Tracking Suspended Solids in Drinking Water Reservoirs During Algal Blooms
Australia's drinking water reservoirs sit at the intersection of intensifying climate stress and increasingly stringent water-quality expectations. Heatwaves that once arrived in February now extend from November through April across much of the eastern seaboard, and with them come the nutrient pulses that drive cyanobacterial blooms. Warragamba, Wivenhoe, the Upper Yarra chain, and Mount Bold have all experienced recurring blue-green alerts in the past decade, prompting utilities such as Sydney Water, Seqwater, Melbourne Water, and SA Water to revisit how they observe source water in real time.
Conventional sampling programs struggle to keep pace with bloom dynamics. A grab sample taken at 9 am may miss a scum layer that forms by midday under calm, sunny conditions. Operators need continuous surrogate measurements that track the particles and biomass that matter to downstream treatment, and optical suspended-solids monitoring has emerged as a practical answer. By measuring near-infrared backscatter at multiple depths, instruments can flag the early rise in particulates that often precedes visible discolouration or taste-and-odour complaints from customers.
The case study presented here draws on a multi-month deployment at a large storage servicing a metropolitan supply system in southeastern Australia. During the deployment, the reservoir experienced two distinct bloom episodes linked to warm, stratified conditions and a bushfire-related ash load that washed into the catchment. The instrumentation captured the rise, peak, and decline of suspended-solids concentrations with a resolution that the utility had not previously been able to obtain, and the resulting dataset informed both immediate operational choices and longer-term planning.
What follows is a technical account of how the system was specified, where the sensors were installed, what patterns emerged in the data, and how the operations team translated those patterns into action. Readers familiar with the submarine pipeline case study will recognise some of the optical principles, though reservoir monitoring introduces its own constraints around biofouling, depth profiling, and long service intervals.
Catchment Pressures Driving the Need for Continuous Data
Australia's drinking water catchments have changed markedly over the last twenty years. Land use around storages such as Lake Burragorang and Lake Wivenhoe now includes peri-urban development, intensified agriculture, and forestry operations that periodically deliver nutrient-rich runoff. After the 2019–2020 bushfires, ash and charred debris entered many catchments across New South Wales and Victoria, releasing nutrients that fuelled cyanobacterial growth well into the following summer.
The Bureau of Meteorology has documented rising air temperatures across southern and eastern Australia, with overnight minima climbing faster than daytime maxima in many inland catchments. Warmer nights reduce the mixing that normally disrupts buoyant algal colonies, allowing them to dominate the photic zone. The result is a reservoir profile that can shift from mixed to strongly stratified within days, concentrating biomass near the surface where raw-water intakes often draw.
Regulators have responded accordingly. The NHMRC Australian Drinking Water Guidelines include guidance on cyanobacteria and their toxins, and state-based instruments such as the Safe Drinking Water Act 2003 in Victoria require utilities to demonstrate due diligence in managing source-water risks. Utilities facing statutory notification obligations when cell counts exceed trigger values have a direct commercial and reputational incentive to detect blooms earlier than visible monitoring allows.
Traditional weekly or twice-weekly sampling cannot meet that standard. Operators need an instrument that reports suspended-solids concentration every few minutes, ideally with multiple measurements across the water column, so that intake selection and pre-treatment chemistry can be adjusted in step with the bloom.
How Optical Backscatter Sits Alongside Standard Water-Quality Metrics
Turbidity, measured in nephelometric turbidity units, has long been the workhorse surrogate for particulate load. However, NTU responds to a broad particle-size distribution and is heavily influenced by fine clays that may be unrelated to biological activity. Optical suspended-solids sensors use near-infrared light at wavelengths such as 850 nm and apply a calibration tied to dry-weight concentration of suspended matter in milligrams per litre.
The distinction matters during algal blooms. Phytoplankton cells, colonial aggregates, and the extracellular organic matter they release all scatter light differently than mineral sediment. By calibrating against gravimetric suspended-solids measurements during pilot work, operators obtain a signal that tracks biomass-related particulates more directly than turbidity alone. Many Australian utilities now record both metrics in parallel, recognising that NTU remains useful for filter performance assessment while suspended-solids data informs source-water management.
Instrument selection also involves considering the difference between single-point and multi-point systems. A single wiper-cleaned probe mounted on a buoy or near the intake gives a snapshot at one depth. A multi-point profiler, lowered through the water column, can reveal vertical structure: the thin surface layer of actively growing cyanobacteria, a deep chlorophyll maximum, and the sediment-laden bottom layer that storm events re-suspend.
Battery and communication architecture is another practical concern. Reservoir installations are remote, often without mains power, and need telemetry that integrates with the utility's existing SCADA platform. Modern instruments offer low-power operation suitable for solar charging and standard industrial protocols that slot into existing telemetry without requiring a parallel data system.
Sensor Placement and Configuration During the Deployment
The reservoir in question supplies a major metropolitan service area, with an off-take tower that draws from several depths depending on water quality and storage level. The deployment team installed four optical probes at fixed depths: one metre below the surface, at the existing intake depth, just above the thermocline during stratification, and two metres above the bed. A fifth probe was fitted to a small service vessel for profile transects during routine surveys.
