- 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 events
There were 65 events notified to the Inspectorate in 2025. Thirty-two of these were single-property incidents, 16 of which related to consumers’ distribution systems and were therefore outside the control of the water company.
Figure 16 – Water quality events in Wales
The most significant event related to regulation 31 approval for Dŵr Cymru Welsh Water, with an Event Risk Index (ERI) score of 5,299. By comparison, the highest scoring event in 2024 had an ERI of 240. Regulation 31 events can attract large ERI scores due to the size of the potentially affected population and the duration of the event, which may extend over a considerable period while unapproved products remain in situ before they are identified and removed.
The second highest ERI score was recorded by Albion Eco, at 657, for a loss of supply event following an upstream mains burst. In this case, the small supplied population of 180 meant that the ERI score was proportionately higher than it would have been for a similar event affecting an incumbent company, as the entirety of the company’s supplied population was affected. Figure 11 shows ERI for Wales in 2025. The bars show the individual company ERI scores, the Wales line is ERI for the companies combined to one, the ERI median line is the median company ERI.
Figure 17 – ERI score per company
Figure 18 – ERI over 5 years
Figure 19 – ERI score for each company in England and Wales
Figures 20 and 21 show the number of events in Wales and across the industry respectively. In Wales there is no discernible trend, however the last two years have seen a peak similar to the previous peak within the charted 25 period. Further analysis of the nature of the events causing these peaks would be required to determine any pattern or commonality.
Figure 20 – Wales ERI and median ERI – Total scores for each company
Figure 21 – ERI scores in Wales compared to the rest of the industry
| Company code | Event name | Cause of event | Final classification | Final event assessment | ERI score |
|---|---|---|---|---|---|
| DWR | Regulation 31 Approval | Use of unapproved materials/products | 3 (Significant) | Prosecution considered | 5,298.682 |
| ALE | Broughton Mains | Mains problem/ damage – Mains – Burst | 3 (Significant) | Recommendations made | 657 |
| DWR | Aston Burst | Mains problem/ damage – Mains – Burst | 3 (Significant) | Enforcement-legal instrument | 245.807 |
| DWR | Bryn Cowlyd WTW Loss of Supply | Mains problem/ damage – Mains – Burst | 4 (Serious) | Recommendations made | 164.335 |
| DWR | Cardiff Discolouration | Depressurisation of network following false high readings on level probe at Cefn Mabley SR leading to level dropping below top of outlet main. Discolouration but no loss of supply. | 3 (Significant) | Enforcement-legal instrument | 30.856 |
| DWR | Llanrumney Burst Main | Mains problem/ damage – Mains – Burst | 2 (Minor) | Enforcement-legal instrument | 3.631 |
| DWR | Carmarthen Loss of Supply | Mains problem/ damage – Mains – Burst | 3 (Significant) | Enforcement-legal instrument | 3.026 |
| DWR | Gower/Swansea Discolouration | Mains problem/ damage – Mains – Burst | 3 (Significant) | Suggestions made | 1.57 |
| DWR | Rhyl Discolouration | Mains problem/ damage – Mains – Burst | 2 (Minor) | Suggestions made | 1.564 |
| DWR | Swansea Discolouration | Mains problem/ damage – Mains – Burst | 2 (Minor) | No recommendations or suggestions made | 0.807 |
Specific events
DWR – Regulation 31 approval
A routine regulation 31 compliance assessment, undertaken in relation to materials intended for use in a planned capital scheme, identified that products supplied and installed within the public distribution network between 12 June 2024 and 27 August 2024 did not hold valid approval during that period. The company confirmed that neither the product composition nor the instructions for use had changed between the lapse in approval and subsequent reapproval. A risk assessment carried out by the company concluded that the associated risk to water quality was therefore low.
To support this conclusion, the company completed surface area calculations for typical fittings across the likely installed size range. These calculations indicated that the fittings met the exemplar criteria set out in Advice Sheet 8, and this was submitted as evidence to demonstrate that the risk to supplied water quality remained low. However, the company was initially unable to determine the exact specifications or installation locations of the fittings but has since carried out retrospective analysis and determined a population of 1,123,195 were impacted out of a total company population of 3,093,158.
