Application of acoustic classification in different sedimentary environments: A case study of the Gaoping slope in the southwest coast of Taiwan

Abstract

• Acoustic remote sensing can be used to determine different types of sedimentary seabeds using quantitative classification. • Three attributes extracted from acoustic waveforms reveal four types of sedimentary seabeds. • The proposed method differentiates sedimentary seabed types with similar grain size distribution. • Core data from the offshore area of southwestern Taiwan validates the method. • This study provides a sensitive and robust quantitative method and reduces the uncertainty of manual identification between different seabeds. The sedimentary seabed properties are direct indicators of its stability, whose measure has several important implications in understanding and forecasting geological events. This study investigates the classification of sedimentary seabed types in different depositional environments using the three attributes: reflection coefficient (RC), similarity index (SI), and amplitude ratio (AR) values extracted from sub-bottom profiling data. Four sedimentary seabed types were evaluated automatically according to the waveforms in the horizontal and vertical directions. To infer the seabed properties, the attribute SI is used to identify the occurrence of homogeneity of sediment transportation in an area, while RC and AR are used to determine the seabed characteristics in terms of the variation of acoustic energy reflected from the sediment. To validate the results of this quantitative analysis method, sixteen seafloor sediment samples collected at different sites in the offshore area of southwestern Taiwan were analyzed. The automatic classification results suggest that the different sub-bottom profiling images are associated with the grain size distributions and compositions, such as the concentrated distribution, sand, silt, and clay percentages, and the mean bulk density. In addition, an inhomogeneous environment with similar grain size distribution can be identified. The proposed seabed classification is an effective method to study seabed characteristics, especially for identifying chaotic sediments.