Properties of Low Frequency Underwater Ambient Noise in the Ocean Sound Channel
Open Access
- Author:
- Nichols, Stephen Matthew
- Graduate Program:
- Acoustics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- March 17, 2017
- Committee Members:
- David Bradley, Dissertation Advisor/Co-Advisor
David Bradley, Committee Chair/Co-Chair
Charles Holland, Committee Member
Thomas B Gabrielson, Committee Member
Diane Marie Henderson, Outside Member - Keywords:
- Ambient noise
CTBTO
Correlation matrices
Array shape estimation - Abstract:
- In the ocean sound channel, where sounds are known to propagate great distances, the ambient noise field is a dynamic mixture of many distinct noise sources. This work is focused on identifying predictable properties of the low frequency (<100 Hz) noise field and improving the capabilities of hydroacoustic systems for sensing and monitoring geophysical, biologic, and anthropogenic activity. In this dissertation, three unique analyses of the properties of ambient noise are examined. First, the composition of an ambient noise field is assessed, by identifying sets of characteristic spectra. Three proposed methods for identifying these spectrum sets are: spectrum correlation matrix sorting, manual replica selection, and Principal Component Analysis. The characteristic spectra identified by each of these techniques are then used to reconstruct an approximation of the measured noise field. The accuracy of each of these methods for reconstructing the measured noise field is assessed by comparing the frequency correlation matrices of the measured and reconstructed noise fields, which characterize the dominant frequency behavior within a dataset, helping to identify the dominant source mechanisms. Second, the wind-driven component of the acoustic spectrum below 20 Hz is compared across measurement locations in three different ocean basins. In a 1993 study by McCreery et al., the authors proposed that the spectrum of wind-driven noise below 5 Hz should follow a characteristic -23 dB/octave slope, which should be constant across all measurement locations, because of its purely geophysical driving mechanism. In testing this hypothesis, the current work identifies strong similarities between the spectra measured at various sites, though small, basin-dependent spectral features arise due to the local seafloor configuration. Lastly, the angular dependence of the apparent sound speed measured by a receiver array is examined. For an array that is small relative to the source distance, the arrival angle and propagation speed of a signal can be computed by assuming that it can be approximated as a plane wave, and comparing the acoustic arrival time differences. In general, it is expected that the sound speed in water should be isotropic, ignoring any effects of local currents and refraction. However, if the sensor coordinates used to derive the arrival angle and sound speed are incorrect, the resulting sound speeds will strongly depend on the source arrival angle. The current work identifies the form of the angular sound speed dependence for incorrect sensor locations, which is subsequently used to estimate the correct distances between sensors in the array. This procedure is further developed to attempt to identify relative motion of the moored hydrophones within local tidal currents. The data used in this study comes from three stations of the Comprehensive nuclear-Test Ban Treaty Organization’s (CTBTO) hydroacoustic monitoring system. These stations have produced multi-year recordings of low frequency ambient noise in the ocean sound channel. While the primary role of these stations is to monitor the world’s oceans for unsanctioned nuclear weapons testing, the resulting recordings provide an excellent data source for underwater acoustics research in the low frequency band. The work presented in this dissertation emphasizes the value of the CTBTO data and its potential contribution to the understanding of changes in ocean-based behavior over time.