- Drinking Water 2025 – Summary of the Chief Inspector’s report for drinking water in Wales
- Foreword
- Introduction
- Water supplies and testing
- Compliance with standards
- Learning from compliance failures
- Learning from events
- Consumer contacts
- Asset health
- Water safety planning and risk assessment
- Poly and perfluoroalkyl substances (PFAS)
- Audit programme
- Recommendations
- Enforcement
- Security and Emergency Measures Direction - SEMD
- Network and Information Systems - NIS
- Materials in contact with drinking water (Regulation 31)
- Research publications
- Whistleblowers
- Working with stakeholders
Learning from compliance failures
In 2025, Wales recorded several microbiological compliance breaches. A consistent theme across these breaches was insufficient asset resilience, particularly the inability to remove critical tanks from supply for timely internal inspection due to the absence of bypass arrangements and/or contingency options. In several cases, investigations were delayed for many months where assets were operationally critical, limiting confidence that consumers were not exposed to ongoing risk at the time of failure. These cases reinforced the need for proactive resilience planning to ensure that microbiological investigations can be completed promptly under all operating conditions.
Failures at service reservoirs and treated water tanks highlighted weaknesses in inspection regimes, asset condition knowledge, and record accuracy. Several breaches demonstrated the importance of maintaining accurate and up‑to‑date schematics, particularly following asset commissioning or decommissioning, to ensure investigations are comprehensive and sampling strategies are appropriate. Notably, issues were identified in both ageing assets and newly commissioned infrastructure, emphasising that construction age alone does not equate to low risk.
At consumers’ taps, a microbiological breach was attributed to inappropriate sampling point selection. The presence of tap attachments compromised the representativeness of samples, leading to misleading results and unnecessary regulatory action. This reinforced the requirement for clear sampling criteria, staff awareness, and assurance that consumer tap samples accurately reflect supply water quality.
Sporadic taste and odour breaches continued in 2025, driven primarily by recurring raw water precursor compounds and treatment‑related by‑products. While operational mitigation measures were in place, several repeat events indicated that initial controls were not sufficiently robust. Improvements to granular activated carbon (GAC) management, including more conservative exhaustion targets, represent a positive step toward more effective and sustainable control of aesthetic risks.
Iron and discolouration failures remained widespread, with Wales continuing to perform poorly relative to companies in England. The majority of iron breaches were linked to deteriorating cast iron mains, low turnover, dead ends, and inaccuracies in asset records, particularly where closed valves were not reflected in GIS systems. While flushing programmes provided short‑term compliance, they frequently failed to address underlying causes, and in some cases sampling undertaken post‑flushing was not representative of typical network conditions. The Inspectorate consistently emphasised the need to shift from reactive flushing toward timely asset renewal and long‑term investment planning.
A turbidity breach at a borehole pumping station highlighted the importance of effective treatment control and monitoring. The absence of a final water turbidity shutdown contributed to the event, prompting improvements to shutdown arrangements, monitor calibration and proactive maintenance. The incident reinforced the need to protect the full disinfection stage and to ensure that any impacts from chemical dosing on turbidity and pH are adequately controlled.
Overall, the 2025 compliance picture in Wales reinforces the need for stronger asset resilience, accurate records, representative monitoring and proactive investment. Recurrent reliance on reactive measures and short-term compliance sampling provides limited assurance of sustained water quality. Addressing root causes through improved governance, long-term planning and preventative maintenance remains essential to reducing repeat failures and protecting consumers.
Microbiological detections at works and service reservoirs
In Wales, there were three microbiological compliance failures at water treatment works in 2025, all at Dŵr Cymru Welsh Water assets. A recurring theme across several failures was the inability to remove assets from supply for timely investigation due to insufficient asset resilience. There has been a steady increase in microbial failures in Wales since 2021 as can be seen in Figure 9. The following paragraphs describe some of these failure scenarios along with lessons learnt.
Treatment works
| Parameter | Current standard | Total number of tests | Number of tests not meeting the standard | Number of tests not meeting the standard by company |
|---|---|---|---|---|
| Clostridium perfringens (sulphite-reducing clostridia) (confirmed) | number/100 mL | 600 | 0 | |
| E. coli (confirmed) | number/100 mL | 12,108 | 0 | |
| Turbidity – indicator | NTU | 12,093 | 0 | |
| Total coliforms (confirmed) | number/100 mL | 12,108 | 2 | DWR (2) |
Table 5. Microbiological tests – The number of tests performed and the number of tests not meeting the standard
Turbidity – indicator is a critical control parameter for water treatment and disinfection.
