Two ponds read the same on a handheld meter and one of them is killing fish. Two reservoirs read the same and one crop responds while the other does not. The reading is not wrong in either case. It is answering a narrower question than the operator is asking, and the gap between the two is where most oxygen troubleshooting goes wrong.
A concentration is a stock. What a root or a gill lives on is a flow.
This post separates the two, then explains why water above the saturation ceiling does not simply hold more oxygen but delivers it on different terms.
What does a DO meter actually measure?
A dissolved oxygen meter reports the oxygen concentration in the water immediately around its probe tip, at that moment. It does not report how fast oxygen is arriving, how much the water can deliver over an hour, or what reaches a root surface or a gill behind its own boundary layer.
Both common probe types respond to the same underlying quantity. An optical sensor reads how strongly oxygen quenches the luminescence of a dye behind a permeable window. A polarographic sensor consumes oxygen at a cathode and reads the resulting current, which is why it needs flow across the membrane and reads low in still water. Neither returns a mass of oxygen in the vessel; both return the oxygen activity at a sensing surface a few millimeters across, then convert it to milligrams per liter using temperature and salinity.
The gap between that reading and what an organism receives is a boundary layer. Around every root and across every gill there is a thin film of water the bulk flow does not scour, and oxygen crosses it by diffusion alone. Consumption at the surface draws that film down below the bulk concentration, so what the tissue actually sees is the bulk value minus a drop that depends on demand, on flow, and on the geometry of the surface. A probe sitting in the bulk never sees that drop.
Timing hides the rest. Demand is not constant: it climbs with temperature, with feeding, with microbial activity, and with the hours after an irrigation event. A single reading taken at a convenient time can miss the daily minimum entirely, which is the only value that matters in a pond where the low point sets the stocking density.
Why doesn't water hold more than about 8 mg/L?
Water holds about 8 mg/L of oxygen only as warm-water shorthand: air-saturated fresh water is nearer 9.1 mg/L at 20 °C and 8.3 mg/L at 25 °C. Henry's law ties dissolved gas to its partial pressure, air is roughly 21% oxygen at one atmosphere, and that sets the ceiling. Warmer, saltier, higher water holds less.
Henry's law is a proportionality, not a mystery: at equilibrium, the concentration of a gas dissolved in a liquid is proportional to the partial pressure of that gas above it. Air at sea level exerts about one atmosphere, roughly a fifth of which is oxygen, and that fifth is the entire driving pressure. Everything a pond, a reservoir, or an open basin does to itself under air happens under that one number.
The temperature term is the one operators feel. Gas solubility in water falls as temperature rises, so the same water body holds noticeably less oxygen on a hot afternoon than at dawn — while the biological demand it has to satisfy is moving the other way. Warm water is not a small correction; it is the mechanism behind most summer oxygen crashes.
What changes the ceiling is what changes the partial pressure. Raise the fraction of oxygen in the gas phase, as an on-site concentrator does by stripping nitrogen out of air, and the driving pressure rises with it. Raise the total pressure, as depth in a deep basin does, and it rises again. The ceiling is not a property of water. It is a property of the gas the water is in contact with, and it moves when that gas does. The classroom section on key physico-chemical properties works through the same relation in more detail.
How can water carry more than 30 mg/L?
Water carries more than 30 mg/L because the pressure driving oxygen into solution is raised, not because the oxygen stays undissolved. A concentrator feeds nearly pure oxygen instead of air's 21%, and Laplace pressure inside each bubble lifts solubility in the water touching it. Suspended gas then resupplies the dissolved fraction as demand draws it down.
Two mechanisms raise the dissolved concentration, and a third keeps it from decaying. The first is feedstock. Henry's law does not care where the gas came from, only what its partial pressure is at the interface, and a pressure-swing concentrator delivers a stream close to pure oxygen rather than the roughly 21% air provides. Multiply the driving partial pressure several times over and the equilibrium dissolved concentration rises with it. The second is curvature. Surface tension puts a small bubble under substantial internal pressure — on the order of 14 atm inside a 200 nm bubble — and Henry's law reads that pressure as a higher solubility in the water immediately touching that interface. Those two are what actually put oxygen into solution above the air-saturation ceiling. The third is inventory, and it is an additional effect rather than the explanation: gas still in suspension has not dissolved yet, and it resupplies the dissolved fraction as demand draws it down. Why the bubbles stay suspended at all is its own post.
