FULL-COVERAGE MONITORING OF FISH DISTRIBUTION IN AQUACULTURE SEA CAGES USING A DUAL TWO-DIMENSIONAL SONAR COMBINATION SYSTEM

Introduction

Monitoring fish behaviour, such as swimming speed and spatial distribution, in aquaculture sea cages is essential for optimizing feeding management and improving fish welfare. Stereo cameras have been widely used to measure swimming speed; however, they rely on machine-learning-based image processing, which demands substantial computational resources and power, and their performance is degraded by low illumination and biofouling of the camera housing (Ikegami et al., 2026). In addition, stereo cameras provide only a local field of view and cannot capture the spatial distribution of fish across the entire cage. Multibeam sonars offer a light-independent alternative, but their limited field of view makes it difficult to cover an entire commercial sea cage with a single unit. Placing the sonar near the cage wall extends the observable area but introduces detection-range limitations (Banno et al., 2025). A mechanically rotating single-beam sonar overcomes both range and field-of-view limitations by scanning 360° from the cage centre (Sauphar et al., 2026); however, the rotation mechanism limits temporal resolution, and swimming speed cannot be derived from this approach.

In this study, we present a dual two-dimensional (2D) sonar combination system that addresses both the field-of-view and the real-time monitoring challenges simultaneously. Each sonar unit has a fan-shaped beam with an opening angle of 120°. By orienting two units perpendicular to each other, the system achieves near-complete coverage of the cage cross-section without any mechanical rotation, enabling continuous monitoring of swimming speed and depth distribution. Furthermore, all the approaches mentioned above provide only two-dimensional observations, whereas Atlantic salmon (Salmo salar) are known to exhibit three-dimensional spatial heterogeneity within cages. To address this, we additionally conducted a preliminary verification of three-dimensional (3D) monitoring using a multibeam sonar equipped with a rotational scanning mechanism.

Materials and Methods

Two field experiments were conducted. The dual 2D sonar experiment was carried out in September 2025 at the NTNU salmon aquaculture research site Gudmundset, on the west coast of Norway, in a commercial sea cage containing Atlantic salmon post-smolts. Two multibeam sonars (FCS-500, FURUNO ELECTRIC, Japan; 120° beam opening angle per unit) were deployed as a twin unit at the centre of the cage just below the water surface, with their fan beams oriented perpendicular to each other; this configuration constitutes the dual 2D sonar combination evaluated here. Recordings were obtained continuously over extended observation periods covering different times of day. From the calibrated volume backscattering strength (Sv), fish density was mapped on the depth–time plane and individual fish were tracked frame-to-frame to estimate the swimming speed. The 3D verification experiment was conducted in December 2025 in a commercial sea cage at Tokushima, Japan, containing yellowtail (Seriola quinqueradiata); a multibeam sonar equipped with a rotational scanning mechanism was installed at the cage centre, and the transducer was automatically rotated to produce three-dimensional snapshots of the spatial distribution of fish within the cage volume.

Results and Discussion

The dual 2D sonar system successfully captured spatial and temporal variations in fish distribution across the cage. Both swimming speed and depth distribution showed clear changes depending on the time of day, demonstrating the system's capability to quantitatively track diurnal behavioural patterns in real time under practical farming conditions. The estimated mean swimming speeds were approximately 0.6 m/s during daytime and 0.3 m/s during night-time. The rotational multibeam sonar confirmed the feasibility of capturing the volumetric distribution of fish within the cage.

The proposed dual 2D sonar system provides a practical, cost-effective, and real-time solution for wide-area behavioural monitoring in commercial sea cages, overcoming the key limitations of existing methods. Further integration of the dual 2D system with the 3D rotational scanning approach is expected to enable comprehensive whole-cage monitoring that accounts for the three-dimensional heterogeneity of fish distribution. Future work will include verification of the 3D approach with Atlantic salmon.

Figure

The echo of the dual 2D sonar and 3D system.

References

Banno, K., Gao, S., Stolz, C., Tuene, S.A., Aas, G.H., Gansel, L.C. (2025) Multi-beam sonars for monitoring the distribution and welfare of fish reared in sea cages – strengths and challenges. Aquaculture 606, 742553. https://doi.org/10.1016/j.aquaculture.2025.742553

Ikegami, A., Takahashi, Y., Komeyama, K. (2026) Spatio-temporal-dependent characteristic evaluation of yellowtail (Seriola quinqueradiata) in aquaculture cages based on stereo camera measurements. Aquaculture 612, 743251. https://doi.org/10.1016/j.aquaculture.2025.743251

Sauphar, C., Bolinches, P., Gygax, L., Tuene, S.A., Aas, G.K.F.H., Gansel, L.C. (2026) Monitoring behavior of post-smolts Atlantic salmon (Salmo salar) during their first month after sea transfer using a mechanical 360-degree single-beam scanning sonar. Aquaculture 620, 743909. https://doi.org/10.1016/j.aquaculture.2026.743909