SEDIMENT CLASSIFICATION BASED ON IMPEDANCE AND ATTENUATION ESTIMATION

This paper presents a remote marine sediment classification model that can be implemented in real time while underway in a survey. The model is based on the estimation of impedance and attenuation of subbottom sediments from normal incident reflection seismograms. A robust impedance inversion model utilizing layer detection was developed which is implemented with significantly less computation when compared to full inverse methods, and hence runs in real time. A new ‘‘weighted least‐squares fitting’’ procedure is proposed for evaluating the impulse response of the sediment column. The acoustic attenuation in the sediment is determined by measuring the frequency shift of the pulse spectrum using an instantaneous frequency method. The impedance inversion model requires an input of the estimated attenuation to account for the loss in signal energy due to absorption. A recently developed model relating sediment acoustic properties to sediment physical properties for a given depositional environment is employed. The constants appearing in the classification model are evaluated using measurements from a few core samples. Impedance and attenuation estimates are used to predict sediment properties such as porosity, density, mean grain size, and sound speed. The reflection data for the present study were acquired by a linear wideband (full spectrum) sonar. It is used because of its linear system components, high resolution, and wide bandwidth. Analysis of acoustic data acquired by the full spectrum sonar demonstrates the feasibility of remote acoustic seafloor sediment classification.