That third mechanism is why the dissolved concentration does not immediately vent itself. Ordinary supersaturation is fragile: force extra gas into water under pressure, release the pressure, and the excess comes back out at the first nucleation site it finds. Water carrying suspended gas is not in that state. The reservoir replaces dissolved oxygen at roughly the rate demand and outgassing remove it, so the reading holds rather than collapsing in one step. Across published research and our own deployments, dissolved oxygen above 30 mg/L is achievable this way.
State the limit as plainly as the mechanism. That figure is what the water can be brought to, not a level any given system holds indefinitely. Wherever treated water meets the atmosphere across a large surface — an open reservoir, an aerated channel — it trends back toward the air-saturation ceiling, and how fast depends on surface area, turbulence, temperature, and demand. Treat it as an achievable concentration to be verified on your own water, not a set point.
Does supersaturated water behave differently, or is it just more oxygen?
Supersaturated water behaves differently, not just abundantly. Oxygen moves down a concentration gradient, so a higher bulk concentration steepens the gradient into every root, gill, and biofilm it touches. That is a flux argument about the receiving surface. It is separate from transfer efficiency, which describes how much injected gas dissolves rather than escaping.
Diffusive flux is proportional to the concentration gradient, so raising the bulk concentration does not only raise the amount of oxygen present — it raises the rate at which oxygen crosses every boundary layer in the system. That is the difference between a stock and a flow, and it is why the same water can be described as holding more oxygen and as delivering it faster without those being two ways of saying one thing.
The transfer-efficiency pair belongs to a different comparison, and it is worth naming which. Oxygen transfer efficiency asks what fraction of the gas injected into a basin ends up dissolved rather than leaving at the surface. That is a property of the injection process — bubble size, basin depth, contact time — and not a statement about oxygen crossing a root or a gill boundary layer, which is a diffusion problem at a surface no injector ever touches.
Both move with depth and geometry: a deep tank gives a rising bubble more contact time than a shallow one. Neither is our own measurement.
Read the pair as a contrast between two delivery regimes rather than as a constant either one carries around, and draw no energy conclusion from it on its own.
Where this becomes concrete is at the receiving surface. A root in a saturated soil pore, a gill in a warm pond, a biofilm in a filter: each is separated from the bulk water by a film that oxygen has to diffuse across, and each of them draws that film down as it consumes. Raising the bulk concentration raises the gradient across that film, which raises the flux through it. What that does in an irrigated root zone and what it does in a pond or a recirculating system are separate questions with separate evidence.
What should I measure instead?
Useful measurement starts with what the process consumes, not with what the tank contains. Track dissolved oxygen over time rather than as a single reading, take it where the demand is rather than where the probe is convenient, and pair it with a rate: oxygen uptake, redox potential, or recovery after a load.
Start with a time series. A logged probe at the point of demand costs little and answers the question a spot reading cannot: when the daily minimum occurs, how deep it goes, and how long the system spends there. In a pond that is the pre-dawn low; in an irrigated field it is the hours during and after a set, when the profile is saturated and gas exchange with the atmosphere has stopped.
Then measure a rate, because a rate is what a concentration cannot tell you.
- Oxygen uptake rate. Isolate a sample, stop the supply, and log how fast the concentration falls — it gives the demand side directly.
- Recovery time. After a known load, it gives the supply side.
- Redox potential. A cheap proxy in soil and sediment where a gas probe is impractical, and it registers the anaerobic conditions a bulk-water reading will not.
Two disciplines make the numbers mean something. Measure at the surface that matters rather than in the convenient bulk, since the boundary layer between them is the whole subject of this post. And measure before and after against an untreated control, because oxygen, temperature, and season all move together and a change in the reading proves nothing on its own. That is the point of a controlled pilot: the physics is not in question, and whether delivery is your binding constraint is.