Shrimp stocking density: look beyond growth

In intensive shrimp farming, stocking density is not simply a yield-per-square-metre decision. A 2026 Scientific Reports study followed Pacific white shrimp (Penaeus vannamei, also widely reported as Litopenaeus vannamei) at three densities in a recirculating aquaculture system (RAS), followed by a common recovery phase. Survival, growth, appendage integrity and selected behaviours did not change or recover at the same rate. For RAS managers, the operational message is that no single measure can provide an adequate early warning that density has become detrimental.

A design that separates crowding from recovery

Paolo Gamberoni and colleagues used twelve 600-litre tanks connected to a treatment system with a biofilter, UV disinfection, protein skimmer, drum filter and ozone generator. After three weeks of acclimation, shrimp weighing about 14.7 g were allocated to surface densities of 1, 2 or 4 kg/m². In this installation, those treatments corresponded to 1.83, 3.67 and 7.33 kg/m³. Each density was replicated in four tanks.

The first phase lasted 21 days. The researchers measured survival and growth, scored antennae, uropods, exoskeleton, rostrum and eyes, and recorded four categories of abnormal behaviour. A camera above each tank captured 30 minutes twice weekly; five minutes from each recording were coded in BORIS using a predefined ethogram. Haemolymph variables, enzyme activities and the expression of genes associated with stress or immune functions completed the assessment.

All groups were then reset to 0.8 kg/m² for another 21 days. This recovery phase is a major strength of the experiment because it distinguishes rapidly reversible responses from effects that continue after the stocking pressure has been reduced.

Survival and growth confirm a problem that is already established

After 21 days, survival was significantly lower at 4 kg/m² than at 1 and 2 kg/m². Shrimp at 1 kg/m² also achieved higher mean weight and growth indices. Mortality and growth therefore remain valuable outcome measures, but their operational weakness is timing: a clear deterioration means that unfavourable exposure has already affected both animals and production.

The recovery results require careful interpretation. Once all treatments were moved to the same low density, mortality ceased and shrimp previously held at high density showed compensatory growth. They did not, however, fully close the weight gap acquired during the first phase. A faster daily gain after corrective action is not evidence that every earlier consequence has disappeared.

These experimental densities are not farm thresholds. Their meaning depends on usable bottom area, tank geometry, shrimp size, moulting stage, feed distribution, oxygenation and water treatment. The study tests one system for six weeks. It supports a monitoring strategy, not a universal regulatory or commercial stocking limit.

Antennae, uropods and movement add complementary warnings

At high density, antenna damage was more severe after the stress phase. Uropod integrity also varied with conditions, whereas rostrum and eye scores did not differ significantly among treatments. The authors discuss repeated contact with hard surfaces, animal-to-animal interactions and vulnerability during moulting as plausible contributors. After recovery, morphological differences were no longer significant and damage had generally decreased.

Behaviour supplied a different layer of information. Abnormal swimming became more frequent as density increased. Loss of balance was also more frequent in the standard- and high-density groups than in the low-density group. Body cramps and strong escape responses, by contrast, did not vary significantly with treatment. A useful monitoring programme must retain this asymmetry: one responsive behaviour does not make every unusual action a validated indicator of crowding.

During recovery, abnormal events declined and no treatment differences remained significant. Behavioural observation may therefore offer a reversible, non-invasive warning, provided that recording time, duration, feeding context, lighting and counting rules are standardised. An occasional video or a general impression cannot replace a repeatable observation protocol.

Biomarkers do not all tell the same story

Haemolymph glucose, lactate and total protein did not differ among treatments. Phenoloxidase activity, an innate immune measure, and superoxide dismutase activity likewise showed no significant treatment effect. In contrast, glutathione peroxidase and HSP70 gene expression were higher at high density after 21 days, while HSP90 increased in the standard- and high-density groups. These expression patterns returned to baseline after recovery.

It would therefore be unsafe to exclude chronic stress from one immune or metabolic panel. Responses depend on the tissue, sampling time and biological pathway being measured. Laboratory tests retain diagnostic value, but they should answer a defined question and be interpreted alongside production, morphology and behaviour.

Turning the findings into a workable RAS dashboard

A practical dashboard can combine four data families without creating an unmanageable workload.

First, record the population at risk, mortality, feed consumption and growth deviations for each cohort. Second, examine a consistent sample at scheduled intervals for antenna length and symmetry, uropod condition, exoskeleton, rostrum, eyes and moulting stage. Third, score short video sequences collected under standard conditions for abnormal swimming, loss of balance, escape responses and tank-space use. Fourth, compare these signals with live biomass, usable surface, feed distribution and water-quality time series, particularly dissolved oxygen, temperature, pH and nitrogen compounds.

Action levels should be developed from the farm’s own baseline. When several indicators shift together, teams should promptly verify biomass estimates, flow, aeration, feed access and accumulation zones. Depending on the cause, the response may include redistributing or reducing the load, improving feed dispersion, changing the lighting schedule, adding cleanable shelters or increasing surveillance. None of these measures is automatically beneficial. For example, enrichment that traps waste or obstructs flow may introduce a new hazard.

This approach also separates monitoring from diagnosis. Abnormal swimming or antenna damage can trigger an investigation, but neither identifies a cause on its own. Water-quality deterioration, nutritional competition, infection, handling, moulting synchrony and equipment failure may produce overlapping observations. The investigation should follow the signal rather than relabel it as a diagnosis.

Limits and points of caution

The study involved one species, juvenile shrimp of a defined size, twelve experimental tanks and a specifically equipped RAS. Density expressed by area and volume does not fully describe the usable space available to a benthic animal. A 21-day exposure represents a relatively short chronic stress period compared with a commercial production cycle. Behaviour was coded from limited video windows, and direct aggression was seen only once; the authors’ discussion of possible night-time cannibalism was not demonstrated by the observations.

The results should also not be generalised to ponds, biofloc systems, other penaeid species or different life stages without validation. Water quality was kept within suitable conditions during the experiment, so the design does not quantify how crowding interacts with oxygen depletion, nitrogenous waste or a disease outbreak.

Finally, morphological or molecular recovery does not necessarily prove complete restoration of every welfare dimension. Compensatory growth remained partial and size differences persisted. Longer commercial studies are needed before the proposed measures can become a fully validated welfare index or an automated alert system.

Conclusion

Stocking density affects production and several dimensions of shrimp condition, but the responses do not appear or resolve together. Mortality and growth confirm an established deterioration; appendage condition and selected behaviours can strengthen earlier warning; several commonly measured biomarkers were not sensitive in this setting. Reducing density supported genuine recovery, although not every effect was erased. A robust RAS programme should therefore combine simple, comparable time series and connect them to predefined decisions.

How Vetofish can help

Vetofish can help shrimp farms design a site-specific monitoring grid, audit stocking and water-quality conditions, standardise video and morphological observations, and define an investigation plan when several indicators drift. The aim is to rank risks and actions without promising a universal threshold or a guaranteed health outcome.