ASA 125th Meeting Ottawa 1993 May

2pUW9. Stochastic analysis of a three-dimensional seismoacoustic field.

Alexander H-D. Cheng

Dept. of Civil Eng., Univ. of Delaware, Newark, DE 19716

Mohsen Badiey

Indra Jaya

University of Delaware, Newark, DE 19716

Based on geological information existing for a site on the New Jersey continental shelf [Badiey et al., J. Acoust. Soc. Am. 90, 2372 (A) (1991)], a three-dimensional geoacoustic data set was generated using various widely accepted empirical relations. A statistical analysis is performed on the acoustic properties which reveals the random, inhomogeneous nature of seabed sediment. The statistical structure in both the horizontal and the vertical directions are quantified in terms of mean, standard deviation, and correlation length. The data set is then used in a theoretical modeling by both the deterministic and stochastic techniques. The deterministic technique utilizes an anisotropic, inhomogeneous, poroelastic model. The sensitivity of the various physical parameters to the seismoacoustic responses is examined. For stochastic analysis, the Monte Carlo technique is applied. Utilizing the statistical moments obtained from analyzing the field data, a large number of random realizations of seabed acoustic properties can be generated via a spectral technique in one or multi-dimension, and uni- or multi-variate conditions. The deterministic model is then applied to study the effect of the variability of the sediment properties on the acoustic wave field. [sup a)]Presently at Office of Naval Res., Ocean Acoustics, Code 1125OA, 800 N. Quincy St., Arlington, VA 22217-5000.