Implantable sensing technologies for monitoring of behavioural and physiological dynamics in farmed Atlantic salmon

• Main Author: Svendsen, Eirik
• Other Authors: Alfredsen, Jo Arve, Føre, Martin, Olsen, Rolf Erik
• Format: Doctoral or Postdoctoral Thesis
• Language: Norwegian Bokmål
• Published: NTNU 2023
• Subjects: VDP::Technology: 500::Information and communication technology: 550::Technical cybernetics: 553; Atlantic salmon
• Online Access: https://hdl.handle.net/11250/3095142

Atlantic salmon aquaculture is a major food industry with 1.5 million tonnes produced for worldwide consumption in 2022. However, in Norway during the same year, current farming methods resulted in a 17.1% mortality rate (56.7 million individuals). Not only does this reduce profitability, it also leaves room for considerable improvement with respect to responsible marine aquaculture. Part of the losses in Atlantic salmon farming can be attributed to lack of objective data for the fish’s behavioural and physiological responses and how they are related to health and welfare in different farming contexts. The main objective of this PhD project has therefore been to enable realtime data collection of fish parameters during full scale farming operations by creating technologies for behavioural and physiological monitoring to uncover their relation to stress and, by extension, welfare. To this end, it was first shown that the connection between measured behavioural and physiological responses and stress could be made using existing off-the-shelf implants measuring heart rate and motion and blood sample analysis. The sensing principle of pulse oximetry was then selected because it would add to the suite of physiological parameters possible to obtain in real-time during full scale farming operations. The sensing principle was validated by measurement of the optical properties of Atlantic salmon blood. These measurements showed that its optical properties were such that pulse oximeters intended for humans can be used for this species. This was further verified by in vitro measurements of oxygenated and deoxygenated whole blood samples using an off the shelf miniaturized pulse oximeter and Monte Carlo simulations. Because welfare considerations indicated that sensors should be implanted in fish, eight different locations within the peritoneal cavity of live, anesthetized Atlantic salmon were tested. The location giving the best signals across individuals in terms of signal quality index was then used as input to hardware ...