Stochastic modeling of seafloor morphology.

Abstract

One of the basic requirements for modeling ocean bottom scattering and acoustic reverberation is a realistic quantitative description of the seafloor. Because of the small scales usually required (say ∼1 m or smaller) for such modeling, deterministic descriptions of the seafloor are often prohibitive. Fortunately, the gap between what is needed by acoustic modelers and what can be delivered by seafloor mappers can be narrowed or eliminated through the application of stochastic modeling. In the methodology presented here, the statistical characteristics of abyssal hill morphology (the predominant ocean bottom morphology) are modeled using parametrized forms of the topographic moments to order four. The first moment, or mean, is considered known and removable. The model for the second moment, or covariance function, contains parameters describing the rms height, characteristic length and width, azimuthal orientation, and fractal dimension. Parametrization of the third and fourth moments provides information regarding the vertical and lateral asymmetry and the degree of peakiness. The ability of the model to interpolate stochastic behavior to scales smaller than the instrument resolution is investigated using coregistered bathymetry from surface and deep-tow instruments.