The Sutherland Shire's water: what's upstream, and what isn't

1. Where the water comes from

Woronora Dam was built between 1927 and 1941 on the Woronora River, just below where Waratah Rivulet joins it and about 24 km upstream of the river's junction with the Georges River between Como and Illawong. The wall is 66 m high and the reservoir holds 71,790 ML at a full supply level of 169 m above sea level (WaterNSW; Heritage NSW). The catchment is 75 km² by WaterNSW's count and 85 km² by Heritage NSW's, and every hectare of it is the Woronora Special Area, which the public can't enter.

Raw water leaves through two 0.9 m pipes in the base of the wall and is pumped to the filtration plant directly below it. There's no stretch of river between the offtake and the plant. The plant belongs to Wyuna Water (IFM Investors and Veolia) and Veolia runs it for Sydney Water until November 2036. It's a 160 ML/day direct-filtration plant: coagulation, filtration, chlorine and fluoride. It has no barrier designed for PFAS or radionuclides (Veolia).

From there the Woronora delivery system supplies Engadine and Heathcote, Loftus and Maianbar, Lucas Heights, Menai and Illawong, and over the escarpment Helensburgh and Stanwell Park. A 27 km pipeline laid in 1942 follows the river north to Sutherland and crosses the Georges River at Como to Penshurst. The network is cross-connected to Warragamba and Upper Nepean water through Prospect, so in a drought or a plant outage the Shire's taps can run on Sydney's main system.

Lucas Heights is on the wrong side of the dam

ANSTO's campus is 7.9 km from the dam wall on a bearing of 034 degrees. The tip is 8.1 km away on a bearing of 023. Both sit on a plateau between 137 and 163 m above sea level, and the reservoir's surface is at 169 m. Water doesn't run uphill, and more to the point the dam is the end of the catchment. Anything that enters the Woronora River below the wall flows north to the Georges River and never meets the drinking water.

The creeks confirm it. Cleanaway's 2024 scoping report for the tip's expansion has Mill Creek, Bardens Creek and Lucas Heights 1 Creek all flowing north, the latter two into Mill Creek and Mill Creek into the Georges River near Alfords Point (GHD, November 2024). ANSTO's own monitoring reports sample "near the junction of Mill and Bardens Creeks, which drain the LFBG and the Lucas Heights urban landfill" (ANSTO E-761). The one creek from this area that does run east to the Woronora, Fire Creek, enters the river below the dam. ANSTO's eastern slopes probably drain east as well, through two small creeks whose courses aren't legible in its published figure; one of them is named for the Main Discharge Pipeline, which used to run east to the estuary. Either way they reach the Woronora 13 km below the dam, in the tidal estuary.

The only links from Lucas Heights to the Woronora were built by hand. From 1961 to 1980 the Atomic Energy Commission piped treated effluent into the estuary under a controlled discharge licence (AAEC-E-563). After that the effluent went to the sewer through a pressure main under the river to Engadine, which cracked in March 2013 and was being replaced from 2019 (The Leader; AusTender). Both reach the estuary well below the dam. Today the whole plateau, ANSTO included, sends its sewage to Sydney Water's Cronulla plant at Kurnell and the Potter Point ocean outfall.

2. ANSTO: controlled releases, a tritium footprint, and a regulator losing patience

ANSTO at Lucas Heights runs the 20 MW OPAL reactor, the nuclear medicine production lines, waste stores, two sets of legacy liquid-waste tanks (buildings B23 and B57), the Little Forest Legacy Site and the old HIFAR reactor, now in early decommissioning (ANSTO Annual Report 2024-25). The Federal Court set aside the Kimba waste-facility site in July 2023 and the government's replacement framework isn't due until early 2027, so the intermediate-level waste stays at Lucas Heights into the 2040s.