Service reservoirs
| Parameter | Current standard | Total number of tests | Number of tests not meeting the standard | Number of tests not meeting the standard by company |
|---|---|---|---|---|
| E. coli (confirmed) | number/100 mL | 20,057 | 0 | – |
| Total coliforms (confirmed) | number/100 mL | 20,057 | 0 | DWR (10), HDC (2) |
Table 6 – Microbiological failures at service reservoirs
Consumers’ taps
| Parameter | Current standard | Total number of tests | Number of tests not meeting the standard | Number of tests not meeting the standard by company |
|---|---|---|---|---|
| Coliform bacteria (indicator) | number/100 mL | 8,703 | 30 | HDC(4), DWR(26) |
| E. coli (confirmed) | number/100 mL | 8,703 | 1 | DWR(1) |
| Enterococci (confirmed) | number/100 mL | 753 | 0 | – |
Table 7 – Microbiological failures at consumers’ taps
E. coli and Enterococci at treatment works, service reservoirs and consumers’ taps
| Company | E. coli in water leaving treatment works | E. coli in water leaving service reservoirs | E. coli at consumers’ taps | Enterococci at consumers’ taps |
|---|---|---|---|---|
| ALE | 0 – 0 | 0 – 0 | 0 – 12 | 0 – 4 |
| DWR | 0 – 11,166 | 1 – 16,189 | 1 – 8,067 | 0 – 638 |
| HDC | 0 – 942 | 1 – 3,868 | 0 – 624 | 0 – 111 |
Table 8 – Detection of E. coli and Enterococci at treatment works, service reservoirs and consumers’ taps
Note: Results are shown as the number of positive tests followed by the total number of tests
Figure 9 – Microbiological failures over last 10 years
In February, Dŵr Cymru Welsh Water reported a single coliform in a sample collected from Court Farm works final water. The company was unable to identify a root cause, and the Inspectorate was therefore unable to conclude that recurrence was unlikely. Further investigation was deferred until contact tanks and final water tanks could be removed from supply for internal inspection. Due to the critical role of the works in maintaining supplies within the South East Wales Conjunctive Use System during a period of drought, this work was not completed until 21 October 2025, approximately eight months after the breach. Roof flood testing identified no points of ingress and all tanks were reported to be in satisfactory condition.
Similarly, in April, a single coliform detected at Alaw South works required inspection of multiple tanks; however, this was delayed until Cefni works could support supply resilience. At Bretton works, resilience limitations were more acute, with a single-compartment contact tank and no bypass preventing removal from supply following detections in May 2024 and May 2025, delaying internal inspection until October 2025. These examples demonstrate how insufficient resilience and lack of operational flexibility can significantly delay investigations into microbiological risks.
Further cases highlight the importance of asset knowledge and comprehensive investigation processes. At Dŵr Cymru Welsh Water’s Pengarnddu service reservoir no. 3, a coliform detection in July 2025 was not fully investigated due to incomplete schematics, which omitted hydraulic links to an adjacent reservoir, and limited resilience prevented removal of the asset from supply. At Porthmadog service reservoir, outdated schematics incorrectly included a decommissioned works and omitted an active source, complicating the investigation following confirmation of E. coli.
At Cruigau service reservoir, where tankering was used to support demand during hot weather, multiple coliform detections occurred in both the service reservoir and associated tankers within the same period; however, tankering was not initially considered within the investigation. Structural and ingress issues were also identified at several sites. At Axton service reservoir, insects entered through a damaged membrane, despite the roof previously passing a flood test. At Birchgrove service reservoir, internal inspection identified significant ingress through roof joints and a nearby sinkhole, neither of which had been evident during the previous inspection in November 2024.
These examples demonstrate recurring weaknesses in inspection regimes, schematic accuracy, and the consideration of operational activities and changing environmental conditions as contributors to contamination pathways. At Hafren Dyfrdwy’s Hirnant service reservoir, inspection identified deteriorated seals, non-compliant hatch arrangements and vegetation overgrowth. These events demonstrate the need for a systematic and resilient approach to managing microbiological risk across water supply systems. Companies should ensure that critical assets can be taken out of service promptly through appropriate bypass arrangements and alternative supply options, enabling timely inspection and investigation.
Accurate and up-to-date maps and schematics are essential to support effective incident response and must be maintained when assets are commissioned, modified or decommissioned. Investigation procedures should consider all credible sources of contamination, including operational activities such as tankering and network changes. In addition, inspection regimes should not rely solely on flood testing, but incorporate comprehensive structural, visual and condition-based assessments to identify latent defects. Strengthening resilience, improving asset knowledge and enhancing investigation practices will reduce delays, improve root cause identification and support the consistent supply of wholesome water under both normal and abnormal operating conditions.
