Optical sensor drift detection in water quality monitoring

Optical instruments used for turbidity and suspended-solids monitoring can deliver stable data for months, then gradually begin to report a biased result. The change may come from a thin film on the optical window, biological growth, sediment abrasion, trapped air, a damaged wiper, or a shift in the electronics. Because the measurement often continues to look plausible, drift can pass unnoticed until it has affected a project decision.

Routine wipe tests and reference checks provide a practical way to detect this movement early. A wipe test examines the condition of the sensing face and the response before and after cleaning, while a reference check compares the cleaned instrument with a known optical condition. Used together, they help distinguish genuine changes in water quality from fouling, optical-window contamination, and gradual sensor instability.

Why optical readings move

A turbidity monitor estimates the scattering or attenuation of light caused by particles in water. A suspended-solids sensor uses a related optical response, although its relationship with concentration depends strongly on particle size, colour, mineral composition, shape, and the local calibration. The sensing path can therefore be affected by both the water and the physical condition of the instrument.

A coating only a fraction of a millimetre thick can alter the amount of light reaching a detector. Fine clay may settle on a window, while algae or bacterial films can develop in warm, nutrient-rich water. In coastal Queensland, an instrument near an estuary may experience salt deposits during drying cycles, followed by organic fouling when it is submerged again. In a dredging plume, abrasive sand can produce fine scratches that permanently change the optical response.

A drift check should also account for bubbles and installation effects. Bubbles attached to the optical face can scatter light in a way that resembles a short turbidity spike. A sensor mounted too close to a wall, pipe, intake, or bed can see reflected light or an unrepresentative particle field. Understanding the underlying optical technology helps operators separate these influences from an electronic fault or a genuine change in suspended material.

What a wipe test reveals

A wipe test is a controlled comparison between the sensor’s reading before cleaning and its reading after the optical surfaces have been cleaned. The instrument should remain in the same water, position, and measurement mode for the comparison wherever practical. If the signal changes materially after wiping, contamination has been influencing the measurement.

The result is most useful when the operator records the raw or displayed value before and after the test, along with time, depth, water conditions, and the cleaning method. A small change may be normal for a lightly fouled instrument. A large change, particularly when repeated at every visit, indicates that the deployment interval is too long or that the anti-fouling arrangement needs attention.

Cleaning technique matters. Use the method specified for the instrument and avoid abrasive pads, harsh solvents, or force that could scratch a window or damage a wiper seal. A soft, lint-free material and clean water are commonly appropriate, but the manufacturer’s instructions take priority. The aim is to remove the film without changing the optical surface or leaving detergent residue.

Using reference checks correctly

A reference check exposes the cleaned sensor to a stable, repeatable optical condition. This may be a certified turbidity standard, a manufacturer-supplied reference device, or another approved check fixture. The chosen reference should match the instrument’s design and measurement range. A check made in an uncontrolled bucket of site water is a comparison, not a true reference check, because the sample may settle or change during handling.

Allow the sensor and reference medium to reach a stable temperature, remove bubbles, and position the instrument consistently. Keep stray light away from the test area and avoid touching the optical window during the measurement. Record the observed value, the reference value, the instrument identification, and the date. Repeating the same process at each service visit makes gradual movement easier to see.

A reference check is not automatically a full calibration. Calibration changes the relationship between the instrument output and a known standard, while a check assesses whether the current response remains within an acceptable range. If a check fails, wiping and repeating the test can identify surface fouling. If the result still fails, the instrument may require calibration, inspection, repair, or replacement.

Building a field routine

A dependable field routine starts before the instrument is removed from the water. Record the live reading and note weather, flow, tide, visibility, recent rainfall, and nearby activity. This is valuable in Australian catchments where a storm in the Blue Mountains can quickly change river sediment, or where a spring tide may alter an estuarine reading without any instrument fault.

Inspect the cable, mounting hardware, wiper, protective guard, and connectors. Look for loose brackets, marine growth, corrosion, cuts, or sediment packed around the sensing head. Take a photograph when practical, especially for long-term projects or compliance monitoring. Visual evidence can explain a trend that is difficult to reconstruct from a data file several months later.

