How to Reduce Noise in Turbidity Data From Wave-Induced Motion
Turbidity measurements collected in rivers, harbors, coastal zones, and dredging areas often contain rapid fluctuations that do not represent an actual change in suspended sediment. Wave action can move a sensor through the water column, tilt its optical path, stir nearby particles, and alter the distance between the instrument and the surrounding water. The resulting trace may look like a series of sediment pulses even when the plume is relatively stable.
The problem is especially important when an optical turbidity monitor is mounted on a buoy, vessel, pontoon, profiling frame, or nearshore structure. Motion-related variation can obscure plume boundaries, distort compliance measurements, and make it harder to compare observations from different deployments. A reliable solution combines mechanical stability, suitable sampling settings, signal processing, and site-specific validation.
The objective is not to remove every rapid change. Short-lived increases can indicate a genuine resuspension event, vessel passage, discharge, or wave-driven bed disturbance. The goal is to identify the difference between physical water-quality change and measurement noise, then preserve the information that matters for the application.
What Wave Motion Does To An Optical Measurement
An optical turbidity sensor estimates suspended material by measuring how particles scatter or absorb light within a defined sensing volume. When the instrument moves, the optical geometry can change relative to the water and particles. A sensor that tilts toward the bed may encounter a denser layer of sediment, while a sensor that rises into clearer water may show a sudden decrease. Neither change necessarily reflects a uniform shift in turbidity around the monitoring location.
Surface waves also create vertical and horizontal orbital motion. In shallow water, this movement can lift fine sediment from the bed and produce a real increase in suspended solids. At the same time, the instrument may oscillate through that concentration gradient. The recorded signal therefore combines environmental variability with the response caused by sensor movement.
Bubbles and splashing create another source of interference. Air passing through the optical path can scatter light strongly and cause isolated spikes. A loose cable, swinging frame, or poorly restrained mounting arm can amplify these effects. Before changing software settings, inspect the deployment hardware and determine whether the sensor is moving, rotating, contacting the structure, or becoming exposed during low water or wave troughs.
Separate Motion From Genuine Turbidity Change
A useful diagnosis begins with synchronized observations. Compare the turbidity record with water level, wave height, pitch and roll, current velocity, or an accelerometer signal when those data are available. Motion noise often appears as a repeated oscillation related to the wave period, while a true sediment event may persist after the sensor settles or appear across several instruments at the same time.
The shape of the signal provides additional clues. Isolated one-sample spikes commonly result from bubbles, electrical interference, or brief obstruction of the optical path. A periodic rise and fall suggests wave-induced movement or a sensor passing through a concentration gradient. A sustained step change may indicate a real plume, a changed water level, fouling, or a shift in calibration. Looking at the raw time series before applying smoothing helps preserve these distinctions.
Use a reference measurement where possible. A second fixed sensor, a water sample, a transmissometer, or a nearby station can show whether the observed variation is local to the moving instrument. The support answers available from D & A Instruments can also help clarify instrument behavior, terminology, and application-specific operating considerations.
Stabilize The Deployment Before Filtering
Mechanical design is usually the most effective first control. Mount the turbidity monitor on a rigid frame with sufficient mass and a low center of gravity. Keep the optical head away from flexible members that can vibrate in current or waves. Secure cables so they cannot pull on the instrument, form loops in the flow, or strike the sensing face.
The orientation of the sensor matters. Install the optical window according to the manufacturer’s recommended geometry, and avoid pointing it directly toward the bed where wave-driven resuspension is strongest unless that is the phenomenon being studied. In areas with significant surface disturbance, place the sensor deep enough to remain submerged through the expected wave cycle while maintaining a representative sampling depth.
A fixed piling or well-designed bottom frame can provide better repeatability than a free-floating package. If a buoy is required, use a damped bridle, streamlined frame, or gimbal arrangement that limits rotation without blocking water exchange. For vessel-based work, consider the effect of hull motion, propeller wash, and station-keeping errors before interpreting short-term fluctuations as plume behavior.
Deployment geometry should also account for the site’s sediment profile. A sensor positioned just above a mobile bed may correctly record wave-driven suspension, while one placed higher in the water column may provide a more representative estimate of the bulk plume. Document the elevation, orientation, water depth, wave conditions, and mounting arrangement so later changes in the signal can be related to field conditions.
| Source of variation | Typical signal appearance | Field check | Appropriate response |
|---|---|---|---|
| Sensor tilt or vertical movement | Repeating oscillation linked to wave period | Compare with motion, level, or wave records | Improve frame stability and mounting depth |
| Air bubbles | Isolated sharp spikes or brief dropouts | Inspect during waves, tides, and vessel activity | Reposition the sensor and remove bubble traps |
| Bed resuspension | Sustained or repeated rise during energetic conditions | Compare with current, wave, and bed conditions | Retain the event and interpret it as environmental data |
| Cable or frame vibration | Irregular bursts, often at specific flow speeds | Observe deployment hardware in the water | Secure cables and stiffen the support |
| Fouling or deposits | Gradual drift, increased baseline, or erratic response | Inspect and clean optical surfaces | Establish maintenance intervals and quality flags |
| Electrical interference | Synchronous spikes across channels or devices | Check power, grounding, and communications | Improve wiring, shielding, and data acquisition |
Match Calibration To The Water And Sediment
Calibration cannot remove movement, but it determines whether the remaining signal has a defensible relationship to suspended solids or turbidity units. Particle size, color, mineral composition, organic content, salinity, and sensor geometry influence optical response. A calibration developed in clear freshwater may perform differently in estuarine or marine water, even when the nominal concentration is the same.
