Integrating echo-sounder and underwater video data for demersal fish assessment

Abstract

Effective management of demersal fish species includes accurate, spatially explicit assessments of their abundance and distribution. Non-extractive techniques, such as echo-sounders and visual census, are of particular importance in no-take marine reserves where direct sampling is restricted. This study is investigating the use of echo-sounders and Baited Remote Underwater stereo-Videos (stereo-BRUVs) in demersal fish assessment. Echo-sounders have the advantage of covering nearly the entire watercolumn and being able to cover vast areas in a relatively short period. However, "ground-truth" data is usually needed to produce species-specific identification and sampling the area close to the seafloor is problematic, particularly for demersal species in complex topography. In contrast, stereo-BRUVs allow identification to species level in most cases, but samples characterise a particular location within the field of view and the area of influence within the bait plume. The combination of co-located bathymetric and habitat maps, with quantifiable acoustic backscatter and species-specific visually ground-truthed relative abundance, holds potential to further these studies and provide a more cost- and labour-efficient sampling regime. The preliminary investigation into the relationship between active acoustic and stereo-BRUVs showed a significant correlation between the relative biomass recorded by the stereo-BRUVs and the acoustic energy recorded by the echosounder.