What it releases, and where

Liquid effluent goes to the Sydney Water sewer under Trade Wastewater Agreement 4423, which requires it to meet drinking-water radioactivity levels at Cronulla once an agreed 25-fold dilution is applied, and from Cronulla to the Potter Point outfall. The last detailed public table, for 2003-04, shows tritium at 3.6 to 6.1 kBq/L and caesium-137 at 2.1 to 3.4 Bq/L in the effluent, under 12% of the agreed limit (ANSTO E-755). Stack releases of argon-41, xenon, krypton-85, tritium and iodine-131 gave the nearest resident a calculated dose of 0.0026 mSv a year in 2003-04, against a public limit of 1 mSv. ARPANSA's own stack checks in 2023-24 put the busiest stack at 3.34% of its notification level (ARPANSA 2023/24).

What has been found off site

Tritium is the one ANSTO nuclide with a footprint in local water. It arrives as rained-out stack vapour. In 2003-04 the Bardens Creek weir ran mostly at 20 to 90 Bq/L with peaks to 350. In 2005-06 the range was 20 to 3,300 Bq/L with a median of 110 (E-755; E-761). The Woronora estuary and Forbes Creek were below the 10 Bq/L detection limit. For scale, the WHO drinking-water guideline for tritium is 10,000 Bq/L. ARPANSA's independent sampling in October 2021 found gross beta activity of 0.205 to 0.461 Bq/L in the Woronora River above, beside and below ANSTO, which it put down to natural potassium-40, and its 2022-23 round gave it "confidence that no transport or migration of alpha or beta-emitting radionuclides has occurred" (ARPANSA 2021; 2022/23).

Little Forest

About 1,600 m³ of waste went into unlined trenches three metres deep, 1.6 km north of the campus, between 1960 and 1968: 6.9 g of plutonium, 59 kg of uranium, 48 kg of thorium, over a tonne of beryllium and roughly 150 GBq of fission and activation products (ARPANSA assessment). Water pooling in the trenches carries about 12 Bq/L of plutonium-239 and 240 and reaches the surface in heavy rain, and soil downslope holds plutonium up to 829 Bq/kg (Payne and others, 2013). The tritium plume at the trench-centre bore has fallen from 4,170 Bq/L in 2003 to 475 Bq/L in 2023. An engineered cap went on in late 2024, which so far is documented only in a 2026 conference abstract. No alpha emitter has ever been found leaving the site in creek water.

The regulator's patience

In September 2025 ARPANSA wrote that ANSTO's "stewardship of its ageing infrastructure does not currently meet ARPANSA's regulatory expectations" and issued an improvement notice (ARPANSA, Jul to Sep 2025). On 22 April 2026 its CEO issued a formal Direction over building B57, where glass-lined tanks hold acidic intermediate-level liquid waste from 1970s molybdenum-99 production. The tanks "were unable to be assessed with confidence", there is "no treatment pathway currently identified", and ARPANSA called the situation "an unacceptable risk to people and the environment" while recording "no detrimental impact" so far (ARPANSA Direction). ANSTO's position: no release, no dose. The risk is a future tank failure, not a leak that has happened.

The transparency gap

ANSTO stopped publishing environmental monitoring reports after 2006-07. Its annual report carries no discharge or dose figure. Its quarterly reports to ARPANSA aren't published. No local paper covered the April 2026 Direction. There's no public word on PFAS at Lucas Heights, and because the campus is Commonwealth land it sits outside the NSW EPA's programs. None of that is evidence of harm. It does mean the public is relying on ARPANSA's yearly verification, which is thin.

3. The tip: licensed wet-weather discharge to Mill Creek, leachate to sewer, and a 2046 expansion

The Lucas Heights Resource Recovery Park is Cleanaway's, bought from SUEZ on 18 December 2021. Veolia, which runs the Woronora filtration plant, runs Sydney's other putrescible landfill at Woodlawn (independent audit, 2024). The site holds Environment Protection Licences 5065 (landfill) and 12520 (organics), took 845,574 tonnes in 2025 against a cap of 970,000 a year, runs out of airspace around 2030 and closes in 2037 under its current consent (annual review, 2025).

