Kyle M. Becker
Graduate Program in Acoust., Penn State Univ., Appl. Res. Lab., Student Area, P.O. Box 30, State College, PA 16804
A problem of interest to ocean environmental acousticians is understanding the relationship between ocean surface characteristics and acoustic backscattering statistics. Recent experimental work has focused on determining surface properties which cause backscattering strength statistics to deviate from exponential. Using the nomenclature of Ogilvy (Theory of Wave Scattering from Random Rough Surfaces, 1991) several scattering surfaces have been fabricated. Surfaces were modeled using a moving average technique with prescribed distribution functions and correlation functions C(R). Representing surface elevation, h, as a continuous random process, the moving average process of order N is given as h[sub n]=(summation)[sub l=0][sup N]w[sub l]u[sub n-1]. Statistical properties of h are determined by the random uncorrelated variables, u[sub n], and the choice of weights, w[sub l]. Weights are determined by the choice of C(R) and corresponding correlation length, (lambda)[sub 0], or distance over which C(R) decays by 1/e. Isotropic and anisotropic surfaces have been constructed for both Gaussian and non-Gaussian surface height distributions. Acoustic backscattering results are presented for three cases: case I---log-normal height distribution in x and y, equal correlation lengths; case II---Gaussian height distribution, correlation length in y equal to ten times that of x; case III---Gaussian height distribution in x, log-normal distribution in y, equal correlation lengths.