Chemical and physical parameters
| Parameter | Units | Total number of tests | Number of tests not meeting the standard | Number of tests not meeting the standard per company |
|---|---|---|---|---|
| Colour | mg/L Pt/Co | 3225 | 1 | HDC(1) |
| Odour | Dilution number at 25°C | 3222 | 5 | DWR(5) |
| Taste (quantitative) | Dilution number at 25°C | 3220 | 5 | DWR(5) |
| Hydrogen ion (pH) – indicator – zone | pH Value | 3220 | ||
| Sulphate | mg/L SO₄ | 460 | ||
| Sodium (total) | mg/L Na | 749 | ||
| Nitrate (total) | mg/L NO₃ | 751 | ||
| Nitrite – consumers taps | mg/L NO₂ | 751 | ||
| Ammonium (zone) | mg/L NH₄ | 929 | ||
| Aluminium (total) | µg/L Al | 3188 | ||
| Iron (total) | µg/L Fe | 3188 | 17 | DWR(16) HDC(1) |
| Manganese (total) | µg/L Mn | 3190 | 1 | DWR(1) |
| Copper (total) | mg/L Cu | 749 | ||
| Fluoride (total) | mg/L F | 460 | ||
| Arsenic (total) | µg/L As | 749 | ||
| Cadmium (total) | µg/L Cd | 749 | ||
| Cyanide (total) | µg/L CN | 459 | ||
| Chromium (total) | µg/L Cr | 749 | ||
| Mercury (total) | µg/L Hg | 458 | ||
| Nickel (total) | µg/L Ni | 749 | ||
| Lead (10) | µg/L Pb | 747 | 2 | DWR(2) |
| Antimony | µg/L Sb | 749 | ||
| Selenium (total) | µg/L Se | 749 | ||
| Pesticides (total by calculation) | µg/L | 725 | ||
| Polycyclic aromatic hydrocarbons (total by calculation) | µg/L | 752 | ||
| Residual disinfectant – free | mg/L | 8897 | ||
| Residual disinfectant – total | mg/L | 8261 | ||
| Electrical conductivity | µS/cm at 20°C | 2997 | ||
| Boron | mg/L B | 458 | ||
| Benzo[a]Pyrene (total) | µg/L | 754 | ||
| Tetrachloromethane (total) | µg/L | 466 | ||
| Trichloroethene & Tetrachloroethene – sum of 2 substances (total by calculation) | µg/L | 464 | ||
| Trihalomethanes (total by calculation) | µg/L | 754 | ||
| 1,2-Dichloroethane (total) | µg/L | 468 | ||
| Benzene (total) | µg/L | 468 | ||
| Bromate | µg/L BrO₃ | 581 | ||
| Gross alpha | Bq/L | 1 | ||
| Gross beta | Bq/L | 1 | ||
| Tritium | Bq/L | 1 | ||
| Total | – |
Hafren Dyfrdwy reported a final water turbidity failure of 1.25 nephelometric turbidity units (NTU) at Llandinam New Borehole Pumping Station on 20 February. The company’s investigation concluded that the most likely root cause was the presence of undissolved solids associated with lime dosing at the treatment works. Analysis of historical data trends, catchment risk assessments and ultrafiltration membrane performance indicated that the failure was not attributable to deterioration in raw water quality from the Llandinam boreholes.
At the time of the breach, no final water turbidity site shut‑down was operational. In response, the company installed a final water turbidity site shut‑down upstream of the lime dosing point. This was initially set at 0.95 NTU and subsequently reduced to 0.75 NTU on 18 March. In addition, the balance tank dosing chamber (high‑lift sump) was scheduled for cleaning to minimise the accumulation of lime deposits.
The company increased the calibration frequency of the final water turbidity and pH monitors and adopted the draining and cleaning of lime storage tanks as an annual proactive maintenance activity at Llandinam BPS.
Companies are reminded that site shut‑down arrangements should be configured to protect the entirety of the disinfection stage, from the point of disinfectant application through to final water sampling. Alarms and shutdowns should be subject to routine testing regimes to provide assurance of operation. Where chemicals are applied within the disinfection stage, their potential impact on downstream disinfection performance, including effects on turbidity and pH, should be continuously assessed and controlled to ensure compliance with company disinfection policy targets.