What reaches Mill Creek

The licence allows one route to waters: wet-weather discharge from the sediment dams to Mill Creek at monitoring point MC1, limited to 2.5 mg/L of ammonia and 50 mg/L of suspended solids, except that the solids limit is waived when the overflow is "caused by a rainfall event". MC1 discharged six times in 2025-26, with suspended solids of 294, 1,900, 216, 132, under 5 and 127 mg/L and ammonia of 0.27 to 2.08 mg/L (surface water data 2025-26). Ammonia never exceeded its limit. Leachate is never meant to reach the creek. It's treated on site and sent to the sewer, about 242,500 m³ a year. Raw leachate sampled on 16 October 2025 carried 1,250 mg/L of ammonia nitrogen and 2,380 mg/L of chloride (leachate data), which is why it goes to Cronulla and not the creek.

Incidents and enforcement

The worst episode on record was March and April 2022, after 448 mm of rain in 19 days: leachate seeping into a stormwater drain, the organics pond overtopping, and damaged drains sending contact water into Mill Creek, with the ammonia exceedances reported only in the annual return. The enforcement record from 2017 to 2025 holds no penalty notice and no prosecution: EPA advisory and show-cause letters, two formal warnings, an odour reduction program, and three Department warning letters over the 2021-22 tonnage overruns. The EPA told the 2024 auditor there were "no outstanding or ongoing compliance issues".

PFAS

No PFAS result has ever been published for the tip's leachate, groundwater or creek. The licence suites don't include it. That changes now. From 1 October 2026 the EPA's Chemical Control Order requires yearly leachate and quarterly groundwater PFAS monitoring at every licensed landfill, and the EPA wants a PFAS risk assessment in the expansion EIS (NSW EPA, 12 June 2026). Because the leachate is sewered, any PFAS in it goes to Cronulla rather than Mill Creek.

The expansion

The Western Extension (SSD-78269209) would add about 9 million m³ of airspace on the old gun-club land, a 500,000 tonne restricted-waste cell, relocated leachate dams and a permanent realignment of Mill Creek, keeping the site at 970,000 tonnes a year until 2046. No EIS had been exhibited and no decision made at 7 October 2026 (Planning Portal). Sutherland Shire Council objected in December 2024, naming the restricted cell beside the diverted creek, then in August 2026 unanimously endorsed a planning agreement worth about $170 million tied to a closure date of 31 December 2046 (The Leader, 27 August 2026). WaterNSW had "no specific comments", which is what you'd expect for a site outside its catchment.

4. The mine: measured damage in the tributaries, a reservoir that isn't leaking, and treatment doing the work

Peabody's Metropolitan Colliery at Helensburgh has mined longwalls inside the Woronora catchment since 1995. Waratah Rivulet, which supplies 29% of the catchment's water, and the Eastern Tributary both enter the reservoir about 10 km above the dam (IEAPM, September 2023). The 2009 project approval, which runs to 2032, already allowed mining under the reservoir. What the Department approved in March 2020 was the extraction plan for longwalls 305 to 307, the first under stored water (The Leader, 3 April 2020). The reservoir didn't leak: mine inflow averaged 0.11 ML a day against a prediction of 0.2 and a tolerable loss of 1 ML a day, and on that basis the Department approved longwalls 308 to 310 on 12 December 2022 (Department letter).

What has been measured

The water lost is negligible by every panel's reckoning: 0.09 ML a day on average from 2009 to 2017. Water quality is where the damage is. Pools in Waratah Rivulet above longwall 23 cracked and drained in the 2000s. The Eastern Tributary breached its performance measure for iron staining in October 2016 and for pool drainage in January 2017; pools ETAG through ETAR lost their water and Peabody still lists the breach as a non-compliance (Peabody 2024 Annual Review). The Department's expert panel accepts the mechanism:

"the diversion of surface water and groundwater through mining-induced rock fractures and subsequent discharge of contaminated water into creeks and then into the reservoir is evident"

Dissolved manganese where the Eastern Tributary enters the reservoir reached about 2.8 mg/L in 2018. The drinking-water guideline's health value is 0.5 mg/L and WaterNSW's raw-water alert level is 0.02. It has sat at the mine's top trigger level since at least 2017.

