Recirculating aquaculture systems

Also known as RAS, Recirculation aquaculture, Land-based closed containment fish farming

Recirculating aquaculture systems (RAS) raise fish in land-based tanks that filter and reuse most of their water. They need large amounts of energy, are costly to build, and have suffered failures and closures.

CommercialDemonstratedReviewed 22 Oct 2026
AI illustration · Illustrative scene, not a photograph of the specific device or experiment.

In plain language

A recirculating aquaculture system, or RAS, is a fish farm built on land, usually inside a building. Fish live in tanks, and the water is pumped out through filters that remove solid waste. Bacteria in a separate filter turn fish waste into less harmful substances, and the water is then oxygenated and kept at a set temperature before it returns to the tanks. Only a small share of the water is replaced with new water.

This gives operators close control over water quality, temperature and exposure to disease, and it means the farm does not need to sit beside a river or the sea. The same dependence on machinery is the main risk. Pumps, blowers and filters must run all the time, and a single failure can kill large numbers of fish in a short period. Reported examples include a filter malfunction at Atlantic Sapphire's Miami site in 2021 and the closure of the New York company LocalCoho in 2025.

Technical explanation

A RAS treats its water in a loop. Mechanical filters remove suspended solids, nitrifying bacteria in biofilters convert ammonia from fish excretion into nitrate, and oxygen and carbon dioxide levels are managed by aeration or gas stripping. Some designs add disinfection or denitrification. Badiola and colleagues (2018), in a review of energy use in RAS, describe these systems as intensive, with reduced use of water and land, but state that their high energy requirement raises operating costs and the impacts of fossil-fuel use. The review also reports that energy is a minor concern for most of the industry, and that renewable energy has potential in RAS. Its authors call for more efficient designs built around optimised unit processes and equipment selection.

Water use depends strongly on design. An industry guide from Eurofish (Bregnballe and Moth-Poulsen, 2017) gives new-water use of about 200 litres per kilogram of fish produced for some super-intensive farms inside insulated buildings, about 3 cubic metres per kilogram for traditional outdoor farms converted to recirculation, and about 30 cubic metres per kilogram for a traditional trout flow-through system. These figures are not peer reviewed, have no stated year, and are not expressed per year.

Carbon footprint and cost are a separate matter. Liu and colleagues (2016) compared a 3,300 tonne per year land-based RAS salmon model in the United States with Norwegian open net pens. Their abstract reports that total production costs were similar, about 10% higher for the RAS model, and that its capital cost was roughly 80% higher. The RAS model's 15-year net present value was negative. The same abstract gives a carbon footprint of 7.41 kg CO2 equivalent per kg for the US RAS salmon, against 15.22 kg for airfreighted Norwegian salmon, and states that feed production dominated the footprint of both systems.

Fruscella and colleagues (2026), in a review article, argue that RAS often confine fish in artificial, sensory-deprived environments that may harm their welfare. They argue that environmental enrichment is needed to meet fish's behavioural, cognitive and physiological needs, and they propose a shift from stress reduction towards an allostatic welfare framework.

What the evidence says

Each statement is typed by the strength of its evidence and linked to the source that supports it.

Demonstrated Shown in a published experiment, test or shipping product.

RAS are intensive fish production systems that reduce water and land use, but their high energy requirement raises operating costs and the impacts of fossil-fuel use.

Sources: [1] · Abstract (read via Semantic Scholar) · medium confidence

Reported, unverified Stated by a company or team but not independently confirmed.

For some super-intensive RAS inside closed insulated buildings, new water use can be as low as about 200 litres per kilogram of fish produced. A traditional outdoor trout flow-through system is given at about 30 cubic metres per kilogram, and an outdoor farm converted to recirculation at about 3 cubic metres per kilogram.

Sources: [4] · Water use section · low confidence

In a model of a 3,300 tonne per year land-based RAS salmon farm in the United States, capital cost was roughly 80% higher than for Norwegian open net pens, total production costs were about 10% higher, and the RAS model had a negative 15-year net present value.

Sources: [3] · Abstract · medium confidence

In the same model, the abstract gives a carbon footprint of 7.41 kg CO2 equivalent per kg for the US land-based RAS salmon, against 15.22 kg for airfreighted Norwegian salmon, and states that feed production dominated the climate impact of both systems.

Sources: [3] · Abstract · medium confidence

In March 2021, Atlantic Sapphire reported that it expected to lose about 500 tonnes (head on gutted) of fish at its Miami RAS facility. The trade press notes the figure can be read two ways and the company declined to elaborate. The company first blamed a filter malfunction and later pointed to an identified design weakness in the RAS system.

Sources: [6] · Article body, citing company stock exchange announcement · medium confidence

LocalCoho, a land-based RAS salmon company in Auburn, New York, was set to cease operations on 31 January 2025 after a six-year effort. The article reports that it struggled with high operating costs, regulatory complexity and market pressures.

Sources: [7] · Article body · medium confidence

In February 2018, Atlantic Sapphire stated that its Danish Bluehouse produces approximately 3,000 tonnes of salmon a year, and that its planned Miami Phase 1 would produce about 10,000 tonnes a year. The Miami figure was a projection at that time, not a measured output.

Sources: [5] · Article body and company correction note · medium confidence

Proposed A design or plan that has not yet been built or tested.