The company has acknowledged that the installation of unapproved products occurred. While the company maintains that the risk to water quality was low, it was unable to fully substantiate this assessment due to the absence of detailed records relating to fitting types and precise installation locations. Recommendations were given requiring a critical review of procedures concerning record keeping, regulation 31 compliance assurance, and the methodology used for surface area calculations to demonstrate exemplar status.
Companies should ensure that systems are in place to prevent the use of products that do not have regulation 31 approval, the critical point in time being when water goes into supply.
ALE – Boughton loss of supply event
Following an upstream burst on a Dŵr Cymru Welsh Water main, the potable supply to Shotton Mill was lost from 14:00 on 14 August 2025 until 20:00 on 16 August 2025, a total interruption of 51 hours. During this period, the company implemented contingency measures to maintain water availability on site by providing an alternative supply from the separate, partially treated, non potable industrial system fed from Dŵr Cymru Welsh Water Ashgrove Water Treatment Works.
The industrial supply is dosed with sodium hypochlorite at Shotton Mill to maintain a free chlorine residual of 0.1 ppm, intended solely to suppress biological growth within the site’s industrial distribution network. Under normal operating conditions, the potable and non potable systems are entirely separate. However, to maintain water availability for welfare facilities during the incident, a single cross connection was installed, and the potable system was temporarily primed with the non potable supply.
The company provided trend data demonstrating total chlorine residuals, although free chlorine residuals were not evidenced, within the system over the period. Outlet turbidity, conductivity and pH data from online instrumentation at Ashgrove works were also supplied. Turbidity and conductivity remained within regulatory limits, while pH tracked marginally above the Schedule 2 minimum value of 6.5.
To protect consumers on site, the company issued Do Not Drink advice, displayed notices at all outlets, and provided bottled water as an alternative drinking supply.
Following restoration of the potable supply on 16 August 2025, the company removed the temporary cross connection (restoring the air gap separation), undertook flushing to achieve eight turnovers of the calculated system volume, and converted the Do Not Drink advice to Boil Water advice on 17 August 2025 pending satisfactory sample results.
Sampling surveys were conducted at site taps on 18 August 2025 for microbiology analysis and on 19 August 2025 for a full suite of water quality parameters. All results were satisfactory.
Companies are reminded that the wholesomeness requirements of regulation 4 apply to all domestic uses of water, including handwashing in welfare facilities, and that suitable sampling should be conducted to verify that supplied water meets this regulatory requirement. Further, the timing of the issuing and rescinding of any restriction of use advice should be based on a dynamic risk assessment verified through appropriate sampling.
DWR – Aston burst main
A burst occurred on the rising main supplying water from Bretton Works to Aston service reservoir (SR). An initial repair attempt was unsuccessful, resulting in the service reservoir emptying and causing a significant loss of supply to customers in the Deeside area. During the incident response, water tankers were deployed; however, no samples were taken from the tankers themselves. Water quality sample failures were subsequently recorded in areas downstream of tanker infusion points, and it remains unclear whether these failures were attributable to the burst event, to tanker operations or both.
The treatment works was restarted, and the service reservoir and wider network were recharged in a controlled manner. Adequate bottled water supplies were provided to affected customers throughout the incident.
A number of deficiencies were identified during the review of the event. These included the absence of a water quality risk assessment for the use of an overland temporary main, insufficient assessment of the robustness of the initial repair, contributing to a secondary burst and an extended loss of supply, and the lack of in‑situ turbidity monitoring during flushing activities. In addition, no in‑situ water quality check records were available, and none of the deployed tankers were sampled; only the fill point and downstream properties were sampled, with downstream failures recorded.
Recommendations have been given to address these deficiencies and strengthen future operational and water quality assurance processes.
The incident highlights the need for robust water quality assurance and risk assessment during emergency response activities, particularly when using tankers or overland supplies. Effective incident management must include thorough assessment of repair robustness to prevent escalation, clear water quality monitoring and recording during flushing and tanker operations, and routine sampling of tankers themselves to ensure risks are adequately controlled and can be clearly distinguished from underlying network failures.
