Ambient noise and signal uncertainties during the summer shelfbreak primer exercise

Abstract

Uncertainties in noise Ievel, andin signallevel after long-range (42 km) acoustic shallow water transmissions, from a pulsed source, are determined from the summer shelfbreak PRIMER experiment. Fluctuations over the 1 0-day period are not stationary, but are rendered so by tracking the wandering of their means. Then narrow-sense stationary probability density functions are obtained of ambient noise and signal peak transmissions from a match-filter output with time-bandwidth product = 1. The data are centered at 400 Hz, in a 100 Hz bandwidth, and analyzed from three individual hydrophones of a vertical line array. The ambient noise fluctuations closely follow the phase-random Log-Rayleigh density, with standard deviation cr = 5.6 dB. Signal peak statistics are determined from demeaned 50 s samples. The signal statistics over an 8-h period are approximately similar, but not identical, to those over the entire 10-day period. The signal has narrower histograms (cr "' 0.8 dB) than the noise. Log Chi-Square densities, fit to the signal histograms, suggest that about 30 equal intensity components contribute to the fluctuations, many more than can be attributed to the idealized modal structure in the shelfbreak duct. This suggests that either strong scattering affects the signal transmission, or the signal process is not fully phase-random.