Lead
Lead is a toxic metal that can dissolve into drinking water when it comes into contact with lead pipes. Consumers are largely protected from exposure through phosphate dosing, a widely used treatment across many zones in Wales that reduces plumbosolvency. This approach is one of the key reasons why few, if any, consumer tap samples fail the lead standard each year. Additionally, the relatively low number of required lead tests under the Regulations contributes to this outcome.
In 2025, there were two lead failures in Wales, both in Dŵr Cymru Welsh Water’s area and both reported at 11 µg/L. The failure in East Anglesey identified lead solder and a lead communication pipe at the property, alongside failures in orthophosphoric acid dosing. Remedial work was undertaken at the property, and the company was advised to investigate the root cause of the dosing failure and review the maintenance regime.
At the second failure, in Barry, the company’s inspection identified that the supply pipe was made of lead and was therefore the likely cause. Flushing advice was provided and the customer was advised to replace the supply pipe.
Taste and odour
Taste and odour in drinking water are key indicators of consumer acceptability and confidence, and are regulated under the requirement that water must be wholesome and clean. Under drinking water legislation, taste and odour are indicator parameters, meaning that while they are not typically associated with direct health risks at the levels encountered, any deterioration may signal underlying issues with source water quality, treatment processes, or distribution system condition.
Unacceptable taste or odour may arise from a range of causes, including algal activity in source waters, the presence of naturally occurring organic compounds, treatment inefficiencies, or the mobilisation of deposits within the distribution system. Chlorine and chlorinous compounds can also influence taste, particularly where dosing is elevated or poorly controlled.
Where concerns are identified, they require investigation under regulation 18 to determine the cause and extent of the issue. Persistent or widespread complaints may indicate deficiencies in treatment optimisation, asset performance, or network management, and require corrective action to restore consumer confidence.
Companies are expected to monitor and manage taste and odour proactively through effective control of treatment processes, maintenance of distribution systems, and appropriate response to customer reports. These parameters are also considered within regulation 27 risk assessments to ensure that water remains not only compliant but acceptable to consumers.
There were five taste and five odour failures in Wales in 2025, a decrease from 12 in 2024, as shown in Figure 8.
Dŵr Cymru Welsh Water’s Bontgoch zone continued to feature prominently among failing zones, with seven of the 10 failures recorded in this zone.
There have been recurring taste and odour breaches, detected by the company laboratory taste and odour panel, in this water supply zone since January 2020. Investigations have identified elevated concentrations of 2,4,6 trichloroanisole in the supply from Bontgoch works, with sources likely to include both naturally occurring compounds in the raw water and formation as a chlorination by-product during treatment. While laboratory analysis has detected taste and odour and confirmed the presence of this compound, consumers at affected properties have not typically reported abnormal taste or odour. The company has engaged external specialists and implemented a range of mitigation measures, including filter disinfection to remove biomass and associated by-product formation potential, and enhanced monitoring linked to pre-second stage chlorination. Third stage filters downstream of the contact tank are also due to be decommissioned and bypassed by June 2026 as part of longer-term treatment improvements.
In the Abergele and Rhyl water supply zone, a taste breach on 28 January 2025 was attributed to elevated concentrations of 2-methylisoborneol (MIB) in raw water feeding Glascoed works. A similar taste and odour event with the same root cause had occurred in May 2024. Although mitigation measures, including powdered activated carbon dosing and blending with Llanerch borehole supplies, were implemented following the earlier event, these controls were not sufficient to prevent the subsequent breach. Investigations identified that granular activated carbon contactors at Glascoed works were operating sub-optimally due to excessive ferric coagulant dosing, which reduced MIB removal efficiency. The company has since identified the need for front-end acid dosing to optimise coagulation and improve granular activated carbon performance, with delivery planned by March 2027. In the interim, enhanced monitoring, blending strategies and revised granular activated carbon management, including more conservative exhaustion criteria, have been implemented to improve control.
The Inspectorate noted that the recurrence of a taste event within eight months of a previous occurrence indicates that initial mitigation measures were not sufficiently robust. However, it also recognised that additional controls and enhanced surveillance have since been implemented to enable more responsive optimisation of treatment in line with changes in raw water quality. The adoption of improved granular activated carbon management processes is expected to strengthen removal of taste and odour compounds across relevant treatment works.
These breaches highlight the importance of implementing robust, sustainable control measures for taste and odour risks, particularly where raw water sources are prone to seasonal or persistent contamination. Companies should ensure that treatment processes are fully optimised and resilient to variation in raw water quality, with a clear understanding of how upstream processes impact the performance of downstream assets. Monitoring programmes should be sufficiently sensitive and responsive to detect deterioration early and inform timely operational adjustments. Where previous events have occurred, mitigation measures should be critically evaluated to ensure they are effective in preventing recurrence. Strengthening process control, asset performance and proactive risk management will support the consistent delivery of acceptable water quality to consumers.