Each fixed probe was mounted on a stainless-steel frame with a wiper mechanism that operated before every measurement cycle. The wiper reduced biofouling, which builds rapidly in productive reservoirs during summer and can otherwise bias the optical reading. Cleaning cycles were scheduled at twenty-minute intervals during the day and hourly overnight, balancing data density against wiper wear.
Telemetry back to the operations centre used 4G cellular coverage that reached the reservoir site, with data pushed every five minutes into the existing historian. Operators received alerts when suspended-solids readings at any depth crossed thresholds aligned with the utility's incident-response protocols. Where cellular coverage was unavailable, a satellite fallback was configured for the highest-priority alarms.
Calibration of the sensors was checked monthly using grab samples collected in parallel with the optical reading. The team filtered known volumes onto pre-weighed glass-fibre filters, dried them to constant mass, and compared gravimetric results against the instrument output. Across the deployment, the sensors held their calibration within eight percent of laboratory values, well within the operational tolerance for source-water management decisions.
Patterns Recorded During the Two Bloom Episodes
The first bloom, in late spring, began quietly. Suspended-solids readings at the one-metre depth started climbing from a baseline around 4 mg/L to 18 mg/L over four days, while readings at the intake depth remained near 5 mg/L. This vertical decoupling is a classic signal of surface-blooming cyanobacteria that have not yet mixed through the water column. The intake operators, alerted by the threshold-crossing alarm, reduced draw from the upper valves and increased draw from deeper ports, buying time before the bloom migrated downward.
The second bloom, triggered partly by nutrient-rich runoff after a dry storm event, was more aggressive. Within forty-eight hours of the rain, the surface probe recorded peaks above 60 mg/L, with periodic scum formation that pushed instantaneous readings past the instrument's linear range. The mid-depth and intake-depth sensors showed a rising trend with a lag of roughly thirty-six hours, giving operations a window to bring coagulant dosing forward in the treatment process.
A diurnal signal was also visible. During calm, sunny days, suspended-solids concentrations at the surface peaked in late afternoon as buoyant colonies accumulated, then dispersed overnight as wind and cooling mixed the upper layer. This daily cycle, invisible to weekly sampling, became a key input to the treatment plant's polymer dosing schedule.
Weather mattered as much as biology. A southerly change during the second bloom event dropped wind speeds and caused a sudden re-stratification, with the surface layer thickening from two metres to nearly four metres within six hours. The multi-point profiling captured this transition in a way that single-point monitoring could not, and the resulting operational briefing explained a treatment upset that had previously been attributed to instrument error.
From Data to Operational Decisions
The utility's incident-management framework treats source-water thresholds as triggers for defined actions. When suspended-solids at the intake exceeded 15 mg/L, the standard response was to switch to a deeper intake valve, notify the treatment plant, and initiate enhanced coagulation. Above 30 mg/L, a second-tier response activated, including in-line powdered activated carbon dosing and a shift to the reserve clarifier train where one was available.
The optical data did not replace the laboratory cell counts required under the NHMRC guidelines. It complemented them, providing hourly resolution that allowed dosing adjustments between grab-sample results. Operators reported that they could respond to a bloom a full day earlier than under the previous regime, and the catchment health team could confirm in real time whether interventions such as barrier curtains were having the desired effect on horizontal distribution of the bloom.
A secondary benefit was communication. Customer-service teams received advance warning of likely taste-and-odour events, allowing them to prepare messaging before complaints peaked. Several Australian utilities have invested in customer portals and SMS alert systems; pairing those channels with predictive source-water data has measurably reduced complaint volume during summer events.
The deployment also generated baseline data for periods without blooms, which is itself valuable. Understanding the normal suspended-solids range, including the seasonal envelope, makes it easier to detect the slow drifts that indicate catchment changes, reservoir sedimentation, or gradual shifts in land use. Without a baseline, every reading is interpreted in isolation, which can lead to either overreaction or complacency.
Lessons Relevant Across the Australian Water Sector
Several findings from this case study carry over to other utilities operating large storage across the Murray–Darling system, coastal Queensland, and southwest Western Australia. The first is that algal blooms rarely arrive without warning at the intake: the surface signal almost always appears first, and an instrument placed near the surface, integrated with vertical profiling, gives the longest possible lead time.
Operational benefits depend on:
- A clear threshold framework agreed between source and treatment teams
- Routine grab-sample verification to maintain calibration trust
- Alarms routed through existing SCADA rather than a parallel system
- Training so that night-shift operators can act on the data without supervision
A second lesson is that suspended-solids concentration is a more sensitive early indicator than turbidity for biological events, while turbidity remains the better metric for sediment-driven events after storms. Utilities that record both, with their different physical meanings clearly documented, gain a richer picture of what is happening in the catchment.
A third is the value of treating monitoring data as operational infrastructure rather than a compliance artefact. The same dataset that triggered an intake switch on day one provided the basis for a strategic review of mixing aerators the following year. The instruments are documented in the downloads library, where datasheets and application notes describe the optical configurations used in this and similar deployments.
Finally, the case study highlights how Australian utilities can meet rising regulatory expectations without committing to ever-larger sampling programmes. Continuous optical monitoring does not replace laboratory analysis, but it does turn a source water into something operators can see in real time, which is a powerful foundation for both routine performance and emergency response.