Range and depth estimation of bowhead whale calls in the Arctic using a single hydrophone

Abstract

Bowhead whales generate low-frequency vocalizations in shallow-water Arctic environments. The propagation of these sounds is adequately described by normal mode theory. Indeed, at low-frequency the environment acts as a dispersive waveguide. The propagated signal can be modeled by a sum of normal modes, and the source position can be inferred from the different modal arrivals. However, this requires to estimate the modes from the received signal. Traditionally, modal arrivals are separated using synchronized hydrophone arrays. Here a nonlinear signal processing method called warping is used to filter the modes on just a single hydrophone. Once modes are filtered, the source depth and range can be estimated using classical Matched Mode Processing. This methodology is illustrated on an experimental bowhead whale vocalization recorded in the Beaufort Sea. Because seabed properties are not well known, it is necessary to jointly perform source localization and geacoustic inversion.