Dissolved Oxygen (DO)

For farmers, investors or anyone who wants the basics: like people, fish need oxygen to live. The difference is that we breathe air while fish take up oxygen dissolved in water across the gills. If dissolved oxygen falls too low, fish struggle to respire - roughly like a human in a very thin atmosphere.

The air we breathe is roughly 78% nitrogen, 21% oxygen gas and about 1% other gases. When water is aerated with ordinary air, only about one fifth of the gas is actually oxygen. Many modern aquaculture facilities therefore use pure oxygen (100% O₂) instead of air alone. That makes higher density possible and keeps oxygen supply stable even at high feed rates and biomass.

What does oxygen do in the fish?

Oxygen drives fish metabolism: swimming, growth, digestion, immunity and reproduction. The more the fish eat and grow, the more oxygen they need.

When oxygen starts to fall you often see increased stress, poorer growth, higher feed conversion ratio (FCR), greater disease susceptibility and fish crowding toward inlets or aeration zones - in the worst case mass mortality. Oxygen is therefore often described as aquaculture's most important water-quality parameter.

For the chemist

Dissolved oxygen (DO) is the concentration of oxygen dissolved in water, usually in mg/L (roughly the same as ppm in water) or as percent saturation. How much oxygen can dissolve at maximum depends mainly on temperature, salinity, air pressure and elevation. Cold water can hold much more oxygen than warm water.

Theoretical oxygen solubility versus temperature

For many people outside water chemistry it is surprising that water in equilibrium with air only holds about 7-13 mg/L oxygen depending on temperature. That is because air is only about 21% oxygen gas; nitrogen does not supply the fish. With pure oxygen the partial pressure of O₂ rises about fivefold, so much more oxygen can dissolve. If oxygen is added under pressure (e.g. oxygen cone or LHO at about 5 bar absolute), theoretical solubility rises further - so inlet water can be produced well above ordinary air saturation. That is a main reason modern intensive culture often uses pure oxygen systems beyond or instead of aeration alone.

But more dissolved oxygen is not always better. When water is saturated (or supersaturated) under pressure in pipes and oxygen units and then released into culture tanks that are open to the air at about 1 atm, a "soda-bottle effect" can appear: gas wants to leave solution as pressure drops. The result can be bubbles, foam, local supersaturation and uneven gas balance right at the inlet and in jets. Excessively high dissolved oxygen and/or sudden pressure relief can stress fish - including gill irritation, behaviour changes and in serious cases gas bubble disease (especially when total gas pressure and nitrogen are also high). So it is not enough to "max out" DO in the pipe; you must understand how water spreads in the tank - inlet direction, velocity and mixing, dead zones vs short-circuiting, and where the fish actually live. Good hydraulics and controlled inlet mixing distribute oxygen evenly and avoid both local hypoxia and harmful gas pockets. Measure DO where it matters (in the tank, not only in the pressurised line), and see also the section on saturation, supersaturation and N₂.

Oxygen measurement

Without reliable measurement, the oxygen strategy is guesswork. In practice three paths dominate: handheld checks (routine, troubleshooting, calibration control), fixed process sensors linked to PLC/SCADA (continuous trend, oxygen/aeration control and alarms), and sometimes portable loggers or temporary points at commissioning and performance tests. Match technology to risk and density: ponds and low intensity may manage with spot checks; raceways, semi-flow and RAS at high biomass usually need continuous DO where the fish live - often with redundancy on critical units.

Two main types of oxygen sensor are used on farms:

• Optical probes (luminescence / fluorescence). The usual choice on new farms. Often need less maintenance, longer intervals between calibrations, and they do not consume oxygen while measuring.
• Electrochemical probes (Clark-type: galvanic or polarographic membrane probes). Often cheaper to buy, but membrane, electrolyte and zero need checking more often.

Both types usually show mg/L and % saturation, if temperature (and sometimes salinity) is compensated correctly. Check two things: that the sensor has proper temperature compensation, and that % sat is calculated against the right reference - fresh or salt water, plus elevation/air pressure.

