Reducing uncertainty in seafloor fluid vent localization

Abstract

Modern water-column-imaging multibeam sonars have been shown to be effective tools for a variety of ocean mapping applications but generate immense amounts of raw data when recording acoustic backscatter over the entire water column. High data acquisition rates can pose logistic, economic, and technical challenges for rapid processing, analysis, and archiving of these data. These limitations in multibeam water column imaging often provide unique challenges in commercial marine seep hunting surveys that routinely acquire large basin-scale high-resolution multibeam datasets that require rapid processing and interpretation required for selection of coring targets for geochemical sampling of seep sediments. Interpreting the seafloor position of gas emissions in multibeam water column data using common commercial software packages is hindered by slow processing due to these large file sizes, a manual “by eye” qualitative assessment of each sonar ping searching for acoustic anomalies, skill and experience of the interpreter, fatigue of the interpreter during field operations, and environmental or acquisition artifacts that can mask the location of gas emission on the seafloor. These restrictions over regional basin-scale surveys create a qualitative data set with varying inherent positional errors that can lead to missed or incorrect observations about seep-related seafloor features and processes. By vertically integrating midwater multibeam amplitude samples over a desired range of depths, a 2D integrated midwater backscatter raster can be generated and draped over bathymetric data, providing a quantitative synoptic overview of the spatial distribution of gas plume emission sites for enhanced seafloor interpretation. We reprocess a multibeam midwater data set from NOAA Cruise EX1402L2 in the northwestern Gulf of Mexico using a vertical amplitude stacking technique. Constructed midwater backscatter surfaces are compared with digitized plume positions interpreted during EX1402L2 for a comparison into assessing uncertainty in mapping approaches. Our results show that the accuracy of manually digitizing gas emission sites varies considerably when compared with the midwater backscatter amplitude maps. This quantitative plume mapping technique offers multiple advantages over traditional geopicking from cost effectiveness, offshore efficiency, mapping repeatability, and ultimately improving the detectability of gas plume emission on the seafloor. This study shows datasets generated from this method can be reliably be used as a geophysical proxy for locating chemosynthetic and related benthic habitats.