Acoustic characteristics of Atlantic salmon feeding behaviour in recirculating aquaculture systems (RAS)

Could sound be a part of the solution to optimising feeding in recirculating aquaculture systems (RAS)? In land-based aquaculture facilities where the water is recirculated, it is crucial to control the feeding to secure good water quality and growth. More data is needed to optimise the feeding, and a potential data source is the sound in the fish tank, which can be referred to as the soundscape. Fish distribution in the tank, swimming patterns and speed, and responsiveness to feed will affect the soundscape in the tank. This behaviour can be related to appetite, which implies that changes in the soundscape can be used to signal hunger. In a RAS facility, the soundscape will also be affected by the design of the farm. Tank design, placement of pumps, and which equipment is used to treat the water can all affect the soundscape and potentially also fish behaviour. This background noise must be known in order to identify the sounds related to fish behaviour. Knowing the specific soundscapes of different RAS farms can help us better understand observed fish behaviour and performance during commercial production. The primary objective of this PhD project is to determine acoustic characteristics of Atlantic salmon behaviour related to feeding in RAS. Through controlled experiments we will seek to identify acoustic characteristics of the tank soundscape related to appetite. Further, data from commercial salmon farms will be collected to compare and investigate whether the same acoustic characteristics can be identified during normal operation in farms with different design. Data will be related to water quality and camera-based observations of behaviour in RAS.

Controlling the feeding is key in operating a recirculating aquaculture system (RAS). Due to the long retention time in RAS, uneaten feed pellets lead to accumulation of organic matter which is detrimental to the biofilter and increases the turbidity. Uneaten feed is also a waste of resources and constitutes a substantial financial cost, which is why new methods for controlling and optimising the feeding is being explored. This project will investigate the potential use of sound as a data source for optimising feeding of Atlantic salmon in RAS. The research regarding sound in RAS is limited, mostly focusing on sound produced by equipment (pumps etc.) and how it affects the fish. However, the fish in a tank also produce sound, and their presence affects the acoustic characteristics of the sound that is present in the tank. The sound in the tank can be referred to as a soundscape, and will be affected by fish behaviour such as feeding, swimming activity, fish density and distribution in the tank. By recording the soundscape of the tank environment in different settings where the fish behaviour is known, it is possible to correlate the acoustic characteristics of sound with behaviour. This project will collect data from controlled experiments with different feeding regimes to determine acoustic characteristics of behaviour related to feeding. To evaluate the potential use of sound to optimise feeding, the effect of different RAS designs on soundscapes will be investigated by performing a controlled experiment with two different treatments (one with ozone and one without). In addition, a survey will be performed at two commercial farms with different RAS designs to determine whether acoustic characteristics related to feeding can be found during normal operation. This project aims to increase our knowledge about sound in RAS and to explore the possibility of utilising sound recordings to optimise feeding of Atlantic salmon in RAS.