Capacity and performance often specified: a range covering both low DO (emergency) and supersaturation at oxygenated inlets (e.g. 0-20+ mg/L or 0-200% sat depending on model), response time suitable for process alarms (seconds to tens of seconds - not minutes if oxygen drives automation), stability in flowing, particle-laden water, and outputs that fit the plant (4-20 mA, Modbus, digital alarms). Placement is as important as sensor type: avoid dead corners, avoid sitting straight in a pure oxygen jet if you want a tank average, and avoid sucking in air bubbles. For oxygen dosing control one often measures both "critical low" in the tank and sometimes inlet after the oxygen unit - knowing that pipe and tank do not share the same DO.

Problems with oxygen probes (regardless of principle) are common and underestimated: fouling by biofilm, feed oils and particles that make readings sluggish or falsely low/high; drying out or damaged membrane/optical face; wrong or forgotten calibration; temperature errors that hit both mg/L and % sat; cable faults and moisture in connectors; air bubbles or stagnant pockets at the sensor; and "zombie values" that look stable in SCADA while the real tank has already dropped. Probes age and drift - compare regularly against a freshly calibrated handheld reference, keep a spare sensor or probe, and treat alarm limits with common sense (fish behaviour, feed response). Good oxygen measurement is maintenance and placement as much as hardware choice. See also monitoring systems and sensors under equipment / AI & monitoring.

Oxygen consumption in biological systems

In biological systems oxygen is consumed mainly by fish respiration, biofilters (nitrification), other microbes and breakdown of organic matter. In a RAS, a substantial share of oxygen use can occur in the biofilter rather than in the fish themselves. Nitrification NH₄⁺ → NO₂⁻ → NO₃⁻ consumes large amounts of oxygen and must be included in the oxygen balance.

There is also a biological paradox: warmer water holds less oxygen, while fish need more oxygen as temperature rises. That makes summer operation and high feed rates the most critical periods on many farms.

For the process engineer

On a commercial farm, oxygen is in practice a mass-balance question. Every kilogram of feed drives fish respiration, biofilter nitrification and oxidation of organics. Oxygen demand can therefore often be estimated from daily feed - but it is not the average that sizes the system. It is the peak load: peak feed.

Peak feed sizes oxygen tanks, LOX, oxygen cones, LHOs, degassers, emergency power, pumps and emergency systems. Always size oxygen and related hydraulics for maximum daily feed (plus safety margin), not for average seasonal operation.

Biology and chemistry: from feed to emission profile

Feed is the dominant mass input on an intensive farm. The fish convert part of the carbon, nitrogen, phosphorus and energy into growth; the rest leaves as respiratory gases, urine/gill excretion, faeces and feed waste. For salmonids (here exemplified with Atlantic salmon / salmonids in grow-out on modern dry feed), process design can therefore be tied to daily feed in kg/day - but the figures are order-of-magnitude for teaching and early sizing, not bankable design. Feed recipe (protein, fat, digestibility), FCR, temperature, size and waste change the profile strongly.

Simplified biology/chemistry: organic carbon in feed is oxidised to CO₂ (respiration) or leaves as particulate organics (faeces/waste); protein nitrogen not built into biomass is excreted mainly as ammonia/ammonium (TAN) across the gills; phosphorus largely follows faeces and dissolved P; solids load filters and the receiving water. In RAS, TAN is also nitrified in the biofilter (extra O₂ demand and alkalinity use) and solids must be removed before they break down into more dissolved COD/BOD.

Key figures per kg feed (salmonids - example)

Approximate design rules of thumb for 1 kg dry feed to Atlantic salmon/salmonids in good growth (illustrative ranges from established RAS/water-quality practice). Always use feed analysis, measured FCR and site data in hard engineering.

From 1 kg feed — simplified mass-flow picture

1 kg feed in dry feed, salmonids

O₂ ≈ 200–300 g fish respiration / kg feed

↓

Weight gain about 0.8–0.9 kg fish at FCR ≈ 1.1–1.2 (good growth)

↓

Solids (sludge) Faeces + feed waste ≈ 200–400 g TSS
N ≈ 10–20 g · P ≈ 4–12 g · C ≈ 80–150 g

Dissolved in the water TAN ≈ 25–45 g N
NO₂⁻-N: normally low (intermediate)
NO₃⁻-N ≈ 25–45 g if TAN is nitrified
Dissolved P ≈ 1–4 g

↓

Biofilter (RAS) TAN → nitrate. Extra O₂ ≈ 140–200 g if all TAN is nitrified (≈ 4.57 g O₂ / g TAN). System total often ≈ 400–700 g O₂/kg feed.

