Using split beam, broadband acoustics to measure behaviour of salmon in normal and sea lice preventative submerged cages with either air dome or bubbles for swim bladder refilling

Abstract

Emerging submerged sea cage technologies reduce salmon lice infestations, creating a vertical mismatch between the salmon host and surface-dwelling parasitic larvae. The method challenges swim bladder refilling at the surface and alternative air supply could be provided from underwater air domes or bubbles. When new technology requires adaptive behaviours, adequate monitoring is needed to ensure welfare. Acoustics have potential to provide automated, quantitative, and non-invasive behavioural observations 24/7, but need to be tested for this specific purpose. Here split-beam broadband echosounders are used on submerged salmon swimming at commercial densities, with aim of detecting differences in individual and group behaviours related to the amount of air in the swim bladder. About 3000 salmon of 0.7-1 kg are observed in three replicated 2000m3 cages, either provided with normal surface access (control) or submerged to 1 m with air supply from an air dome or air bubbles released at 10m depth (treatments). Transducers were positioned under each cage, pointing upwards, observing 1 treatment (3 out of the 9 cages) at the time. Swimming speed and Target Strength (TS) on individual level and vertical distribution and volume backscatter (Sv) on group level was measured continuously with broadband and narrowband signals. Analysis shows that a stocking density of 1.2 kg m-3, with typically observed schooling densities of 8-16 kg m-3, was too high to track individuals in the main school. Thus, only individuals swimming below the densest areas were studied. These showed behaviours connected to swim bladder fullness: Fish receiving air bubbles had 75% less air in their swim bladders (and some water), reflected in lower TS (-30dB) compared to control TS (-26 dB) and little diurnal swim speed variation (6% slower at day vs night). Fish in the dome treatment had TS values (-26 dB) reflecting a normally filled swim bladder, equal to control fish, and swam 20-30% slower at night compared to day, and diurnal vertical distributions. Sv co-varied with TS and was clearly lower in the bubble cages than dome treatment and control. Swimming speeds measured with camera at day were 6-48% lower than from echo sounders, possibly due to unresolved technical issues in the algorithm and difficulties in tracking at high densities. Submerged salmon seem to cope with air supplied from air domes, but not bubbles. Overall, using broadband split-beam echosounders for tracking behaviour of individual salmon in submerged cages for welfare monitoring has potential, but needs further development and automation to be used in commercial farms at higher densities.