Prediction of ocean ambient sound generated from geophysical signals

Abstract

About 90 buoy months of ocean ambient sound data are collected using Acoustic Rain Gauges in different open-ocean locations. Distinct ambient sound spectra are identified through a series of discrimination process. Some distinct features are described as follows: (1) Excluding effects of rain, observed ambient sound spectra, generated exclusively by wind, exhibit a constant frequency spectral slope in 1–35 kHz. (2) Drizzle produces a prominent spectral peak at 15 kHz, and the magnitude of the spectral peak is very sensitive to the wind speed. These are consistent with earlier findings [Vagle et al. (1990) and Nystuen (1993)]. (3) Preliminary analysis shows that the spectral slope between 1–10 kHz decreases linearly with the rainfall rate. Therefore, the rainfall rate may be reliably predicted by both the spectral slope and the spectral magnitude in 1–10 kHz frequency range. The comparison of these two independent predictions provides the quality check of the estimate of the rainfall rate. An analytical spectral model of the ocean ambient sound from 1–50 kHz is constructed using two input parameters, rainfall rate and wind speed, based on the existing observations. This analytical model is aimed at predicting the ambient sound spectra at varying rainfall rates and wind speeds.