Introduction
Ensuring fish welfare is a key aspect of modern aquaculture, as it impacts both ethical considerations and production performance. Environmental factors and routine operations on aquaculture farms can influence welfare status and lead to physiological and behavioural changes in fish, highlighting the need to assess both these responses. In this context, the development of welfare monitoring tools has become essential within the framework of precision aquaculture, facilitating decision-making on farms (Føre et al., 2017). Among the available technologies, acoustic telemetry allows the recording of fish behaviour under real farming conditions, such as in floating sea cages (Muñoz et al., 2020). However, most studies have focused on salmonids, and knowledge remains limited for Mediterranean species such as European seabass (Dicentrarchus labrax), particularly in open-sea net pens. Furthermore, hidden Markov models (HMMs) offer a useful tool for inferring behavioural states from telemetry data. These models have been widely used in wild fish populations, but their application in aquaculture studies remains limited. Consequently, the aim of this study is to analyse the behavioural patterns of seabass in open-sea net pens across different timescales, as well as to characterize their behavioural states in relation to environmental conditions using acoustic telemetry and HMMs.
Materials and Methods
The activity and vertical distribution of seabass were monitored in four experimental net-pens belonging to IRFAP-LIMIA (Spain). A total of 2,200 adult seabass (weight ≈ 1 kg) were distributed across the four cages, which measured 5.5 m in diameter and 6 – 8 m in depth (N = 550 individuals per cage). Five fish were randomly selected and implanted with acoustic transmitters (ADP-LP7; Thelma Biotel Ltd.). After 7 days of recovery in monitored indoor tanks, fish were returned to the cages, where seabass acceleration and depth were recorded for 5 months between July and November 2023. The data were recorded by 3 acoustic receivers (TBR 800; Thelma Biotel Ltd.) positioned between the cages. A total of ~1.2 million detections were processed. To account for structural differences between cages, depth was calculated relative to the centre of each cage. Linear mixed models (LMMs) and generalized additive models (GAMs) were applied to asses behavioural patterns and to explore the influence of environmental variables on the behaviour of seabass. To characterize behavioural states, a HMM with four states was developed: resting, deep swimming, surface swimming and high activity, estimating the transition probabilities between states whilst taking environmental covariates into account. All procedures were approved by the Animal Experimentation Ethics Committee (CEEA-UIB, Spain; Ref. 206/12/22) and were carried out in accordance with European Directive 2010/63/EU and Royal Decree 53/2013.
Results and Discussion
Acoustic transmitters enabled a successful detection of swimming patterns and vertical distribution of seabass reared in open-sea net pens, which showed changes throughout the five-month experimental period. On one hand, swimming activity decreased slightly over the study period, whilst the fluctuations observed in vertical distribution proved to be sporadic, with the fish remaining stable most of the time in the middle zone of the water column. Clear circadian patterns were identified: swimming activity increased significantly during the day, peaking around midday, particularly after feeding. Vertical distribution also varied throughout the day, with greater use of deeper zones during the daylight hours and a greater presence at the surface at night. Furthermore, during the summer months, an increase in activity and greater use of depth were observed. GAM models revealed a significant effect of environmental variables on the behaviour of seabass. Solar radiation showed a strong positive correlation with both parameters, and wave height and wind also affected vertical distribution. The HMM allowed the identification of the four behavioural states and the characterization of their dynamics throughout the day. A clear daily pattern was observed: at night, resting and surface swimming predominated, whereas deep swimming increased after dawn. Subsequently, following feeding, deep swimming behaviour decreased, giving way to an increase in high-activity behaviour, which peaked 3-4 hours after feeding. At dusk, a transition towards lower-activity states was observed. Transitions between states were strongly modulated by solar radiation and the daily cycle, which favoured the shift from low- to high-activity states and promoted the occurrence of active states during daylight hours. Taken together, these results are consistent with previous studies indicating that seabass behaviour in open-sea net pens is strongly regulated by environmental factors and daylight cycle (Toomey et al., 2025). The integration of acoustic telemetry and machine learning models allows for the characterization of behavioural states and their dynamics under farming conditions. This approach offers a robust framework for improving welfare monitoring and promoting welfare-oriented management of open-sea aquaculture facilities.
Acknowledgment
The authors acknowledge the staff at LIMIA-IRFAP for their assistance. Furthermore, P. A-L. was supported by a Ramón y Cajal postdoctoral grant (Ref. RYC2020-029629-I) from the State Research Agency of the Spanich Government.
References
Føre, M., Frank, K., Norton, T., Svendsen, E., Alfredsen, J. A., Dempster, T., ... & Berckmans, D. (2018). Precision fish farming: A new framework to improve production in aquaculture. Biosystems engineering, 173, 176-193.
Muñoz, L., Aspillaga, E., Palmer, M., Saraiva, J. L., & Arechavala-Lopez, P. (2020). Acoustic telemetry: a tool to monitor fish swimming behavior in sea-cage aquaculture. Frontiers in Marine Science, 7, 645.
Toomey, L., Carbonara, P., Mente, E., Troianou, E., Troianou, D., Spedicato, M. T., ... & Alfonso, S. (2025). Precision fish farming: A sensor-based study on Dicentrarchus labrax in a sea cage environment. Aquaculture Reports, 41, 102691.