Benthic habitat mapping using multibeam sonar systems

Abstract

The aim of this study was to develop and examine the use of backscatter data collected with multibeam sonar (MBS) systems for benthic habitat mapping. Backscatter data were collected from six sites around the Australian coastal zone using the Reson SeaBat 8125 MBS system operating at 455 kHz. Benthic habitats surveyed in this study included: seagrass meadows, rhodolith beds, coral reef, rock, gravel, sand, muddy sand, and mixtures of those habitats. Methods for processing MBS backscatter data were developed for the Coastal Water Habitat Mapping (CWHM) project by a team from the Centre for Marine Science and Technology (CMST). The CMST algorithm calculates the seafloor backscatter strength derived from the peak and integral (or average) intensity of backscattered signals for each beam. The seafloor backscatter strength estimated from the mean value of the integral backscatter intensity was shown in this study to provide an accurate measurement of the actual backscatter strength of the seafloor and its angular dependence. However, the seafloor backscatter strength derived from the peak intensity was found to be overestimated when the sonar insonification area is significantly smaller than the footprint of receive beams, which occurs primarily at oblique angles. The angular dependence of the mean backscatter strength showed distinct differences between hard rough substrates (such as rock and coral reef), seagrass, coarse sediments and fine sediments. The highest backscatter strength was observed not only for the hard and rough substrate, but also for marine vegetation, such as rhodolith and seagrass. The main difference in acoustic backscatter from the different habitats was the mean level, or angle-average backscatter strength. However, additional information can also be obtained from the slope of the angular dependence of backscatter strength.It was shown that the distribution of the backscatter. The shape parameter was shown to relate to the ratio of the insonification area (which can be interpreted as an elementary scattering cell) to the footprint size rather than to the angular dependence of backscatter strength. When this ratio is less than 5, the gamma shape parameter is very similar for different habitats and is nearly linearly proportional to the ratio. Above a ratio of 5, the gamma shape parameter is not significantly dependent on the ratio and there is a noticeable difference in this parameter between different seafloor types. A new approach to producing images of backscatter properties, introduced and referred to as the angle cube method, was developed. The angle cube method uses spatial interpolation to construct a three-dimensional array of backscatter data that is a function of X-Y coordinates and the incidence angle. This allows the spatial visualisation of backscatter properties to be free from artefacts of the angular dependence and provides satisfactory estimates of the backscatter characteristics.Using the angle-average backscatter strength and slope of the angular dependence, derived by the angle cube method, in addition to seafloor terrain parameters, habitat probability and classification maps were produced to show distributions of sand, marine vegetation (e.g. seagrass and rhodolith) and hard substrate (e.g. coral and bedrock) for five different survey areas. Ultimately, this study demonstrated that the combination of high-resolution bathymetry and backscatter strength data, as collected by MBS, is an efficient and cost-effective tool for benthic habitat mapping in costal zones.