A selection of articles

Gas bubble trauma


Gas bubble trauma or Gas bubble disease

Some time ago I received this email:

Hi - I am from Malaysia and a fairly avid koi hobbyist. I would like to seek clarification of DO saturation limits. In page 89 of your book, you have stated 300% as saturation limit with minimum as 6 ppm. Other literature mention the possibility of gas bubble disease when DO exceeds 110% of optimum eg 8.6 at 25C. 

I share my answer below in the hope that it may clear up some of the questions you and other koi keepers might have with gas bubble disease. Do some slow reading.


Supersaturation of pond water is a problem in aquaculture that is reported to be caused:

  1. when cold, saturated surface water is heated underground under pressure and is sourced from a borehole, well or spring;
  2. when cold, gas-rich ground water is heated up for use;
  3. when a body of water becomes supersaturated due to a large and continuous temperature change;
  4. when leaking mechanical water pumps sucks air at the intake side and “crunches” the air into the water;
  5. due to poor, incomplete aeration;
  6. due to poor, incomplete mixing of thermal layers;
  7. due to photosynthesis of macrophytes life in clear pond water;
  8. due to photosynthesis of zooplankton life in culture ponds;
  9. vigorous aeration of pond water.


We are of the opinion that causes 1 to 6 are more likely the culprit in case of gas bubble trauma (GBT), cause 7 normally in association with cause 5-6, while 8 and 9 will rarely, if ever cause problems to fish life. In fact 9 is a cure rather than a cause.


One of the problems with reports on gas bubble trauma is the seemingly conflicting reports that emanate from condensed, incomplete or inconsistent data in the popular press. Readers get confused and neurotic about what oxygen can do or not do to their koi collection. It is not always clear whether we are talking about air or oxygen and whether it is a suburban concrete koi pond, a mud pond for fish culture or a natural lake being referred to. We refer to terms total gas pressure (mm Hg or percent), percent saturation, oxygen tension (mm Hg), oxygen concentration (mg/liter or even ml/liter). These are all legitimate physical measures in their own right, but we do not seem to find consistent comparative data of real cases of gas bubble disease and the circumstances under which it occurred.


The difference between the atmospheric or barometric pressure (BP) and the total gas pressure (TGP) at a given point is called ΔP with the relationship: TGP = BP + ΔP. More frequently TGP is expressed as a percent of the barometric pressure at that point to give: TGP % = 100 x (BP + ΔP) / BP which is a measure of the percent saturation of the water with atmospheric gasses.


We can approximate that at one meter depth water ΔP = 74mm Hg (=1 m H2O) and that if we had a small bubble at that point, the air would be subjected to a total gas pressure of 760 + 74 = 834 mmHg or 109.7% saturation.


The air in the bubble is 20.94% oxygen and thus has a partial pressure, PO2 = 834 x 0.2094 = 174.6 mmHg. The oxygen in the air in the bubble will on contact enter the water until the pressure in the water is same as that in the air, namely 174.6 mmHg, or converted 174.6/17.13 = 10.19 mg/liter at 20 deg C.


At sea level and 20 deg C the solubility of oxygen in fresh water is about 9.08 mg/liter, so we increase oxygen concentration by 1.11 mg/ liter to 10.19 mg/liter maximum or 100 x 10.19/9.08 = 112% oxygen saturation at the point where we are injecting a bubble (1 m depth).


[At 3 meters deep the result will be 984 mmHg air resulting in 206.04 mmHg oxygen pressure resulting in 12.02 mg/liter or 132% saturation maximum. At different temperatures the solubility of gases also changes, eg, for reference: at 7 degrees C the solubility of oxygen is 12.13).]


So first thing to realize is that no matter how much aeration we apply by blowing in air at one meter depth, we cannot exceed 112% at one meter (or 132% at 3 meter) and it will always be less because as the bubble travels upwards it expands again and a certain amount of degassing will take place. (any of the gasses in the water depending on their solubility and relative abundance) In the end the DO will only increase fractionally above the solubility of oxygen at that barometric pressure and temperature.


Because the bubbles move up towards the surface, an upward current is created that replaces surface water with bottom water. If not interfered with (pumping, etc.) and allowed to gain momentum, a large body of water can be set in motion in this manner. Under such conditions some sort of equilibrium will be reached between the volume of air put in, the oxygen dissolved at the surface, the amount of oxygen consumed in the pond/biomass and the amount of degassing that takes place. The more oxygen that is demanded, the more will be available. It is as simple as that and very good for the pond health. Under these conditions I fail to see how vigorous aeration with air stones can cause gas bubbles to appear on the gill or gas emboli in the gill lamella, blood and other cavities of the fish. I have also never witnessed it. Degassing will rather be the norm.


One should also understand that under conditions of high temperature, or high altitude or simply high fish load, we really aerate to supply enough oxygen for the demand of the living things in the pond, whether bacteria, plants or fish, and to maintain DO levels at what the local temperature and barometric pressure allow.


Koi keepers in tropical regions sometimes chill their pond water from above 30 down to 25OC or so to increase the DO above the perceived minimum of 8 mg/liter. This practice invites trouble in case of a prolonged power failure. Both chiller and air pump stop working in such conditions. Without proper mixing and degassing, the temperature will quickly rise causing conditions of troublesome supersaturation. Power backup to maintain an air pump is essential to avert a possible disaster.


Conditions can be speculated that can give rise to supersaturation with air stones when bubble size is too small and the pressure too low to cause a substantial column of upward flowing bubbles or when air stones are too deep, (3m = 129%, 5m = 148%). This peculiar situation could occur with a malfunctioning or underrated diaphragm or piston air pump.


