CO2 Accumulation and Degassing in RAS

Where carbon dioxide in recirculating aquaculture comes from, which concentrations sources treat as harmful, how stripping columns and aerators remove it, and how CO2 interacts with pH.

Carbon dioxide stripping is one of the five core processes of a recirculating aquaculture system (RAS), alongside circulation, solids removal, biofiltration and aeration (SRAC Publication 453). Unlike ammonia, CO₂ is not converted by the biofilter; it has to be physically driven out of the water or shifted chemically. This guide covers where it comes from, published tolerance limits, removal devices and its link with pH.

Sources of CO₂ in a recirculating loop

Respiration of the fish and of bacteria in the system produces the CO₂ (SRAC 452; FAO guide). Eshchar, Mozes and Fediuk (2003) cite an excretion of about 15 moles, or 660 g, of CO₂ per kilogram of feed consumed by gilthead sea bream (Sparus aurata), depending on fish weight and temperature. Nitrification adds more: stoichiometry tabulated in a United Nations University Fisheries Training Programme review gives 5.85 g of CO₂ per gram of ammonium nitrogen oxidized. SRAC 453 notes that surface water normally holds about 0.5 mg/L of CO₂, whereas a poorly designed RAS can reach 50 to 100 mg/L.

Oxygenation strategy decides whether this becomes a problem. Because the partial-pressure gradient for CO₂ is much steeper than for oxygen, aeration with atmospheric air that meets the oxygen demand also tends to meet CO₂ removal needs (Eshchar et al.). Pure oxygen systems reduce the air–water interface available for gas exchange, so CO₂ can accumulate to toxic levels; SRAC 453 states that a pure oxygen design must be paired with a CO₂ stripping unit. In a super-intensive marine system in Eilat, further intensification at 94 kg/m³ was limited by high CO₂.

Effects and published thresholds

According to SRAC 452, fish begin to show stress above 20 mg/L because CO₂ interferes with oxygen uptake; affected fish gather at the surface around aerators, become lethargic and sharply reduce feeding. SRAC 453 adds that high CO₂ lowers pH and can stop nitrifying bacteria from working, so ammonia or nitrite rise. Published limits vary with species, size and water chemistry:

FAO guide: normal 10 to 15 mg/L, unfavourable above 15 mg/L.
Eshchar et al. (2003), citing earlier work: a general recommendation of below 10 mg/L; mature salmonids reported to tolerate 10 to 20 mg/L, catfish and tilapia considerably more.
Mota and co-workers (2018), Atlantic salmon (Salmo salar) post-smolts in brackish-water RAS at 12 ppt, six levels from 5 to 40 mg/L for 12 weeks: growth was significantly lower at 12 mg/L and above, and skin dermis was thinner at 40 mg/L; maximum growth occurred below 12 mg/L.
Summerfelt and co-workers (2018), Atlantic salmon post-smolts in freshwater RAS for 384 days at 14.1 °C and alkalinity of 237 mg/L as CaCO₃: 20 mg/L versus 8 mg/L made no significant difference to harvest weight, feed conversion or survival.

Stripping towers and cascade columns

SRAC 453 describes the unpressurized packed column, or spray tower, as the widely used stripping device. Water falls through packing while a blower pushes air, usually drawn from outside the building, upward against it. The units are designed to maximize the gas-to-liquid ratio, typically 0.67 to 1.33 cubic feet of air per minute for each gallon per minute of water, which converts to roughly 5 to 10 volumes of air per volume of water. In the FAO guide the same function is performed by a trickling filter or degasser: water enters over a perforated distribution plate and falls through stacked plastic media to maximize turbulence and contact. FAO also notes that degassing removes free nitrogen gas as well as CO₂.

Moran (2010) measured a 1.7 m cascade column with block-type plastic packing at 15 °C and influent CO₂ of 10 to 60 mg/L. A single pass removed 67 to 89 percent of the CO₂, with efficiency rising as influent concentration rose. Countercurrent forced air did not improve stripping in that column, probably because passive airflow was already sufficient. Efficiency was lower in saline water at equivalent influent levels, and the author concludes that CO₂ removal is more problematic in seawater systems.

Aerators, airlifts and blowers

Blown air: tilapia producers value its simultaneous CO₂ stripping. Moving bed biofilters also contribute because of the air that keeps their media moving (SRAC 453).
Aeration wells: they move water and strip gases at the same time but remove gases less efficiently than a trickling degasser (FAO guide).
Airlifts: Moran (2010) found lower mass transfer than in the cascade column; raising lift height raised the gas-to-liquid ratio and contact time. Eshchar et al. cite stripping efficiencies of 20 to 130 g CO₂ per kWh for airlift pumps.
Paddlewheel versus submerged aerator in a 100 m³ seawater tank: the CO₂ transfer coefficient was 1.55 per hour versus 0.76 per hour; the paddlewheel removed up to 1,200 g CO₂ per hour at about 1.2 kg per kWh (Eshchar et al.).

Air source matters. SRAC 453 reports that most CO₂ problems in blown-air systems occur in winter, when blowers are moved indoors and ventilation is restricted to save heat: the building air becomes CO₂-enriched, CO₂ diffuses into the water, and the result is low pH with high ammonia and nitrite. Keeping blowers outside, or splitting them between inside and outside, limits this.

Interaction with pH and alkalinity

CO₂ reacts with water to form carbonic acid and drives pH down (SRAC 452). For a given total carbonate carbon, a higher pH means less free CO₂, so CO₂ can be controlled by stripping, by pH adjustment or by both (Eshchar et al.). SRAC 452 presents a pH management diagram for 25 °C that plots pH against alkalinity: low pH with adequate alkalinity calls for more aeration, while low pH with low alkalinity calls for sodium bicarbonate together with aeration. Stripping also raises pH locally, which is why calcium carbonate scale forms on air diffusers (SRAC 453).

The effect is not entirely negative. Lower pH reduces the share of toxic un-ionized ammonia in total ammonia nitrogen, and respiratory CO₂ has been described as a way to hold NH₃ at a safe level, even though low pH also reduces nitrification (Eshchar et al.). Degassing design therefore has to be considered together with the pH set point chosen for the biofilter.