Aquaculture
Water quality monitoring in aquaculture
Dissolved oxygen, ammonia, pH and temperature - what to measure continuously, what to measure by hand, and how to build an alarm chain that survives the power cut that causes the problem.
Fish live in their own waste. Everything in aquaculture water quality follows from that, and from a second fact: the parameter that kills fastest is the one that gives the least warning.
The parameters, ranked by how quickly they hurt you
Dissolved oxygen - hours. A pond that was fine at midnight can be killing fish by dawn, because respiration continues all night while photosynthesis stops. This is the parameter that justifies continuous monitoring with alarms, and it is where redundancy belongs.
Temperature - hours to days. Directly sets metabolic rate, oxygen demand and how much oxygen the water can hold. Cheap to measure, and every dissolved oxygen sensor measures it anyway.
Ammonia and nitrite - days. Rise when the biofilter is undersized, disturbed or cold, or when feeding outruns the system’s capacity. Slow enough for daily or twice-weekly test kits, fast enough to matter.
pH - days to weeks. Matters in itself, and matters more because it governs how much of the total ammonia is present in the toxic un-ionised form. A pH rise turns an acceptable ammonia reading into a dangerous one without the ammonia figure changing at all.
Nitrate, alkalinity, hardness - weeks. Slow-moving background chemistry. Alkalinity deserves more attention than it usually gets, because nitrification consumes it and a system that runs out loses pH stability abruptly.
Alarm design is the actual product
A monitoring system’s value is entirely in whether it wakes someone up. Three principles matter more than sensor specification.
The alarm path must not depend on what fails. A system that alerts over site wifi is useless in a power cut, which is precisely when oxygen fails. At least one alarm route needs independent power and a cellular or SMS path.
Two thresholds, not one. A warning level that gives time to respond and a critical level that means act now. A single threshold set low produces panic; set high it produces alert fatigue and gets muted.
Test it on a schedule. An untested alarm chain is an assumption. Pulling a probe into air once a month and confirming that the message actually arrives on the phone of the person who is on call takes five minutes and is the only thing that verifies the whole chain.
The calibration discipline
Every probe drifts, and drift is undetectable from the data alone because a drifting sensor produces perfectly plausible numbers. The countermeasure is a handheld meter, calibrated independently, walked past the fixed sensors on a fixed schedule and logged.
When the two disagree by more than a small margin, the fixed sensor is recalibrated. When they keep disagreeing, the sensor is replaced. This is unglamorous and it is the difference between a monitoring system and a source of false confidence.
Related
Oxygen measurement is only useful if you can act on it, which means aeration. In closed systems the same parameters are governed by biofilter behaviour, covered in recirculating aquaculture systems. Species-specific tolerance figures are well documented in the open literature - finding agricultural research covers how to locate them.
Dissolved oxygen sensor types
Partner programmes for this category are not in place yet, so no product links are shown. The comparison is by sensor technology.
- 01
Optical / luminescent DO sensors
Sensor class
The current standard for continuous monitoring. A luminescent coating is excited by light and the decay time of its emission varies with oxygen concentration. No oxygen is consumed by the measurement.
Strengths
- Very stable, calibration intervals measured in months
- No flow dependence, works in still water
- Not poisoned by hydrogen sulphide
Limitations
- Higher purchase price than membrane types
- Sensing cap is a consumable, typically annual
- Slightly slower response than a polarographic probe
No partner link for this product yet - the comparison is editorial only.
- 02
Polarographic / galvanic membrane sensors
Sensor class
The older electrochemical method. Oxygen diffuses through a membrane and is reduced at a cathode, producing a current proportional to concentration.
Strengths
- Low purchase cost
- Fast response
- Well understood, widely serviced
Limitations
- Consumes oxygen, so requires water movement across the membrane
- Frequent calibration, membranes and electrolyte need regular replacement
- Drifts noticeably between services
No partner link for this product yet - the comparison is editorial only.
- 03
Handheld multi-parameter meters
Instrument class
Portable spot-checking across DO, pH, temperature and conductivity. The reference against which fixed installations are verified.
Strengths
- Verifies fixed sensors and catches drift
- Covers ponds and tanks that are not instrumented
- Essential for troubleshooting
Limitations
- Provides no overnight coverage, which is when oxygen fails
- Only as good as its own calibration discipline
- Not a substitute for continuous monitoring
No partner link for this product yet - the comparison is editorial only.
Frequently asked questions
What dissolved oxygen level is safe?
It depends on species and temperature, but as a general working rule warm-water species should stay above 4 mg/l and cold-water species such as trout and salmon above 6 mg/l, with growth suppression beginning well before distress is visible. Saturation matters as much as concentration: warm water simply cannot hold as much oxygen, so the same milligram figure represents a tighter margin in summer.
How often should probes be calibrated?
Optical dissolved oxygen sensors typically hold calibration for one to three months; membrane types need it every one to two weeks. pH probes drift faster than most operators expect and should be checked weekly against two buffers. The interval matters less than having one written down and logged - drift is invisible without a record.
Do I need ammonia sensors or are test kits enough?
For most operations, test kits are enough for ammonia and nitrite, because these change over hours to days rather than minutes. Continuous ion-selective ammonia sensors are expensive and need frequent maintenance. Dissolved oxygen is the parameter that justifies continuous instrumentation, because it can go from adequate to lethal overnight.