ArticlePDF Available

Abstract and Figures

Introduction: Monitoring large carnivores is a central issue in conservation biology. The wolf (Canis lupus) is the most studied large carnivore in the world. After a massive decline and several local extinctions, mostly due to direct persecutions, wolves are now recolonizing many areas of their historical natural range. One of the main monitoring techniques is the howling survey, which is based on the wolves' tendency to use vocalisations to mark territory ownership in response to howls of unknown individuals. In most cases wolf howling sessions are useful for the localisation of the pack, but they provide only an aural estimation of the chorus size. We tested and present a new bioacoustic approach to estimate chorus size by recording wolves' replies and visualising choruses through spectrograms and spectral envelopes. To test the methodology, we compared: a) the values detected by visual inspections with the true chorus size to test for accuracy; b) the bioacoustic estimations of a sample of free-ranging wolves' replies developed by different operators to test for precision of the method; c) the aural field estimation of chorus size of a sample of free-ranging wolves' replies with the sonogram analysis of the same recordings to test for difference between methods. Results: Visual inspection of the chorus by spectrogram and spectrum proved to be useful in determining the number of concurrent voices in a wolf chorus. Estimations of chorus size were highly correlated with the number of wolves counted in a pack, and 92 % of 29 known chorus sizes were recognized by means of bioacoustic analysis. On the basis of spectrographic evidence, it was also possible to identify up to seven concurrent vocalisations in a chorus of nine wolves. Spectral analysis of 37 free ranging wolves' replies showed a high correlation between the chorus size estimations of the different operators (92.8 %), but a low correlation with the aural estimation (59.2 %). Conclusions: Wolf howling monitoring technique could be improved by recording wolves' replies and by using bioacoustic tools such as spectrograms and spectral envelopes to determine the size of the wolf chorus. Compared with other monitoring techniques (i.e., genetic analysis), bioacoustic analysis requires widely available informatic tools (i.e., sound recording set of devices and sound analysis software) and a low budget. Information obtained by means of chorus analysis can also be combined with that provided by other techniques. Moreover, howls can be recorded and stored in audio file format with a good resolution (i.e. in "Wave" format), thus representing a useful tool for future listening and investigations, which can be countlessly employed without risks of time deterioration.
Content may be subject to copyright.
Visualizing sound: counting wolves by using a
spectral view of the chorus howling
Passilongo et al.
Passilongo et al. Frontiers in Zoology (2015) 12:22
DOI 10.1186/s12983-015-0114-0
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
M E T H O D O L O G Y Open Access
Visualizing sound: counting wolves by
using a spectral view of the chorus howling
Daniela Passilongo
1*
, Luca Mattioli
2
, Elena Bassi
1
, László Szabó
3
and Marco Apollonio
1
Abstract
Introduction: Monitoring large carnivores is a central issue in conservation biology. The wolf (Canis lupus) is the
most studied large carnivore in the world. After a massive decline and several local extinctions, mostly due to direct
persecutions, wolves are now recolonizing many areas of their historical natural range. One of the main monitoring
techniques is the howling survey, which is based on the wolvestendency to use vocalisations to mark territory
ownership in response to howls of unknown individuals. In most cases wolf howling sessions are useful for the
localisation of the pack, but they provide only an aural estimation of the chorus size.
We tested and present a new bioacoustic approach to estimate chorus size by recording wolvesreplies and
visualising choruses through spectrograms and spectral envelopes. To test the methodology, we compared: a) the
values detected by visual inspections with the true chorus size to test for accuracy; b) the bioacoustic estimations
of a sample of free-ranging wolvesreplies developed by different operators to test for precision of the method;
c) the aural field estimation of chorus size of a sample of free-ranging wolvesreplies with the sonogram analysis of
the same recordings to test for difference between methods.
Results: Visual inspection of the chorus by spectrogram and spectrum proved to be useful in determining the
number of concurrent voices in a wolf chorus. Estimations of chorus size were highly correlated with the number
of wolves counted in a pack, and 92 % of 29 known chorus sizes were recognized by means of bioacoustic analysis.
On the basis of spectrographic evidence, it was also possible to identify up to seven concurrent vocalisations in a
chorus of nine wolves. Spectral analysis of 37 free ranging wolvesreplies showed a high correlation between the
chorus size estimations of the different operators (92.8 %), but a low correlation with the aural estimation (59.2 %).
Conclusions: Wolf howling monitoring technique could be improved by recording wolvesreplies and by using
bioacoustic tools such as spectrograms and spectral envelopes to determine the size of the wolf chorus. Compared
with other monitoring techniques (i.e., genetic analysis), bioacoustic analysis requires widely available informatic
tools (i.e., sound recording set of devices and sound analysis software) and a low budget. Information obtained by
means of chorus analysis can also be combined with that provided by other techniques.
Moreover, howls can be recorded and stored in audio file format with a good resolution (i.e. in Waveformat), thus
representing a useful tool for future listening and investigations, which can be countlessly employed without risks
of time deterioration.
* Correspondence: dpassilongo@uniss.it
1
Department of Science for Nature and Environmental Resources, University
of Sassari, via Muroni 25, I-07100 Sassari, Italy
Full list of author information is available at the end of the article
© 2015 Passilongo et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Passilongo et al. Frontiers in Zoology (2015) 12:22
DOI 10.1186/s12983-015-0114-0
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Introduction
Monitoring and managing the recovery of large wide-
ranging mammalian carnivores are major issues for both
conservation biologists and wildlife managers [1]. In the
light of the continuous habitat loss and direct persecu-
tion of these species, large carnivores conservation has
become a pressing need [2]. Often described as charis-
matic species[3], these carnivores actually have a high
ecological [2, 4], economic and social impact [3, 57]
and, therefore, an in-depth knowledge of their ecology
and behaviour is highly desirable for the development of
an effective conservation strategy.
At the beginning of the 1990s, after a massive decline
caused by centuries of direct persecution, deforestation
and overhunting of its natural prey [8, 9], the wolf
(Canis lupus), became almost extinct in Western Eur-
ope, United States and Mexico [4]. At the present,
thanks to legal protection and socio-economic changes
in Europe, as the improvement in habitat quality and the
presence of large populations of wild ungulates [10]; the
relocation effort [11] and areas with vast public lands
[12] in US, wolves have reoccupied 67 % of their histor-
ical range worldwide [4]; in this context, monitoring
wolf population is a crucial issue for wolf conservation
efforts.
One of the main wolf monitoring tools is the howling
survey [13]: given the tendency of resident wolves to re-
spond to extraneous vocal stimuli in order to defend the
resources in their territories and to avoid encounter with
neighbour packs, this approach consists in the acoustic
stimulation produced through human simulation or
playback of actual wolf howls [1416]. This method was
employed in several studies for monitoring and census-
ing wolf packs [15, 1719] and other species with similar
vocal behaviour [20, 21]. However, in most cases, suc-
cessful field wolf howling sessions end with the pack lo-
calisation and an aural estimation of chorus size, and
only few attempts to determine chorus size through
complex sound and statistical analysis have been made
[2226]. In this regard, several studies have already
found both individual [2628] and group vocal signature
[29] by using spectrographic analysis, thus emphasising
the high potential of bioacoustic tools to improve the
knowledge on this species. Nevertheless, no study has
yet used available bioacoustic softwares and visual spec-
trographic inspection of the howling to estimate chorus
size.
In most cases chorus size is estimated aurally, by
counting each individual as it joins the chorus [15, 30].
