Eigenvector Subspace Projection Applied to the Analysis of Shallow Water Array Noise Gain Data

The method of eigenvector subspace projection is applied to data collected on a bottom-mounted horizontal line array located on the U. S. continental shelf east of Ft. Lauderdale, FL, in order to investigate the impact of the ocean environment on array noise gain (ANG). Array sidelobe response associated with the continual passage of ships into nearby Port Everglades tends to obscure the impact of environmental forcing mechanisms on ANG. In this work, we demonstrate the application of dominant eigenvector nullspace projection to coherently filter the discrete noise spectrum in order to investigate the impact of wind-driven surface noise on ANG statistics using the continuous noise spectrum. In particular, we examine endfire beam ANG data concurrent with measured wind speed during a period sampling two wind speed extremes, ≤ 5 kn and ≥ 15 kn, on the same day in September, 2007 for both lowand highshipping densities. As expected, when the discrete noise spectrum dominated the noise field, ANG fluctuations correlated with the periodic passage of local ships near the array. When nullspace projection was employed, the ANG stabilized, and a correlation was revealed between endfire ANG and wind speed as it cycled between low and high extreme over a 16-hour period. We review the implementation of the nullspace projection technique for decomposing array response. The analysis will illustrate that 1) array noise gain in the South Florida continental shelf environment is almost never 10log10N due to the anisotropy of the noise field, and 2) using concurrent meteorological data, it is possible to empirically validate the expected link between wind-generated surface noise and array noise gain. The results illustrate the role that eigenvector projection can play in the understanding of array noise gain and the impact of environmental forcing mechanisms on array performance.