Land-based fish farming grows aquatic species entirely on shore, inside engineered facilities that treat and recirculate the same water rather than drawing continuously from a river, lake, or ocean. The most common design is a recirculating aquaculture system, usually called a RAS, which passes water through mechanical filters, biological filters, and gas-exchange units in a continuous loop. This approach decouples fish production from natural water bodies, giving operators precise control over temperature, oxygen, and waste chemistry while sharply reducing the volume of water consumed and the effluent released into the environment. The engineering required is substantial, but the concept is straightforward: keep the water clean enough that fish thrive in a tank the way they would in a healthy body of water, and do it without needing an actual body of water.
How Water Stays Clean in a Closed Loop
Fish excrete ammonia through their gills and generate solid waste from uneaten feed and feces. In a river or open ocean pen, dilution handles much of that load. In a closed tank, concentrations rise fast enough to become toxic within hours if nothing intervenes. A RAS addresses this with a chain of treatment steps, each targeting a different class of pollutant.
Solids removal comes first. Most commercial systems use micro-screen drum filters that physically strain suspended particles out of the water as it flows through a rotating mesh. A study examining solids loading in a RAS found that the sedimentation and drum-filter combination captured the vast majority of biochemical oxygen demand and suspended solids in the system’s waste stream, even when only about a fifth of the recirculating flow actually passed through the drum filter screen itself.
After solids are removed, the water enters a biological filter, or biofilter. This is the heart of a RAS. Inside the biofilter, specialized microbes convert toxic ammonia first into nitrite and then into relatively harmless nitrate, a process called nitrification. The biofilter medium is usually some form of plastic carrier, either fixed in place or kept tumbling by aeration in a moving-bed design. A full-scale Atlantic salmon facility described in a recent study used moving-bed biofilters alongside drum filters and a dedicated carbon-dioxide degasser in each rearing module.
Carbon dioxide is the often-overlooked third problem. Fish respire CO₂ into the water, and because a RAS recirculates, CO₂ accumulates if it is not actively stripped out. A degassing column or packed tower forces air through a thin film of water, allowing dissolved CO₂ to escape into the atmosphere. Without this step, elevated CO₂ depresses the water’s pH and can stress or suffocate fish even when dissolved oxygen readings look fine.
The Microbes That Make It All Possible
The biofilter’s performance depends on living communities of bacteria and archaea that colonize the filter media. Getting those communities established, a process operators call “acclimating” or “maturing” a biofilter, is one of the trickiest parts of starting a new system. Research on biofilter acclimation has shown that complete nitrification can establish within about seven days, signaled by falling ammonia, rising nitrate, and high abundances of the key microbial families responsible for each conversion step.
The microbiology turns out to be more complex than early textbooks suggested. In both freshwater and marine RAS, ammonia-oxidizing archaea often outnumber the ammonia-oxidizing bacteria that were historically considered the primary workhorses. A study of a shrimp RAS found archaeal ammonia-oxidation genes were more abundant than their bacterial counterparts in every compartment of the system.
A freshwater RAS biofilter study found an even more striking picture: archaea outnumbered the traditional bacterial ammonia oxidizers by more than three orders of magnitude, and a subset of the Nitrospira bacteria present appeared capable of performing the entire ammonia-to-nitrate conversion on their own, a recently discovered lifestyle called complete ammonia oxidation.
What this means in practice is that biofilter health is not just about one or two bacterial species. These filters harbor over a thousand bacterial genera alongside archaea, and their composition shifts with water depth, temperature, and routine operational changes like cleaning or adjusting flow rates. Operating conditions like ammonia concentration also shape the community: biofilters acclimated at higher ammonia loads develop a larger proportion of nitrifiers but lower overall microbial diversity, while those acclimated at lower loads maintain higher diversity that may benefit broader nutrient cycling.
How Land-Based Systems Compare to Open-Water Farming
The most direct comparison is between a RAS and a flow-through system, where water enters from a natural source, passes through the fish tanks once, and exits. Flow-through systems are simpler and cheaper to build, but they offer limited control over temperature and are wholly dependent on the source water’s quality and volume.
A welfare study comparing rainbow trout raised in a French RAS with trout in a flow-through system fed by lake water found that when water quality in the RAS was kept at optimal levels, the fish showed no measurable welfare disadvantage. Basal stress-hormone levels were equivalent, swimming behavior was similar, and learning performance in behavioral tests was comparable. The RAS trout actually grew heavier, likely because operators could maintain warmer water temperatures year-round.
A separate comparison of seahorse culture in a zero-exchange RAS versus a traditional flow-through found that production in the RAS was roughly 23% higher and survival about 16% higher, even though individual growth rates did not differ significantly between the two systems. The gains came from tighter environmental control reducing mortality rather than from making the animals grow faster individually.