However, precise estimation of the number of wolves is
difficult with this method because only the first two or
three wolves enter in the chorus as a staggered basis
followed by the rest of the pack en masse [15]. More-
over, the packs often reply before the end of the
stimulus, and, in particular, packs with more adults tend
to reply more quickly [15, 22, 30]. Additionally, since
subordinate adultsand pupshowls consist in a rapid
frequency modulation, they add complexity to the
chorus and even experts may encounter difficulties in
counting them (see [15]). Counting ability in humans
has been tested also in the case of human voices [31]
and music instruments [32]. In both cases, human aural
perception generally failed to count more than three
concurrent sources, however, during the test on the de-
numerability of music instruments, musicians performed
about 20 % better than non-musicians [32], thus
highlighting the variability level associated to individual
background and expertise. Consistent with these find-
ings, studies on the aural denumerability of chorus size
also recognised about 3 wolves [15, 22].
For this reason, researchers conservatively set a fixed
pack size to their data collected through howling survey
(see [33, 34]), and, if possible, determine the pack size
and the aggregation rate of the packs of their target pop-
ulations using other, often more expensive and time-
consuming techniques (genetic: [34, 35], tracks: [33],
VHF/GPS collars: [36, 37]; camera traps: [38]). Since
during the reply some individuals of the pack could stay
silent [16] or be temporarily absent [37], chorus size
does not necessarily correspond to the actual pack size;
even so, chorus size can be used as a proxy of the mini-
mum pack size. As for wolves, the pack is the basic so-
cial unit [39] and pack size is correlated with several
ecological traits such as hunting efficiency and prey se-
lection [40, 41], pack size estimation is a key-issue for
wolf monitoring, research, and conservation purposes.
In this paper, we present a bioacoustic approach based
on the visual inspection of the sound to estimate chorus
size of the pack. We validated the method by evaluating
both its accuracy (comparison of the chorus size esti-
mated through the method with the real chorus size)
and its precision (comparison of the chorus size esti-
mated through the method by two operators) The effi-
ciency of the method was then proved by comparing the
chorus size estimated through visual inspections with
the estimation of the chorus size based on an aural
estimation.
Therefore, the aim of this study was to test the bene-
fits of using bioacoustic analysis to estimate wolf chorus
size by recording and analysing wolveschoruses during
the wolf howling monitoring sessions, thus highlighting
a further connection between behavioural and conserva-
tion issues.
Results and discussion
Accuracy of the method
Visual estimation by spectrogram and spectrum pro-
vided an accurate evaluation of the chorus size as a
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 2 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
strong positive correlation was found between real and
estimated values in both our tests of known chorus size
(Human simulated howling, HSH: n= 20, Spearmans
rho = 0.90, p< 0.0001; and Wolf downloaded howling,
WDH: n= 9, Spearmans rho = 0.97, p< 0.0001), (Fig. 1)
with no difference between real and estimated means
(Wilcoxon signed rank test, HSH: n= 20, Z= 2.12, p=
0.057; WDH: n=9, Z= 1.13, p= 0.501). The most com-
mon error was the underestimation (false negative) for
one individual (five and two cases in HSH and WDH),
followed by the underestimation of two individuals (two
and one cases) (Fig. 1). Overestimation (false positive)
occurred twice, with four concurrent voices estimated in
a chorus size of three (in HSH) and six concurrent calls
estimated in a chorus of five individuals (in WDH)
(Fig. 1). Choruses composed by two individuals were al-
ways correctly estimated and it was even possible to rec-
ognise up to six different voices in a chorus of eight
contemporaneous voices (in HSH test) and up to seven
different voices in a chorus of nine elements (in WDH
test). In the two cases of overestimation, the detection of
the phantom voicewas probably due to the misinter-
pretation of a mix of non-linear (deterministic chaos,
subharmonics) [42, 43] and environmental phenomena
(reverberation, echoes) [44]. In fact, while harmonic
overtones were easily recognised because of their shape
(the same as F0) and frequencies (integer multiple of F0)
(Fig. 2) and, therefore, hardly misinterpreted as a differ-
ent fundamental frequency (one more individual), other
phenomena could affect and limit chorus visualisation.
However, repeated measures in different parts of the
chorus could effectively limit overestimation errors.
Considering the overall sample (test HSH and WDH),
92 % of the total number of voices were recognised by
the visual analysis of the choruses, thus highlighting the
potential and wide applicability of this methodology.
In the last few years, like all the inherent computer
science technologies, analysis tools such as spectrograms
and spectral envelopes based on the digital Fourier
transform [45] have become accessible to a broad range
of researchers [46]. This was possible thanks to the
spread of several highly interactive software (e.g., Raven,
made available by Cornell University [47], and widely
accessed open source software environments and pro-
gramming languages such as R [48], with its Seewave
package specifically dedicated to time series (i.e. sounds)
visualisation and analysis [49] (see Fig. 3 and Additional
file 1 and Additional file 2 for an example of spectro-
graphic and spectral visualisation of a wolves chorus).
Precision of the method
Results from the visual estimation performed by another
independent operator after the training on the method-
ology, also showed high correlation values for both test
(Human simulated howling, HSH: n= 20, Spearmans
rho = 0.92, p< 0.0001; and Wolf downloaded howling,
WDH: n= 9, Spearmans rho = 0.80, p= 0.009); and no
statistical difference between real and estimated means
(Wilcoxon signed rank test, HSH: n= 20, Z= 1.73, p=
0.147; WDH: n=9, Z= 0.74, p= 0.495) confirming the
accuracy and showing the precision of the bioacoustic
method.
Similarly, analysis of free ranging wolvesreplies showed
a high correlation between bioacoustic estimations of
chorus sizes by independent operators (Spearmansrho=
0.93, p< 0.0001) with no difference between operators
means (Wilcoxon, n=37, Z=1.898, p= 0.112). Chorus
size estimation by visual inspection was the same for both
operators in 30 replies out of 37 (88 %), and operator 2
agreed with operator 1 in 131 out of 138 total voices clas-
sification. Maximum difference between operators was of
two individuals in only one case and one individual in six
Fig. 1 Bioacoustically predicted versus Real chorus size. Scatter plots representing real versus bioacoustically predicted chorus size for Human
Simulated (HSH) (panel a) and Wolf Downloaded Howling (WDH) (panel b) tests. Radiuses are proportional to the number of cases. Estimation
was exact in 12 cases out of 20 for HSH and 5 cases out of 9. The diagonal represents the 1:1 correlation
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 3 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
cases (Fig. 4a), thus showing a high precision of spectro-
gram analysis.
Efficiency of the method
Comparison of aural estimations of the 37 free ranging
wolvesreplies with the bioacoustic estimations (Fig. 4b)
showed low correlation in both cases (vs. Operator 1:
Spearmans rho = 0.59, p= 0.0001; and vs. Operator 2:
rho = 0.63, p< 0.0001), which thus proved the two
methods do not provide the exact same results.
Aural and bioacoustic estimations showed similar
overall mean pack size, i.e. comparable accuracy, but
very low concordance that is different precision. Mean
chorus size obtained from bioacoustic and aural method
did not differ statistically (Wilcoxon, n= 37, Operator 1
vs. Aural: Z= 0.446; p= 0.689; Operator 2 vs. Aural: Z=
Fig. 2 Spectral components of a wolf chorus. Narrow band spectrogram (DFT size: 2048 samples; Hanning window; frequency grid: 21.5 Hz;
time step: 10 ms; bandwidth: 37.5 Hz) showing 3.5 s of a wolf chorus emitted by a free ranging pack and recorded during the howling survey.
