Spectral analysis of littoral water ambient noise in the tropical region

Abstract

Increasing ambient noise levels in the oceans has been a cause of concern due to its impact on sonar performance deployed for military as well as non-military applications as they present sub-optimal behavior in poor SNR conditions. Classical noise filtering techniques offer limited utility in tropical littoral waters because of the random fluctuation of the ambient noise. The repeated surface and bottom interactions of the acoustic signal in littoral waters and diurnal and seasonal variation of the surface parameters of the tropical regions are the reasons of these fluctuations. The paper highlights the challenges in spectral analysis of ambient noise recorded in the tropical littoral waters off the west coast of India in the Arabian Sea for enhancing sonar performance. The spectral analysis aspects attempted in this work include, the power spectral density (PSD) characterization using suitable Probability density function (PDF) fitting tool, the statistical analysis of power spectra using moments obtained from PSD, with respect to diurnal and month wise variations, the spectral comparison presents the difference in low shipping hours and high shipping hours based on local shipping traffic timings and the differentiation of the dominant source of noise namely wind and shipping noise using correlation coefficient as a directive. The analysis has been undertaken in very shallow waters closer to the coast in the west coast of India using multiple sensors placed laterally apart. The timing of the data recording was such that it captured the diurnal environmental variation of a tropical region and also the recording area was closer to a busy port to capture the shipping noise characteristics. Minimal information is available in the open source regarding spectral analysis of the ambient noise in the tropical littoral waters of the Indian coast.