No abstract

High-resolution bistatic images of a typical abyssal hill on the western flank of the Mid-Atlantic Ridge are made with a low-frequency towed-array system operating remotely at 1/2 convergence zone ͑ϳ33.3 km͒ stand-off. Comparison with modeled images, generated from high-resolution supporting bathymetry sampled at 5-m intervals, roughly the wavelength scale, reveals that steep scarps return the strongest echoes because they project the largest area along the acoustic path from the source to receiver. Prominent returns deterministically image scarp morphology when the cross-range axis of the system's resolution footprint runs along the scarp axis. Statistical fluctuations inherent in the scattered field prevent the system from distinguishing smaller-scale anomalies on the scarps, such as canyons and gullies ͑ϳ100-200 m scale͒, that would otherwise be resolvable in range, in certain bistatic geometries. The mean bi-azimuthal scattering distributions of the two major scarps on the abyssal hill are identical and have strengths equal to the constant Ϫ17 dBϮ8 dB. This suggests that long-range reverberation from prominent geomorphological features of the world's mid-ocean ridges can be adequately modeled as Lambertian with albedo /10 1.7 , given supporting bathymetry sampled with sufficient frequency to resolve the projected area of these features.

A model is developed for the calculation of the spatial properties of the surface-generated noise in a three-dimensional ocean. This is an extension of the work of Kuperman and Ingenito [J. Acoust. Soc. Am. 67, 1988–1996 (1980)], which used a normal-mode representation of the noise field in a stratified ocean. Noise fields are simulated for both point receivers and vertical line receivers. These examples show how the spatial and directional characteristics of the noise field are affected by the ocean environment. For example, as is apparent in ambient noise data, surface noise propagating at high angles over a sloping ocean bottom is deflected into shallower angles. Also, matched-field processing simulations in three-dimensional ocean environments can be done in a consistent manner: signals and surface-generated noise are modeled by propagating through the same environment with the same theory.

An approach for avoiding the problem of environmental uncertainty is tested using data from the TESPEX experiments. Acoustic data basing is an alternative to the difficult task of characterizing the environment by performing direct measurements and solving inverse problems. A source is towed throughout the region of interest to obtain a database of the acoustic field on an array of receivers. With this approach, there is no need to determine environmental parameters or solve the wave equation. Replica fields from an acoustic database are used to perform environmental source tracking [J. Acoust. Soc. Am. 94, 3335-3341 (1993)], which exploits environmental complexity and source motion.

A matched-field processing algorithm based on the eigenvectors of the covariance matrix is developed for localization problems involving interference from multiple sources and ambient noise. Since the multivalued Bartlett processor is related to the Bartlett processor, it should be more robust in practice than eigenprocessors that are related to the maximum likelihood processor. This eigenprocessor also has the advantage that it provides well-defined estimates of the locations of multiple sources. An example involving moving sources and ambient noise illustrates that matched-field eigenprocessing requires source motion in order to be reliable when there are multiple sources.

Bottom profiling traditionally uses broadband signals received on a line array at long ranges to estimate the bottom layer structure and thickness. In this paper, a subbottom profiling method is developed and applied to a ship-towed line array using the same ship towed source to estimate the subbottom layer structure and thickness. A ship towed line-array system can be used to estimate bottom properties using geoacoustic inversion and can cover a wide area in a short time. It needs some prior information about the subbottom structure and layer thickness, without which the solution can be ambiguous and even erratic when resolving parameters over a wide area. It is shown that the required subbottom information can be obtained from the time-angle relation by beamforming the same acoustic signal data used for geoacoustic inversion. The time-angle analysis is used to expose the prevalent physics intrinsic to geoacoustic inversion. One finds that the tau-p relation of the bottom and the bottom reflection coefficients, sampled at discrete angles associated with bottom and multiple surface-bottom returns, are often adequate, for many practical applications, to uniquely determine the geoacoustic bottom at low (< or =1 kHz) frequencies.

Geoacoustic inversions with a towed horizontal array are of interest for rapidly characterizing sediment properties over changing regions. To be of practical value, inversions must yield consistent and reliable results for consecutive or neighboring data. A method of determining inversion reliability a priori is delineated using an empirical approach and confirmed with inversion results in terms of consistency. Geoacoustic parameter hierarchy and resolvability are empirically analyzed using two different methods: one requires knowledge of the source function and the other does not. Inversion results using the two methods are compared using both synthetic data and experimental data from MAPEX2000. The inversions employ a global optimization technique which navigates the parameter space in directions aligned with valleys of the cost function, increasing inversion algorithm efficiency and disclosing parameter correlations and hierarchy.

Interference from noise and uncertainties in the environmental parameters are arguably the two most serious limitations in matched-field processing (MFP). Among the techniques that have been developed for handling these difficulties are the noise-canceling processor [M. D. Collins, N. C. Makris, and L. T. Fialkowski, “Noise cancellation and source localization,” J. Acoust. Soc. Am. 96, 1773–1776 (1994)] and focalization [M. D. Collins and W. A. Kuperman, “Focalization: Environmental focusing and source localization,” J. Acoust. Soc. Am. 90, 1410–1422 (1991)]. The noise-canceling processor is a generalization of the Bartlett processor that is based on matching the covariance matrix of the data with replica covariance matrices of the signal and the noise. Simulations are presented to illustrate the performance of the noise-canceling processor when there are errors in the noise replica. Focalization is a generalization of MFP in which environmental parameters are included along with source parameters in the search space. An implementation of this approach that is suitable for applications is developed and tested. The noise-canceling processor and focalization are used to simulate the localization of a source buried in noise in an uncertain environment.