Spectrogram (a) and its spectral envelope (b) at the second 1.7 are presented. Three different howls recognizable because of the different shape
of F0 and harmonic structure, are present at the same time. Other amplitude picks are due to background noise and echoes. Legend: W = wolf ;
F0 = fundamental frequency; F1 = first harmonic; F2 = second harmonic
Fig. 3 Three wolveschoral howls. Chorus howls of at least three different wolves (free-ranging) recorded during the howling survey in 2007.
Spectrogram (a) and spectrum (b) (window length: 4026 samples) were computed by Seewave, an open source R project package dedicated to
the sound analysis. Colours (from red to blue) represent amplitude degradation. See also Additional file 1 and Additional file 2 for analysis of
wolves choruses with Seewave
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 4 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.362; p= 0.186) as a consequence of bi-directional dif-
ferences between aural and bioacoustic estimations.
Aural and visual pack size estimations were identical in
only 12 cases (32 %); a difference of one individual was
found in 16 cases, a difference of two individuals was
found in eight cases. Finally, three more individuals were
estimated aurally with respect to the bioacoustic estima-
tion in one case (Fig. 4b).
For polyphonic sounds, visualisation is a source separ-
ation method [46] and, below certain sizes, good spec-
trographic exams of wolf choruses can realistically reveal
a number of different contemporaneous sources. How-
ever, as the number of contemporaneous sources in-
creases (79 sources, depending on the quality of the
recordings and the chorus modulation), the overlap be-
tween howls generates a high-density spectrogram (and
spectral envelope), in which the different components of
the sound can hardly be discriminated.
Here we showed a high correlation between bioacous-
tic estimation performed by visual inspection of the
chorus and real chorus size and between the estimations
of different operators, thus highlighting the precision of
the technique. Sound editing can also improve listening
quality: indeed, most of the factors affecting aural field
estimation (high background noise) can be totally (chick-
adees) or partially (in case of overlap with the target fre-
quencies - e.g., airplanesengines) removed (filtered out)
during the sound analysis phase.
Conversely, we showed.discrepancies between aural
and bioacoustic chorus size estimations. Previous works
[15, 22] also showed that aural field estimation of wolf
packs commonly failed to count more than three wolves
correctly, consistent with human general failure to rec-
ognise more than three concurrent sources in a poly-
phonic contest, both in the case of human voice [31]
and music instruments [32].
Although chorus size does not necessarily correspond
to pack size [16, 37] and further studies are needed in
order to better determine the age class of the members
and the overall active space of the choruses, we believe
that this method represents a very useful tool and a rela-
tive low-budget technique for the estimation of mini-
mum wolf pack size.
Conclusions
In this paper, we presented a bioacoustic approach based
on visual spectrographic inspections of wolveshowling
to handily estimate the chorus size of the pack. We
demonstrated that the process of recording and analys-
ing wolvesreplies collected during the howling survey
has several advantages. Once recorded, a sound is avail-
able for a potentially infinite number of listening ses-
sions. Moreover, in order to improve the quality of the
sound and, consequently, of the chorus size estimation,
these recordings can also be edited (i.e., filtered, ampli-
fied) and saved in modified versions. For this reason, we
suggest to save the original files for possible future stud-
ies. Also, prerequisite for a good spectrographic analysis
and thus for a good estimation of the chorus size is the
recording quality: an ultra-directional microphone, un-
compressed waveor aifformat and an adequate sam-
pling frequency ( 22,000 or 44,000 Hz) are essential
characteristics for recording the howls correctly.
Recordings of wolvesreplies and bioacoustic analysis
can help determine chorus size, thus allowing to count
different wolves in a chorus in a more precise and less
sensitive way than with the more subjective, extempor-
aneous field estimation by ear.
However, in order to be effectively used, the applica-
tion of the methodology for regular monitoring of wolf
populations requires a preliminary training and, for this
reason, researchers will have to instruct and guide
Fig. 4 Comparison between estimated chorus size bioacoustically predicted from two different operators and between aural and bioacoustic.
Scatter plots showing comparison between bioacoustic estimation performed by two operators and aural and bioacoustics estimations of 37
chorus. Bioacoustic estimations by visual inspection of independent operators (Panel a) were highly correlated. Aural and bioacoustic estimations
(Panel b) were poorly concordant. Radiuses are proportional to the number of cases. The diagonal represents the 1:1 correlation
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 5 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
managers in the use of the technique described. A bio-
acoustic approach to the howling survey can help wild-
life and conservation biologists monitor wolves as well
as other canid species with similar acoustic territorial
marking behaviour, such as the coyote (Canis latrans),
and the golden jackal (Canis aureus), offering a realistic,
objective (because based on the spectral components of
the choruses) estimation of any pack size of up to six-
eight members and thus, the possibility to track this
trend.
Since this source of information can be combined with
other sources (e.g., genetic samples [50], snow tracks
[34], remote photo-videos [51]) in a cross-modal moni-
toring, we believe that chorus size estimations by means
of bioacoustic analysis can help biologists monitor wild
populations of vocal animals, through non-invasive
methods and properly inform administrators on the con-
servation strategies required, on the basis of a realistic
estimate of the status and trends of these predators.
Materials and methods
Tests procedure
We evaluated the reliability of the chorus size estimation
by visual inspection of spectrogram and spectrum. First,
we measured the accuracy of the estimated chorus size.
During a howling survey there is generally no visual access
to the replying pack. We thus analysed bioacoustically
choruses of known size, either simulated by humans (i.e.
Human simulated howlingtest, HSH; n= 20), or of
real wolves (i.e. Wolf downloaded howlingtests,
WDH; n= 9). Second, we estimated the precision of the
technique by comparing bioacoustically estimated values
of 37 free-ranging wolveschoral replies from two inde-
pendent operators. A training on bioacoustic analysis fol-
lowing the procedure highlighted in this paper was
performed by operators before the tests.
Finally, we tested whether aural estimations and esti-
mations from visual inspection of spectrogram and
spectrum provided comparable results. Using the same
37 wolvesreplies we compared aural chorus size estima-
tions obtained in the field during the howling survey
(not necessarily by the same field operators) with the
bioacoustics estimations of the same choruses.
Data acquisition
Human-simulated howls (HSH) were recorded in sum-
mer 2012 and 2013, by groups of 2 to 8 volunteers who
were asked to howl together after training on howling
simulation. Breaking and flat howls were alternated, with
at least 5 howls per trial, and one individual entering in
the chorus as a staggered basis, following Harrington
and Mech [14]. Human howls were recorded in Fonte
del Baregno (43°62N, 11°93E), within the protected
area called Alpe di Catenaia in the Apennine Mountains,
in the North-East of Tuscany, Italy. Distance between
source and recorder was 100 m.
Choruses from the internet (Wolf Downloaded Howls,
WDH) were downloaded from YouTube as video file
(.flv) format with VSO downloader 2.9.12 [52] after a re-
search with keywords such as wolf,howlsor similar
terms. We selected and downloaded 9 videos in which it
was effectively possible to count the howling wolves and
all the howling wolves were well visible. So as to be sure
that the chorus size corresponded to the pack size, 8 out
of the 9 choruses used in the WDH were recorded in
captivity (especially in the zoo). The videos were then
converted from the original video format (MP4, .mov
or .flvtypes) into audio format (2 channels, Wave for-
mat, 44,100 KHz and 16 bit format) with the software
4Free Video Converter [53] . All the links to the original
files are in Additional file 3.