Fruscella and colleagues argue that RAS often confine fish in artificial, sensory-deprived environments that may harm their welfare, and that environmental enrichment is needed to meet their behavioural, cognitive and physiological needs.

Sources: [2] · Abstract (full text not accessed) · medium confidence

Limitations and open problems

  • Energy demand is high. Badiola and colleagues (2018) identify energy as a significant cost and sustainability issue for RAS, although the same review reports that energy is a minor concern for most of the industry.
  • The modelled salmon comparison by Liu and colleagues (2016) found capital cost roughly 80% higher and a negative 15-year net present value for the RAS model, with total production costs about 10% higher. The model reflects one farm size and location.
  • The water-use figures in the Eurofish guide are from an industry source, have no stated year, and vary widely with system design. Partly recirculating systems use far more new water than closed systems.
  • Welfare evidence for RAS is still developing. Fruscella and colleagues (2026) argue that current welfare assessment does not capture behavioural and cognitive needs. This profile did not access the full text of that review.
  • Single points of failure in filtration, oxygen supply or pumping can kill large numbers of fish in a short time. The Atlantic Sapphire Miami incident in 2021 was reported to be linked to a filter malfunction and a design weakness.

Possible future uses

Speculative

Ideas, not predictions. None of these has been demonstrated.

  • Smaller RAS plants close to cities that shorten the distance fresh fish travels to shops and restaurants.
  • RAS powered mainly by renewable electricity, which could narrow the emissions gap with net-pen salmon that Liu and colleagues modelled.
  • Enriched RAS tanks designed around fish behaviour, if commercial trials show that welfare gains outweigh added maintenance and hygiene costs.

Timeline

  1. 2016Liu and colleagues publish a techno-economic and carbon footprint comparison of a 3,300 tonne land-based RAS salmon model and Norwegian open net pens in Aquacultural Engineering. Sources: [3]
  2. 28 Feb 2018Atlantic Sapphire states that its Danish Bluehouse produces about 3,000 tonnes of salmon a year and that its planned Miami Phase 1 would produce about 10,000 tonnes a year. Sources: [5]
  3. May 2018Badiola and colleagues publish a review of energy use in RAS in Aquacultural Engineering, volume 81. Sources: [1]
  4. Mar 2021Atlantic Sapphire reports a mass mortality at its Miami RAS facility, with about 500 tonnes (head on gutted) of fish expected to be lost. The company did not elaborate on the figure. Sources: [6]
  5. 31 Jan 2025LocalCoho, a land-based salmon RAS company in Auburn, New York, is set to cease operations, after a six-year effort to cultivate coho salmon. Sources: [7]
  6. 17 Feb 2026Fruscella and colleagues publish a review article arguing that environmental enrichment is needed to improve fish welfare in RAS. Sources: [2]

Futures this could help bring about

AI illustration
How far away?

Most fish farmed on land

Pilots · 1 of 6 steps done

Glossary

Recirculating aquaculture system (RAS)
A fish farm that filters and reuses most of the water in its tanks, replacing only a small share with new water.
Biofilter
A filter in which bacteria convert ammonia from fish waste into less harmful nitrogen compounds.
Stocking density
The amount of fish held in a given volume of water or tank space.
Head on gutted (HOG)
A product weight measure for fish with the head left on and the guts removed.
Net present value
The value today of a project's future cash flows, after discounting for the time value of money.

Sources

  1. peer reviewed paper
    Energy use in Recirculating Aquaculture Systems (RAS): A review
    M. Badiola, O.C. Basurko, R. Piedrahita et al.. Aquacultural Engineering, May 2018. DOI: 10.1016/j.aquaeng.2018.03.003. Accessed 22 Oct 2026.
  2. review article
    Fish Welfare in Recirculating Aquaculture Systems (RAS): The Imperative for Environmental Enrichment (EE)
    Lorenzo Fruscella, Annamaria Passantino, Benz Kotzen. Animals (MDPI), 17 Feb 2026. DOI: 10.3390/ani16040635. Accessed 22 Oct 2026.
  3. peer reviewed paper
    Comparative economic performance and carbon footprint of two farming models for producing Atlantic salmon (Salmo salar): Land-based closed containment system in freshwater and open net pen in seawater
    Yajie Liu, Trond W. Rosten, Kristian Henriksen et al.. Aquacultural Engineering, 2016. DOI: 10.1016/j.aquaeng.2016.01.001. Accessed 22 Oct 2026.
  4. other
    Guide to Recirculation Aquaculture
    Jacob Bregnballe, Thomas Moth-Poulsen. EUROFISH Magazine (eurofish.dk), 18 Jul 2017. Accessed 22 Oct 2026.
  5. news
    Atlantic Sapphire nears pivotal Danish harvest
    Salmon Business, 28 Feb 2018. Accessed 22 Oct 2026.
  6. news
    Atlantic Sapphire takes action after mass mortality
    Fish Farming Expert, 29 Mar 2021. Accessed 22 Oct 2026.
  7. news
    Coho crash: why this land-based trailblazer had to close its doors
    Salmon Business, 28 Jan 2025. Accessed 22 Oct 2026.

Revision history

3 changes since this profile was created.

Changes on 22 Oct 2026
Published on owner instruction (owner requested space, food and water topics)dc4cd89
Sources and claims re-checked against originals409337d
Initial profile drafted from primary sourceseaac9eb