Fish welfare: 11 indicators for aquaculture farms

Fish welfare cannot be reduced to mortality records or a single stocking-density value. It emerges from the interaction between the animals, their environment, husbandry practices and health status. Published in late 2025, the EU Aquaculture Assistance Mechanism’s Code of good practices on fish welfare and fish welfare indicators identifies a focused set of eleven general indicators designed for practical use. A major peer-reviewed review published in 2026 complements this framework with welfare toolboxes for five important European farmed fish species.

Together, these resources support a practical shift from occasional welfare checks to continuous prevention: observe a manageable set of signals, interpret them in combination and intervene before a deviation develops into injuries, disease or production losses.

From occasional checks to operational monitoring

A welfare indicator is a measurable feature that provides information about an animal’s state. Some indicators are animal-based, such as behaviour, body condition and visible injuries. Others describe the environment that directly shapes welfare, including dissolved oxygen.

The European Code differentiates rapid, minimally invasive operational welfare indicators from measures that require capture, detailed examination or physiological sampling. This distinction matters on commercial farms. Daily monitoring must be realistic for farm staff, while more invasive assessments should be used during scheduled reviews or when an abnormality requires investigation.

The Code is guidance rather than a new legal requirement. It also stresses that general indicators must be adapted to the species, life stage, production system and local conditions. The goal is therefore not to impose identical thresholds across farms, but to establish meaningful baselines for each site and stock.

The 11 practical indicators

Indicator	What it may reveal in practice
Food anticipatory behaviour	Disrupted routines, excessive hunger, stress or loss of learned responses
Feeding response	Reduced appetite, environmental stress, disease or inappropriate feed delivery
Condition factor or emaciation	Nutritional imbalance, competition, chronic disease or poor growth
Body injuries	Traumatic handling, unsuitable equipment, aggression or poorly managed density
Use of space	Avoidance of a zone, unsuitable flow, local hypoxia, lighting or habitat problems
Dissolved oxygen	Immediate respiratory risk and an aggravating factor for many other stressors
Respiration rate	Early response to hypoxia, temperature, stress or gill impairment
Swimming activity	Lethargy, agitation, escape behaviour or difficulty maintaining position
Skin and fin condition	Abrasion, aggression, handling damage, water-quality problems or infection
Surface activity	Respiratory distress or signs of consciousness after stunning
Vestibulo-ocular reflex	An indicator for assessing stunning and killing, not routine grow-out monitoring

No single observation provides a complete welfare diagnosis. A temporary reduction in feed intake may follow a normal temperature change, but it may also be the first sign of deteriorating water quality or infection. The warning value rises when several indicators change together or when a deviation persists over time.

Designing a dashboard that staff will actually use

Overly complex scoring systems are rarely sustainable. A better approach is to select a small core set of indicators for each production stage.

For a grow-out stock, a daily panel might include feeding response, spatial distribution, swimming activity, apparent ventilation rate, visible injuries, mortality and critical water-quality parameters. More detailed measurements — such as condition factor, fin scoring, weighing, sampling or laboratory tests — can be scheduled at appropriate intervals or triggered by an alert.

Three principles make monitoring more reliable:

• Establish a baseline. A score is meaningful only when compared with the normal pattern for the species, life stage, stock and facility under stable conditions.
• Track trends rather than isolated values. A gradual change over several days may be more informative than a single reading.
• Link every alert to an action. An abnormal observation should trigger checks of water quality, equipment, feeding, stocking conditions, recent handling and health status.

A simple traffic-light system — normal, watch and act — can support rapid decisions. Thresholds should be developed from farm history and reviewed with the veterinarian and production team. Where automated sensors are used, calibration, alarm testing and backup systems remain essential. Poor-quality digital data can be more dangerous than no data because it creates false reassurance.

Read water quality and fish behaviour together

The EU Code places water quality at the centre of fish welfare. Recirculating aquaculture systems depend on multiple technical and biological processes and therefore generally require more intensive monitoring of a wider range of parameters than many open systems. Dissolved oxygen, temperature, pH, unionised ammonia, nitrite, carbon dioxide and, where relevant, salinity or suspended solids should be interpreted against species- and life-stage requirements.