Beyond RAS: Other Land-Based Approaches
Recirculating systems are the dominant model for high-value commercial species, but they are not the only way to farm fish on land.
Biofloc technology takes a fundamentally different approach to waste management. Instead of filtering ammonia out of the water through a biofilter, operators add a carbon source (like molasses or wheat flour) to encourage heterotrophic bacteria to assimilate dissolved nitrogen directly into microbial biomass. The resulting “biofloc” particles serve as supplemental food for the cultured species, turning a waste product into a feed source. The method works best for species that tolerate or even benefit from turbid, particle-rich water, like shrimp and tilapia.
Aquaponics couples a fish-rearing system with hydroponic plant beds. Nutrient-rich water from the fish tanks irrigates plant roots, and the plants strip out nitrogen and phosphorus before the water returns to the fish. The concept produces two crops from one water supply. In practice, fish and plants rarely have perfectly matched nutrient demands, and supplementation strategies are an active area of research to close the gap between what the fish produce and what the plants need.
Integrated multi-trophic aquaculture, or IMTA, takes the nutrient-recycling idea further by adding multiple species at different levels of the food web. A land-based IMTA trial combining finfish, mussels, and macroalgae found that the mussels and seaweed together reduced ammonia in the water by over 90%, nitrite by 75%, and nitrate by about 45%, while the macroalgae gained a protein-content boost of roughly 13% from the nutrient-rich environment.
The Off-Flavor Problem
One of the more frustrating challenges specific to RAS is off-flavor. Compounds called geosmin and 2-methylisoborneol (MIB), produced by certain bacteria and algae in the system water, accumulate in fish flesh over time and give it a muddy or earthy taste. People can detect these compounds at extremely low concentrations, so even trace amounts create a product-quality problem.
Research tracking off-flavor accumulation in rainbow trout and European whitefish in a RAS found that both compounds built up steadily in fish tissue over months, with geosmin accumulating faster in the fattier rainbow trout fillets after roughly 140 days. Water concentrations reached tens of nanograms per liter, enough to taint the fish.
The standard industry remedy is depuration: transferring fish to clean water before harvest and letting the compounds flush out of the flesh. A depuration trial with European whitefish showed that after 16 days in clean water, both geosmin and MIB concentrations in fillets dropped below detection limits. That depuration period adds time, labor, and tank space to the production cycle, all of which affect costs, but it works reliably.
Choosing the Right Species
Not every fish adapts equally well to tank life. Species that tolerate high stocking densities, accept formulated feed readily, and grow efficiently at controlled temperatures are natural candidates. Atlantic salmon currently dominates the large-scale land-based sector because the market price is high enough to justify the capital investment. But operators and researchers are constantly evaluating alternatives.
Atlantic cod has been tested in land-based RAS in Europe, where a feasibility study found that while the biology was workable, electricity cost was one of the most significant economic hurdles, particularly for temperature control in warmer climates. In New Zealand, hāpuku (a type of groper) is being developed specifically for land-based production. Researchers have established methods to spawn, raise juveniles, and grow the fish to harvest entirely on land, where water quality, pathogens, and genetic interactions with wild populations can all be tightly managed.
Selective breeding programs are beginning to tailor fish genetics to indoor conditions. Breeding goals for land-based species may increasingly prioritize general robustness, selecting for animals that resist a wider range of diseases and tolerate the environmental variability inherent in engineered systems, rather than optimizing for a single trait like growth rate.
Stocking Density and Fish Welfare
How many fish you pack into a tank matters enormously for both animal welfare and economic returns. A study on juvenile gibel carp in a land-based RAS compared low and high stocking densities and found that fish in the lower-density group grew significantly larger, had better feed-conversion efficiency, and showed fewer molecular markers of physiological stress. The high-density group experienced lower dissolved oxygen, higher nitrite levels, and activated cellular stress-response pathways. Return on investment was also lower in the high-density group, because the degraded conditions ate into growth efficiency more than the extra fish compensated for.
The takeaway is that overstocking a RAS is not just a welfare concern but a financial one. Operators have a direct economic incentive to find the density sweet spot where they maximize biomass per cubic meter without triggering the cascade of poor water quality and impaired growth that pushes costs up per kilogram of fish produced.
The Economics of Growing Fish Indoors
Land-based RAS facilities require considerably more upfront capital than conventional open-water farms. A detailed economic comparison of a land-based closed-containment RAS and an open net-pen system, both producing 3,300 metric tons of head-on gutted Atlantic salmon, found that the RAS capital cost was approximately $54 million versus $30 million for the net-pen system, a difference of 80%.