After the initial observation, perform the wipe test and record the post-cleaning result. Follow it with the reference check under controlled conditions. If the reference response is acceptable but the before-and-after wipe difference is growing, schedule more frequent cleaning. If the cleaned response remains outside the permitted range, keep the result flagged and follow the relevant service procedure rather than quietly applying an adjustment.

Reading drift signatures in the data

The shape of a time series can reveal the source of a problem. A slow upward movement during a deployment, followed by an abrupt return after cleaning, is a classic fouling signature. Regular saw-tooth patterns may indicate repeated accumulation and servicing. Short isolated spikes that coincide with rainfall, vessel movement, or pump operation may be real events rather than drift.

A stable reference result with unusual field data points towards installation, sampling, or environmental effects. A failing reference result after the optical face has been cleaned suggests a sensor or electronics issue. A reading that changes when the cable is moved can indicate connector damage or water ingress. These interpretations should be checked against maintenance records and, where possible, a second measurement method.

Thresholds should be based on the instrument specification, the project’s data-quality objectives, and the consequences of an incorrect result. A research deployment may tolerate a different uncertainty from a dredging monitoring programme linked to an environmental approval. Trend limits, repeatability limits, and maximum allowable wipe correction should be documented before the data are reviewed, so decisions are not shaped by hindsight.

Managing Australian field conditions

Australian deployments span tropical rivers, temperate reservoirs, remote inland waterways, and exposed coastal works. In Darwin or Far North Queensland, warm water can accelerate biofouling and make a fortnightly service interval inadequate. In the Murray–Darling Basin, low flows, fine sediment, and strong seasonal changes can produce a different balance between deposition and biological growth. Site-specific service planning is more reliable than adopting a single interval across a national programme.

Dredging projects around Port Hedland, Gladstone, Newcastle, or Port Melbourne may combine high suspended solids with vessel movement, tides, and heavy equipment. A sensor can be clean and functioning correctly while the plume changes rapidly around it. Operators often need to compare the instrument with a depth-integrated sample, another monitoring point, or a laboratory result before deciding that an unusual reading is drift.

Remote work also affects quality control. A crew travelling to a station near Kununurra or across a long outback access route may have limited access to replacement parts and certified standards. Keeping a clean reference kit, spare wiper components, suitable lint-free wipes, and a protected field log can prevent a simple fouling issue from becoming a long data gap. Local contractors should receive the same written procedure as the central project team, with Australian spelling, units, and site names used consistently in records.

Connecting checks with instrument management

Maintenance records should make each result traceable. Include the instrument serial number, firmware or configuration where relevant, deployment location, reference material identification, reference value, observed value, wipe correction, operator, and follow-up action. Store the record with the time series so an analyst can distinguish corrected data from original observations.

A practical record can use three status levels: acceptable, review, and action required. “Acceptable” means the reference response and physical inspection meet the project limits. “Review” indicates a developing wipe correction, an unusual installation condition, or a result that needs comparison with another measurement. “Action required” covers a failed reference after cleaning, visible damage, repeated instability, or a result outside the manufacturer’s operating range.

For equipment selection, integration, and replacement planning, the current instrument range provides useful context about the available monitoring systems and their intended applications. D & A Instruments equipment is now supported by Campbell Scientific, so product-management and contact information should be checked through the current support pathway before arranging service or sourcing components.

Keeping data defensible

The strongest drift-detection programme treats a wipe test and a reference check as part of measurement assurance, rather than as occasional maintenance chores. The wipe test shows how much the field condition is influencing the sensor. The reference check shows whether the cleaned instrument still responds as expected. Together, they create evidence for accepting, flagging, or investigating a period of data.

Terminology should be consistent across field teams, laboratories, consultants, and asset owners. Terms such as turbidity, suspended solids, fouling, calibration, verification, drift, and reference standard can be used differently between organisations. A shared water-quality glossary helps reduce ambiguity when procedures, reports, and regulatory submissions are prepared.

When the routine is applied consistently, drift becomes easier to manage before it compromises a decision. A documented wipe correction, a repeatable reference response, and a clear escalation rule give Australian monitoring teams a defensible basis for interpreting optical data across rivers, estuaries, dredging sites, reservoirs, and marine installations.