Broadband matched field processing (MFP)-derived estimates of 3D source location using data from hydrophone line arrays deployed in various geometries, i.e., vertical, horizontal, and tilted 45° from vertical, are presented in this paper. These data were collected in two shallow water (100–200 m) experiments off the coast of San Diego. Results show that estimates of source range and depth remain surprisingly coherent in the presence of large mismatch in bathymetry, but are offset from the true position by as much as 100%. The offsets are independent of array geometry. In contrast, for estimates of source azimuth, bathymetry mismatch typically causes a degradation in MFP correlation rather than an appreciable offset. However, errors in the assumed tilt from vertical of an array can lead to large offsets in the estimated source azimuth, particularly as the nominal angle of the array from vertical becomes smaller. Predictions from a simple analytical model based on adiabatic normal modes in ideal waveguides provide good fits to the broadband MFP results.

No abstract

No abstract

Detailed seafloor environmental modeling of a moderately complex shallow water area, coupled with multitone matched-field processing, produced reliable and unambiguous source detection and tracking. Frequency averaging of the matched-field output sufficiently reduced sidelobe ambiguities. The seabed geologic model is a gridded database containing water depth, sediment grain size, sediment thickness, and acoustic basement type. Grid cells are squares of side length equal to 2 arc s. Software, separate from the geologic model, computes a geoacoustic model for any desired grid square.

The detection performance of low spectral-level, broadband signals from a submerged acoustic source in the presence of surface interference is described and illustrated with examples obtained in the May 1996 Shallow Water evaluation cell Experiment (SWellEx-96). The experiment was carried out west of Point Loma in 200-m water of complex bathymetry. Adaptive broadband matched-field processing (MFP) was performed on a 240-m aperture horizontal line array. Detection and localization in range, depth, and azimuth is illustrated at all bearings except near the array broadside. The effects of horizontal aperture on MFP robustness and sidelobe levels are also discussed.