What the panels and WaterNSW say

The 2019 Independent Expert Panel found "no evidence" that the no-connective-cracking conditions had been breached (IEPMC 2019). Its successor, in September 2023, found that the consultant letters Peabody relied on to conclude "not exceeded" were "not based on sufficient data and analysis", and wrote that "the Panel cannot rule out the possibility that the Metropolitan Mine has had a significant adverse impact". WaterNSW wrote to Peabody three times in 2022 and 2023 about the manganese, and in its 2025 submission on the latest modification called avoiding Swamp 106 altogether "the only effective protection" (Illawarra Flame, 8 September 2025). A former WaterNSW hydrogeologist's claim in March 2023 of unreported fracture planes under the reservoir is disputed; the Department said it had no such reports (ABC).

The Hacking River prosecution

In September and October 2022, Turkeys Nest Dam at the mine's pit-top overflowed coal-laden water into Camp Gully Creek in the Royal National Park. On 21 March 2025 the Land and Environment Court fined Metropolitan Collieries $196,560 and ordered $304,806 in costs (EPA, 24 March 2025). That was pollution by a mining company, but of Port Hacking's river. The pit-top drains east. The longwalls run west, under the Woronora catchment.

What's undecided

Modification 4 would reconfigure longwall 317, add longwall 318 under Swamp 106 and the headwaters of Honeysuckle Creek, and extend mining from 2029 to 2032. It drew 216 submissions in August 2025, 140 of them objections, including a unanimous one from Sutherland Shire Council (The Leader, 9 September 2025). The Commonwealth's Independent Expert Scientific Committee advised on 9 November 2025 that "the only effective way to minimise or mitigate potential impacts is to avoid longwall mining beneath all swamps and waterways" (IESC). The Department asked its own panel for advice in April 2026, and the file sat at "More Information Required" on 7 October 2026 (Planning Portal). Then on about 7 September 2026 Peabody lodged a scoping report for a Stage 2: 22 new longwalls inside the Special Area, 31 million tonnes, mining to 2044, with the full application due between 2027 and 2029 (ABC, 8 September 2026). The pattern so far: every extraction plan inside the 2009 footprint has been approved, with panels trimmed rather than refused.

What the plant does about it

Iron and aluminium mostly arrive as particles and leave with the turbidity in coagulation and filtration. Manganese arrives dissolved and has to be oxidised with permanganate to keep treated water under 0.02 mg/L, the level above which it stains pipes and laundry. The cost has been operational. From mid-2022, flood sediment, organic matter and metals held the 160 ML/day plant at its 40 ML/day minimum for over a year, with Sydney Water running the Kurnell desalination plant and paying contract penalties. The tap stayed safe. But "treatable" isn't the "negligible" the 2009 approval promised, and nobody can yet say how much of the 2022-24 deterioration was record rain and how much was the mine, because the load and sediment studies the panel asked for in 2023 haven't been delivered.

5. Cover-ups and controversies: one confirmed failure of candour, no concealment of Woronora data

The 2024 PFAS episode confirmed

On 15 June 2024 Sydney Water told ABC Radio there were "no known PFAS hotspots in our drinking water catchments". At that point it tested one plant routinely and relied on a desktop risk assessment for the rest. Sampling that month found PFOS at 0.0164 µg/L in the Cascade plant's water at Katoomba, and nothing at Woronora (ABC, 20 August 2024). The NSW Legislative Council's Select Committee on PFAS reported on 11 September 2025. Finding 1: Sydney Water "did not perform an appropriate level of due diligence" before making the claim. Finding 12: nothing requires a utility to report PFAS compounds that sit outside the guidelines (report). Emails read at the hearing showed agencies agreeing "in a show of unity" to decline an ABC interview, wording Sydney Water's managing director called "probably not completely appropriate". Whether that amounts to a cover-up is disputed; the committee didn't use the word. Campaigner Jon Dee's "clearly a cover up" of 12 August 2025 is an allegation, and it's about the Blue Mountains.

