Conference PaperPDF Available

Analysis of acoustic signals emitted from a nursery school of sperm whale (Physeter macrocephalus) recorded off Ventotene Island (Southern Tyrrhenian Sea, Italy)

Authors:
  • Marine Life Conservation
22
ANALYSIS OF ACOUSTIC SIGNALS EMITTED FROM A HERD OF SPERM WHALES
(PHYSETER MACROCEPHALUS) ENCOUNTERED OFF VENTOTENE ISLAND
(SOUTHERN TYRRHENIAN SEA, ITALY)
M. Azzali1, A. Impetuoso2, B. Mussi3, A. Miragliuolo3, M. Battaglia4, W. Antollovich2, and R. Dernowski1
1CNR -ISMAR Acoustic Repart of the Fishery Division, Largo Fiera della Pesca, 1 60125 Ancona, Italy
2Marine Life Conservation onlus, Via Salgari 9, 10154 Torino, Italy
3Studiomare, Via D’Abundo 82, 80075 Forio DIschia (NA), Italy
4ENEA Gruppo Frascati, via E. Fermi 27, 00044 Roma, Italy
INTRODUCTION The first recordings of underwater sperm whale sounds have been made on 1952
(Watkins, 1980). Half part of a century has spent and a lot of studies have been done on this type of sounds. Four
main categories have been classified (Whitehead and Weilgart, 1990; Goold and Jones, 1995; Jaquet et al., 2001):
(1) regular clicks have an interclick interval (ICI) of about 0,5 to 1 s and are produced for echolocation purposes; (2)
slow clicks have an ICI of about 5 to 7 s and may have several different functions; (3) creaks are series of very rapid
clicks with up to 220 clicks per second and represent short-range echolocation when sperm whales are closing in on
their prey; (4) codas are short, patterned series of clicks with irregular repetition rates and are mainly produced
during social interactions, playing a role in communication.
Among these sounds emitted by sperm whales, the sequences of clicks, named creaks, stand out as conspicuous
signals, indicative of peculiar behaviours. Sequences of clicks were recorded continuously for 17 minutes from a
herd estimated acoustically in 12 sperm whales, included a juvenile, found in the archipelago Pontino – Campano off
the Ventotene island in the October 2002. This area have been intensely studied since 1991, and Sperm Whales,
among other seven species of cetaceans, have been frequently encountered (Airoldi et al., 2003; Impetuoso et al.,
2004). The rhythmic fashion of the sequences emitted from all the members of the group are presented together with
the strips of high noise level, probably produced by the vessel. Some sequences emitted by eight different individuals
are examined in detail. The spectra of the selected sequences are shown and discussed.
MATERIALS AND METHODS Data were collected on board of StudioMare research vessel "Jean Gab",
a 17.70 m wooden cutter equipped for underwater listening with towed hydrophones synchronized with an
underwater camera (Panasonic CCD Camera WV-KS152). However the present study focuses only the acoustic
signals emitted by the whales, while the biological and social context of the click emission is not investigated.
Signals were recorded from an hydrophone (3 32000 Hz) with pre – amplifier into a DAT (Sony TCD D100).
The sperm whale herd, subject of the study, was encountered off the cost of Ventotene the 8th October 2002 at 12:45.
However the recording presented in this study started only at 17:20, when the sperm whales approached the vessel,
and lasted 17 min. At least four whales were clearly sighted including at least one juvenile. It was definitely a
juvenile that got more and more close to the vessel until it broke the hydrophone (17:37). However the number of
whales, counted visually, was scarcely reliable and surely underestimated. From the analysis of the features of the
recorded signals, carried out in laboratory, 12 different individuals were identified. The acoustic recognition of the
whales and the processing of the signals emitted by them was based on Rainbow Click (Vers. 3.00.0002 copyright
IFAW -1998-2002) programme and on a SW packet developed originally by C.N.R. of Ancona on MATLAB
platform. . The 17 min of recording were divided into 51 equal intervals of 20 s (Table 1). The resulting bins of
identical length were then inspected and the sequence of clicks falling into each bin classified on the basis of its
features. In total 41 sequences were analysed, classified and attributed to 12 different whales. Each row in Table 1
contains the series of click sequences emitted by the whale, identified by the number in the first column. Within the
bins of a row are reported the beginning and the end of each sequence (i.e. the sequence length) emitted by the
animal related to that row. The noisy and the empty bins of each row are also shown in Table 1.