DWR – Bryn Cowlyd works loss of supply
An event in North Wales in January 2025 left more than 114,000 people without a water supply for seven days. Dŵr Cymru Welsh Water’s Bryn Cowlyd works supplies the North Wales coastal area, including Conwy, Llandudno and Colwyn Bay. On 13 January, a potential leak was identified on the works outlet main. A walk of the main identified a leak under the Afon Ddu and the company began planning the repair. Before this could take place, a power outage in the early hours of 15 January resulted in an automatic shutdown of the works. When a restart was attempted, the outlet main failed to reach the required pressure and it was concluded that the works could not be restarted until the main had been repaired. The areas normally supplied by Bryn Cowlyd works subsequently experienced loss of supply over seven days while the main was repaired and the system refilled. This affected a population of 114,562 people, or 47,734 properties, supplied either directly from the gravity system from Bryn Cowlyd final water tank or by service reservoirs also supplied by the works.
Due to its location beneath a river, the repair of the main was challenging and required a large excavation to provide safe access. Once exposed, the company identified a pinhole leak that had eroded a coupling, creating a hole of approximately 50–75 mm.
Alternative supplies in the form of tankers, Arlington tanks and bottled water were deployed to provide a water supply to the affected population until normal supplies were restored.
The loss of supply generated local and national media interest, with reports of consumers collecting sea water to flush their toilets as they were unable to access sufficient alternative supplies for this purpose.
Once the main had been repaired and the works restarted, the company encountered further complications with elevated turbidity in the final water tanks, which served to provide required contact time to achieve full disinfection. In addition, samples collected from the works final sampling point prior to return to supply were deemed unrepresentative of the water from the tank. The company received a recommendation to collect a suite of samples on return to supply, to validate its risk assessment that the water was wholesome and adequately disinfected. Other recommendations were also made in relation to sampling, including sufficient sampling for Cryptosporidium, as the one sample collected during the event was unable to be analysed due to debris on the slide.
During the event, the company used 27 tankers to retain or restore supplies to consumers over the seven-day period, with 390 discharges made. Nine samples were collected from these tankers during the event, significantly below the requirements of regulation 6 of the Regulations, which requires tankers to be sampled at the point of discharge and during each subsequent 48-hour period. Prior to the event, the Inspectorate had served a regulation 28(4) notice on the company for alternative supplies, due in part to procedures and resources being insufficient to allow adequate sampling. While the company is working towards compliance with regulation 6 and the notice, events will continue to score higher on the Event Risk Index. The Inspectorate’s enforcement team is monitoring the company’s progress against the legal instruments.
DWR – Cardiff discolouration
On the evening of 11 March 2025, pressure and flow alarms were activated in the Cardiff area downstream of the Cefn Mabley service reservoir. Subsequent investigation identified that the water level in compartment 2 of the reservoir was 2.5 m, significantly lower than the 3.5 m indicated by telemetry. No associated reservoir level alarms, rate‑of‑change alerts, or flow deviation alarms had been generated. The reduced reservoir level resulted in partial depressurisation of a downstream trunk main and discolouration within the affected water quality zones.
At the time of the incident, compartments 1 and 1A of Cefn Mabley service reservoir were out of service for maintenance in accordance with the company’s tank inspection notice (ref. DWR‑2021‑00002). Telemetry data analysis enabled the company to trace the issue back to the service reservoir within two hours of the initial alarm activation.
Customer communication was undertaken through mass text messaging, telephone calls, and website updates. Sampling surveys were deployed across the affected area to assess water quality impacts, and bottled water was proactively supplied to vulnerable consumers. The defective level probe at Cefn Mabley service reservoir was replaced, a secondary level probe installed, and manual level monitoring implemented for the duration of the incident.
A company‑wide review of service reservoir level probes was subsequently completed, with remediation or contingency plans established for any probes found to be outside tolerance or inaccessible. The investigation identified a discrepancy between the written SR inspection procedure and the inspection form used on site, specifically regarding the requirement to verify level probe accuracy during inspections. This procedural inconsistency, combined with the absence of level verification and water quality checks during the earlier transfer of water from in‑service compartment 2 to out‑of‑service compartment 1, represented potential missed opportunities to prevent the incident.
The incident demonstrates the importance of reliable reservoir level monitoring and fully aligned inspection procedures, as failures in telemetry accuracy, alarm configuration, and verification processes—particularly during periods of reduced operational resilience—can lead to undetected loss of service reservoir volume, downstream water quality impacts, and missed opportunities for early intervention.
