Environmental inversion using the singular value decomposition modes of multiple frequency vertical line array data

Abstract

A method is presented for geoacoustic inversion that uses the depth-dependent normal modes extracted from vertical line array (VLA) data. The normal modes for multiple frequencies are obtained by forming a cross-spectral density matrix (CSDM) for each frequency of VLA data, measured as a source moves outward in range. The singular value decomposition (SVD) of each CSDM yields the depth-dependent normal modes of the ocean environment at the time the signal was recorded [T. B. Neilsen and E. K. Westwood, J. Acoust. Soc. Am. 101, 3025 (1997)]. The data-extracted mode functions are then used to invert for the environmental parameters. The Levenberg-Marquardt nonlinear optimization method is used to find the best fit between the data-extracted and modeled mode functions by adjusting the environmental parameters. The algorithm requires computation of the first partial derivatives of the mode functions with respect to the varied parameters, which is achieved using the ORCA normal mode model. The method is useful for determining the sound speed profile and the characteristics of the uppermost bottom layers. Results using simulated data will be presented. [Work supported by U. S. Army Research Office, AASERT Grant number DAAH04-95-1-0486, and ONR.]