Detection and Clustering of Acoustic Backscatter from Hydrothermal Vent Plumes

Abstract

Detection, localization, and characterization of active hydrothermal vent sites may be accomplished as part of a seafloor survey conducted in a single sortie of an autonomous underwater vehicle (AUV) equipped with a forward-looking sonar (FLS) and non-acoustic environmental sensors. This concept was investigated by processing and analyzing 396 kHz acoustic backscatter from hydrothermal plumes recorded by the Cabled Observatory Vent Imaging System (COVIS). COVIS was primarily a multibeam sonar system mounted on a stationary seafloor platform in 2197 m of water depth near a hydrothermal vent cluster on the Endeavour Segment of the Juan de Fuca Ridge (Bemis et al., 2015). Consistent plume detections were obtained with a 12-dB threshold applied to images of range-independent signal-to-clutter ratio formed directly at the output of the sonar beamformer. Given that the volume scattering strength (VSS) at nominal acoustic frequencies of 200 kHz to 400 kHz is between −60 and −75 dB re 1 m−1 for acoustic backscatter due to the temperature-driven turbulence of vent fluids in the water column, a secondary VSS threshold of −60 dB re 1 m−1 was used to discriminate against potential echoes from demersal fish or the seafloor. The resulting detected samples were clustered with a k-medoids partitioning algorithm, and each cluster was fitted with an ellipse whose map-relative coordinates can be saved into a 6-element feature vectors. At the end of a preprogrammed seafloor survey, the AUV can then direct itself to the highest mode of a 2-D histogram of cluster centers in the map-relative frame. The vehicle can then reacquire the target with the FLS and proceed with characterization of the presumed vent site with non-acoustic sensors.