RESULTS The Table 1 shows the click sequences emitted by each component of the herd identified
acoustically. The click sequences have different lengths from about 1 to about 14 seconds (average length = 4.34 s;
SD = 3.35 s) and contain from dozens to hundreds of clicks in the frequency range from few up to eight thousand
Hertz. An interesting aspect of the click sequences is their apparent rhythmicity. It seems that there is a sort of
synchronization between the sequences emitted by the different whales of the herd (Table 1). All the animals are
involved, almost cyclically, in producing click sequences as in accordance with a certain formalized repertoire. Only
whales 2, 7 and 9 emitted two consecutive sequences, before receiving apparent answers from other components of
the group. Moreover some animals seem to be more active (whale 2 emitted 7 sequences in 17 min.) than others
(whales 3, 4, 5 and 6 emitted only 2 sequences in 17 min.). In the bins with noise no sequence was detected, but only
few separated short clicks. It seems that the whales were almost silent during the noise intervals.
23
The mean spectra of a selected number of sequences, emitted by eight whales (Nr. 1, 2, 3, 5, 6, 9, 11 and 12) and
indicated in Table 1 by bold type, are shown in Fig. 1. The fundamental frequencies differ in each animal. The
spectra of whales 5, 6 and 11 have only one clear fundamental frequency, at 2.9, 1.9 and 1 kHz respectively; the
other frequencies in the spectra are small in comparison with the dominant ones. The spectra of the whales 1, 2, 3,
and 9 have from two to four peaks, close in amplitude, in the range 2- 4 kHz. The spectrum of the whale 12,
identified visually as juvenile, is characterized by about ten harmonics falling in the range 2- 4 kHz. These
harmonics superimposed on the spectrum of whale 12, as well as the ones visible on the spectra of whales 1, 2, 5, 6,
9 and 11, are probably the effect of pulse repetition frequency (RPF) of the click signals that are contained in the
analyzed sequences. However the actual PRF of whale 12 is very peculiar for the high number and the intensity of
the harmonics that appear in the spectrum. It may be related to the high interest shown by the juvenile towards the
hydrophone, as described above.
CONCLUSIONS The investigation on the click signals produced by a herd of sperm whales, included at
least one juvenile, in 17 minutes of continuous recording indicate that:
The fundamental frequencies in the spectra differ in each animal. This has allowed identification of the number (12)
of animals in the herd that is greater than the one deduced visually (at least 4 individuals).
The spectra of 11 whales are characterised by few clear peaks falling at frequencies different in each animal.
The spectra of the juvenile are characterised by many harmonics, superimposed on a spectrum that seems to have no
fundamental frequency. These harmonics, probably due to the high pulse repetition frequency of the click signals
contained in the sequences, are interpreted as a high interest of the juvenile towards the hydrophone.
The sequences of clicks were emitted almost cyclically by the whales during the intervals free of noise: apparently a
sender waited for the answer of one or more listeners before emitting a new sequence of clicks.
During the intervals with noise the whales were almost silent.
REFERENCES
Airoldi, S., Fasano, D., Gavazzi, N., Miragliuolo, A., Mussi, B., and Mariani, M. 2003. Diving behaviour and reactions to tagging
of long – finned pilot whales in the central Mediterranean Sea. (this volume).
Goold, J. C. and Jones, S. E. 1995. Time and frequency domain characteristics of sperm whale clicks. J. Acoust. Soc. Am, 98(3):
1279-1291.
