Geoacoustic inversion

点击翻译按钮获取中文摘要

GEOACOUSTIC INVERSION: Seabed geoacoustic parameters, such as sediment layering structure, sound speed, density, attenuation and scattering strength, strongly affect the propogagion of ocean acoustic fields. Hence, knowledge of these parameters is required to understand and model sound propagation in shallow-water environments, as required for many sonar applications (as well as geotechnical and geological purposes). Estimating seabed parameters from ocean acoustic measurements provides an economical and convenient in situ alternative to direct sampling (e.g., coring), with a sensitivity relevant to sonar. However, this requires solving a strongly nonlinear inverse problem that is inherently nonunique (i.e., a range of solutions can acceptably fit the data). A variety of geoacoustic inversion approaches have been applied, with varying degrees of success, including the inversion of full-field acoustic measurements, acoustic reverberation, seabed reflectivity, ambient noise, and ship noise. These approaches differ in their ability to estimate various geoacoustic parameters and to resolve parameter variations as a function of depth and/or range; however, the efficacy of these approaches has not been examined in a systematic or rigorous way. The overall objectives of this proposal are to develop and quantitatively study and compare the above-mentioned approaches to geoacoustic inversion. A key component of quantitative analysis of geoacoustic inversion is estimating the uncertainties of the recovered parameters, which indicates the information content of the data and provides an objective basis for comparison of different data sets. Further, only with quantitative uncertainty estimation can issues such as resolving parameter spatial variability or frequency dependence be addressed in a meaningful manner (e.g., differentiating uncertainty from variability). Parameter uncertainty estimation in nonlinear inversion is itself a challenging research problem. This proposal is based on a Bayesian inference formulation and a numerical (statistical) sampling approach, under development, that builds up probability distributions for the unknown seabed parameters based on the data and available physical constraints (prior information).