Woronora's own numbers confirmed

The one Woronora-specific allegation came from Friends of the Earth at the inquiry's hearing on 5 February 2025: that about 98% of the PFAS detected at Woronora, mostly a short-chain compound called PFBA, had gone unpublished because Sydney Water reported only three compounds (transcript). The monthly spreadsheets Sydney Water began publishing in January 2025, after FoE's freedom-of-information requests, bear it out. PFBA at Woronora was 0.02 µg/L on 3 June 2025, 0.065 µg/L on 7 July and 0.032 µg/L on 11 August, then below the reporting limit by August 2026 (Sydney Water). The allegation was right, and publication resolved it. The point that remains: Australia has no guideline value for PFBA, WaterNSW doesn't test raw water for it, and the June 2025 drinking-water guidelines say nothing about it (NHMRC).

The mining approvals disputed

Critics called the March 2020 approval "scandalous" and said it was made "under cover of coronavirus", and the Sutherland Shire Environment Centre says WaterNSW's warnings were brushed aside. Those warnings were on the public record, in submissions and in a June 2020 parliamentary debate. The dispute is that the Department approved despite them, not that anyone hid them. The nearest thing to an official finding that harm was understated is the September 2023 panel advice quoted above. There is no official finding of concealment.

The May 2023 plant shutdown confirmed

The Woronora plant was shut down in May 2023 for high turbidity in its filtered water after an unapproved change to the plant's control software altered chemical dosing. Sydney Water notified NSW Health, backwashed the filters and re-routed supply (NSW Health to IPART, August 2023). There was no news coverage. It surfaced months later in a regulatory submission, as did the year the plant spent at minimum output. That's the pattern across all of this: not concealment, but disclosure through channels nobody reads. No boil-water alert, E. coli detection or Cryptosporidium incident was found for the Woronora system between 2020 and 2026.

6. Testing it yourself

What's already published

Sydney Water publishes fortnightly PFOS, PFHxS, PFOA and PFBS results for Woronora at reporting limits of 0.1 to 0.5 ng/L, monthly spreadsheets of 41 to 42 PFAS compounds since January 2025, and a quarterly report of about 70 characteristics by address lookup, none of them radiological (Sydney Water). WaterNSW publishes quarterly raw-water PFAS at the dam, an annual catchment report with five Woronora sites, and a gross alpha and beta screen every three years, last run in March 2025 with every site under 0.5 Bq/L. Cleanaway publishes quarterly surface-water, groundwater, leachate and gas data for the tip (Cleanaway). ARPANSA publishes its yearly verification of ANSTO. Read those first. Most of what a sample could tell you is already there.

Which result would point where

Tritium is ANSTO's fingerprint. Nothing else in the area makes it, so a creek reading clearly above the 10 to 20 Bq/L detection floor downstream of the campus would be a signal. Gamma spectrometry for caesium-137, cobalt-60 and iodine-131, and a gross alpha and beta screen at the guideline's 0.5 Bq/L, complete the set. The tip shows up as ammonia above about 0.5 mg/L together with conductivity, chloride, potassium and alkalinity in an otherwise dilute creek. PFAS in Mill Creek would be new information; the national guideline for freshwater ecosystems is 0.02 µg/L for PFOS at the 99% protection level. The mine shows as iron, manganese and aluminium, but sandstone streams are naturally acidic and iron-stained, the affected tributaries are inside the no-entry Special Area, and the reservoir and plant strip the metals before the tap. A sample below the dam proves nothing either way about the mine.

Labs and costs

Envirolab in Chatswood runs a 28-compound PFAS suite, Sydney Water Laboratory Services at West Ryde takes external clients for metals and PFAS, and ALS at Smithfield does both. None publishes prices. Radionuclides are specialist work: AWQC in Adelaide runs gross alpha and beta and radium, and ANSTO's and ARPANSA's laboratories run tritium and gamma spectrometry, with no public price list and no confirmation they take private samples. As a rough guide, metals cost $60 to $200 a sample, PFAS $300 to $600, and radionuclides several hundred. PFAS bottles must be polypropylene or HDPE, never PTFE-lined, and dissolved metals need filtering through 0.45 µm in the field (Envirolab).