Free-ranging wolvesreplies were collected from 2008
to 2014 during a wolf howling monitoring program (fol-
lowing the Habitat Directive on priority species [92/43/
EEC]) carried out in the Province of Arezzo (3230 km2),
Eastern Tuscany, Italy.
Wolf howling survey was performed in summer (from
July to October), when the pack activity was focused in
the home-sites, because of the pups presence and the rate
of response was consequently higher [14, 15, 54]. Sam-
pling sites were chosen so as to cover the whole study
area, following the method described by Harrington and
Mech [15] as saturation censusand adapting it to local
requirements/topography to maximise the range of audi-
bility and minimise sound dispersion [54]. Following the
standard procedure suggested by Harrington and Mech
[15]: i) no session was conducted during rainfalls nor with
strong wind; ii) wolf howling was performed overnight, to
minimise the anthropogenic noise; iii) two trials were con-
ducted per site.
Wolves respond to the howling of unfamiliar individ-
uals in six different basic ways, from retreating silently
to remaining and replying with/by vocal approach [14]
in relation to their resources (e.g. fresh prey), and social
context (e.g. presence of pups) and to the stimulus [16];
moreover, they respond to human simulated howling as
well as to playbacks [13, 14]. For these reasons, our
stimuli always consisted in a chorus howls emitted by
two individuals (howling playback by a captive pair of
wolves (duration: 1.20 min) or by human simulated
howling (duration: circa 1 min)). Playback of recorded
chorus howls was emitted by an exponential horn with
high emission directionality (120° horizontal coverage
and 60° vertical).
Three minutes after the first stimulus, if no answer
had followed, a second trial (higher in volume to cover a
bigger area) was attempted, after which the operators
left the site. In case of response, reply bearing, times and
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 6 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
an extemporaneous estimation of the chorus size by ear
(from the operator without headphone and microphone)
were recorded for each answer. For a better localization
of the pack we repeated one or more trials from a place
closer to the presumed site of response or concurrent
sessions were performed by two groups of operators.
Real pack sizes were unknown for the free-ranging
wolves.
Humansand free-ranging wolveshowls were cap-
tured with a Sennheiser directional microphone fitted
with a windshield (ME67 head with K6 power module
frequency response: 5020,000 Hz) and saved on a
hand-held M-Audio Microtrack 24/96 II digital recorder,
in uncompressed Wave format with a 44,100 Hz sam-
pling rate and 16 bits amplitude resolution.
Bioacoustic analysis
Acoustic signals were analysed with Raven pro 1.4, de-
veloped by Cornel Lab of Ornithology) [47], and with
the open source Seewave package [49] in R v. 2.9.0 [48]
to implement the spectral view.
To estimate chorus size by visual inspections, spectro-
grams and spectral envelopes were computed for each
audio file (Figs. 2 and 3). Spectral envelope (or spectrum)
represents the sound at a given instant, showing the fre-
quencies on the horizontal axis and the sound pressure
(or amplitude) on the vertical axis [47, 55, 56]. The spec-
trogram of a sound represents instead a sequence of
spectra, showing time on the horizontal axis, frequency
on the vertical axis, and the sound pressure as a
greyscale or different colour scale (Fig. 3) [47, 55, 56].
Spectrogram and spectrum are based on the mathemat-
ical function Fourier transform [47, 55, 56], and the ver-
sion of this function which is used to represent
digitalised/discrete signals is called discrete Fourier
transform (DFT) [47]. DFT size represents the length of
the analysis window (the window size), and thus the
number of frames sampled to compute each spectrum of
the spectrogram, while the window function (i.e., Han-
ning, Gaussian) determines how to taper the abruptness
of the onset and offset of a segment [56]. A narrow-
band spectrogram (high window size values) results in a
spectrogram with frequencies which clearly differ from
one another. To analyse wolves choruses, parameters
were set as follows: DFT size: 2048 samples; Hanning
window; frequency grid: 21.5 Hz; time step: 10 ms,
where frequency grid = (sampling frequency)/DFT size,
while time step was taken to be the distance between
the centre of subsequent samples. In case of noise in the
recordings (anthropogenic: cars, planes, high music from
villages, human voices, bells; natural: wind, rivers, other
animals), a band-pass flat filter (1002000 Hz) was ap-
plied to delete noise and thus to improve the audibility
of wolvesreplies.
Every single howl emitted by a wolf appears as a fun-
damental frequency (F0) and its harmonic overtones, or
harmonics (Fig. 2). The fundamental frequency is the
glottal pulse rate and determines the pitch of the voice
[56], while harmonic overtones are integer multiples of
the fundamental frequency (F0*2; F0*3;, F0*N) [56].
Chorus size was then estimated by counting the num-
ber of different howls (viewed as the fundamental plus
harmonics) visualised at the same time (Fig. 2)), assum-
ing that one wolf cannot produce multiple fundamental
frequencies at a given time. Harmonic overtones were
easily recognised because of their shape (same as F0)
and frequencies (integer multiples of F0); since the dif-
ference between howls (inter-harmonic space) doubles
from the fundamental frequencies to the first har-
monics, they can also help recognise different howls
(Fig. 2).
Statistical analysis
All statistical analyses were carried out using R v. 2.9.0
[48]. Spearmans rank correlations were computed to
compare the real and bioacoustically predicted chorus
size, the bioacoustic estimations of the chorus sizes per-
formed by two different operators and the bioacoustic
and aural estimation of the chorus size. Hypotheses of
no mean differences between the real and bioacoustically
predicted chorus size, the bioacoustic estimations of the
chorus sizes performed by two different operators and
the bioacoustic and aural estimation of the chorus size
were tested by Wilcoxon matched-pairs signed-ranks
test.
Additional files
Additional file 1: R Script for wolf choruses analysis with Seewave
package in R environment. (R 1 kb)
Additional file 2: Audio file of free-ranging wolf howling.
(WAV 7962 kb)
Additional file 3: Links to the videos used for WDH test.
(DOCX 12 kb)
Competing interests
The authors declare that they have no competing interests.
Authorscontributions
Conceived and designed the experiments: DP, LM, MA. Performed the
experiments: DP, LM, EB, LS. Analysed the data: DP, LM, EB, LS. Wrote the
paper: DP, LM, EB, MA. All authors read and approved the final manuscript.
Acknowledgements
We want to thank Regione Toscana and Arezzo Province for logistical and
financial support. We are also grateful to Andrea Gazzola, Emanuela
Donaggio for advice and for helping during the fieldwork. We are grateful to
two anonymous referees for providing useful comments. We also thanks
Foreste Casentinesi National Park Rangers for helping in data collection and
all students and volunteers involved in this study.
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 7 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Funding
This work has been supported by Regione Toscana, University of Sassari and
by the Provincial Government of Arezzo (Italy). The funders had no role in
study design, data collection and analysis, decision to publish, or preparation
of the manuscript.
Author details
1
Department of Science for Nature and Environmental Resources, University
of Sassari, via Muroni 25, I-07100 Sassari, Italy.
2
Servizio Piano Faunistico,
Provincia di Arezzo, Piazza della Libertà 3, I57100 Arezzo, Italy.
3
Szent István
University, Institute for Wildlife Conservation, Páter K str 1, Gödöllő, Hungary.
Received: 17 March 2015 Accepted: 1 September 2015
References
1. Mladenoff DJ, Sickley TA, Wydeven AP. Predicting gray wolf landscape
recolonization: logistic regression models vs. new field data. Ecol Appl.
1999;9:3744.
2. Dickman AJ, Macdonald EA, Macdonald DW. A review of financial
instruments to pay for predator conservation and encourage human-
carnivore coexistence. Proc Natl Acad Sci. 2011;108:198366.
