We have carried out a
high-statistics study of the “B” call of the blue whale (Balaenoptera
musculus), as observed from the Pioneer Seamount
underwater observatory, 90 km off the California coast. Data were collected in several contiguous blocks,
via an underwater cable, over a period of 13 months. The continuous operation and the
non-invasive nature of the observations make this a homogeneous and
unbiased sample. We find a remarkable
degree of consistency in the “B” calling patterns of the blue
whale: The calling interval is constant
to 4% accuracy; and the frequency of calls is adjusted to be
the same, over long periods of time, and for all calling individuals, to an
accuracy of 0.5 %. The calling
frequency itself is apparently changing with time, with all
individuals visiting Pioneer Seamount tracking the change. Two substantially different types
of calling sequences are observed, each at the same frequency, each
with its well-regulated (but different) calling interval.
A number of questions
concerning blue-whale acoustic activity are raised. What types of behavior correspond to these two
calling sequences? What is the function
of the precisely regulated call frequency?
How do individuals of different size adjust their call to the
same frequency? And what sense is able
to perceive frequency to this remarkable accuracy?
And finally, is there any information
transmitted by acoustic calling which is so remarkably invariant?
Help and advice is gratefully acknowledged from: Dave Mellinger,
Chris Fox, Jonathan Klay, Andy Lau, and Haru Matsumoto, NOAA-PMEL Jim Mercer
and Lyle Gullings, APL, University of Washington; Ching-Sang
Chiu, Monterey Naval Postgraduate School; and John Bourg and Jim Lockhart, San
Francisco State University
Underwater topography near Pioneer Seamount. (Graphic courtesy of
PMEL/NOAA)
ROMBERG TIBURON CENTER
Pioneer Seamount
Observatory
D
D
D
D
D
D
D
D
D
D
BREAK
cable
cable
01010101
010101001010101010101010101010
A
B
A
A
A
A
A
A
A
B
B
B
B
B
B
B
A
B
B
B
B
B
B
B
A
A
A
A
A
A
A
B
B
B
B
B
B
B
A
A
A
B
B
B
THE
CONSTANT-FREQUENCY “B’ CALL
The “B” call of the blue
whale is a contact call – it seems to be used by the whales to stay in
contact with each other. This call is
remarkable in at least two respects.
(1) All whales “tune” to the
same frequency, with a pitch discrimination far better than that of
humans. (2) This frequency chosen
by the group has steadily decreased over recent years. The determination of this frequency is
described below. The average frequency over
the 13 month period studied here is 16.01 Hz, with a variance of the sample of
0.09 Hz. (This variance may be
dominated by measurement error.)
From our data alone we determine a rate of decrease in this frequency
of 0.08 Hz, in rough agreement with published observations.
Distributions in central “B”-call frequency and chirp rate. For
this 13-month period the average frequency
was 16.01 Hz, with a call-to-call variance
of 0.09 Hz. For the musically inclined,
16 Hz is exactly 4
octaves below Middle C, and a
deviation of 0.09 Hz corresponds to an interval of 0.09 half-tones!!
Decline in call frequency over 13
months at Pioneer Seamount. The rate of decrease from this data alone is estimated at 0.08 Hz/year
Decline in call frequency over the
last 10 years. The green line is the trend from Pioneer Seamount data alone.
The “B” call consists of a repeating waveform of extremely
predictable frequency, about 16 Hz. The
power at multiples of 16 Hz consists of harmonics of the fundamental
– the waveform departs increasingly from a perfect sinusoid as the power in
the harmonics increases. The harmonic
structure, and the envelope of the signal (power versus time)
vary in our data sample – especially in the “B-only” sequences. However, the fundamental frequency at the
center of the call and the (slight) rate of decrease of the
frequency during the call are extremely reproducible, as described below.
SEASONAL AND DIEL VARIABILITY
The clearest trends in these data are the absence of
calls from mid-April to mid-July,
followed by plentiful calling through mid-November. There
were more calls in 2001 than 2002, and the calls were fainter in 2002. One simple explanation would be that the whales stayed closer to shore in 2002, though we
have no other evidence for this
hypothesis.
[The ‘grayed-out’ points show the number of calls triggering the matched-filter
detection. The black points show the number of calls remaining after least-squares
fitting. A manual scan of spectrograms was carried out in addition for the
2002 data.]