Where to sample, and what's legal

Five sites, sampled in paired dry and wet rounds (seven days without rain; 24 to 48 hours after 20 mm) with a duplicate and a blank: a flushed cold tap on the Woronora system; Mill Creek below Cleanaway's discharge point where public land begins; Bardens Creek from the Heathcote Road reserve; the Woronora River at Woronora or Bonnet Bay at low tide; and Heathcote Creek or Kangaroo Creek as a sandstone control. Everything draining to the dam is a Schedule 1 Special Area with fines up to $22,000 for a person and $44,000 for a company (WaterNSW). The river below the dam is outside it. ANSTO's buffer land and Cleanaway's site are private, so sample from public road reserves.

Why a Geiger counter won't help

Tritium's beta particle has a maximum energy of 18.6 keV and travels about a micrometre in water. The mica window of a Geiger-Müller tube stops it (arXiv 2202.11068). The gamma emitters of interest occur at hundredths of a becquerel per litre: a one-litre bottle at the guideline screening level emits half a decay a second, against a tube background of 0.2 to 0.5 counts a second. Only liquid scintillation counting and shielded gamma spectrometry work at those levels. The one thing a counter does read, ambient dose rate, is already published at around 0.1 µSv an hour.

The cheapest test is a letter

Friends of the Earth's freedom-of-information request to Sydney Water surfaced the Woronora PFBA results that no creek sample could have found. A GIPA request to Sydney Water for the Woronora system's full analyte list, one to WaterNSW for the appendix of its radiological screen, and a Commonwealth FOI request to ANSTO for its quarterly creek tritium data cost nothing and would close most of the gaps below.

7. Five things to watch

Each is a yes-or-no question with a public source that will settle it.

1. Modification 4 decided. The Department records a determination of Metropolitan Coal's Mod 4, approval or refusal, on or before 30 June 2027. Settles on the Planning Portal.
2. Cleanaway's EIS exhibited. The Western Extension EIS goes on public exhibition on or before 31 December 2027. Settles on the project page.
3. PFOS at Woronora reaches 1 ng/L. Any fortnightly Sydney Water result for Woronora treated water reports PFOS at or above 0.001 µg/L between November 2026 and the end of 2027. The highest so far is 0.0006. Settles on Sydney Water's results page.
4. Lucas Heights PFAS published. Cleanaway publishes a PFAS result for the tip's leachate or groundwater on or before 30 June 2027, as the new Chemical Control Order requires. Settles on Cleanaway's monitoring page.
5. National waste framework released. The Department of Industry, Science and Resources publishes the final Australian Radioactive Waste Management Framework on or before 30 June 2027; it's "expected in early 2027". Settles on industry.gov.au.

8. What isn't known, and what would change the answer

The gaps that matter: the courses of ANSTO's two eastern creeks (its published figure isn't legible); ANSTO's creek and sediment data after 2006-07, which sit in unpublished quarterly reports to ARPANSA; Sydney Water's quarterly treated-water tables for the Woronora system, which are behind an address lookup; the metal loads from the Eastern Tributary and Waratah Rivulet, the reservoir sediment cores and the hydrodynamic model the expert panel asked for in 2023; Cleanaway's first PFAS results; and the 75 versus 85 km² catchment figure.

What would change the answer: a mine water make above the 0.5 ML/day trigger, or a reversed hydraulic gradient at the reservoir bores; ANSTO nuclides other than tritium above detection in Mill or Bardens Creek water, which no measurement since 2003 has shown; treated-water manganese or iron above the guideline at a Woronora tap; a health-based guideline for PFBA below 65 ng/L; a described southward groundwater pathway from the Lucas Heights plateau, which the tip's groundwater management plan doesn't contain; or the Department refusing Mod 4, which would break a 17-year pattern of approval.