Modelling of Atlantic salmon (Salmo salar L.) behaviour in sea-cages : A Lagrangian approach

We constructed a Lagrangian (individual-based) model of the behaviour of Atlantic salmon in sea-cages. The behaviour of individual fish was affected by a set of environmental factors and the other individuals in the population. The model was based on behavioural parameters derived from literature and comparisons between model performance and published quantitative data sets. Simulations demonstrated that the model was able to predict the circular schooling patterns familiar to caged Atlantic salmon based on social interactions among the individuals and a motivation to keep a certain distance to the cage wall. The model accurately predicted the vertical distribution patterns of salmon over 24 h within a sea-cage when verified against three detailed observational data sets from autumn, implying that the model closely replicated behavioural responses toward the combined effects of temperature and light. Deviations seen in the comparisons were possibly due to unmodelled features in the feeding response. Potential uses of the model include: 1) increasing understanding of how environmental factors affect fish behaviour; 2) coupling with measurements in a model-based estimator structure to accurately estimate biomasses within sea-cages; and 3) predicting the welfare status of fish in sea-cages under specific environmental conditions.