Investigation of acoustic vector sensor data processing in the presence of highly variable bathymetry

Abstract

Data has been collected on acoustic vector sensors mounted on autonomous underwater gliders in the Monterey Bay during 2012–2013. Previous processing work computed the acoustic vector intensity to estimate bearing to impulsive sources of interest. These sources included small explosive shots deployed by local fishermen and humpback whale vocalizations. While the highly impulsive shot data produced unambiguous bearing estimations, the longer duration whale vocalizations showed a fairly wide spread in bearing. In this work, causes of the ambiguity in bearing estimation are investigated in the context of the highly variable bathymetry of the Monterey Bay Canyon, as well as the coherent multipath interference in the longer duration calls. Sound speed data collected during the previous experimental effort, along with a three-dimensional bathymetric relief of the Monterey Bay Canyon, are incorporated into a three-dimensional version of the Monterey-Miami Parabolic Equation Model. Propagation results are computed over a frequency band from 336–464 Hz in order to provide predictions of pulse arrival structure. This data is analyzed using conventional pressure plane-wave beamforming techniques in order to highlight horizontal coupling caused by the canyon bathymetry. The data is also analyzed using the previously developed acoustic vector intensity processing string and shown to exhibit a qualitatively similar spread in the estimated bearing.