Search

This paper presents the technical design and validation of an autonomous surface vehicle (ASV) system capable of robotic search, localization and autonomous tracking of free-ranging fish carrying acoustic transmitter tags. The system consists of three identical differential-thrust catamaran vehicles equipped with time-synchronized acoustic telemetry receivers capable of accurate timestamping and decoding of signals received from individually coded tags. Multi-vehicle signal receptions allow time-difference of arrival (TDoA) measurements and are used in position estimators to enable collaborative fine-scale tracking and mapping of the movements of individual fish. The performance of the system was validated through several sea trials culminating in a deployment over several days in a Norwegian fjord where migrating wild Atlantic salmon smolts and sea-run brown trout had previously been tagged using acoustic transmitters in a nearby river. The three vehicles were able to operate autonomously over the entire deployment, carrying out coordinated searches along shorelines, keeping stations at hot-spot locations, and undertaking formation control while pinpointing fish positions. For the total of 5112 tag detections made during the deployment, 17 unique tag identifiers (fish) were registered at least twice, contributing significantly to the ongoing study of migrating salmonids in the area. Design and validation of a system of autonomous fish tracking vehicles

Over the last few decades, acoustic fish telemetry has developed into a viable approach for remote monitoring of fish behaviour in the marine environment. Simultaneously, unmanned surface and underwater vehicles have found extensive use as scalable, persistent and cost-effective platforms for ocean observation. Equipping such robotic vehicles with fish telemetry receivers constitutes an emerging approach with the potential to significantly push the current operational limits of fish movement studies at sea. Here, we present an energy-autonomous robotic fish telemetry platform realised through the integration of a real-time acoustic receiver into an ocean-going wave- and solar-powered unmanned surface vehicle. The vehicle frame and energy harvesting solutions are based on the commercially available AutoNaut USV, while the vehicle’s control and communication systems were developed using open-source software and standard hardware components. The open architecture permitted deep integration of the acoustic receiver as a system-level payload making fish detections and other sensor data available in real-time to the vehicle’s onboard control system. The vehicle is thus prepared with local situational awareness to support autonomous control during vehicle-fish encounters, as well as conventional interfaces for remote piloting and data management through long-range wireless communication links and the Internet. The vehicle concept was investigated theoretically and experimentally in an acoustic range test and a full-scale sea trial. When driven passively by waves, tag detection performance was comparable to that of traditional moored receivers, while activation of the auxiliary electric thruster caused a reduction in detection radius of more than 50%, confirming wave-power as the ideal mode of propulsion. Finally, by deploying the AutoNaut over a period of several days at the outskirts of a Norwegian fjord during the seaward migration of Atlantic salmon post-smolts, we demonstrated that the vehicle was able to detect an acoustically tagged post-smolt into the open ocean beyond the reach of the fjord’s stationary receiver grid, while using only wave- and solar energy harvesting to power its operation. The ability to observe small individual fish in the ocean environment using an energy-autonomous robotic vehicle creates novel and unprecedented opportunities for scientific inquiry in fish behaviour and movement ecology studies at sea.

In this study the performance of the Internet of Fish (IoF) concept, a real-time acoustic positioning and fish monitoring system, was assessed in a commercial marine fish farm in Norway. Central to the IoF concept is the Synchronisation and LoRa Interface Module (SLIM), which is a battery operated surface unit that provides distributed time synchronisation and LPWAN support to a submerged digital acoustic receiver. Six SLIM/acoustic receiver pairs were placed inside a fish cage with acoustically tagged fish at a link-length of 200m from a centralised gateway. All nodes achieved a Packet Error Rate of less than 8% and a position accuracy of 1.5m. This work is licensed under CC BY-NC-ND

A study of migration timing and marine migration behaviour of Atlantic salmon (Salmo salar) and brown trout (Salmo trutta) smolts/post-smolts was conducted using acoustic telemetry in the Stryneelvaeelva and the adjacent Nordfjord fjord system, Western Norway. Juvenile fish abundance and macroinvertebrate communities were also sampled in the rivers Stryneelva, Erdalselva and Hjelledøla. Juvenile salmon abundance was moderate in Stryneelva, low in Hjelledøla and not present in Erdalselva. Presence of 0+ salmon juveniles in Hjelledøla, upstream of the lake Strynevatnet, indicates successful spawning at this location. Low abundance of juvenile brown trout (0+: 98 % survival probability/km was calculated in all fjord sections. The resulting total mean fjord-migration survival probability was 34 % (±6.2 % SE), with much higher survival probability of early vs late migrants (60 % vs 27 %). Brown trout post-smolts exhibited large among-individual variation in fjord area use. Of the 34 migrants, 5 (14.7 %) migrated to outer fjord areas (>55 km), while 23 (67.6 %) remained in the inner fjord. Seven brown trout smolt individuals (20.6 %) were recorded as return migrants by November, with an average sea residency of 107.7±31.7 days (min-max: 55–159 days). Depth use was predominantly the upper 2 m of the water column, with less individual variation than for salmon. This study has provided novel knowledge about the migratory behaviour of salmon and brown trout smolt from Stryneelva. The results show that a large fraction of the salmon smolt (the early migrants) migrated to the outer fjord areas two weeks earlier than previously thought and used as guidance for both smolt-sea lice monitoring and also for assessment of sea lice infection pressure from local aquaculture industry. We also want to express concern about the population situation for brown trout in Stryneelva. The low juvenile densities documented in this project combined with low and decreasing sea trout catches and spawning-stock counts strongly indicate a critical population status.

Atlantic salmon with initial body weight 0.6 kg kept in 3.3 m3 tanks was fed one meal daily, either by dropping the feed from one point, or by spreading the feed over the water surface, for one month. The relative feed intake (% of body weight per day) in salmon fed without spreading or with spreading of the feed was 0.63±0.05 and 0.64±0.02 %, respectively. The growth rate was identical in salmon fed without or with spreading of the feed, and no difference in variance in final weight was found. Fin damage was given a score at termination of the trial as a measure of competitive behaviour during feeding. No difference in fin damage for salmon fed without or with spreading of the feed was revealed. In conclusion, the data show that for the conditions used in the trial, spreading of feed had no influence on feed intake or growth of salmon.