Shipping Noise Predictions: Capabilities and Limitations

Abstract

This paper addresses shortcomings in the ability to predict either current levels of the ambient noise generated by shipping or future trends in those levels that might result from changes in the world's shipping fleet. In particular, experimental evidence is presented that predictions of increases in the sound generated by the world's ships based on increases in their speeds and lengths are not justified. This is because, contrary to the classical model of shipping source levels (Ross, 1976), there is a negligible correlation between the source levels of an ensemble of ships and the speeds and lengths of those ships. We also present two examples of noise model predictions that result in large errors. The first shows that two state-of-the-art noise models can yield significantly different noise predictions (5 dB) for the same deep-water, open ocean site. About two dB of this difference is attributed to an approximation inherent in the acoustic propagation model of one of the noise models that is acceptable in some deep-water regions, but not in others. The remaining discrepancy is attributed to differences in the acoustic environmental databases (sound speeds, bathymetries, and geo-acoustic models) between the two noise models. The second example shows that neglecting the local shipping component in a littoral region near a port can result in a noise prediction that is over 15 dB less than a measured value. Taken together, these examples indicate that large errors can result because of inappropriate propagation models or incomplete or inaccurate shipping and environmental data bases. [Work supported by ONR.]