Physicochemical data do not replace animal observation. An acceptable average may conceal an unfavourable zone, a rapid fluctuation or poor access to the most oxygenated flow. Changes in fish distribution, ventilation or swimming may therefore provide an earlier warning than a fixed alarm threshold.

This combined approach also improves health surveillance. Fin damage, reduced appetite or increased ventilation cannot identify a pathogen on their own, but they can prompt earlier clinical examination, targeted sampling or investigation of system performance.

Species-specific interpretation remains essential

The 2026 review by Noble and colleagues proposes welfare indicator toolboxes for Atlantic salmon, rainbow trout, European sea bass, gilthead sea bream and common carp. It emphasises that the meaning and sensitivity of an indicator may vary with species, age, season, production system and the procedure being assessed.

The objective is not to collect every possible measure. It is to choose indicators that are relevant, repeatable and sensitive enough to detect deterioration. Observer training is therefore critical. Illustrated criteria, short video examples and agreement exercises between staff members can reduce subjectivity and improve continuity between shifts.

Digital tools can help record and analyse trends, but they should support rather than replace informed observation. A camera or sensor may detect a change; interpreting its significance still requires knowledge of the stock, the system and recent events.

Turning welfare data into preventive action

A useful welfare dashboard should connect observations with farm records. Feeding changes, grading, transport, vaccination, treatment, equipment failures, water-source changes and unusual weather may all explain or contribute to a deviation. Keeping these events on the same timeline makes patterns easier to identify.

Escalation procedures should also be explicit. For example, a persistent reduction in feeding response combined with increased ventilation could trigger immediate water-quality verification, inspection of pumps and oxygenation, a clinical review and, where indicated, diagnostic sampling. This prevents teams from relying on informal judgement at the very moment when rapid, coordinated action is most valuable.

Conclusion

The eleven indicators highlighted in the European framework are not universal standards. They are a practical starting point for turning fish welfare into a documented process linked to daily decisions. Their value increases when they are combined with water-quality data, mortality records, husbandry events and veterinary surveillance.

A good monitoring system is concise, site-specific, consistently recorded and connected to clear escalation rules. Used in this way, welfare assessment becomes a preventive management tool rather than a compliance exercise.

How Vetofish can support aquaculture professionals

Vetofish can help farms select species- and system-appropriate indicators, develop scoring sheets, define alert levels and formalise response procedures. Support may include husbandry and welfare audits, staff training, joint interpretation of welfare and water-quality data, and integration of the monitoring plan into veterinary health and biosecurity programmes.

References and supporting documents

• Noble C., Abbink W., Alvestad R., Ardó L., Bégout M.-L., Bloecher N., et al. (2026). “Welfare Indicators for Aquaculture Research: Toolboxes for Five Farmed European Fish Species.” Reviews in Aquaculture, 18(1), e70109. DOI 10.1111/raq.70109.
• Arechavala López P., Bassett D., Llorente Manzanares M., Pacchiarini T., Xandri Royo P. M. (2025). Code of good practices on fish welfare and fish welfare indicators. Publications Office of the European Union. DOI 10.2926/4420264.
• Barreto M. O., Rey Planellas S., Yang Y., Phillips C. J. C., Descovich K. (2022). “Emerging indicators of fish welfare in aquaculture.” Reviews in Aquaculture, 14(1), 343–361. DOI 10.1111/raq.12601.
• European Food Safety Authority (EFSA), López Baquero A., Millán Caravaca C., Fabris C., Van der Stede Y., Candiani D. (2025). “Fish husbandry systems: exercise of the EFSA AHAW Network (animal welfare topic).” EFSA Supporting Publications, 22(12), EN-9821. DOI 10.2903/sp.efsa.2025.EN-9821.

Read the EU Code of good practices on fish welfare and fish welfare indicators.