3. Primack RB. Essentials of conservation biology. Fifth: Edition; 2010.
4. Ripple WJ, Estes JA, Beschta RL, Wilmers CC, Ritchie EG, Hebblewhite M,
et al. Status and ecological effects of the worlds largest carnivores. Science.
2014;343:1241484.
5. Walpole MJ, Leader-Williams N. Tourism and flagship species in
conservation. Biodivers Conserv. 2002;11:5437.
6. Kleiven J, Bjerke T, Kaltenborn BP. Factors influencing the social acceptability
of large carnivore behaviours. Biodivers Conserv. 2004;13:164758.
7. Berger KM. Carnivore-Livestock conflicts: effects of subsidized predator
control and economic correlates on the sheep industry. Conserv Biol.
2006;20:75161.
8. Promberger C, Schroder W, editors. Wolves in Europe: status and
perspectives. Ettal, Germany: Munich Wildlife Society; 1993.
9. Breitenmoser U. Large predators in the Alps: the fall and rise of mans
competitors. Biol Conserv. 1998;83:27989.
10. Chapron G, Kaczensky P, Linnell JDC, von Arx M, Huber D, Andrén H, et al.
Recovery of large carnivores in Europes modern human-dominated
landscapes. Science. 2014;346:15179. doi:10.1126/science.1257553.
11. Ripple WJ, Beschta RL. Trophic cascades in Yellowstone: The first 15 years
after wolf reintroduction. Biol Conserv. 2012;145:20513. doi:10.1016/
j.biocon.2011.11.005.
12. Fritts SH, Bangs EE, Gore JF. The relationship of wolf recovery to habitat
conservation and biodiversity in the northwestern United States. Landsc
Urban Plan. 1994;28:2332. doi:10.1016/0169-2046(94)90040-X.
13. Pimlott DH. The use of tape-recorded wolf howls to locate timber wolves.
Toronto: Twenty-second Midwest Wildlife Congress; 1960. p. 15.
14. Harrington FH, Mech LD. Wolf howling and its role in territory maintenance.
Behaviour. 1979;68:20748.
15. Harrington FH, Mech LD. An analysis of howling response parameters useful
for wolf pack censusing. J Wildl Manag. 1982;46:68693.
16. Harrington FH. Aggressive howling in wolves. Anim Behav. 1987;35:712.
17. Fuller TK, Sampson BA. Evaluation of a simulated howling survey for wolves.
J Widl Manag. 1988;52:603.
18. Nowak S, Jędrzejewski W, Schmidt K, Theuerkauf J, Mysłajek RW,
Jędrzejewska B. Howling activity of free-ranging wolves (Canis lupus) in the
Białowieża Primeval Forest and the Western Beskidy Mountains (Poland).
J Ethol. 2006;25:2317.
19. Blanco JC, Yolanda C. Surveying wolves without snow: a critical review of
the methods used in Spain. Hystrix. Ital J Mammal. 2012;23:3548.
20. Szabó L, Heltai M, Szucs E, Lanszki J, Lehoczki R. Expansion range of the
golden jackal in Hungary between 1997 and 2006. Mammalia. 2009;73:30711.
doi:10.1515/MAMM.2009.048.
21. Gese EM, Ruff RL. Howling by coyotes (Canis latrans): variation among social
classes, seasons, and pack sizes. Can J Zool. 1998;76:103743.
22. Harrington FH. Chorus howling by wolves: Acoustic structures, pack size
and Beau Geste effect. Bioacoustics. 1989;2:11736.
23. Dugnol B, Fernández C, Galiano G. Wolf population counting by
spectrogram image processing. Appl Math Comput. 2007;186:82030.
24. Dugnol B, Fernández C, Galiano G, Velasco J. Implementation of a diffusive
differential reassignment method for signal enhancement: An application to
wolf population counting. Appl Math Comput. 2007;193:37484.
25. Dugnol B, Fernández C, Galiano G, Velasco J. On a chirplet transform-based
method applied to separating and counting wolf howls. Signal Process.
2008;88:181726.
26. Passilongo D, Dessi-Fulgheri F, Gazzola A, Zaccaroni M, Apollonio M. Wolf
counting and individual acoustic discrimination by spectrographic analysis
[Abstract]. Bioacoustics. 2012;21:789.
27. Tooze ZJ, Harrington FH, Fentress JC. Individually distinct vocalizations in
timber wolves, Canis lupus. Anim Behav. 1990;40:72330.
28. Palacios V, Font E, Marquez R. Iberian wolf howls: acoustic structure,
individual variation, and a comparison with North American populations.
J Mammal. 2007;88:60613.
29. Zaccaroni M, Passilongo D, Buccianti A, Dessi-Fulgheri F, Facchini C, Gazzola
A, et al. Group specific vocal signature in free- ranging wolf packs. Ethol
Ecol Evol. 2012;24:32231.
30. Joslin PWB. Movements and home sites of timber wolves in Algonquin Park.
Am Zool. 1967;7:27988.
31. Huron D. Voice Denumerability of Homogeneous Timbres. Music Percept
Interdiscip J. 2010;6:36182.
32. Stöter F-R, Schoeffler M, Edler B, Herre J. Human ability of counting the
number of instruments in polyphonic music. Proc Meetings Acoust.
2013;19:0350344.
33. Boitani L. Patterns of homesites attendance in two Minnesota wolf packs. In:
Harrington FH, Paquet PC, editors. Wolves of the World: Perspectives of
Behavior, Ecology and Conservation. New York: Noyes, Park Ridge; 1982. p.
15872.
34. Apollonio M, Mattioli L, Scandura M, Mauri L, Gazzola A, Avanzinelli E.
Wolves in the Casentinesi Forests: insights for wolf conservation in Italy
from a protected area with a rich wild prey community. Biol Conserv.
2004;120:24960.
35. Stenglein JL, Waits LP, Ausband DE, Zager P, Mack CM. Estimating gray wolf
pack size and family relationships using non invasive genetic sampling at
rendezvous sites. J Mammal. 2011;92:78495.
36. Burch JW, Layne GA, Follmann EH, Rexstad EA. Evaluation of Wolf
Density Estimation from Radiotelemetry Data. Wildl Soc Bull.
2005;33(4):122536.
37. Iliopoulos Y, Youlatos D, Sgardelis S. Wolf pack rendezvous site selection in
Greece is mainly affected by anthropogenic landscape features. Eur J Wildl
Res. 2013;60:2334.
38. Galaverni M, Palumbo D, Fabbri E, Caniglia R, Greco C, Randi E. Monitoring
wolves (Canis lupus) by non-invasive genetics and camera trapping: A small-
scale pilot study. Eur J Wildl Res. 2012;58:4758. doi:10.1007/s10344-011-0539 -5.
39. Mech LD. The Wolf: The Ecology and Behaviour of an Endangered Species.
New York: The Natural History Press, Garden City; 1970.
40. Jedrzejewski W, Schmidt K, Theuerkauf J, Jedrzejewska B, Selva N, Zub K,
et al. Kill rate and predation by wolves on ungulate populations in
Bialowieza primeval forest (Poland). Ecology. 2002;83:134156.
41. Shmidt Mech LD. Wolf pack size and food acquisition. Am Nat.
1997;150(4):5137.
42. Peters G, Tembrock G. Subharmonics, biphonation, and deterministic chaos
in mammal vocalizations. Bioacoustics. 1998;9:17196.