DWR – Llanrumney burst main
At approximately 15:00 on 3 September 2025, Dŵr Cymru Welsh Water was alerted, via pressure and flow alarms, to a burst on a 15‑inch trunk main in the Llanrumney area of Cardiff. The company implemented its established emergency planning arrangements, drawing on existing area and activity action plans, and acted swiftly to mitigate the impact on consumers.
Within the first twelve hours of the incident, two bottled water stations were established, bottled water was distributed to vulnerable consumers, tankers were deployed for network infusion, repairs were initiated, and rezoning measures were implemented to minimise the extent of the supply interruption. The repair was completed and the main recharged during the early hours of 4 September 2025.
The incident attracted local media interest, and the company provided regular progress updates to the press. Some deficiencies were identified in the alternative supplies’ response, particularly regarding tanker sampling in accordance with Regulation 6. However, overall, the company’s operational response effectively limited the impact on consumers and restored supplies in a timely manner.
To provide assurance against the requirements of SEMD, companies should focus on creating and exercising appropriate emergency plans for large-scale trunk main bursts of this nature. These plans should be designed to minimise supply disruption, protect consumer confidence and control water quality impacts.
DWR – Carmarthen loss of supply
A significant loss of supply event affecting the area east of Carmarthen occurred between 12 and 14 July 2025. The event was initiated by a burst on a 21-inch trunk main from Brondini service reservoir, detected at 04:49, with the burst located shortly afterwards. Initial risk assessments indicated up to 6,500 properties could be affected; however, operational mitigation, including rezoning, tankering and deployment of an overland main, reduced the number of properties without supply to 290 on the first day and 59 by 13 July. Repairs were completed on 14 July and supplies were fully restored. A total of 96 ‘no water’ contacts and one water quality contact were reported during the event.
The Inspectorate concluded that, although supply interruptions were managed to minimise consumer impact, there were deficiencies in water quality control and event reporting. It is likely that water supplied during the event may have been unwholesome, with elevated iron concentrations and reports of a ‘bitter’ taste identified in tanker samples, albeit unaccredited taste tests. There were also procedural gaps, including incomplete sampling of tankers discharging into the network and a failure to promptly withdraw a tanker following an adverse test result. In addition, information relating to sampling, tanker operations, and overland main commissioning and disinfection was not fully reported.
Companies should ensure that robust and consistently applied operational controls are employed during emergency supply arrangements, particularly where tankers and overland mains are deployed. Companies should also ensure that all alternative supply assets are fully risk assessed, disinfected, sampled and documented, and that adverse results trigger immediate and decisive action. Accurate and complete event reporting, including consumer contacts and operational data, is essential to demonstrate compliance and support learning. Strengthening these areas will reduce the risk of supplying unwholesome water and improve resilience during loss of supply events.
DWR – Gower/Swansea discolouration
A significant discolouration event affecting the Gower area of Swansea occurred between 15 and 16 October 2025, following a burst on a 20-inch steel trunk main. A small leak identified on 9 October progressively worsened and resulted in a major burst, which was isolated in the early hours of 15 October. The burst occurred as a result of a longitudinal split adjacent to a previously installed repair collar. Network flushing and controlled flow reduction were implemented to manage the impact, and the main was fully repaired by 20:00 on 16 October. While no loss of supply occurred and positive pressure was maintained, 45 consumer contacts relating to discolouration were received. Sampling identified failures for iron, turbidity and manganese at four properties on 15 October, although resamples taken the following day were satisfactory, demonstrating that water quality was restored promptly.
The Inspectorate concluded that the company supplied water that exceeded prescribed concentrations for turbidity, iron and manganese, rendering it unwholesome and in breach of regulation 4(2)(c) of the Regulations. However, it noted that the operational response was timely and effective, with coordinated actions between network and production teams reducing discolouration risk and facilitating repair. Mitigation measures included controlled reductions in treatment works output, flushing of affected mains, and in situ checks (although turbidity monitoring was not undertaken during flushing). The Inspectorate welcomed the proactive mitigation approach and effective verification prior to returning the network to supply, while also recommending that historic repairs are more explicitly considered within leak detection, risk assessment and prioritisation processes.
