Tracking Sediment Plumes from a River Restoration Dam Removal

Removing a derelict weir or small hydroelectric dam from a recovering waterway is rarely a clean affair. Within hours of the first concrete block being lifted, the reservoir sediment that has lain dormant for decades begins to move, sometimes in dense, fast-travelling clouds that can choke a downstream creek for kilometres. A team working on a Snowy Mountains tributary recently chose to follow those clouds in real time using optical backscatter and turbidity sensors, rather than relying on the slower rhythm of manual grab sampling. Their instruments, supplied by the team at D & A Instruments, produced a continuous picture of how the river responded as the structure came down in stages.

The work was driven by a joint brief from the local water authority, a state environment agency, and a freshwater ecology group attached to a regional university. They needed to know whether the released sediment would exceed ecological thresholds, how long the plume would persist, and whether the project would trigger compliance action under the relevant water-quality guidelines. Optical instruments offered the spatial and temporal resolution to answer those questions, and the resulting dataset has since become a reference point for similar removals being scoped across the Murray–Darling Basin and Tasmania.

The site and the structure coming down

The waterway in question is a fifth-order stream that tumbles out of the Snowy Ranges into a broader valley historically prized for its platypus and native fish populations. The structure blocking it was a 14-metre concrete diversion weir built in the 1950s to feed a now-closed hydroelectric scheme. After decades of disuse, the weir had silted up to within a metre of its crest, leaving an estimated 28,000 cubic metres of fine sediment stored upstream. The decision to remove it came after a state-level rivers and catchments review flagged the structure as a priority barrier to native fish movement, particularly for the locally protected Macquarie perch.

Removing the weir in a single blow was never on the table. The construction crew, working under a tight environmental window set by the regulator, lifted the structure in three controlled stages over ten days. Each stage released a different volume of impounded material, and the project team expected a stepped response in the river downstream. The challenge was to capture that response in enough detail to satisfy both the environmental manager on site and the compliance officer back in the capital.

Choosing optical sensors for a highly variable fluvial system

Selecting the right instrument for a job like this is rarely as simple as reaching for the most accurate laboratory turbidimeter on the market. The field team needed sensors that could sit in fast, debris-laden water for weeks at a time, log data on their own internal power, and survive being hauled out for cleaning by a field tech wearing waders. Optical backscatter sensors, suspended-solids probes, and a multi-parameter water-quality sonde were ultimately deployed in a paired arrangement at four sites spaced from 200 metres to 6.5 kilometres downstream of the weir.

The arrangement allowed the team to distinguish between sediment released from the reservoir itself and material re-mobilised from the river bed as flows picked up. A separate near-infrared optical unit was installed on a mid-channel pontoon to provide a continuous surface reading for cross-checking against the in-stream sensors. This combination of fixed and mobile platforms is becoming standard practice in Australian fluvial monitoring, where crews often need to cover long reaches of remote waterway without easy access to mains power or telemetry infrastructure.

Calibration in coloured, organic-rich water

Calibration is where many field campaigns quietly fall apart, and the team was keen to avoid the well-known trap of trusting a factory curve in water that bears little resemblance to the standard formazin reference. The Snowy tributary runs through peat-rich forest, which gives the base flow a strong tea-like colour and a measurable background absorbance in the near-infrared band that several optical sensors use. Pre-deployment calibration was therefore carried out in a darkened lab using locally collected sediment, sieved and diluted, alongside the standard formazin series described in the turbidity sensor calibration guide.

Each sensor was then re-checked in situ using a portable secondary standard and a field-filtered sample comparison. The team deliberately ran the calibration at three temperatures spanning the expected seasonal range, since the optical response of the suspended-solids probe had been observed to drift in cold water during earlier work in Tasmania. Recording those offsets in the metadata of every data file made the post-processing far more straightforward when the high-turbidity events began.

Watching the plume unfold in real time

The first stage of the removal produced a sediment plume that was visible from the bank within twenty minutes of the concrete being lifted. Within two hours, optical readings on the nearest downstream sensor had risen from a base of around 15 NTU to a peak of 1,840 NTU, with suspended-solids concentrations briefly exceeding 4,200 milligrams per litre. The data streamed via a small 4G telemetry unit to a project dashboard, allowing the on-site environmental officer to make hour-by-hour calls about whether to pause the next stage of demolition.

By the second and third stages, the team had refined their approach. They used the optical record to identify the leading edge of each new plume, then deployed a small boat-mounted ADCP and a towed optical package to map the plume cross-section as it passed a fixed reference cross-section. The combination let them build a near-continuous record of suspended-sediment flux rather than relying on isolated point measurements, which would have missed the rapid rise and fall of each event. The plume was tracked for eleven days, by which time background turbidity had returned to within 5 NTU of pre-removal levels.

Turning optical data into a sediment budget

Raw optical readings on their own are useful for an operator on the day, but the regulator required something more formal: a sediment budget reconciling how much material left the reservoir, how much was deposited on the bed downstream, and how much exited the study reach. The team built that budget by integrating the suspended-solids time series at each monitoring site, weighting the readings by velocity from the simultaneous ADCP transects, and applying a simple conversion factor validated against the field-collected sediment samples.

The exercise showed that roughly 62 per cent of the impounded material exited the study reach during the eleven-day monitoring window, with the remainder deposited across the first 1.2 kilometres of the former reservoir bed and the immediate downstream riffle. These results are now archived alongside the technical documentation for the project, where they can be compared against outputs from similar removals being run through the same processing chain. The data files include the calibration certificates, the raw optical streams, and the conversion scripts used in the analysis, which is a format the Murray–Darling Basin Authority has signalled it would like to see more often.

Ecological recovery in the months that followed

Optical monitoring did not stop once the sediment cloud had dispersed. The team left a subset of sensors in place for a further six months to track how the river recolonised the formerly impounded reach. Macroinvertebrate drift samples were collected at fortnightly intervals, and a fixed underwater camera recorded fish passage events past the former weir site. Within nine weeks, the first small galaxiids and a handful of adult Macquarie perch had been recorded moving upstream through the former reservoir footprint, something that would have been impossible while the weir was intact.

Turbidity remained slightly elevated for the first three months, driven largely by resuspension of fine material from the freshly exposed reservoir bed during high flows. The optical record showed that these post-project spikes rarely exceeded the thresholds set in the project's environmental management plan, which the regulator later used to justify issuing a closure notice on the environmental approval. The ability to demonstrate compliance with hard numbers, rather than anecdotal field observations, made the difference between a smooth sign-off and a protracted review.

Lessons for future Australian river restoration

The campaign has since become a working example within a national network of restoration practitioners convened by a coalition of universities and water agencies. Three lessons stand out for anyone planning a similar project. First, optical sensing pays for itself the moment a sediment event becomes unpredictable, because the alternative is a frantic scramble for grab samples that almost always arrive too late. Second, calibration in coloured or organic-rich Australian waters is not optional, and teams that skip the local verification step tend to underestimate suspended-solids loads by a factor of two or more.

Third, the value of a well-organised dataset extends well beyond the project itself. The same processing chain, sensor pairing, and reporting format can be lifted into a weir removal on the Murray, a coastal lagoon opening in Queensland, or a small barrier clearance on a Tasmanian creek, with only minor changes to the local calibration constants. As more Australian water authorities look to remove redundant barriers over the coming decade, the optical monitoring template developed for this project is likely to become a familiar sight in approval documents and post-project reports.