43. Riede T, Herzel H, Mehwald D, Seidner W, Trumler E, Böhme G. Nonlinear
phenomena in the natural howling of a dog-wolf mix. J Acoust Soc Am.
2000;108:143542.
44. Richards DG, Wiley RH. Reverberations and Amplitude Fluctuations in the
Propagation of Sound in a Forest: Implications for Animal Communication.
Am Nat. 2008;115:38199.
45. Harris F. On the Use of Windows for Harmonic Analysis with the Discrete
Fourier Transform. Proc IEEE. 1978;66:5183.
46. Herbst CT, Herzel H, Svec JG, Wyman MT, Fitch WT. Visualization of system
dynamics using phasegrams. J R Soc Interface. 2013;10:20130288.
47. Charif RA, Waack AM, Strickman LM. Raven Pro 1.3 Users Manual. Ithaca,
New York: Cornell Laboratory of Ornithology; 2008.
48. Core Development Team R. R: a language and environment for statistical
computing. Vienna, Austria: R foundation for statistical computing; 2011.
49. Sueur J, Aubin T, Simonis C. Seewave: a free modular tool for sound analysis
and synthesis. Bioacoustics. 2008;18:21326.
50. Marescot L, Pradel R, Duchamp C, Cubaynes S, Mrboutin E, Choquet R, et al.
Capture recapture population growth rate as a robust tool against
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 8 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
detection heterogeneity for population management. Ecol Appl.
2011;21:2898907. doi:10.1890/10-2321.1.
51. Morgan TW, Elliott CL. Comparison of remotely-triggered cameras vs.
howling surveys for estimating coyote (Canis latrans) Abundance in central
Kentucky. J Ky Acad Science. 2011;72(2):8490.
52. VSO Downloader. Copyright VSO Software SARL. VSO-Software SARL Bat A,
1er Etage,109 Avenue de Lespinet, 31400 Toulouse.
53. 4Free Video Converter. 4 Free Studio. Copyright© 2000~2015 4Free Video
Converter Inc. a Multimedia Utility Company.
54. Gazzola A, Avanzinelli E, Mauri L, Scandura M, Apollonio M. Temporal
changes of howling in south European wolf packs. Ital J Zool. 2002;69:15761.
55. Boersma P, Weenink D. Praat: doing phonetics by computer. 2009.
http://www.praat.org.
56. Hopp S, Owren MJ, Evans CS, editors. Animal Acoustic Communication:
Sound Analysis and Research Methods. Berlin: Springer; 1998.
Submit your next manuscript to BioMed Central
and take full advantage of:
Convenient online submission
Thorough peer review
No space constraints or color figure charges
Immediate publication on acceptance
Inclusion in PubMed, CAS, Scopus and Google Scholar
Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
Passilongo et al. Frontiers in Zoology (2015) 12:22 Page 9 of 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
... This area presumably acts as a source for the expansion of the species to the Northern Apennine and the Alpine region [32,39,40]. To estimate the minimum number of individuals and packs' number and location in the region, we combined field data, obtained from wolf-howling [41][42][43] and camera trapping [36,44], with molecular data derived from the microsatellite genotyping of both invasively and non-invasively collected samples [37,44]. Additionally, we performed a Population Viability Analysis (PVA [45]) to build demographic models to better estimate the actual population size, reconstruct previous trends, and predict future population trends. ...
... Summer wolf-howling surveys were conducted annually between July and October (following Gazzola et al. [41] and Passilongo et al. [42]), but we used an opportunistic approach by focusing on the previously identified or putative packs' homesites. The sonographic analysis of chorus howls provided information on the number of packs and reproduction events [42,64]. ...
... Summer wolf-howling surveys were conducted annually between July and October (following Gazzola et al. [41] and Passilongo et al. [42]), but we used an opportunistic approach by focusing on the previously identified or putative packs' homesites. The sonographic analysis of chorus howls provided information on the number of packs and reproduction events [42,64]. ...
Article
Full-text available
Simple Summary Wildlife conservation effectiveness relies on the accuracy of population data used to build management plans. Estimating population size and dynamics of large vertebrates living on wide areas is a challenging task and results can vary depending on the methods adopted. In the present study, the size of a wolf population was estimated both by field monitoring and statistical modeling on demographic data. Results showed discrepancies with respect to previous bibliographic estimates which resulted significantly lower. Moreover, statistical techniques on demographic data allowed a deeper knowledge of population’s trend, pointing out the importance of mortality rates in driving population growth. We suggest combining statistical modeling and field monitoring to achieve a satisfactory knowledge of population size and dynamics useful to foster reliable management and conservation plans. Abstract We estimated the current size and dynamics of the wolf population in Tuscany and investigated the trends and demographic drivers of population changes. Estimates were obtained by two different approaches: (i) mixed-technique field monitoring (from 2014 to 2016) that found the minimum observed pack number and estimated population size, and (ii) an individual-based model (run by Vortex software v. 10.3.8.0) with demographic inputs derived from a local intensive study area and historic data on population size. Field monitoring showed a minimum population size of 558 wolves (SE = 12.005) in 2016, with a density of 2.74 individuals/100 km². The population model described an increasing trend with an average annual rate of increase λ = 1.075 (SE = 0.014), an estimated population size of about 882 individuals (SE = 9.397) in 2016, and a density of 4.29 wolves/100 km². Previously published estimates of wolf population were as low as 56.2% compared to our field monitoring estimation and 34.6% in comparison to our model estimation. We conducted sensitivity tests to analyze the key parameters driving population changes based on juvenile and adult mortality rates, female breeding success, and litter size. Mortality rates played a major role in determining intrinsic growth rate changes, with adult mortality accounting for 62.5% of the total variance explained by the four parameters. Juvenile mortality was responsible for 35.8% of the variance, while female breeding success and litter size had weak or negligible effects. We concluded that reliable estimates of population abundance and a deeper understanding of the role of different demographic parameters in determining population dynamics are crucial to define and carry out appropriate conservation and management strategies to address human–wildlife conflicts.
... For more information of this approach in AMT, readers could refer to [51,50]. [45]. Analyzing the wolf howling signal is an efficient approach to evaluate how many wolves are there in the field under survey. ...
Preprint
We propose to combine cepstrum and nonlinear time-frequency (TF) analysis to study mutiple component oscillatory signals with time-varying frequency and amplitude and with time-varying non-sinusoidal oscillatory pattern. The concept of cepstrum is applied to eliminate the wave-shape function influence on the TF analysis, and we propose a new algorithm, named de-shape synchrosqueezing transform (de-shape SST). The mathematical model, adaptive non-harmonic model, is introduced and the de-shape SST algorithm is theoretically analyzed. In addition to simulated signals, several different physiological, musical and biological signals are analyzed to illustrate the proposed algorithm.
... Yet far more information that species' distributions may be attainable. Efficient identification of wolf or coyote vocalizations can facilitate estimates of group size (Passilongo et al. 2015) and may be helpful in identifying individuals (Larsen et al. 2022). Because wolf and coyote pups make distinctive sounds, researchers could also use PAM to identify demographically critical events like successful reproduction. ...
