Modelling the coastal ambient noise field in time, frequency, and space

Abstract

Natural ambient noise levels in time, frequency, and space, in coastal regions are modeled based on local environmental forcing and propagation conditions. Continuous audio recordings were taken between April 18th and May 12th, 2018 off Sooke Inlet, BC, a shallow water region with complex bathymetry and persistent vessel traffic, for model parameterization. The received noise levels were decimated to their hourly minimums, and spectral component filtering was used for source classification. An ambient noise model in the form of a linear combination of wind speed and rain rate source terms, modulated by a tidally driven amplitude term, was developed. Model-data comparisons of monthly sound levels from April 2018 to February 2019 show less than 5 dB error above 1 kHz, and less than 3dB error above 20 kHz. A computational sound propagation model is used to compute the source level per unit of area of the natural noise generating mechanisms.