Identification and abundance estimation of mesopelagic fish in the Namibian Benguela using multifrequency acoustics

Abstract

Mesopelagic species are one of the least studied components of the ocean ecosystem, with major uncertainties about their biomass. High abundances of fish, jellyfish, krill and other plankton often occur in mixed aggregations in the Benguela Current region, which challenges conventional acoustic target identification. Multifrequency acoustic techniques were developed to distinguish mesopelagic fishes from other major co-occurring sound-scattering organisms, such as jellyfish, krill and micronekton (here ‘plankton’). Frequency responses were extracted from acoustic reference data collected by the RV Dr. Fridtjof Nansen in the Namibian Benguela in April–May 2019. Acoustic targets were sampled using pelagic sampling trawls and verified with an in-trawl camera system. The acoustic classification algorithm was applied to data collected during surveys in the northern Benguela region in 2017 and 2019 to estimate mesopelagic fish biomass, and its performance was evaluated by visual inspection of echograms, comparing trawl catches with acoustic estimates, and comparing estimates derived from daytime and night-time acoustic data, with and without the assistance of trawl verification. Biomass estimates of the target category mesopelagic fish from the different approaches ranged from 8.0 × 10⁵ to 1.4 × 10⁶ tonnes in 2017, and from 1.0 × 10⁶ to 1.3 × 10⁶ tonnes in 2019. This study demonstrates that the use of frequency responses for mesopelagic fish, krill and ‘plankton’ may lead to improved acoustic biomass estimates in this region.