Article
Passive acoustic monitoring has emerged as a scalable, noninvasive tool for monitoring many acoustically active animals. Bioacoustics has long been employed to study wolves and coyotes, but the process of extracting relevant signals (e.g., territorial vocalizations) from large audio datasets remains a substantial limitation. The BirdNET algorithm is a machine learning tool originally designed to identify birds by sound, but it was recently expanded to include gray wolves (Canis lupus) and coyotes (C. latrans). We used BirdNET to analyze 10,500 h of passively recorded audio from the northern Sierra Nevada, USA, in which both species are known to occur. For wolves, real-world precision was low, but recall was high; careful post-processing of results may be necessary for an efficient workflow. For coyotes, recall and precision were high. BirdNET enabled us to identify wolves, coyotes, and apparent intra- and interspecific acoustic interactions. Because BirdNET is freely available and requires no computer science expertise to use, it may facilitate the application of passive acoustic surveys to the research and management of wolves and coyotes, two species with continental distributions that are frequently involved in high-profile and sometimes contention management decisions.
... One could also push automation even further by having the algorithm filter out more accurate information, as suggested by (Passilongo et al. 2015): number of individuals, classification if juveniles are howling or not, identification of individuals. ...
Article
Full-text available
Wolves are spreading in the Alpine region at an increasing rate, which leads to human--wolf conflicts. In order to reduce those and to perform an active wolf management, solid information about the presence of wolves is required. Getting this information is challenging since wolves are nocturnal, have sharp senses, and large territories. A monitoring method is to detect wolf howling, which can be heard over several kilometres and therefore simplifies finding them. Current acoustic methods, however, are very labour-intensive as a memory card has to be fetched from the device in the field and then the recordings have to be checked for howling manually. We present a novel approach to acoustic wolf monitoring using a convolutional neural network that runs on an embedded system in the wolf territory. Thus, we obtain accurate real-time information about the presence of wolves. On our data set, we achieve an F1-score of 0.61, thus outperforming previous systems by far. We develop prototypes and conduct two field test: first in a zoo, where we even achieve an F1-score of 0.8, and then in a wolf territory, where we successfully detect wolf howling.
... New technical tools such as acoustic sensors (e.g., the "Songmeter") allow the systematic recording of spontaneous howling. Although the effort for fieldwork is considerably reduced since the acoustic sensors automatically record the sounds (Passilongo et al. 2015;, hours of sound recordings must then be analysed in order to find the wolf howls. However, considering the increasing automation of these types of data analysis (e.g., Stowell and Plumbley 2014), this method has a promising future . ...
Technical Report
Full-text available
Monitoring standards and strategy to optimise the integrated monitoring of the status of the wolf alpine population. On behalf of the LIFE WolfAlps project
... We now show a real world data analysis. An important issue in conservation biology is to estimate the number of wolves in the field [33,17]. In this final example we show the numerical result with a wolf howling signal. ...
Article
Analysis of signals with oscillatory modes with crossover instantaneous frequencies is a challenging problem in time series analysis. One way to handle this problem is lifting the 2-dimensional time-frequency representation to a 3-dimensional representation, called time-frequency-chirp rate (TFC) representation, by adding one extra chirp rate parameter so that crossover frequencies are disentangled in higher dimension. The chirplet transform is an algorithm for this lifting idea, which leads to a TFC representation. However, in practice, we found that it has a strong “blurring” effect in the chirp rate axis, which limits its application in real-world data. Moreover, to our knowledge, we have limited mathematical understanding of the chirplet transform in the literature. Motivated by the need for the real-world data analysis, in this paper, we propose the synchrosqueezed chirplet transform (SCT) that enhances the TFC representation given by the chirplet transform. The resulting concentrated TFC representation has high contrast so that one can better distinguish different modes with crossover instantaneous frequencies. The basic idea is to use the phase information in the chirplet transform to determine a reassignment rule that sharpens the TFC representation determined by the chirplet transform. We also analyze the chirplet transform and provide theoretical guarantees of SCT.
... With low quality recordings and low-cost userfriendly software, it is also possible to identify wolves, whereas more costly software such as MATLAB, may limit the access for scientists as funding and knowledge of programming is necessary to use such software [55]. Further, different studies use different software [31,40,42,43], which makes comparison of data and replication difficult [55]. ...
Article
Full-text available
Simple Summary: This study evaluates the use of acoustic devices as a method to monitor wolves by analyzing different variables extracted from wolf howls. By analyzing the wolf howls, we fo-cused on identifying individual wolves, subspecies. We analyzed 170 howls from 16 individuals from the three subspecies: Arctic wolves (Canis lupus arctos), Eurasian wolves (C.l. lupus), and Northwestern wolves (C.l. occidentalis). We assessed the potential for individual recognition and recognition of three subspecies: Arctic, Eurasian, and Northwestern wolves. Abstract: Wolves (Canis lupus) are generally monitored by visual observations, camera traps, and DNA traces. In this study, we evaluated acoustic monitoring of wolf howls as a method for monitoring wolves, which may permit detection of wolves across longer distances than that permitted by camera traps. We analyzed acoustic data of wolves' howls collected from both wild and captive ones. The analysis focused on individual and subspecies recognition. Furthermore, we aimed to determine the usefulness of acoustic monitoring in the field given the limited data for Eurasian wolves. We analyzed 170 howls from 16 individual wolves from 3 subspecies: Arctic (Canis lupus arctos), Eurasian (C. l. lupus), and Northwestern wolves (C. l. occidentalis). Variables from the fundamental frequency (f0) (lowest frequency band of a sound signal) were extracted and used in discri-minant analysis, classification matrix, and pairwise post-hoc Hotelling test. The results indicated that Arctic and Eurasian wolves had subspecies identifiable calls, while Northwestern wolves did not, though this sample size was small. Identification on an individual level was successful for all subspecies. Individuals were correctly classified with 80%-100% accuracy, using discriminant function analysis. Our findings suggest acoustic monitoring could be a valuable and cost-effective tool that complements camera traps, by improving long-distance detection of wolves.
... Furthermore, wolf populations are often monitored through the use of howl surveys (Joslin 1967;Harrington and Mech 1983;Fuller and Sampson 1988;Thiel et al. 2009), mainly to locate packs (Harrington and Mech 1983;Fuller and Sampson 1988), detect successful reproduction (presence of pups, Palacios et al. 2016), and estimate the average number of wolves that participate in a chorus (Palacios et al. 2017;Passilongo et al. 2015). We will discuss how our results could be applied to improve the reliability of wolf surveys based on acoustic signals. ...
Article
Wolf packs perform group vocalizations called chorus howls. These acoustic signals have a complex structure and could be involved in functions such as strengthening of social bonds, territory advertisement, or spacing between packs. We analyzed video recordings of 46 chorus howls emitted by 10 packs of wolves held in captivity, in order to investigate whether sex, age, social status, pack, or individual influence the way wolves participate in a chorus. We found that, during a chorus, wolves vocalized 63% of the time, with the howl being the most common vocalization (36% of the chorus duration), followed by woa (13.5%), other vocalizations (11.8%), and bark (1.7%). The main factor affecting the vocal behavior of wolves was age, since young wolves vocalized less and uttered shorter acoustic signals than adults. The discriminant analysis carried out with the wolves of Cañada Real pack assigned 89.3% of the cases to the correct individual, which is much better than the assignment expected by chance, suggesting that individuals could have a unique vocal usage during a chorus howl, mainly due to the use of howls and woa-woa howls. Based on our results, we propose that in the context of a chorus the woa-woa howl is important, although further research is needed to address this issue properly.
