Better insights during smolt production lead to safer transfer to sea
By Kristin Aase, Insight Manager at Aquabyte
Kristin has a Master's degree in Aquaculture Biology from the University of Bergen, and works closely with Aquabyte's customers to analyze data, generate insights, and identify opportunities for improvement.
A Norwegian version of this article has been published by LandbasedAQ
The foundation for a successful sea transfer
As a smolt or post-smolt producer, there are many major activities to plan, carry out, and document while the fish are at the facility: transfers, grading, vaccination, light management, and smoltification.
Ultimately, however, they all serve the same purpose: ensuring that the fish are as robust and well-prepared as possible when they undergo one of the most demanding physiological transitions in the entire production cycle – the transition from freshwater to seawater.
In Norway, 43.3 million salmon weighing more than three grams were reported lost during the freshwater production phase in 2025, and the challenges do not end after sea transfer. According to Fiskehelserapporten 2025 (The Fish Health Report 2025), mortality in Norway remains relatively high among fish weighing less than 0.55 kg at sea – a weight category that largely represents the first few months of the seawater production phase.
Against this backdrop, I believe we as an industry should pay greater attention to knowledge, documentation, and sound procedures during this critical transition. Our goal must be to produce more robust smolt and facilitate safer transfers to sea.
Findings from Fiskehelserapporten 2025
In the Norwegian Veterinary Institute's survey for Fiskehelserapporten 2025, incomplete smoltification and loser-fish syndrome are highlighted as important causes of reduced growth in salmon at sea. These are also identified as growing concerns:
"Among the general comments on fish health and welfare in the Norwegian aquaculture industry, there are calls for greater attention to freshwater production, with an emphasis on robust smolt whose health status and physiological condition take priority over the time pressures of intensive operations and the pursuit of growth."
(Fiskehelserapporten 2025, translated from Norwegian.).
Inadequate size grading is also identified as one of the main causes of uneven feed distribution, deteriorating water quality, and high stress levels – factors that can significantly affect growth, welfare, and mortality during the early stages of seawater production.
There is clearly good reason to collect reliable and relevant data about smolt before sea transfer. The question we as an industry must ask is how well today's manual sampling practices are suited to this purpose.
Major decisions based on a small number of fish
Today, decisions made before sea transfer are primarily based on manual spot checks. A limited number of fish are removed from the tank, anesthetized, weighed, measured, and visually assessed, often with more extensive sampling shortly before delivery. These methods are well established, but they have limitations we should be aware of.
Individual fish can provide a misleading picture of the overall population. In groups containing several hundred thousand fish, individual specimens can easily have a disproportionate influence on sampling results. If the group varies in size or is unevenly distributed within the tank, sampling may also introduce systematic bias.
Visual assessments are vulnerable to variation. Lighting conditions, the person conducting the assessment, how the fish are netted, and where in the tank they are collected all influence the results. Scoring of smoltification index, wounds, and fin damage involves subjective judgment, which varies between individuals, facilities, and over time.
Frequent manual handling increases stress. Even when performed correctly, handling causes additional stress during a phase when smoltification is already energy-intensive. The stress response can persist long after handling, affecting not only the fish that were actually netted.
At freshwater production facilities, internal transfers and handling rank highly among the causes of reduced smolt welfare.
Identical data points on land and at sea
The data points used for monitoring and documentation at sea are largely the same as those that are important on land: average weight and biomass, size distribution and growth rate, wounds, scale loss, and fin damage. These data can be collected using a camera-based solution adapted to the conditions in a smolt production tank.
Our own solution is one example: a modified version of the Hammerhead camera, combined with algorithms developed for smaller fish and higher stocking densities than those typically found at sea. The system monitors fish from 50 grams and continuously collects data on weight, size distribution, and welfare indicators such as wounds, scale loss, and fin damage.
This makes it possible to monitor developments without repeated manual handling or exposing the fish to additional stress.
The greatest benefit of camera-based smolt monitoring lies in the ability to collect continuous data and identify trends over both short and long periods. Has the group grown steadily, or did growth level off three weeks ago? Is weight variation increasing? Did fin damage develop gradually or suddenly – and if so, following which event?
These are questions that manual spot checks can never adequately answer.
Below is an example that I believe illustrates the value of continuous monitoring. The graphs show a clear increase in dorsal fin damage, scores 2 and 3, among smolt during the final weeks before sea transfer. The continuously collected data reveal a clear negative trend that individual manual samples would not be able to capture.
Continuous monitoring at sea, but not on land
In recent years, we have seen a significant increase in the use of camera technology and AI-based image analysis to monitor fish during seawater production.
In Fiskehelserapporten 2025, the Norwegian Veterinary Institute summarizes the benefits of this technology, highlighting its ability to monitor larger numbers of fish, provide real-time monitoring, and perform continuous scoring using consistent methods. The report emphasizes that this "reduces human subjectivity."
It also states:
"Automatic welfare monitoring is, in itself, a positive measure for improving welfare, as it reduces the need for physical handling of fish."
