A method for identifying Sound Scattering Layers and extracting key characteristics

Abstract

Mid-trophic level water-column (pelagic) marine communities comprise millions of tonnes of zooplankton and micronekton that form dense and geographically extensive layers, known as sound scattering layers (SSLs) when observed acoustically. SSLs are ubiquitous in the global ocean, and individual layers can span entire ocean basins. Many SSLs exhibit clear diel vertical migration behaviour. Vertical migrations contribute substantially to the 'biological pump', such that SSLs have important global biogeochemical roles: SSLs are important conduits for vertical energy and nutrient flow. Ship-based remote sensing of SSLs using acoustic instruments (echosounders) enables their shape and density to be quantified, but despite SSLs being discovered in the 1940s, there is no consistent method for identifying or characterising SSLs. This hampers ecological and biogeographical studies of SSLs. We have developed an automated and reproducible method for SSL identification and characterisation, the sound scattering layer extraction method (SSLEM). It functions independently of echosounder frequency and the spatial scale (vertical and horizontal) of the data. Here we demonstrate the SSLEM through its application to identify SSLs in data gathered to a depth of 1000 m using 38 kHz hull-mounted echosounders in the south-west Indian Ocean and Tasman Sea. SSLs were identified in the water column as horizontally extensive echoes that were above background noise. For each identified SSL, a set of 9 quantitative 'SSL metrics' (describing their shape, dynamics and acoustic backscattering distribution) were determined, enabling inferences to be made concerning the spatial arrangement, distribution and heterogeneity of the biological community. The method was validated by comparing its output to a set of visually derived SSL metrics that were evaluated independently by 8 students. The SSLEM outperformed the by-eye analysis, identifying three times the number of SSLs and with greater validity; 95% of SSLs identified by the SSLEM were deemed valid, compared to 75% by the students. In the same way that data obtained from satellites have enabled the study and characterisation of global phytoplankton distribution and production, we envisage that the SSLEM will facilitate robust, repeatable and quantitative analysis of the growing body of SSL observations arising from underway-acoustic observations, enhancing our understanding of global ocean function.