Jaquet, N., Dawson, S., and Douglas, L. 2001. Vocal behaviour of male sperm whales: Why do they click?. J. Acoust. Soc.Am,
109(5): 2254-2258.
Impetuoso, A., Antollovich, W., Battaglia, M., Dernowski, R., Miragliuolo, A., Mussi, B., and Azzali, M. 2004. Comparison of
acoustic signals among three odontocete species (Grampus griseus, Physeter macrocephalus, Globicephala melas) recorded in a
natural environment.Pp. 36-39. In: European Research on Cetaceans – 15. Proc. 15th Ann. Conf. ECS, Rome, Italy, 2001.
Watkins, W. A. 1980. Acoustics and the behaviour of Sperm Whale. In: Animal Sonar System, pp. 283-289.
Whitehead, H. and Weilgart, L. 1990. Click rates from sperm whales. J. Acoust. Soc. Am, 87(4): 1798-1806.
24
Table 1. Click sequences emitted by a herd of 12 sperm whales in 17 minutes of continuous recording. Each row
contains the series of click sequences emitted by the whale, identified by the number in the first column. Within the
bins of a row are reported the beginning and the end of each sequence emitted by the animal related to that row.
The noisy and the empty bins of each row are also shown .
Whales
Rhythm of click sequences
25
Sperm Whale 1 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 10:44 to
10:46 min.
Sperm Whale 2 Mean Spectrum
frequency Hz
Rel. Power s
p
ectral den sit
y
(
V2
)
Time DAT:
from 03:26 to
03:34 min.
Sperm Whale 3 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 03:41 to
03:50 min.
Sperm Whale 5 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 09:02 to
09:10 min.
Sperm Whale 6 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 14:18 to
14:26 min.
Sperm Whale 9 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 10:05 to
10:12 min.
Sperm Whale 11 Mean Spectrum
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 30 to 52
sec.
Sperm Whale 12 Mean Spectrum
Hz Fre
q
uenc
y
Rel. Power s
p
ectral densit
y
(
V2
)
Time DAT:
from 07 to 10
sec.
Bin 33
Sequence:
10:44-10:46
Bin 11
Sequence:
03:26-03:34
Bin 14
Sequence:
04:22-04:36
Bin 28
Sequence:
09:02-09:10
Bin 43
Sequence:
14:14-14:26
Bin 31
Sequence:
10:05-10:11
Bin 2
Sequence:
00:30-00:32
Bin 1
Sequence:
00:07-00:10
Fig. 1. The mean spectra of a selected number of sequences, emitted by eight whales, included a juvenile. The spectra
of whales 1, 2, 3, 5, 6, 9 and11 are characterized by few clear peaks falling at frequencies different in each animal.
The mean spectrum of the juvenile (whale 12) is characterized by about ten harmonics, falling in the range 2- 4 kHz
Article
Full-text available
Off Kaikoura, New Zealand, we recorded individually identified male sperm whales (Physeter macrocephalus) for entire dive cycles in order to investigate vocal behavior of individual whales and to examine possible functions of sperm whale clicks. In our study, sperm whales were almost always silent at the surface. They consistently started clicking within 25 s after fluking-up and diving. During the first 10 s of clicking, interclick intervals were significantly correlated with water depths at the location of fluke-up. The first "creak" was produced on average 7.5 min into a dive. Interclick intervals usually decreased substantially before clicks turned into "creaks." The highest click rate recorded in this study was 90.9 click/s, and clicks-within-creaks were much shorter than "usual clicks" (mean of 3.6 ms versus 17 to 30 ms). The number of creaks per minute of dive and the length of a dive were significantly correlated. On average, sperm whales were silent for the last 3.6 min before surfacing. Short sequences of "surface clicks" (3 to 8 metallic clicks with mean interclick interval of 5.5 s) were often produced at the end of a dive (in 57% of the dives), but their function remains puzzling. The results of this study suggest that usual clicks and creaks are both used for echolocation purposes, the former to gather information about acoustically reflective features and the latter to detect prey.