Companies are encouraged to consider the importance of proactive and risk-based management of trunk mains, particularly where historic repairs and emerging leaks coincide. Companies should ensure that risk assessments incorporate asset history to prioritise intervention and prevent escalation to burst events. Operational responses should include comprehensive real-time monitoring, including turbidity, alongside coordinated network control and verification prior to returning assets to service. Maintaining sufficient critical repair materials and embedding learning from trunk main reviews will strengthen resilience and reduce the likelihood and impact of discolouration events.
DWR – Llanrumney do not drink
A significant event occurred in Llanrumney, Cardiff on 17 October 2025, where ‘do not drink’ advice was issued to 20 properties. The event followed a sediment removal flush during which a valve, previously recorded incorrectly on the GIS system, was opened without appropriate risk assessment. This allowed stagnant water to enter the distribution main, leading to multiple consumer contacts reporting discolouration, unusual tastes and odours, and one report of gastric illness. The valve was re-isolated on the same day and restrictions remained in place until 29 October 2025, when satisfactory sampling results enabled the advice to be lifted.
The Inspectorate concluded that the company supplied water that exceeded prescribed concentrations for a range of parameters including metals, turbidity, hydrocarbons and polycyclic aromatic hydrocarbons, and therefore the water supplied was unwholesome. Deficiencies were identified in operational control and incident response, including failure to undertake a suitable risk assessment prior to valve operation, delays and limitations in the initial sampling strategy, and a lack of early recognition of potential polyaromatic hydrocarbon (PAH) contamination despite consumer reports indicating petrol-like odours. The sampling approach, which followed mains flushing, reduced the likelihood of capturing representative worst-case water quality, and this was considered a breach of regulatory requirements. The Inspectorate also identified shortcomings in determining the extent of the affected area and in the timely and consistent application of ‘do not drink’ advice to consumers.
The event highlights the need for robust risk assessment and control of network operations, particularly when altering valve configurations or undertaking flushing activities. Companies should ensure that asset condition, operational history and data accuracy are fully considered before interventions, and that sampling strategies are designed to capture representative water quality at the earliest stage of an event. Early recognition of contamination risks, particularly where multiple consumer indicators suggest chemical contamination, should trigger an appropriately escalated response. Strengthening these areas, alongside improved processes to define and protect the affected population, will reduce the risk of supplying unwholesome water and improve the effectiveness of response to water quality incidents.
HDC – Market Street Llangollen do not drink
A significant event occurred in Llangollen on 11 December 2025, involving the issue of ‘do not drink’ advice to six properties following the detection of elevated lead concentrations. The event developed from earlier sampling in June and July 2025, which identified lead within a shared supply serving eight properties. Flushing advice was initially provided; however, further sampling in December identified an elevated 2-minute flush result at one property, triggering the escalation to ‘do not drink’ advice. Investigations identified that the shared communication pipework comprised mixed materials, including MDPE, copper and galvanised pipe, with some properties remaining on restrictions into January 2026 due to ongoing access and replacement challenges.
The Inspectorate concluded that water supplied during the event exceeded the prescribed concentration for lead and was therefore unwholesome in breach of regulation 4(2)(b) of the Regulations. While the company had undertaken risk assessments and mitigation actions, these did not fully capture the extent of the lead risk or prompt sufficiently protective responses at earlier stages of the investigation. Delays were identified in confirming pipework materials and in sampling all properties on the shared supply. The Inspectorate also highlighted that the potential contribution of galvanised pipework to lead exceedances, including via galvanic action, had not been fully considered. Recommendations were made to improve procedures for pipe material verification, strengthen lead risk assessments, and consider clearer triggers for escalation to ‘do not drink’ advice.
Companies are encouraged to ensure a precautionary and systematic approach to managing lead risks, particularly on shared supplies with uncertain or mixed pipe materials. Companies should ensure that all connected properties are promptly identified, sampled and protected where elevated results are detected, and that physical verification of pipework materials is undertaken at the earliest opportunity. Risk assessments should explicitly consider mechanisms such as galvanic corrosion and the presence of galvanised pipework, and clear escalation criteria should be in place to trigger protective advice. Strengthening these practices will reduce the risk of prolonged consumer exposure to lead and improve the effectiveness and timeliness of incident response.