Premises for digital twins reporting on Atlantic salmon wellbeing

Many species of fish, birds and mammals commonly live in human captivity; Atlantic salmon Salmo salar is one of them. The international legal status of the welfare of captive animals is slowly developing and still requires rigorous specification. For example, even though fish have complex cognition and elements of sentience, The United Nations' animal welfare principles still take a functional health-centred perspective overlooking the cognitive-affective component. Wellbeing problems remain a major source of slow growth and high mortality in intensive aquaculture of Atlantic salmon. The value system for decision making in vertebrates is based on expectations of emotional wellbeing for the options available and is linked with the individual's assessment of its future. We propose a new approach for monitoring and improving the welfare of salmon (or any other captive or wild vertebrate) based on modelling the salmon's wellbeing system by digital twins, which are simulation models that implement major bodily mechanisms of the organism. Indeed, predictions on boredom, stress and wellbeing can all be captured by a computational evolutionary model of the factors underlying behaviour. We explain how such an agent-based model of salmon digital twins can be constructed by modelling a salmon's subjective wellbeing experience along with prediction of its near future and allostasis (the bodily preparation for the expected near future). We attempt to identify the building blocks required in digital twin models to deliver early warnings about escalating issues that could eventually lead to negative effects on salmon health in aquaculture. These models would provide critical insights for optimizing production processes and could significantly reduce the reliance on animal experiments. Overall, reports of a population of digital twins could support the implementation of 3Rs - replacement, reduction, refinement - by offering actionable information to fish farmers as well as consumers, voters, politicians and regulators on relevant issues as well as guide experimental work on animal wellbeing across species.