by Marco G. Pistrin
The call almost always sounds the same. Dissolved oxygen is sitting low, the fish are off their feed and something must be done before the weekend. And the reflex answer is almost always the same: jam more oxygen in.
It is the most reasonable response and on a great many number of farms, it is potentially aimed at the wrong problem.
To be fair, the industry earned that reflex. Oxygen supply is a solved problem and solving it took real work. LOX delivery is reliable and priced sensibly (well, not always), on-site generation has come down far enough that a mid-sized farm can own its supply outright, and cones, oxygenators and low-head and medium head units are mature engineering sold by people who know what they are doing. Twenty years ago, a farm could genuinely run out of gas. Today almost none do and that achievement does not get said often enough.
But supply and transfer are two different variables, and only one of them has a display.
A generator tells you what it produced. A flowmeter tells you what left the manifold. Nothing on a typical farm tells you what crossed the interface into the water and that is the only number the fish ever experience. But between the gas you paid for and the oxygen in the tank, there’s a physical process with its own set of rules and those rules don’t really are about what you spent upstream.
Here is where it turns uncomfortable. Wagner and colleagues tested a low-head oxygenation system across five gas-to-liquid ratios, from 0.10 to 0.83 percent and measured mean absorption efficiencies between 67 and 91 percent. Efficiency peaked at a gas-to-liquid ratio of 0.20 percent, near the bottom of the range tested and diminished as gas flow rose beyond it. Vinci et al., two decades later and in marine systems, found the same wall approached from the other side: the harder a unit is driven toward a high outlet saturation, the less of the supplied gas it captures. At 150 percent outlet saturation, their unit transferred 97 percent of the oxygen delivered to it. Pushed to 230 percent, that fell to 72, and in fresh water, to 58.
Now read that against the reflex stated above. On a unit already working past its efficiency peak, which is where most of them sit at production gas flows, opening the valve raises the gas-to-liquid ratio and lowers the proportion that dissolves. More oxygen does reach the water. It reaches it by wasting a larger share of every cubic metre bought. We are not solving the problem. We are buying past a symptom at a steadily worsening exchange rate. And in turn, oxygen is being produced for the birds, so to speak!
What governs transfer is not the size of the supply. It is fall height, hydraulic loading and submergence; it is contact chamber height, orifice size, the number of stages. Not one of those appears on the generator’s display and not one of them improves because something was turned up.
I have stood on a farm where the on-site generator could produce several times what the fish were consuming. We opened the units up and watched the flowmeters climb and the water barely answered; it would not hold setpoint through the afternoon feed regardless of what was pushed into it. And this was a farm on river water, which hands you a night-time oxygen sag for nothing. Put a stress event on top of that sag, a beaver falling into the head of a raceway, or anything that sends the fish into a panic and their demand through the roof and the shortfall stops being an inefficiency and becomes a mortality. There was never a shortage of oxygen on that site. There was a shortage of contact, distributed unevenly across a row of LHO units that nobody had ever measured individually.
Which is precisely the diagnostic that almost nobody ever runs and it costs only a morning. Simply take dissolved oxygen in and out of every unit separately and reconcile the sum against total flow. And it is worth noting that a fouled unit still shows gas moving cheerfully through its flowmeter, as does a short-circuited one. An uneven split between parallel units is invisible from the header and obvious the moment you measure the resultant flow.
The cost of not knowing is paid three times over. First in gas, where the spread between 91 percent absorption and 65 is a further quarter of everything you purchase, vented, hour after hour, for the life of the installation. Second in capital, because the honest answer to a transfer problem is never a larger generator, or a bigger unit (LHO or whatever you have) and yet that is frequently what gets bought. And third, most expensively, in the fish; the suppressed appetite, the feed conversion that drifts over a season, the growth curve that flattens without ever tripping an alarm, because the animals are living in a sag that nobody resolved.
Then comes the day the system is asked for everything at once; a grading, a warm spell, a pump down and the margin you assumed you had was never margin. It was inefficiency you had been paying for and reading as headroom.
So, before you accept the next quote for a new shiny oxygenation ‘thing’: the farm almost certainly makes plenty of oxygen. The number nobody has is how much of it reaches the water and which unit is losing it. That number costs a morning with a meter and it has never once come back flattering.
You are not buying oxygen. You are buying contact. Nobody has ever sold a farmer a cubic metre of contact time, which is precisely why it goes unmeasured and precisely what the fish are short of.
