Characterisation and analysis of sperm whale clicks
Electronic versions
Documents
39.3 MB, PDF document
Abstract
Regular clicks of diving sperm whales Physeter macrocephalus were recorded in deep oceanic water off the Azores, and subsequently sampled to computer disk for analysis. A total of 8540 clicks were marked and analysed. Simple temporal analysis of inter-click intervals revealed mean click rates for male sperm whales of 1.1713 s- 1 , and for female sperm whales of 1.9455 s- 1. Fourier analysis showed distinctive peaks in the spectra of bull male sperm whale clicks at approximately 400 Hz and 2 kHz, which were stable over extended periods of up to 20 minutes. The clicks of smaller female sperm whales showed similar spectral peaks, shifted to approximately 1.2 kHz and 3 kHz, but which were less pronounced and less stable with time than the male peaks. All clicks contained broadband components up to at least 15 kHz. The previously reported multiple pulsed
structure of sperm whale clicks is confirmed, and waveform filtering shows the structure to emerge only above certain threshold frequencies, approximating to estimated air sac dimensions in the sperm whale head. It is considered that the function of sperm whale clicks is echolocation.
Sound velocity was measured in samples of spermaceti oil from the spermaceti sac of a 15.6 m male sperm whale Physeter macrocephalus under varying temperature and pressure regimes. Measured sound velocities were in the range 1390 to 1530 ms-1 between the limits 22°C to 38°C and 0 to 9.1 MPa. Sound velocity was later measured in castor oil, spermaceti oil from the melon of the dwarf sperm whale Kogia simus, and a final sample of the original Physeter spermaceti oil. Measured sound velocities in castor oil were in the range 1490 to 1560 ms-1 between the limits 6°C to 40°C and Oto 9.1 MPa.
Measured sound velocities in Kogia melon core spermaceti oil were in the range 1395 to 1670 ms-1 between the limits 9°C to 38°C and 0 to 9.1 MPa. Measured sound velocities in Kogia melon peripheral spermaceti oil were in the range 1430 to 1510 ms-1 between the limits 7°C to 38°C and 0 to 9 .1 MPa. Measured sound velocities in the final Physeter spermaceti oil sample were in the range 1420 to 1520 ms-1 between the limits 12°C to 38°C and 0 to 9 .1 MPa. In all cases, sound velocity increased linearly with increasing pressure, but decreasing non-linearly with increasing temperature.
Waveform autocorrelation and cepstrum analysis techniques were used to measure intrapulse intervals from sequences of male and female sperm whale clicks. Autocorrelation was generally successful provided that click signals were bandpass filtered, to reveal the high frequency pulsed structure, prior to analysis. Cepstrum analysis revealed that sperm whale click spectra contain ripples with periods equal to the reciprocal of the intra-pulse interval, in line with theoretical expectations. Cepstrum analysis was successfully used to assess intra-pulse intervals, the resulting data having less scatter than autocorrelation results. Neither autocorrelation or cepstrum analysis were sufficiently robust methods for making spot measurements of intra-pulse intervals from individual clicks, but realistic trends were obtained by analysing clicks over individual sequences of 6 minutes. With both autocorrelation and cepstrum analysis, a 50 point moving average was applied to the resulting data to visualise the trends in intra-pulse interval with time. Overall cepstrum analysis produced results with less scatter than autocorrelation analysis, and reasonable estimates of sperm whale body lengths were extrapolated from the data.
structure of sperm whale clicks is confirmed, and waveform filtering shows the structure to emerge only above certain threshold frequencies, approximating to estimated air sac dimensions in the sperm whale head. It is considered that the function of sperm whale clicks is echolocation.
Sound velocity was measured in samples of spermaceti oil from the spermaceti sac of a 15.6 m male sperm whale Physeter macrocephalus under varying temperature and pressure regimes. Measured sound velocities were in the range 1390 to 1530 ms-1 between the limits 22°C to 38°C and 0 to 9.1 MPa. Sound velocity was later measured in castor oil, spermaceti oil from the melon of the dwarf sperm whale Kogia simus, and a final sample of the original Physeter spermaceti oil. Measured sound velocities in castor oil were in the range 1490 to 1560 ms-1 between the limits 6°C to 40°C and Oto 9.1 MPa.
Measured sound velocities in Kogia melon core spermaceti oil were in the range 1395 to 1670 ms-1 between the limits 9°C to 38°C and 0 to 9.1 MPa. Measured sound velocities in Kogia melon peripheral spermaceti oil were in the range 1430 to 1510 ms-1 between the limits 7°C to 38°C and 0 to 9 .1 MPa. Measured sound velocities in the final Physeter spermaceti oil sample were in the range 1420 to 1520 ms-1 between the limits 12°C to 38°C and 0 to 9 .1 MPa. In all cases, sound velocity increased linearly with increasing pressure, but decreasing non-linearly with increasing temperature.
Waveform autocorrelation and cepstrum analysis techniques were used to measure intrapulse intervals from sequences of male and female sperm whale clicks. Autocorrelation was generally successful provided that click signals were bandpass filtered, to reveal the high frequency pulsed structure, prior to analysis. Cepstrum analysis revealed that sperm whale click spectra contain ripples with periods equal to the reciprocal of the intra-pulse interval, in line with theoretical expectations. Cepstrum analysis was successfully used to assess intra-pulse intervals, the resulting data having less scatter than autocorrelation results. Neither autocorrelation or cepstrum analysis were sufficiently robust methods for making spot measurements of intra-pulse intervals from individual clicks, but realistic trends were obtained by analysing clicks over individual sequences of 6 minutes. With both autocorrelation and cepstrum analysis, a 50 point moving average was applied to the resulting data to visualise the trends in intra-pulse interval with time. Overall cepstrum analysis produced results with less scatter than autocorrelation analysis, and reasonable estimates of sperm whale body lengths were extrapolated from the data.
Details
Original language | English |
---|---|
Awarding Institution |
|
Supervisors/Advisors |
|
Thesis sponsors |
|
Award date | 1998 |