Benchmarking range-dependent seismo-acoustic propagation problems

Abstract

Developing the capability to accurately solve range-dependent problems in ocean acoustics has been going on for many years. A milestone in this development was the ASA benchmark effort in 1990 [J. Acoust. Soc. Am. 87, 1497–1545 (1990)], where a set of models was applied to generic range-dependent acoustic fluid problems to establish a set of benchmark solutions for model verification purposes. A further benchmarking effort, which included shear waves in the bottom, was published in 1997 [IEEE J. Ocean. Engn. 22, 226–236 (1997)] involving two seismo-acoustic codes based on spectral-integral techniques and a boundary-element code. Other codes have become available since then for solving general range-dependent elastic problems, either based on marching schemes that ignore backscattering (or adopt a single-scatter approximation) or techniques that solve the full boundary-value problem. In an attempt to establish a set of benchmark solutions for seismo-acoustic problems, we solve several low-frequency elastic problems involving changing bathymetry in multiple layers, and coupling between various wave types, including interface waves. The models in use or used are a parabolic equation model (RAM) and a commercial finite-element model (FEMLAB).