Sensor Drift and Recalibration Schedules for Aquarium Probes

Why pH, oxygen, temperature and conductivity probes drift, how USGS separates drift from fouling, which deviations justify recalibration, and how to verify and log calibrations.

Drift versus fouling

Every immersed sensor loses accuracy in two distinct ways, and the U.S. Geological Survey (USGS) separates them deliberately. Fouling is a coating on the sensing surface: silt, oil, algae, bacterial film or precipitates. Its error is measured as the difference between the reading before and after the sensor is cleaned. Calibration drift is a change in the sensor or its electronics and is measured after cleaning, by comparing the sensor with a standard of known value. A recalibration performed on a dirty probe bakes the fouling into the new calibration. USGS also notes that temperature and conductivity sensors are much less affected by fouling than dissolved-oxygen, pH and turbidity sensors, which sets the order of attention in an aquarium.

Drift mechanisms by sensor type

pH glass electrodes

A pH probe is a combination electrode: a proton-selective glass bulb filled with a pH 7 buffer and a silver/silver-chloride reference. According to the USGS National Field Manual, a properly working electrode drifts about 0.1–0.2 mV per hour, its zero potential varies by about ±2 mV from day to day, and after conditioning the slope stays within ±0.2 percent of the theoretical Nernst response (59.16 mV per pH unit at 25 °C). Sensitivity loss comes from a partly clogged reference junction or a change in filling-solution concentration; the remedy is reconditioning or replacement, not more frequent calibration. USGS uses slope as the health indicator: a slope below 94 percent of theoretical signals deterioration, and an electrode at 90 percent or less must not be used. Wikipedia adds that equilibration time grows as the electrode ages.

Dissolved-oxygen sensors

Membrane-type (galvanic or polarographic) DO sensors drift through the membrane itself: it fouls, loses elasticity, and the potassium-chloride electrolyte behind it degrades. USGS expects the membrane to need replacement every 2–4 weeks in the field and warns of anode poisoning by sulfide, visible as a gray or black anode and confirmed by chronically low readings that persist after a new membrane. Because oxygen solubility and membrane permeability both change with temperature, temperature compensation is essential. Optical (luminescent) DO sensors have no membrane to deplete, do not depend on flow, show few interferences when clean and are described as having long-term stability, but they still foul.

Temperature

USGS describes thermistors as reliable, accurate and durable sensors that need little maintenance. Their drift is small but not zero: USGS requires accuracy within ±0.2 °C, an annual five-point calibration over 0–40 °C in a water bath against a NIST-traceable reference thermometer, and two-point checks at the expected annual minimum and maximum three or more times a year.

Conductivity and salinity

Conductivity cells are among the least fouling-prone sensors. Readings are compensated to 25 °C, and salinity is calculated from compensated conductance rather than measured directly, so a temperature error propagates into a salinity error. Routine cleaning is rinsing with deionized water; platinum-coated electrodes must never be brushed.

Calibration intervals from instrument literature

There is no universal interval. Wikipedia's summary of metrology practice lists the inputs: the manufacturer's recommended interval, how heavily the instrument is used, and the results of previous calibrations. An instrument that repeatedly passes can have its interval lengthened; one that repeatedly fails needs a shorter one. The reference standard should carry no more than one quarter of the uncertainty of the instrument being checked.

USGS continuous-monitor practice offers a working schedule. Service visits are weekly or more often where accuracy matters, and as frequent as every third day in warm, nutrient-rich water, which is a fair description of a stocked aquarium. At each visit the probe is checked against standards and recalibrated only if the deviation exceeds the calibration criteria: ±0.2 °C, ±0.2 pH unit, ±0.3 mg/L DO, and ±5 µS/cm or 3 percent for conductance. Smaller deviations are left alone, because adjusting within the instrument's own noise does not improve the data. Data are discarded when the required correction exceeds the maximum allowable limits: ±2 °C, ±2 pH units, ±2 mg/L or 20 percent DO, ±30 percent conductance. A schedule derived from these sources:

pH: verify weekly against two buffers; recalibrate when the check deviates by more than 0.2 pH unit; record the slope; recondition or replace below 94 percent.
Membrane DO: inspect weekly; expect membrane and electrolyte replacement every 2–4 weeks, followed by recalibration.
Optical DO: clean and verify weekly against an independent meter; recalibrate on a deviation above 0.3 mg/L.
Temperature: two-point check against a reference thermometer several times a year; act on a deviation above 0.2 °C.
Conductivity and salinity: verify against a standard solution weekly to monthly; recalibrate above 5 µS/cm or 3 percent.

Verification with reference solutions

A calibration check is only as good as the standard. USGS requires pH buffers certified traceable to NIST Standard Reference Material, and two buffers that bracket the expected sample pH: pH 7 plus either 4 or 10. Buffers are labelled with the date opened and the expiration date, discarded on expiry, never poured back into the stock bottle, and brought to the sample temperature before use, because buffer pH is itself temperature dependent (pH 10 changes most). Carbon dioxide from the air affects pH 10 buffer more than pH 7, and pH 7 more than pH 4, so bottles stay capped. For temperature the reference is a NIST-traceable thermometer. For DO and conductivity USGS reads an independent, freshly calibrated field meter next to the installed probe before and after cleaning, which reveals fouling and drift separately.

Documenting calibrations

USGS keeps a pH meter log book with the Nernst slope and the millivolt readings in the pH 7 and pH 4 buffers after every calibration. Field notes record the initial reading before cleaning, the reading after cleaning, the calibration-check values and whether a recalibration was made. UF/IFAS asks recirculating farms to keep signed logs of water-quality measurements, system checks and actions taken. The value of the record is diagnostic: USGS states that a record needing frequent, large corrections points either to heavy fouling or to a sensor with calibration-stability problems, which is the signal to shorten the interval, recondition or replace the probe.