OBSERVATION
OF BLUE WHALE CALLS FROM PIONEER SEAMOUNT
Michael D. Hoffman,1 Newell Garfield,2 and Roger Bland1
1.Physics
and Astronomy Department and Romberg Tiburon Center for Environmental Studies,
San Francisco State University
2. Department of Geosciences and Romberg Tiburon Center for Environmental Studies, San Francisco State University
2. Department of Geosciences and Romberg Tiburon Center for Environmental Studies, San Francisco State University
http://www.physics.sfsu.edu/~seamount
COMMUNICATION BY BLUE WHALES???
CALLING PATTERNS
Other
researchers have found several distinct calling patterns for Eastern Pacific blue
whales. The most prominent in our data
are patterns of repeated “A”-”B” call pairs (AB calling)
and sequences of “B” calls alone
(B-only calling). For spectrograms
of these sequences, see the GALLERY to the far right.AB
calling sequence (day 285, 2001). Note
pattern of equally spaced AB pairs, separated by breathing intervals
“B-only” calling sequence (day 284, 2001). Note the closer spacing in the absence of
the “A” call.
THE BLUE-WHALE “A” AND
“B” CALLS
Blue-whale sonorization consists mainly
of two calls, referred to as the “A” call (a pulsing call) and the “B” call (a
“moan” of steadily decreasing frequency.) Other, rarer calls are shown in the GALLERY
to the right.
The
“A” call consists of a series of very regular pulsations, with two components,
at different frequencies and out of phase in time. The high-frequency component peaks
at 90 Hz and shows little harmonic structure.
The low-frequency component has most of its power near 16 Hz, the
fundamental frequency of the “B” call.
Most “A” calls seem to have short sections before and
after the pulsations sounding very much like the “B” call..
A
90 Hz
16 Hz
B
48 Hz
16 Hz
32 Hz
A?
A?
B?
B?
Time interval between B calls. The time to the following peak is shown on
the x axis, and the time on
to the peak after that, on the y axis.
The peaks are found by automatic pattern recognition, so many peaks are missing. The prominent bands show the peak spacing of
about 125 sec for AB calling
sequences and of about 50 sec, for B-only calling.
Histogram of time intervals. The main peak corresponds to the spacing
between peaks in AB sequences,
with a peak at twice its value from cases where an intermediate peak is missed by the pattern
recognition. The lower peak shows the peak spacing for B-only calling. Both times are extremely well defined.
AB calling: T= 128 sec,
s=8 sec
B-only calling: T = 50 sec, s=3 sec
Spectrogram obtained from averaging waveforms for a
sequence of loud calls. We have represented the central 6 seconds of this call by a
sinusoid with frequency decreasing with time. The
waveform from each of the 6984 calls identified using a matched filter was fit
using least squares to
determine its central frequency (f0) and rate of decrease of the frequency (alpha).
The results are shown to the right.
Here the hour of the day (UT) is plotted vertically
against the date, horizontally, for all 6984 triggers from the matched-filter detection. There is no striking change in the day-night
variation over the 13-month period
of this study.
Diel variability of the rate of “B” call triggers. While calling is observed at all hours of
the day, the plot peaks
in those hours corresponding roughly to the daylight hours in the Pacific
Daylight time zone. However, there are various possible biases
which could lead to such an effect which we have not so far excluded (for instance, a day-night
difference in the ambient acoustic noise).
Contour map showing the location of Pioneer Seamount. (Graphic courtesy of
the Monterey Bay National Marine Sanctuary.)
Fitted frequency and sweep rate for the day 285, 2001
sequence of very loud calls
(red points) compared to a larger sample of calls from the entire year’s data. The tighter grouping of this sample
probably reflects the smaller
errors associated with the larger signals.
The distribution of red
points is centered at (f0,a) = (16.03 Hz, 0.06 Hz/sec), with widths of (s f0, s a)
= (0.04 Hz, 0.01 Hz/sec).
observatory
off line
6
pm PST
6
am PST
Calambokidis
et al., 2002, CRITTERCAM data.
Lunges through krill layer, spacing 80 sec
dives
spaced 500 sec apart
McDonald, Calambokidis, Teranishi and Hildebrand, JASA 109, 1728. Near Santa
Barbara, Ca, Oct. 1997. “AB” sequences
spaced 125-130 sec apart;
“B-only” sequences spaced 45-55 seconds apart.