustain unique avifaunal assemblages. There is a growing need for biodiversity inventories for Afromontane forests, especially through the utilisation of unskilled observers. Acoustic surveys are a potential aid to this, but more comparisons of this technique with that of traditional point counts are needed. We conducted a systematic survey of the ANP avifauna, assessing whether acoustic and traditional surveys resulted in different species richness scores, and whether this varied with habitat and species characteristics. We also investigated the role of habitat and elevation in driving variation in species richness. The ANP provides habitat types including scrub, moorland, montane, hagenia and bamboo forests. Overall, the surveys yielded 101 identified species. The acoustic method resulted in higher species richness scores compared to the traditional method across all habitats, and the relative performance of the two methods did not vary with habitat type or visibility. The methods detected different species, suggesting that they should be used together to maximise the range of species recorded. We found that habitat type was the primary driver of variation in species richness, with scrub and montane forest having higher species richness scores than other habitats. Résumé fr Les forêts afromontane, comme celles du parc national d’Aberdare (ANP) au Kenya, présentent des assemblages d’avifaune uniques. D’où, le besoin croissant d’inventaires de la biodiversité de ces forêts afromontanes, notamment réalisés par des observateurs non qualifiés. Les relevés acoustiques en sont une aide potentielle, mais il est nécessaire d’avoir un plus grand nombre de comparaison entre les relevés acoustiques et le dénombrement par point traditionnel afin d’évaluer la meilleure méthode à utiliser. Ainsi, nous avons effectué une étude systématiques de l‘avifaune de l’ANP, en évaluant si les relevés acoustiques et traditionnels entraînaient des résultats de richesse d’espèces différents et s’ils variaient en fonction des caractéristiques d’habitat et d’espèce. Nous avons également étudié le rôle de l’habitat et de l’élévation dans la variation de la diversité des espèces comme l’ANP contient différent types d’habitats, tells que de la brousse, la lande et des forêts subalpines, d’Hagenia et de bambou. Pour tous les habitats, la méthode acoustique a donné des résultats plus élevés de richesse d’espèces que la méthode de dénombrement par point traditionnelle, qui a dans l’ensemble, a permis d’identifier 101 espèces d’oiseaux. La performance relative des deux méthodes n’a pas variée en fonction du type d’habitat ou de la visibilité. Les deux des méthodes ont détecté différentes espèces d’oiseaux, ce qui suggère qu’elles devraient être utilisées ensemble afin de maximiser la gamme d’espèces enregistrées. Nous avons constaté que le type d’habitat était le principal moteur de variation de la richesse des espèces et que la brousse et la forêt subalpine en particulier, donnent des résultats de richesse en espèces plus élevés que les autres habitats.
Article
Overdispersion is a problem encountered in the analysis of count data that can lead to invalid inference if unaddressed. Decision about whether data are overdispersed is often reached by checking whether the ratio of the Pearson chi-square statistic to its degrees of freedom is greater than one; however, there is currently no fixed threshold for declaring the need for statistical intervention. We consider simulated cross-sectional and longitudinal datasets containing varying magnitudes of overdispersion caused by outliers or zero inflation, as well as real datasets, to determine an appropriate threshold value of this statistic which indicates when overdispersion should be addressed.