Interpreting turbidity spikes during dam flush operations

A dam flush is a controlled release used to clear fine sediment accumulated behind a dam or diversion structure. The resulting plume can raise turbidity quickly at the outlet, travel downstream, and settle over a period ranging from hours to several days. An episodic sediment release of this kind differs from ordinary catchment runoff because its timing is linked to gate operations, reservoir level, and the physical behaviour of newly mobilised bed material.

A sharp turbidity spike does not automatically indicate a water-quality incident. Turbidity monitors measure light scattering rather than sediment mass directly, so readings in nephelometric turbidity units, or NTU, cannot be converted to suspended sediment concentration without site-specific calibration. Fine silt, coarse sand, organic particles, entrained bubbles and mineral coatings can produce very different optical responses under the same flow condition.

Australian conditions add several useful clues when interpreting these events. Alpine catchments around the Snowy Mountains can deliver highly variable flows, while Tasmania’s wet catchments may produce sustained sediment movement after intense rain. Downstream users in the Murray–Darling Basin also need to distinguish operational releases from irrigation-quality changes and storm-driven sediment inputs.

The objective is to identify what caused each rise, estimate the sediment plume’s likely movement, and provide defensible information to water managers. D & A Instruments developed turbidity monitors, suspended-solids sensors, hydrology systems and groundwater profilers for environmental research, defence, dredging-plume monitoring and OEM integration; that product line is now supported by Campbell Scientific.

Why a flush creates a sharp signal

Reservoir drawdown can expose loosely deposited silt and clay on slopes near the outlet. When released through low-level gates, this material forms a density current that travels along the reservoir floor before entering the downstream channel. The initial concentration may be low, but increasing flow can erode the deposit and create a concentrated sediment front within minutes.

The shape of the hydrograph and turbidity trace provide the first clues. A genuine sediment pulse often begins close to a gate-opening event, peaks before maximum discharge, and recedes more slowly than the flow. A rainfall-driven increase is more likely to coincide with catchment rainfall and affect several tributaries at once. A pump or treatment-plant discharge may instead produce a recurring daily pattern.

Sediment transport is also nonlinear. A modest increase in velocity can mobilise a large quantity of fine bed material, so suspended sediment concentration may rise faster than discharge. Operators should therefore avoid interpreting turbidity from discharge alone. Gate position, reservoir level, upstream and downstream gauges, and observations at intermediate sites help establish whether the spike represents a moving plume rather than a local sensor effect.

A representative Australian flush event

Consider a representative operation at a hydro reservoir in the Snowy Mountains. The reservoir has accumulated fine sediment near its outlet following several dry years, and crews plan a controlled flush before the onset of winter irrigation demand. A monitoring station 300 metres downstream normally records 6–12 NTU during stable baseflow. An upstream reference gauge and a rain gauge provide additional context.

During the release, the downstream turbidity rises from 10 NTU to 65 NTU within fifteen minutes and reaches approximately 190 NTU as outlet velocity increases. Grab samples collected by the monitoring crew show suspended sediment concentration increasing from around 15 to 95 milligrams per litre. The rainfall record remains dry, gate telemetry confirms the discharge, and the upstream reference sensor stays close to baseline. Together, these observations support attribution to mobilised reservoir sediment rather than rainfall.

The turbidity falls rapidly after the main pulse, but a lower, longer tail remains for several hours. That tail matters for irrigation offtakes and river users because withdrawal timing can affect both sediment concentration and deposited material. In the Murray–Darling Basin, such timing may involve irrigation entitlements, environmental release conditions and communication between dam operators, water authorities and affected landholders.

This example also shows why a single threshold is inadequate. The 190 NTU peak looks severe in isolation, yet it occurs for a limited period, coincides with a known operation and is corroborated by laboratory samples. The same value during an unexamined storm could have a very different cause and consequence.

Reading the optical sensor response

Optical instruments are valuable because they provide continuous measurements through changing flows and can reveal short-lived features that periodic sampling misses. Their readings, however, depend on particle size, shape, refractive index, colour and sensor geometry. A pulse dominated by clay may scatter light differently from one dominated by coarse quartz sand, even when the two samples contain similar suspended sediment concentrations.

Sensor condition must also be considered. Biofouling, dust, scratched windows and deposits from previous events can change sensitivity or cause erratic readings. Very fine sediment may blind a sensor during the peak and make the measured turbidity lower than the actual value. Turbulent outlet flow can entrain air bubbles, which scatter light and may create an apparent sediment spike with little change in suspended solids.

Field turbidity readings should be checked against laboratory-filtered samples and discharge data rather than treated as interchangeable sediment measurements. Terms such as turbidity, suspended sediment and optical backscatter have specific meanings in monitoring guidance; the water-quality glossary provides a useful reference when preparing reports or comparing instruments.

Separating sediment from other causes

Several non-sediment processes can resemble a flush event. A thunderstorm may produce simultaneous rises at tributary sites, while a dust storm can carry fine material into an inland river and generate a broad, shallow turbidity increase. Warm, shallow reservoirs can experience algal blooms, and high-velocity releases may introduce entrained air. Each requires a different interpretation and may prompt different management actions.

A robust attribution process combines temporal, spatial and physical evidence. Timing is established by aligning sensor clocks with gate records, rainfall observations and sample collection. Spatial evidence comes from comparing upstream, near-outlet and downstream stations. Physical evidence includes turbidity, discharge, suspended sediment samples and observations of water colour or visible plume boundaries.

Useful working rules are based on agreement among evidence sources. A rise recorded by one sensor without supporting discharge or sample data remains unconfirmed. A coordinated increase across several instruments, together with changing sediment concentration, is more likely to represent an environmental release. Thresholds for notification should reflect the sensitivity of downstream users, the sensor’s validated range and the consequences of false alarms, rather than a universal NTU value.

Field checks for monitoring crews

Australian crews commonly describe a controlled reservoir release as a “flush”, including the associated gate movements, downstream sampling and public notifications. Before an event, the monitoring record should establish the conditions needed to separate normal variability from an unusual signal. The following checks are suitable for dam operators, environmental consultants and irrigation monitoring teams:

• Record a stable pre-release baseline at each gauge, including its variability rather than relying on one reading.
• Align instrument clocks with gate telemetry, rainfall records and the sampling schedule.
• Inspect optical windows, wipers and flow paths for fouling, scratches or sediment deposits.
• Confirm sensor range, low-flow limits, calibration history and event-trigger settings before release.

Once discharge begins, the priority is to preserve a coherent event record. Rapid decisions depend on data that can be compared later, especially when operators are balancing irrigation supplies, environmental conditions and downstream notifications. These field actions help during and immediately after the release:

• Flag rapid increases only when turbidity, discharge and available sample evidence show a consistent pattern.
• Preserve raw readings alongside filtered or processed values so later QA can reproduce the event.
• Collect timed samples near the peak, during recession and after visible settling has begun.
• Record rain, dust, algal material, bubbles, debris and unusual upstream activity in the event log.

Current technical and product-management enquiries for the Campbell Scientific-supported D & A range can be directed through the contact page. For hydrology, environmental research, defence or OEM applications, the most useful case history combines instrument behaviour with rainfall, operating and laboratory records. That combined evidence turns a turbidity spike from an isolated number into an interpretable account of sediment movement and its operational consequences.