Figure 12 – Taste & odour failures since 2016
Iron
Figure 13 – Iron breaches since 2018
In the Dŵr Cymru Welsh Water Barry and Sully water quality zone, an iron breach was linked to a closed valve that was incorrectly recorded as open on the company’s mapping system, creating a dead leg and allowing sediment to accumulate. A similar issue was identified in the Malpas, Caerleon and Cwmbran zone, again caused by incorrect asset status on GIS records following valve operations. In response, the company has introduced a standardised, web-based process to improve communication between operational and office teams and ensure network records reflect the live status of assets. These cases demonstrate the importance of accurate asset information in preventing avoidable water quality failures. Several failures were associated with deteriorating cast iron mains and low network turnover. In the Sketty and Gower area, an iron exceedance followed a burst and was linked to heavy tuberculation in an unlined main with low flow, requiring ongoing flushing and investigation. In Cardiff and Rumney, repeated flushing and temporary measures, such as trickle caps, were used to restore compliance. However, samples taken following these interventions were not representative of underlying network conditions at the time of the failure, compromising root cause investigation. Internal inspections confirmed extensive corrosion in a 1912 cast iron main, yet no renewal has been prioritised. Similar issues were identified in Hereford, Newport West and North Anglesey, where sediment accumulation at dead ends or within ageing trunk mains caused repeated or high-risk failures. In Hereford, repeated failures persisted despite flushing, prompting a business case for longer-term investment, while in North Anglesey an oversized, corroding trunk main continues to present an ongoing risk.
Other cases further emphasise the limitations of short-term operational responses. In Pencoed and the Bridgend Valleys, iron failures were attributed to sediment build-up in trunk mains, with flushing undertaken while longer-term investment options were considered. In Malpas, Caerlon and Cwmbran zone a manganese failure was due to a shut valve creating a dead end in the distribution network promoting sediment build up.
Across these events, flushing has commonly been used as the primary mitigation measure, often supported by monitoring and maintenance programmes. In several cases, low-velocity maintenance flushing programmes have been put in place to increase local main turnover. Dŵr Cymru Welsh Water has recently committed to a greater focus on sediment removal flushing, using higher velocities to systematically clear accumulated sediment from its networks. It should be noted that sediment removal and scouring velocities are not always achievable with the available assets and operational conditions, for example where internal tuberculation has reduced original internal pipe diameters. While these actions can provide temporary improvement, they do not address the underlying causes, particularly where mains are structurally degraded or poorly configured hydraulically. In some cases, flushing may also disturb deposits and increase short-term risk if not carefully controlled and validated through representative sampling.
These breaches demonstrate the need for a more proactive and strategic approach to managing discolouration risk. Companies should ensure that asset records accurately reflect operational status at all times, supported by robust processes for updating systems following field activities. There should be clear short, medium and long-term strategies for addressing iron risks, including timely consideration of asset renewal where infrastructure condition limits the effectiveness of operational controls. Reliance on reactive flushing and compliance sampling alone does not provide assurance of sustained water quality, particularly where samples are taken under artificially improved conditions. Effective risk management requires representative monitoring, understanding of hydraulic conditions, and targeted investment to remove root causes, ensuring the consistent supply of wholesome and acceptable water to consumers.
Figure 15 – Dŵr Cymru Welsh Water iron breaches
Colour
Colour in drinking water is an important indicator of aesthetic quality and treatment performance. It can originate from the raw water source, particularly where natural organic matter such as humic substances is present, as well as from metals like iron and manganese or disturbances within the distribution system. While not a direct health concern, elevated colour may indicate underlying operational issues and can affect consumer confidence, and it is therefore expected to be maintained at levels acceptable to consumers.
A colour breach was reported in Hafren Dyfrdwy’s concessionary supplies zone. The property is supplied from a spring source, with water piped to the property and treated by a point-of-use ultraviolet treatment system. A permanent boil water notice is in place at the kitchen tap. Although the company services the ultraviolet system, the disinfection process is not monitored or validated. The company recognises that deterioration in raw water quality, including elevated colour, may reduce the effectiveness of point-of-use ultraviolet treatment. The same property also recorded a colour breach in February 2024. Concessionary supplies in this zone are due to be connected to a mains supply in 2026 under notice HDC-2022-00001, and the company has reported that it remains on track to complete this work. As a precautionary boil water notice remains in place, and permanent connection to mains supply is due by the notice completion date.