155
features. (3) Intrinsic dynamics of the populations (e.g. body size, sexual selection)
can also affect the acoustic repertoire produced by its individuals.
As we hinted earlier, characterization of the acoustic repertoire of animals at sea
may be challenging, but it is necessary as a first step to conduct passive acoustic
monitoring (PAM) of cetaceans, by means of which acoustic recordings can be classified and attributed to a particular species without the need of a visual match to
verify the identity of the sound producing animal.
For some species, much of the information available comes from captive animals. While it may be the best approximation we can get to the species living in the
wild, it is important to consider that the signals recorded and associated behavior
displayed may not be representative of the behavior recorded of wild animals. For
example, captive animals usually produce sounds with lower amplitude and frequency emphasis than animals in the field (Au 1993, Madsen and Wahlberg 2007).
In order to characterize the acoustic repertoire of a given species, we need to
keep in mind what our goal is, but at the very least we need to be able to identify the
species with a high degree of certainty; this is, either by using a hydrophone array
that allows us to correlate the location of the animal with our observations, or we
need to be in an area at a specific time in which there is only our species of interest.
In some fortunate cases, however, sympatric species may produce very different
sounds and we may be tempted to believe which sound is produced by which species. A typical example in our region would be studying the repertoire of Peale’s and
Dusky dolphins, two sympatric species that produce very different sounds: Peale’s
dolphins produce narrowband high-frequency (NBHF) clicks and it is believed that
they do not whistle; while Dusky dolphins produce broadband (BB)
1
clicks and
whistles are a major component of their acoustic repertoire. The risk may be that we
associate all whistles to Dusky dolphins (because that’s what we know) instead of
trying to determine whether Peale’s dolphins also produce whistles under certain
circumstances. This assumption may hinder us from learning something new and
promote spreading wrong information.
When it comes to the recordings themselves, a typical underwater recording system nowadays consists of a hydrophone, a preamplifier, and a band-pass filter connected to a digital recorder. The hydrophone is a transducer that converts sound
pressure into voltage variations which is then amplified by the preamplifier. The
band-pass filter reduces low frequency noise and filters out any signal component
above the Nyquist frequency (half the sampling frequency of the digital recorder)
which is the theoretical maximum frequency that can be digitized unambiguously.
For example, if our digital recorder has a sampling frequency of 192 kHz, the
recorded signals will be limited to frequencies below 96 kHz (Nyquist frequency),
making this setup unsuitable to characterize properly the acoustic repertoire of
species that produce sounds that exceed 96 kHz or to record NBHF clicks at all,
1 NBHF refers to clicks produced by some species of odontocetes which have all their frequency
content above 100 kHz, longer duration and a narrower frequency bandwidth than the short BB
clicks produced by most delphinid species.
6 Bioacoustic Techniques Applied to Odontocete Conservation and Management…
features. (3) Intrinsic dynamics of the populations (e.g. body size, sexual selection)
can also affect the acoustic repertoire produced by its individuals.
As we hinted earlier, characterization of the acoustic repertoire of animals at sea
may be challenging, but it is necessary as a first step to conduct passive acoustic
monitoring (PAM) of cetaceans, by means of which acoustic recordings can be classified and attributed to a particular species without the need of a visual match to
verify the identity of the sound producing animal.
For some species, much of the information available comes from captive animals. While it may be the best approximation we can get to the species living in the
wild, it is important to consider that the signals recorded and associated behavior
displayed may not be representative of the behavior recorded of wild animals. For
example, captive animals usually produce sounds with lower amplitude and frequency emphasis than animals in the field (Au 1993, Madsen and Wahlberg 2007).
In order to characterize the acoustic repertoire of a given species, we need to
keep in mind what our goal is, but at the very least we need to be able to identify the
species with a high degree of certainty; this is, either by using a hydrophone array
that allows us to correlate the location of the animal with our observations, or we
need to be in an area at a specific time in which there is only our species of interest.
In some fortunate cases, however, sympatric species may produce very different
sounds and we may be tempted to believe which sound is produced by which species. A typical example in our region would be studying the repertoire of Peale’s and
Dusky dolphins, two sympatric species that produce very different sounds: Peale’s
dolphins produce narrowband high-frequency (NBHF) clicks and it is believed that
they do not whistle; while Dusky dolphins produce broadband (BB)
1
clicks and
whistles are a major component of their acoustic repertoire. The risk may be that we
associate all whistles to Dusky dolphins (because that’s what we know) instead of
trying to determine whether Peale’s dolphins also produce whistles under certain
circumstances. This assumption may hinder us from learning something new and
promote spreading wrong information.
When it comes to the recordings themselves, a typical underwater recording system nowadays consists of a hydrophone, a preamplifier, and a band-pass filter connected to a digital recorder. The hydrophone is a transducer that converts sound
pressure into voltage variations which is then amplified by the preamplifier. The
band-pass filter reduces low frequency noise and filters out any signal component
above the Nyquist frequency (half the sampling frequency of the digital recorder)
which is the theoretical maximum frequency that can be digitized unambiguously.
For example, if our digital recorder has a sampling frequency of 192 kHz, the
recorded signals will be limited to frequencies below 96 kHz (Nyquist frequency),
making this setup unsuitable to characterize properly the acoustic repertoire of
species that produce sounds that exceed 96 kHz or to record NBHF clicks at all,
1 NBHF refers to clicks produced by some species of odontocetes which have all their frequency
content above 100 kHz, longer duration and a narrower frequency bandwidth than the short BB
clicks produced by most delphinid species.
6 Bioacoustic Techniques Applied to Odontocete Conservation and Management…