Teaching sketch in the spirit of older Swedish EPA (Naturvårdsverket) nutrient-flow figures for fish farms. Not a permit calculation — use feed analysis, measured FCR and current guidance in licensing.

© 2026 DABCE AB

Teaching approximations for salmonids. Not a substitute for mass balance against actual feed and life stage.

Example (worked illustration, not a fixed answer): peak feed 1,000 kg/day × 0.25 kg O₂/kg feed ≈ 250 kg O₂/day for fish metabolism alone. Add nitrification (e.g. 35 g TAN/kg feed × 4.57 ≈ 160 g O₂/kg feed → ~160 kg O₂/day) and system demand quickly sits around 400+ kg O₂/day before efficiency, losses and safety margin - which is why LOX/PSA, cones/LHO and peak-feed thinking are central in commercial salmonid RAS.

Full emission profile (simplified)

Per kg feed in, one can think of a complete "emission and residue profile" that the farm must handle - either in water treatment (RAS/semi-flow) or via the receiving water (flow-through/cages) within permit limits:

• Gases: O₂ is consumed; CO₂ is produced (respiration). N₂ issues relate more to source water/total gas than to feed - but tank gas balance is still affected.
• Nitrogen: TAN from gills; then NO₂⁻/NO₃⁻ if nitrification occurs; faecal organic N.
• Solids: faeces + feed waste (TSS) - mechanical removal, sludge, receiving water.
• Dissolved/fine organics: BOD/COD, turbidity, fouling on gills and UV.
• Phosphorus and other nutrients: dissolved + particulate - important for discharge permits and eutrophication.
• Heat: metabolism and equipment - relevant in closed halls.
• Biomass out: harvested fish (the goal) - FCR links feed to sellable growth.

1 kg feed in
↓
Growth (depends on FCR)
+
O₂ consumed (fish + biofilter)
+
CO₂ out (respiration)
+
TAN → (NO₂⁻) → NO₃⁻
+
TSS / sludge (faeces + waste)
+
P and nutrients
+
Heat / residual organics

The design point: do not only count "how much oxygen". Count the whole chain - oxygen, CO₂ stripping, solids, nitrogen and sludge - against the same peak feed. See also carbon dioxide (CO₂), ammonia, sludge handling, FCR and FlowFarm for concept sizing.

Delta DO (ΔDO)

One of the most important design parameters is the difference between oxygen into and out of the fish unit. Example: inlet 10 mg/L, outlet 7 mg/L gives ΔDO = 3 mg/L. Larger allowed ΔDO → less water flow needed; smaller ΔDO → more flow. That directly affects pump size, energy use, pipe sizes, tank hydraulics and CAPEX. In many modern RAS facilities, optimising ΔDO (together with oxygen transfer and CO₂ stripping) is among the main levers to cut energy cost.

Common oxygenation techniques

• Low-pressure aeration, diffusers, paddlewheels (especially ponds).
• Oxygen cones and LHOs (Low Head Oxygenators) with pure oxygen.
• Venturi, microbubbles and nanobubbles in more specialised systems.

Pure oxygen is often used at high biomass because transfer efficiency is much higher than with ordinary air - see also the sections on oxygen cones, LHO, aeration and oxygen demand under equipment and oxygen & gas.

Oxygen and CO₂ must be seen together

Oxygen can never be viewed in isolation. Intensive systems must also manage oxygen supply, carbon dioxide stripping, alkalinity, pH, ammonia load and hydraulics. Many high-intensity facilities are limited by carbon dioxide long before they are limited by oxygen supply. A farm can show good DO values while fish performance is limited by elevated CO₂. See the carbon dioxide (CO₂) section.

Design rules

• Keep oxygen near saturation or the desired supersaturation on inlet to culture units.
• Avoid chronically low levels in the tanks - measure where the fish actually are.
• Size the oxygen system for maximum daily feed (peak feed), not the average.
• Build in redundancy for power loss, oxygen failure and pump trips.
• Optimise both oxygen transfer and CO₂ stripping - and think ΔDO and flow together.

Oxygen is in practice the fish farm's equivalent of fuel in an engine. Without enough oxygen, neither the fish, the biofilter nor the production system works as intended. For many modern facilities, oxygenation is therefore one of the most important technical, biological and economic design parameters.