The production-cost gap, however, was much smaller. Total cost per kilogram in the RAS was about $5.60 compared to $5.08 in the net pen, roughly a 10% premium. When interest and depreciation were excluded, operating costs were nearly identical at around $4.30 to $4.37 per kilogram. The implication is that the financial barrier to land-based farming is primarily the cost of building the facility, not the cost of running it day to day.
Energy is the largest operational expense specific to RAS. Pumps run continuously, biofilters need aeration, and temperature control can require heating or cooling depending on local climate. Electricity costs vary dramatically by location, which is one reason site selection for land-based farms has become a strategic exercise. Operators look for cheap, reliable power, ideally from renewable sources, along with access to cold water (which reduces cooling costs for cold-water species like salmon) and proximity to markets to minimize freight.
Monitoring and Automation
Because a RAS is essentially a life-support system, continuous monitoring is non-negotiable. If a pump fails or ammonia spikes overnight, fish can die within hours. Traditional monitoring relies on probe-based sensors measuring dissolved oxygen, pH, temperature, and ammonia, often connected to alarms that alert operators to critical deviations.
Computer vision is increasingly layered on top of conventional sensors. Cameras trained on fish tanks can detect behavioral changes, like abnormal swimming patterns or reduced feeding activity, that signal stress or disease before water-chemistry readings flag a problem. One deep-learning system designed for RAS achieved roughly 96% accuracy in distinguishing five different behavioral states of fish schools from video footage. Another approach using object-detection models for disease identification in fish reached 98% accuracy in laboratory testing, though real-world performance in turbid, variable RAS water remains a harder challenge. A dataset collected from a commercial tilapia RAS in Mexico paired underwater video footage with physicochemical measurements to help train visual models that account for real farming conditions, where lighting, turbidity, and fish density all fluctuate.
These systems are not replacing human operators yet, but they are extending the operator’s senses. A camera that watches the tank 24 hours a day catches things a person walking the facility floor twice a day will miss.
Environmental Trade-Offs
Land-based farming eliminates several of the environmental problems associated with open-water aquaculture. There is no direct nutrient discharge into wild waterways, no risk of farmed fish escaping and interbreeding with wild populations, and no sea-lice transmission to migrating wild stocks. The waste stream is concentrated and can be captured and treated or, in integrated systems, converted into fertilizer or secondary crops.
The carbon footprint is more ambiguous. A RAS burns significant energy, and unless that energy comes from low-carbon sources, the greenhouse-gas intensity per kilogram of fish produced can rival or exceed that of well-managed open-water operations. Aquaculture in general produces lower carbon emissions than livestock, especially beef, but the advantage narrows for energy-intensive indoor systems. Operators pursuing net-zero or low-carbon goals typically need to pair their facilities with renewable energy or waste-heat recovery from industrial neighbors.
Land-based IMTA systems offer a partial answer by turning waste nutrients into productive biomass rather than treating them as a disposal problem. The nutrient-removal efficiencies demonstrated in multi-trophic trials, with ammonia reductions above 90% when mussels and seaweed are integrated, suggest that thoughtful system design can reduce both the treatment burden and the carbon cost of handling waste.
What Consumers Think About Fish Raised on Land
Consumer acceptance is an underappreciated bottleneck. Many shoppers are unfamiliar with land-based salmon or other RAS-produced fish, and their willingness to buy depends heavily on how the product is presented. An experiment testing different communication strategies for land-based salmon in the UK found that providing detailed information highlighting environmental benefits significantly increased participants’ perceptions of eco-friendliness and healthiness compared with a bare-bones description. A narrative-style description written in a chef’s tone further boosted willingness to buy and trust in the stakeholders involved. Framing mattered more than most operators might expect: participants who received the richer, more narrative information were measurably more supportive of land-based production than those given a neutral product description.
Microbiome Engineering and Disease Management
As land-based systems scale up, disease outbreaks in densely stocked tanks represent one of the highest-stakes risks. Traditional responses lean on antibiotics or chemical treatments, but there is growing interest in managing the microbial environment itself as a form of disease prevention. Approaches grounded in ecological theory, such as designing microbial communities that competitively exclude pathogens or using techniques like fecal microbiota transplantation and synthetic microbial consortia, are being explored to stabilize beneficial microbiomes in RAS water and on fish surfaces. The idea is to engineer the system’s microbiology so that disease-causing organisms cannot easily gain a foothold, rather than treating infections after they appear. This is still an emerging field, but it aligns with the broader trend in land-based aquaculture: if you control the environment tightly enough, you can prevent problems rather than reacting to them.