Collect representative samples across the expected concentration range and under the conditions in which the monitor will operate. Include low, moderate, and high turbidity rather than relying on a single point. If waves are likely to resuspend bed material, include samples from that condition or use a separate relationship for the resuspension regime. Record whether samples are collected near the sensor, at the same depth, and at the same time as the optical readings.
The distinction between turbidity and suspended-solids concentration should remain clear. Turbidity is an optical response, while suspended solids are a mass concentration determined through laboratory analysis. A monitor may provide excellent repeatability in turbidity units without producing a universal solids concentration. For a detailed discussion of site-specific calibration, consult this calibration guidance when planning freshwater, brackish, or marine deployments.
When wave conditions change the local particle population, recalibration may be necessary even if the instrument itself remains stable. A storm can introduce coarser particles, organic debris, or a different mineral fraction. Treat these shifts as part of the monitoring model rather than assuming that one calibration curve applies indefinitely.
Select Sampling Settings That Preserve Meaning
Sampling frequency should be high enough to capture the environmental process of interest, but excessive raw data can make wave-related variation look more important than it is. If the objective is to detect a dredging plume over minutes, recording many measurements per second may provide little operational value unless motion diagnostics are also being collected. If the objective is to study wave-cycle resuspension, a higher rate may be essential.
Averaging is useful when applied deliberately. A short moving median can suppress isolated spikes from bubbles, while a moving mean can represent average exposure over a defined interval. Median filters are generally less affected by single extreme values, but they can flatten genuine short pulses if the window is too long. Choose the window in relation to wave period, sensor response, travel time of the plume, and the reporting interval required by the project.
Avoid filtering raw data irreversibly. Preserve the original measurements, quality flags, and processing parameters. Produce a separate derived series for reporting so that analysts can revisit unusual events and test another filter if the project interpretation changes. A simple quality-control record should identify missing data, saturation, implausible jumps, communication errors, cleaning events, and periods when the sensor was out of the water.
Motion-aware processing can improve confidence. If an accelerometer or tilt sensor is available, flag turbidity measurements collected during excessive movement rather than automatically deleting them. A motion flag can distinguish “high turbidity during stable deployment” from “high turbidity during severe oscillation.” This approach retains transparency and prevents a smoothing algorithm from concealing conditions that may be important to site operations.
Build A Practical Noise-Reduction Workflow
A repeatable workflow makes it easier to compare deployments and explain decisions to project managers, regulators, or research partners. Begin with a physical inspection, then examine raw data alongside environmental and motion records. Establish acceptance limits before reviewing the final results whenever possible, because post hoc decisions can unintentionally favor a preferred interpretation.
Use the following controls as a field and data-processing checklist:
- Secure the sensor, cable, and frame against rotation, vibration, drag, and impact.
- Confirm that the optical head stays submerged and remains at the intended depth through the wave cycle.
- Record wave, water-level, motion, or current data that can support later diagnosis.
- Apply a short, documented filter only after checking raw data and selecting a window suited to the monitoring objective.
- Retain raw files, calibration records, maintenance notes, and quality flags with every processed dataset.
For long-term stations, review baseline behavior after installation, cleaning, storms, major maintenance, and changes in water level. A stable instrument should show a comparable response during comparable conditions. Gradual drift or increasing scatter may point to fouling, wear, cable damage, or a changed mounting arrangement rather than increased environmental variability.
Dredging projects benefit from operational thresholds that account for confidence in the measurement. A single filtered spike should rarely trigger a major decision by itself. More defensible rules may require persistence over several intervals, agreement with a second station, or confirmation that the sensor remained within its movement limits. The appropriate rule depends on whether the system is used for research, process control, environmental compliance, or an OEM monitoring package.
Move From Noisy Traces To Defensible Measurements
Reducing wave-related noise is a measurement-design task rather than a software shortcut. A stable mount limits unwanted motion, appropriate placement reduces exposure to bubbles and extreme gradients, and synchronized diagnostics reveal when the instrument is behaving differently from the surrounding water. Calibration and filtering then support interpretation instead of compensating for preventable deployment problems.
D & A Instruments’ optical sensing experience covers marine and freshwater monitoring, dredging plume observation, hydrology, defense applications, and OEM integration. For specifications, application material, and documentation relevant to a particular system, review the available technical downloads and use the current Campbell Scientific product-management and contact information for support.
Apply these controls during the next deployment: stabilize the sensor, document its geometry, capture motion context, validate the calibration with representative samples, and preserve the raw record. That process turns an unstable-looking turbidity trace into data that can support confident engineering, environmental, and research decisions.