Extended detection of shallow water gas seeps from multibeam echosounder water column data

Multibeam echosounder water column data provides a three-dimensional image of features between the water surface and the seafloor. Although this swath of acoustic data can be collected over a wide range of angles, most of the data, at least beyond the range to the first seafloor return, is contaminated by noise created by receiver array sidelobe interference. As a result, the water column data beyond the minimum slant range commonly is excluded from analysis. This paper demonstrates a method to consistently filter and extract targets comprising a gas seep feature across the multibeam swath, including targets within the areas dominated by receiver array sidelobe interference. For each sample range, data are filtered based on the mean plus a certain number (k) of standard deviations of the sample values along that range. The filtering is coupled with a morphological classification to retain only targets of interest while excluding background data and noise. Data were collected over a shallow water artificial gas seep using two different flow rates and at three different vessel speeds. Using the proposed method, 119 of 124 test seeps were identified correctly. Seep targets were identified at all angles across the water column fan up to beam pointing angles of 55°, with 19 of 23 seeps being correctly identified at angles greater than 50°. This method demonstrates that features can be extracted and geolocated in the sidelobe noise when the interference is appropriately filtered. These results will improve the areal extent of multibeam surveys and increase the utility of acoustic data in capturing information on water column targets directly above the seafloor.