Article
Full-text available
After a long period of human persecution that led it to extinction in most of its distribution range, the wolf is undergoing a fast recovery. Despite being described as an elusive species only living in remote areas, wolves are recently occupying also human‐dominated landscapes, increasing the frequency of direct contacts with humans. Nevertheless, it is unclear whether this situation is only caused by a numerical increase or partially facilitated by a higher tolerance of wolves with respect to human proximity. We focused on a European region offering an abundant, widespread and long‐term monitored wolf population to analyse wolf pack distribution and its relationship with human presence across areas recolonized across four different time periods (1972, 1996, 2005 and 2016). Throughout areas recolonized during different periods, wolves were initially located in mountains and hills, occupying plains only in the recent past. Although they consistently tended to be located as far as possible from urban settlements, especially from those with over 5000 inhabitants, in 2016, 70% of packs included at least one urban settlement within the expected home range. Moreover, the distance kept by wolves from the nearest urban settlement was more constrained in areas of recent recolonization (2005 and 2016) and in the mountainous altitudinal range, likely due to a reduced availability of territories. We showed that wolves tend to keep as far away as possible from humans, but they can also permanently occupy human‐dominated landscapes to cope with the lower availability of space induced by their remarkably successful recolonization. Our results shed light on an upcoming scenario for the conservation of large carnivores.
Article
Full-text available
Parent-offspring interactions are essential to interpret animal social evolution and behavior, but their role in mediating acoustic communication in animals that interact vocally is still unclear. Increasing evidence shows that primate vocal communication is way more flexible than previously assumed, and research on this topic can provide further information on how the social environment shaped vocal plasticity during the evolution of the Primate order. Indris communicate through elaborated vocal emissions, usually termed songs. Songs are interactive vocal displays in which all members of the family group alternate their emissions, taking turns during chorusing events. We aimed to understand whether specific rules regulate the turn-taking of different group members and investigate the flexibility of indris’ vocal behavior when co-singing with their offspring. We found that social factors can influence the turn-taking organization in a chorus, as offspring were more likely to drop out from the parents’ duet than join in, and we speculate that overlap might signal competition by members of the same-sex. The duet between the reproductive pair was the most common type of singing organization, followed by a duet between mothers and sons and the triadic interaction between mother, father, and son. Interestingly, parents’ solo singing seems to stimulate offspring to vocalize, and we also found that mothers and fathers simplify, at least in part, song elaboration when chorusing with offspring. Our results indicate that indris can perform short-time adjustments to the number of co-emitters and their identity: our approach is advantageous in highlighting the multilevel influences on primate vocal flexibility. Moreover, it provides evidence that some aspects of our vocal plasticity were already present in the lemur lineage.
Article
Full-text available
Wolves (Canis lupus) are difficult to survey, and in Most countries, snow is used for identifying the species, counting individuals, recording movements and determining social position. However, in the Iberian peninsula and other southern regions of its gobal range, snow is very scarce in winter, so wolves must be surveyed without snow. In Spain and Portugal, wolves are surveyed through estimating number of wolf packs in summer by means of locating litters of pups when they are at rendezvous sites. Packs are confirmed when pups are observed or respond to simulated howling. We make a critical review of this method, exploring the sources of error when estimating number of packs, the constraints of the simulated howling method, the sources of uncertainty caused by variations in effort, in observer experience and in other variables. We stress the difficulty of assessing average pack size and percentage of wolves not included in packs (pairs and solitary wolves), which can exceed 30% of the population. These restrictions make this method unaccurate and unable to detect moderate or even large population size variations. At the same time, indices based on abundance of wolf tracks and scats is hampered by the lack of snow and the problem to distinguish them from those of dogs. We conclude that accepting the limitations of these wolf surveys and highlighting the uncertainty of the figures they provide is more realistic and will encourage a more prudent approach to wolf management.
Article
Full-text available
The Casentinesi Forests, in the northern Apennines, harbour a rich community of wild ungulates, with the wolf representing the largest predator in the area. Between 1993 and 2000, wolf pack distribution in the area was monitored and estimates of pack size were obtained by wolf-howling surveys, snow-tracking, and occasional observations. Three to five packs were detected yearly, with sizes averaging 4.2 � 0.9 wolves (maximum 7). The overall density in the area was 4.7 wolves per 100 km 2 with an average distance between adjacent packs of 11.1 km. The high wolf density in the Casentinesi Forests is mostly related to abundance and size of wild prey. In this, like in other areas at low latitudes, wolf density depends mainly on the number of packs, as pack size is rather small and recruitment limited by early dispersal and high mortality. Three homesites used in several years by resident packs were discovered. Homesite fidelity and pack reproductive success were higher in fully protected rather than harvested areas. Establishing a network of protected areas with high ungulate diversity and abundance is proposed as the main factor for allowing a full recovery of the wolf population in Italy. 2004 Elsevier Ltd. All rights reserved.
Article
Full-text available
We review Seewave, new software for analysing and synthesizing sounds. Seewave is free and works on a wide variety of operating systems as an extension of the R operating environment. Its current 67 functions allow the user to achieve time, amplitude and frequency analyses, to estimate quantitative differences between sounds, and to generate new sounds for playback experiments. Thanks to its implementation in the R environment, Seewave is fully modular. All functions can be combined for complex data acquisition and graphical output, they can be part of important scripts for batch processing and they can be modified ad libitum. New functions can also be written, making Seewave a truly open-source tool.
Article
Full-text available
Wolf (Canis lupus) packs were surveyed using simulated howling survey (Harrington and Mech 1982) in a 1,400km21,400-\text{km}^{2} area of northcentral Minnesota with a known density of radio-marked packs. Mean distance that wolf howls could be heard was 2 km (n = 74 trials). Howling at 55 sites indicated the presence of 5 packs and resulted in a density estimate (7.5 packs/1,000 km2) that was imprecise (95% CI = 2.5-16.4/1,000 km2) and 1.7 × greater than the actual density (4.3/1,000 km2). Because of logistical and statistical constraints, the technique is not practical for surveying large (e.g., state- or province-wide) areas, but simulated howling is useful for locating packs in smaller areas.
Book
The last decades have brought a significant increase in research on acoustic communi­ cation in animals. Publication of scientific papers on both empirical and theoretical aspects of this topic has greatly increased, and a new journal, Bioacoustics, is entirely devoted to such articles. Coupled with this proliferation of work is a recognition that many of the current issues are best approached with an interdisciplinary perspective, requiring technical and theoretical contributions from a number of areas of inquiry that have traditionally been separated. With the notable exception of a collection edited by Lewis (1983), there have been fewvolumes predominatelyfocused on technical issues in comparative bioacoustics to follow up the earlyworks edited by Lanyon and Tavolga (1960) and Busnel (1963). It was the tremendous growth of expertise c:()ncerning this topic in particular that provided the initial impetus to organize this volume, which attempts to present fundamental information from both theoretical and applied aspects of current bioacoustics research. While a completely comprehensive review would be impractical, this volume offers a basic treatment of a wide variety of topics aimed at providing a conceptual framework within which researchers can address their own questions. Each presentation is designed to be useful to the broadest possible spectrum of researchers, including both those currently working in any of the many and diverse disciplines of bioacoustics, and others that may be new to such studies.
Article
Gray wolves (Canis lupus) were studied from April 1972 through April 1974 in the Superior National Forest in northeastern Minnesota by radio-tracking and simulated howling. Based on replies during 217 of 456 howling sessions, the following recommendations were derived for using simulated howling as a census technique: (1) the best times of day are dusk and night; (2) July, August, and September are the best months; (3) precipitation and winds greater than 12 km/hour should be avoided; (4) a sequence of 5 single howls should be used, alternating "flat" and "breaking" howls; (5) trials should be repeated 3 times at about 2-minute intervals with the first trial at lower volume; and (6) the trial series should be repeated on 3 nights as close to each other as possible. Two censuses are described: a saturation census and a sampling census.