Farmed shrimp welfare: observe before judging

Does a shrimp that feeds quickly have good welfare? Is an animal remaining under shelter calm, unwell or simply approaching a moult? Behaviour provides shrimp farms with information that can be collected without handling or sampling every animal. It is also easy to overinterpret movement, escape or aggregation. A 2025 review of behavioural testing in decapods describes a promising but immature field: many measures can support an investigation, yet few have validated thresholds that can independently classify the welfare of a commercial group.

The practical response is not to wait for a perfect indicator. Farms can build repeatable observations and interpret them alongside water quality, feed delivery, stocking density, moult stage and production outcomes. Behaviour then becomes an early warning and a question-generating tool rather than a stand-alone diagnosis.

Observable behaviour does not have one fixed meaning

Shrimp explore, search for feed, groom, occupy shelters, interact and respond to light changes or perceived threats. These activities depend on species, size, sex, moult stage, time of day, substrate and previous experience. The same movement may therefore mean different things in different settings.

Campbell and Lee review preference, motivation, activity, risk-response, aggression and learning tests. Choosing a shelter or substrate shows that an animal uses a resource when it is available. It does not automatically quantify how important that resource is or prove that its absence impairs health. Similarly, time taken to resume feeding after a stimulus may capture part of a risk response, but repeatability must be established and hunger, moult stage and acclimation must be considered.

Individual variation is not mere noise. Animals exposed to the same tank can display consistently different strategies. Labelling one pattern as normal and another as abnormal without further evidence may hide useful information about the distribution of responses across the population.

What a density experiment can—and cannot—show

A 2021 experiment illustrates the need for careful interpretation. Fifty-six juvenile Pacific white shrimp, Litopenaeus vannamei, with a mean mass of 7.91 g were allocated to 6.2, 12.4 or 24.8 animals per square metre. Four groups were tested at each density. Twenty-minute feeding sessions were recorded on eight consecutive days and analysed using automated tracking.

Relatively stable dominance hierarchies were observed. Subordinate animals spent longer in the feeding area, while dominant shrimp explored more of the arena. At the highest experimental density, behavioural differences related to dominance were smaller. Relative feed consumption increased even though the shrimp spent less time on the crowded feeding tray.

The experiment does not establish 24.8 shrimp per square metre as a welfare optimum. It involved small groups, a test arena and short observations. Its principal outcome was the organisation of feeding behaviour, not long-term health, injury, survival or affective state. High feeding activity may reflect strong motivation, but it may also accompany competition. Commercial interpretation must include growth variation, contacts, damage, water quality and the geometry and operating conditions of the actual system.

Building a useful farm-level observation system

A field ethogram should be short enough to use consistently. A team might define five to eight visible items: distribution in the unit, locomotor activity, access to feed, recovery after disturbance, use of shelter or substrate, aggressive interactions, grooming and abnormal responses. Each item needs an operational definition so that different observers record the same event.

Observations should be made at comparable times: before and after feeding, under the same light phase, for a fixed duration and from a consistent viewpoint. In turbid or deep systems, video, passive acoustics or viewing windows may supplement direct observation. Automated tracking must be validated in the target installation because bubbles, reflections, overlap and high density can produce confident-looking but incorrect tracks.

Behaviour does not replace environmental monitoring. Reduced activity becomes more informative when it coincides with declining oxygen, changing salinity, rising temperature, a new feed or accumulating nitrogenous waste. Moult patterns and recent interventions should also be logged. A sudden population-wide shift warrants an immediate check of equipment and water before a complex behavioural explanation is proposed.

Treating enrichment as a testable hypothesis

A shelter, substrate, darkened area or novel feeding method is not beneficial by definition. It may trap waste, impede water exchange, obstruct inspection or create a monopolised resource. A change should therefore be compared with a baseline using outcomes selected in advance.

A proportionate assessment combines preference, sustained use, absence of health deterioration and observations over several weeks. If animals use a structure but aggressive contacts, injuries or water conditions worsen, the design needs revision. Conversely, no dramatic improvement in growth does not prove that choice is worthless; the intended benefit may be greater control over the environment or reduced disturbance.

Scientific limits and defensible decisions

The review highlights limited pharmacological and cross-context validation for tests adapted from other animals. Biology also differs substantially among caridean shrimp and dendrobranchiate prawns. Thresholds should not be transferred between species, life stages, laboratories and farms without testing.

Current farms can track internal trends rather than claim universal standards. An indicator becomes more credible when it is repeatable, changes in response to a relevant challenge, moves back after correction and agrees with independent measures. Differences between comparable units are often more actionable than a single absolute score.

Conclusion

Shrimp behaviour deserves a place in routine surveillance, provided that context is retained. Activity, feeding, shelter use and interactions can offer early warnings, but none proves welfare status alone. A small standardised ethogram linked to water, moult and production data can already improve decisions while the evidence base develops.

How Vetofish can help

Vetofish can help farms select species- and system-specific indicators, pilot video protocols, connect behavioural drift with health and environmental data, and train teams in repeatable observation. The goal is a field-ready dashboard that can be refined as validation evidence grows.