Environmental and signal models for acoustic models for acoustic propagation through a variable ocean

Abstract

Acoustic signals propagating through the ocean are refracted, scattered, and attenuated by the ocean volume and boundaries. Although much about the ocean environment and its effect on acoustic propagation is understood, there is always variability associated with both environment parameters and the received signal, and uncertainty in our knowledge of both. Environment and signal models (e.g., sound speed, bottom loss, and received level) often include only mean values, but they can include the higher order moments that parametrize variability. Interest in relating environmental and signal models and variability is motivated by recent efforts to capture ocean environmental uncertainty that have enhanced understanding and models of environmental parameters, and by advances in ocean acoustic signal processing that indicate how to exploit knowledge of signal variability in order to improve sonar performance. These two developments can be connected if the relationship between environmental and signal parameters (or models) is understood. We show that when a signal prediction must reflect realistic, environment-driven variability over a particular space-time interval, the environmental models used to construct the signal model must contain the environmental variability present during this same interval. [Work supported by ONR Undersea Signal Processing Code 321US.]