Chapter
The first good recordings of underwater sperm whale sounds that we have were made 27 years ago (22 April 1952, R/V Caryn at 38°29′N, 69°29′W) and five years later the sounds were identified definitely and described (Worthington and Schevill 1957). At every opportunity since then, we have stopped with these whales and listened underwater (Watkins 1977). Repeated observations have gradually sorted out many of the sounds. The overlapping clatter of click series that are usually heard in the presence of these whales have given way with time to acoustic sequences that apparently have characteristics identifiable with individuals. The unorganized welter of sound can be simplified to series of pulses that are related and that can sometimes be traced for hours. Whales underwater have been located and tracked by their own sounds. Yet with all this, the sounds are still not well enough correlated with the whales’ actions to provide anything but glimpses into the role that acoustics plays in the behavior of sperm whales.
Article
The rate of production of clicks by groups of (mainly female) sperm whales off the Galápagos Islands was dependent on two principal factors: the number of whales present and the behavioral state of the group. When the whales were in their principal (occupying about 80% of their time) behavioral state, diving deep for prolonged periods and usually being seen at the surface singly or in pairs, each whale made trains of clicks with an interclick interval of about 0.5 s about 70% of the time. About 10% of the time, the groups remained at or near the surface, with individual members forming clusters containing greater than five individuals, and being generally silent. Group behavior was sometimes intermediate between these extremes, with some whales silent at the surface in medium‐sized clusters, and others clicking at depth. Click rate can be used as an indicator of the behavior of sperm whales, the size of a group, and/or the number of groups present. In our studies, groups associated often, and more frequently in 1985 than in 1987, perhaps because of changes in oceanographic conditions. The overall click rate, needed to calibrate acoustic censuses, was estimated to be 1.22 clicks s− 1/animal.
Article
Regular clicks from diving sperm whales, both large bull males and smaller females, were recorded in deep oceanic water off the Azores and subsequently sampled to computer disks for digital analysis. A total of 8540 clicks were marked and analyzed. Simple temporal analysis of the interclick intervals during feeding dives revealed mean click rates for male sperm whales of 1.1713 s-1 and 1.9455 s-1 for females. Fourier analysis showed distinctive peaks in the spectra of bull male sperm whales at 400 Hz and 2 kHz which were stable over extended periods of up to 20 mins. The clicks contained higher frequency components with energy ranging up to at least 12 kHz but not concentrated at any sharply defined frequency. The clicks of smaller female sperm whales showed similar spectral peaks, shifted to 1.2 and 3 kHz, respectively, but these peaks were less pronounced than those in the male click spectra and less stable with time. Higher frequencies were also present up to at least 15 kHz. The previously reported multiple pulse structure of sperm whale clicks is confirmed, but digital filtering reveals this structure to be frequency dependent. Analysis using the short-time Fourier transform confirms the complex time-frequency structure of individual clicks. The frequencies at which the multiples emerge in male and female clicks supports the idea of air cavities in the sperm whale head acting as sound reflectors, although the magnitude of the second pulse at high frequencies suggests some form of off axis distortion. It is also possible that air cavity resonance in the head of the sperm whale may act to reinforce the high-frequency components of the click, and that such components may have superior range and resolution performance in terms of echolocation.
Diving behaviour and reactions to tagging of long -finned pilot whales in the central Mediterranean Sea
  • S Airoldi
  • D Fasano
  • N Gavazzi
  • A Miragliuolo
  • B Mussi
  • M Mariani
Airoldi, S., Fasano, D., Gavazzi, N., Miragliuolo, A., Mussi, B., and Mariani, M. 2003. Diving behaviour and reactions to tagging of long -finned pilot whales in the central Mediterranean Sea. (this volume).