Water quality and wastewater measurement

What is a sonde in water quality monitoring?

• One deployment, several parameters. Measured at the same point and the same instant.
• Simultaneity is the real benefit. Many parameters only make sense read against each other.
• Biofouling is the main enemy. Growth on the optics changes readings before it looks dirty.
• Drift is silent. A drifting sensor keeps producing plausible numbers that are simply wrong.

Why combine sensors at all

The obvious benefit is practical. One deployment, one cable, one power supply and one maintenance visit instead of several, which matters a great deal in a chamber, a borehole or a watercourse where access is difficult.

The more important benefit is that the readings become interpretable. Several water quality parameters are close to meaningless in isolation.

Dissolved oxygen is the clearest case: its saturation depends on temperature, so an oxygen figure without a simultaneous temperature reading cannot be converted into percentage saturation, which is usually the ecologically relevant number. Ammonia toxicity depends on pH and temperature. Conductivity is conventionally corrected to a reference temperature. In each case the correction requires a measurement taken at the same place and moment, which is precisely what a sonde provides.

Biofouling, the dominant failure mode

Any surface left in natural water grows a biofilm. On a water quality sensor that film sits directly on the measuring surface, and it changes the reading long before the probe looks visibly dirty.

The effects are specific rather than general. On an optical dissolved oxygen sensor the film consumes oxygen itself, so the sensor reports the depleted conditions inside the biofilm rather than the water it is sitting in. On a turbidity sensor, growth on the optical window scatters light and produces a rising apparent turbidity that has nothing to do with the water.

The characteristic signature is a slow drift over days to weeks that resets abruptly to a different value after a maintenance visit. That step change at cleaning is the diagnostic, and it is a strong argument for logging maintenance events alongside the data.

Mitigations include wipers that mechanically clean the optical face on a schedule, copper components which inhibit growth, and siting the sonde in flow rather than in a stagnant pocket. None of them removes the need for a maintenance interval; they extend it.

Drift, and why it is more dangerous than failure

A sensor that fails outright is a manageable problem. The data stops, the gap is obvious, and somebody investigates.

A drifting sensor is worse, because it keeps producing numbers in a believable range. They are simply wrong, and nothing about them announces it. Decisions get made on them, trends get reported from them, and the error is only discovered at the next calibration, by which point the entire intervening record is suspect.

The practical defences are unglamorous. Calibrate on a defined schedule with traceable standards. Record the as-found reading before adjusting anything, because the as-found value against the standard is what tells you how much drift occurred and therefore whether the preceding data can be trusted. Where a parameter is critical, a second sensor provides a cross check, since two independent sensors rarely drift identically.

For any record intended to support a compliance position, the calibration history is as much part of the evidence as the readings themselves. Data without it is difficult to defend.