The same physical realities should hold for venturies, but I have not done any calculations on them and simply recommended that they should preferably not be deeper that 30 cm below the surface for a 20 m3/hour flow rate. Under these conditions we know it is safe.


If we now consider a high pressure centrifugal pump. At the outlet of a typical 0.45 kW centrifugal pump, ΔP can be as much as 14 x 74 mm Hg = 1036 mm Hg potentially causing a saturation of 236% if sucking in air at the inlet. (and 275% for a 1.1 kW pump). A bubble leaked inward can be squashed to zero, so that there is no bubble to expand normally as low pressure areas are reached. Due to the natural cohesion of water molecules, bubbles will only appear when a certain threshold is breached and then the first places to happen will be on points of unevenness. If you put a dry hand and arms in such supersaturated water, bubbles will instantly form on your skin, and particularly on the hair of your arms. If you heat the water you will see the same bubbles appearing on the inner surface of the vessel.


Such supersaturated water is deadly to fish and will form emboli in the capillary vessels of the gills which may block them and consequently the tissue dies.


Some large air stones in such a pond will probably get rid of all the supersaturation by degassing and we will be none the wiser.


Also note that as we aerate with air we also bring lots of nitrogen into the picture, and it is the nitrogen component of delta-P that may cause problems for the fish. (expand)


Consider a body of water at 15 deg C. If it is saturated with O2 at should have 10.07 mg/liter O2 (at sea level) If the temperature should increases to 20 degrees saturation level will be 9.08 mg/liter if the excess oxygen can escape. If the water is not agitated the water at the surface will quickly degas, but lower down, the water will remain supersaturated by 100x10.07/9.08 = 110.9% and this will approximately hold for nitrogen as well.


Further reading

Colt (1986) estimated that ΔP values of as little as 22 mmHg within the fish could cause bubbles to form. ΔP values of 25 to 50 on a continuous basis may cause GBT to develop, while 50 to 200 mmHg would cause acute GBT that Fidler and Miller (1994) summarised as:

-          Bubble formation in the cardiovascular system, causing blockage of blood flow and death

-          Overinflation and possible rupture of the swim bladder in young (or small) fish, leading to death or problems of overbuoyancy.

-          Extracorporeal bubble formation in gill lamella of large fish or in the buccal cavity of small fish, leading to blockage of respiratory water flow and death by asphyxiation

-          Sub-dermal emphysema on body surfaces, including the lining of the mouth. Blistering of the skin of the mouth may also contribute to the blockage of respiratory water flow and death by asphyxiation.


What is yet not clear to me is what does it takes for ΔP > 20 mmHg inside the fish to happen and how does it happen. Why would it happen?

So (if we remember from above) according to Colt (1986), trouble can be expected with GBT with less than 110% supersaturation (of air). We can see how it can happen with mechanically supersaturated water, and we can also understand how it can happen if a body of water with little movement heats up with 5 degrees. We can understand this will never happen in a well aerated pond with lots of water movement.

Boyd (1990) quotes literature about ponds with carp exposed to 150% oxygen saturation had a greater frequency of disease that those with 100 to 125% saturation, and that carp died when 300% oxygen saturation was reached.

Carp fry died from GBT in a hatchery during rapid phytoplankton growth with DO levels of 20 mg/liter (thus say 220% saturation) (Faruqui, 1975)

Goldfish died in a pond where photosynthesis and solar heating cause pO2 values of 450 mmHg (that is 290% oxygen saturation, assuming 20deg C) (Takashi and Yoshihiro, 1975)

Boyd (1990) are of the opinion that in culture ponds self shading prevents supersaturation from phytoplankton blooms. He and co-workers sampled surface water in channel catfish ponds. In the morning ΔP values were –20 to –200 mmHg and in the afternoons above 100 mmHg but sometimes above 300 mmHg. No GBT was observed.

He concluded that only in shallow ponds with clear water and abundant macrophytes, the entire water volume can become supersaturated and there will be no safe haven for the fish.

Boyd (1990) concluded that GBT has not been carefully studied and we want to agree with him.

We want to add lack of agitation (by whatever means) and temperature change as factors that will trigger or exacerbate the condition of GBT.


Now back to your questions after a long detour:

While koi will survive at 3 mg/liter oxygen, 6 mg/liter can be regarded as a minimum to ensure growth (metabolizing). (Boyd, 1982 and Michaels, 1988). (Small carp will even acclimatize to as little as 0.3 mg/liter!) Having more than 6 mg/l available for your fish can not add anything more to their growth or well being other than limiting the chances of the DO getting below 5 mg/liter, but what you must have is the means (by aeration or otherwise) to replenish continuously the DO used up by all the fish and biological process in the pond.

The 300% oxygen saturation we quoted come from two references (Boyd, 1990 and Schlotfeldt and Alderman, 1995, from the EU Association of Fish Pathologists). Their reasoning is that the problem of GBT is not so much with oxygen than with nitrogen, and that fish can tolerate such high oxygen levels. The problem comes with very specific conditions.

I admit that the quoted value may be misleading and, based on all of the above, could lead to GBT.

It is also correct that GBT is quoted as being possible if DO or air saturation is above 110% (delta-P > 46 mmHg). This is all based on Colt’s work, which apparently needs further study.

My feeling is yes, GBT can happen, but it is a function of the time available for adjustment. Just as a diver (or a fish) will perish if brought quickly from depths to the surface, likewise a fish will need time to handle the transition. Very specific conditions have to coincide to cause trouble.

I hope it answers most of you question.





Colt, J. 1986. Gas supersaturation - impact on the design and operation of aquatic systems. Aquacultural Engineering. 5: 49-85.

Colt, J., and H. Westers. 1982. Production of gas supersaturation by aeration. Trans. Am. Fish. Soc. 111: 342-360.