Stephanie Fried

Extracting the local Green’s function on a horizontal array from ambient ocean noise

Using only ocean ambient noise recordings it is possible to approximate the local time domain Greens function (TDGF) and extract the time delays associated with different ray path between the elements of a bottom hydrophone array. Comparing the strength of the noise correlation function taken over increasing time windows with residual fluctuations points to an optimum time window to use in the noise correlation function. Through comparison with computer simulations the resulting time series is shown to accurately approximate noise responses in the environment. Analysis of the TDGF gives accurate environmental detail, specifically the critical angle at the water-sediment interface.

On the coherent components of low-frequency ambient noise in the Indian Ocean

This letter demonstrates that the dominant coherent component of low-frequency (1 Hz < f < 20 Hz) ambient noise propagating between hydrophone pairs of the same hydroacoustic station, deployed in the deep sound channel of the Indian Ocean, is directional and mainly originates from Antarctica. However, the amplitude of the peak coherent noise arrivals, obtained using a 4-month-long averaging interval, was relatively low given the small hydrophones spacing hydrophones (<2 km). Hence, extracting similar coherent arrivals between two distinct hydroacoustic stations separated instead by thousands of kilometers for noise-based acoustic thermometry purposes seems unlikely, even using a year-long averaging.

Measuring the effect of ambient noise directionality and split-beam processing on the convergence of the cross-correlation function

Measurements of ambient noise have been used to infer information about the ocean acoustic environment. In recent years the correlation of ambient noise has been shown to give estimates of the travel time of acoustic paths between the sensors recording the noise. A number of issues affect the results of the noise correlation. This paper presents the results of noise correlation of the two horizontally separated arrays of sensors in the 2010 ambient noise experiment. Using the experimental data, the effects on the convergence of the noise correlation are examined with respect to the size and shape of the arrays, the length of time used, and the directionality of the noise field.

Using hydrophones as vector sensors

Hydrophone arrays with spacing much less than an acoustic wavelength can be converted to vector sensors. Subsequent vector sensor signal processing can then be applied. Two particular applications are presented: The first is converting very low frequency acoustic data to seismic type data that contain polarization information and the second is getting directional information from sub wavelength acoustic arrays. We start with a review of the simple theory followed by some illustrative simulation examples. We then apply these signal processing methods to ocean acoustic data.

Using an autonomous underwater vehicle to track the changing arctic ambient noise field in real time

The ambient noise field, particularly the directionality of the noise, can provide a wealth of information about the local environment. Changes in the ambient noise field often reflect changes in the physical environment. Accurate calculation of the noise field, though, can be a challenge. Because of their maneuverability autonomous underwater vehicles (AUVs) provide novel capabilities not only for measuring and analyzing the local noise field, but also for continuous tracking of changes to the noise field and thus the environment. Of interest is the measurement and analysis of the ambient noise in arctic environments. By integrating models for arctic ambient noise into an AUV simulation, this paper analyzes the use of AUVs in real-time autonomous tracking of the three-dimensional changing arctic ambient noise field.

Measuring the spatial characteristics of the ambient noise field from an autonomous underwater vehicle

For autonomous underwater vehicles (AUVs), the primary method of sensing the local environment is through acoustics. The local noise field contains a wealth of information the AUV uses - from target tracking to communication to general understanding of the environment. An assessment of the spatial composition of the ambient noise field can provide details about the physical environment as well as information for the AUV to incorporate into its control decisions. The challenge is in accurately measuring the directionality of the noise field from a single line array while continuously updating this measure to reflect changes in the environment and additional information as the AUV moves. This analysis presents a method for continuously assessing the spatial characteristics of an ocean ambient noise field measured by an AUV with a towed hydrophone array.

Low-frequency broadband noise correlation processing in deep-water

The Comprehensive Nuclear-Test-Ban Treaty Organization operates an International Monitoring System (IMS). The IMS includes hydroacoustic stations composed of hydrophones deployed in the ocean deep-sound-channel in a two-kilometers-side triangular configuration(referred to as triad) in the horizontal plane. Data are continuously recorded on hydrophone triads (with a sampling frequency of 250 Hz) and have been archived during the last decade. Previous experimental studies have demonstrated that coherent waveform can be extracted from broadband coherent processing of ocean ambient noise, typically above f > 100 Hz [e.g., see Roux et al., J. Acoust. Soc. Am. 116(4), 1995–2003 (2004)] . We investigated here the emergence of coherent arrivals from the correlation processing of the low-frequency broadband ambient noise recorded during the years 2006–2007 on IMS hydrophones located in the Southern Hemisphere. This low-frequency acoustic ambient noise includes various components from anthropogenic and biological s...

Cross‐correlation of diffuse surface noise between horizontally separated sensors.

The use of ambient noise in the ocean to extract the time of arrival structure of the time domain Green’s function (TDGF) between horizontally separated sensors was first demonstrated in deep water using shipping noise [Roux et al., J. Acoust. Soc. Am. 116, 1995 (2004)]. Subsequently, surface noise recorded along a vertical array was used in the passive fathometer application [Siderius et al., J. Acoust. Soc. Am. 120, 1315 (2006)]. A noise field from primarily volumetrically distributed biological noise sources recorded along a horizontal array was also shown to extract a shaded TDGF in very shallow water [Fried et al., J. Acoust. Soc. Am. 124, EL183 (2008)]. This same noise correlation function (NCF) processing is now demonstrated along horizontally separated hydrophones using diffuse, low sea‐state surface noise, where the spectrogram of the noise reveals no significant shipping component. These results are presented along with a discussion of the relevant signal processing as well as the theory and sim...

Incorporating array processing to improve resolution of ambient noise data cross‐correlation along a horizontal array.

Numerous authors have shown that the cross‐correlation of ambient noise recordings at two points can produce an approximation of the Green’s function between those points. The resolution of the cross‐correlation is fundamentally limited by the characteristics of the noise field and the changing environment. In order to make this technique the most effective, we wish to limit the time needed to extract the desired environmental information. To that end, array processing techniques can be adapted to work with the noise cross‐correlation function taken along a horizontal array to decrease the time needed to resolve multiple individual returns. For a given noise field, beam forming between horizontal arrays is effective in reducing the time needed to identify both the direct and single‐bounce surface reflection paths.

Comparing time domain Green's functions with simulated noise and ambient noise data cross‐correlation for a horizontal array

Previous work has shown that an approximation of the Green’s function can be extracted from ambient noise data through cross-correlating the received signals along an array. The resulting Green’s function approximation gives accurate time-of-arrivals for the multipaths between hydrophones but can only approximate the magnitude of the arrivals in the time domain Green’s function. Nevertheless, some useful information can be obtained from the relative amplitudes of the correlated returns assembled. Further, a Monte Carlo noise model simulation for a similar environment for which noise data was collected reproduces the same crosscorrelation arrival structure for the processed ambient noise data and the theoretical time domain Green’s function arrival structure. [Research supported by ONR].