(Translated from Norwegian.)
What strikes me when reading this is how little attention is given to the freshwater production phase. In Fiskehelserapporten 2025, camera-based welfare monitoring is discussed exclusively in the context of seawater production. The chapters addressing welfare challenges in freshwater production do not mention this technology at all.
Perhaps they should?
From reactive to proactive monitoring
Today, fish health professionals are typically involved only when problems have already become apparent, such as reduced appetite, increased mortality, wounds, or visible fin damage. As a result, monitoring and follow-up are largely reactive.
In my experience, this is where one of the greatest strengths of continuous camera-based monitoring lies. For example, being able to see that the proportion of fish with scale loss has increased from 2% to 10% over ten days provides far more information than simply recording 10% scale loss in a single sample.
In the first case, we can see not only the level but also the development and direction of the trend. This provides an early warning and a basis for investigating potential causes before the situation deteriorates further – whether related to stocking density, water quality, previous handling, or technical equipment.
In practice, the difference is between identifying and addressing a negative trend early and having to investigate the cause and find a solution to a problem that has already developed.
Technology for decision support
Adopting technologies such as AI-based camera monitoring is not about allowing algorithms to make biological decisions. It is about ensuring that professionals are better informed and have access to reliable decision support when they need it. I believe this places important requirements on the technology itself, something that Fiskehelserapporten 2025 also makes very clear.
AI-based camera solutions used to support professional decisions must be transparent and verifiable. This is an essential requirement, and one that should be expected. For example, the images and trend analyses underlying human decisions must be accessible to those responsible for making them.
In this way, the technology provides a documented basis for decision-making, while assessments and final decisions remain the responsibility of professionals at the smolt production facility.
Better information when receiving smolt at sea
A sea farming site preparing to receive smolt has a great deal to organize and needs more information than average weight and a health certificate to ensure the best possible start.
Regulations already establish specific requirements in this area: risk-based health checks under the Norwegian Aquaculture Operations Regulations, documented smoltification through appropriate testing, and – something I believe is particularly worth noting – representative sampling.
The question is how we document compliance with these requirements. Many will recognize the familiar maze of Excel spreadsheets, forms, folder systems, and emails.
Camera technology and AI-based monitoring during the smolt production phase create opportunities for greater standardization. And I believe there is considerable potential here.
Today's digital tools make it possible to create standardized sea transfer reports that follow fish groups from land to sea. These reports can include a comprehensive status assessment at the time of transfer, along with information about the smolt's development over previous weeks and months. This may include size distribution, growth trends, wounds, scale loss, fin condition, and behavioral data. The information can also be supplemented with production data such as handling history, mortality, and biological test results.
Benefits for both fish and farmers
When camera-based monitoring systems are used both on land and after sea transfer, comprehensive documentation from the freshwater production phase can follow the fish through the transition and be compared with developments after transfer. Before sea transfer, the data can support decisions about grading, feeding plans, and net mesh size.
After transfer, they provide a reference point for assessing what happened during transport and the first period at sea.
For the fish, the benefit is a better-planned, gentler, and more precise transition from land to sea. Insights into size distribution make it easier to provide appropriate feeding from day one, while also reducing competition and aggression. Continuous weight measurements at freshwater production facilities can also enable more effective grading, allowing sea farming sites to receive groups with more uniform size distributions.
Clear information about skin health and fin condition before the fish arrive also enables better planning and precautions during transport and handling in the initial period at sea.
For the fish farmer, the main benefit is predictability: a better basis for feeding plans and monitoring during the first few weeks, improved risk assessments before the first handling operation at sea, and more fact-based communication between freshwater and seawater production teams when problems arise.
As things stand today, we often lack the information needed to determine whether challenges originated with the fish before transfer or resulted from events that occurred afterward.
Unfortunately, some smolt groups do have challenges. But when fish farmers receive sufficiently detailed information as early as possible, they can focus on making the best of the situation.
Ultimately, this is what matters
Today's camera technology and AI make it possible to move from occasional spot checks to continuous insights and a deeper understanding of fish during the freshwater production phase. The technology does not eliminate the need for all physical sampling, nor does it replace the veterinarian or fish farmer observing the fish from the tank edge.
However, it can enable fish farmers to detect and respond to changes between sampling events. At the same time, the technology can simplify the management and transfer of detailed smolt data when fish are moved to sea.
The goal is not to collect more data for the sake of collecting data. The goal is to establish more standardized collection and sharing of important documentation. This can provide a stronger foundation for better decisions at an earlier stage, helping ensure the best possible transition from land to sea, with reduced mortality and fewer fish welfare challenges.
Sources:
Mattilsynet (The Norwegian Food Safety Authority). Helsekontroll og dokumentasjon ved utsett eller flytting av fisk. mattilsynet.no
Moldal, T., Wiik-Nielsen, J., Oliveira, V. H. S., Svendsen, J. C. & Sommerset, I. (2026). Fiskehelserapporten 2025. Veterinærinstituttets rapportserie nr. 5a/2026. Veterinærinstituttet.
