The UK water infrastructure sector is adapting to changes in reservoir safety and environmental legislation, particularly concerning aquatic invasive non-native species such as zebra mussels (Dreissena polymorpha) and quagga mussels (Dreissena rostriformis bugensis). Their impacts on large raw water reservoirs, including sedimentation, corrosion, and ecological degradation, necessitate integrated biosecurity and mitigation strategies across clients and designers. Conventional mitigation measures including protective coatings, high-pressure water jetting, scraping tools, abrasive brushing, specialised pigging systems, Regulation 31-approved molluscicides, nature-based catchment interventions, and operational abstraction controls are widely used to manage mussel colonisation within water infrastructure assets. However, their effectiveness and feasibility are often constrained by the accessibility of the structure, the extent of biological fouling, and the safety risks involved with confined or submerged working environments. These limitations highlight the need for innovative, safer, and more adaptable solutions for challenging hydraulic environments. The use of a remotely operated cleaning machine (ROCM), combined with the early integration of biosecurity, access, and maintainability requirements within asset design, introduces such a step change in dam safety practice. The ROCM provides efficient, diver-free cleaning of culverts, intake tunnels, and other restricted hydraulic assets through remotely controlled, hydraulically driven brushes and shredding mechanisms, significantly reducing operational risk while enhancing reliability and long-term maintainability.
The Flood Estimation Handbook (FEH) rainfall depth-duration-frequency (DDF) model is routinely used in combination with event-based rainfall-runoff models when estimating design floods with return periods up to 10 000 years. The FEH22 model is the latest version, superseding FEH13 and the original FEH model. This paper examines the difference between design rainfall and peak flow for the 150-year, 1000-year and 10 000-year events across 71 of Welsh Water’s reservoir catchments. Rainfall depths (for the catchment specific critical duration) were found to be up to 37% lower when using the FEH22 model compared with the FEH13 model. This increased to 44% when estimating the 10 000-year peak flow, using the default parameters in ReFH2.3. This illustrates how significant the choice of DDF model can be when carrying out flood conveyance assessments and designing schemes to comply with reservoir safety legislation/guidance. As the consequences of a dam failure could be catastrophic, a precautionary approach is often applied within reservoir safety. Care should however be taken to ensure the FEH13 DDF model is still reliable, especially where erroneous data was known to have been used for calibration. Using the FEH22 model and adding an allowance for uncertainty and/or climate change may be more appropriate.
This briefing note summarises the presentations held at the International Commission on Large Dams 2026 conference regarding the new Technical Bulletin 209 – ‘Methods for the Improvement of Data Quality and Management of Dam Performance Monitoring Data’ which is currently being produced by the Technical Committee on Dam Surveillance. The presentations summarised the different chapters within the bulletin, which address the foundational steps essential for effective surveillance and risk management.
Permanent siphons are increasingly being employed to increase the discharge capacity at existing reservoirs to ensure the available drawdown provision is in line with current guidance. Routine ‘wet’ testing of reservoir drawdown systems, during which flows are discharged from the system, is considered fundamental to reservoir safety to ensure that there is confidence that such systems can be relied upon to operate in emergency situations. This is particularly true for siphon systems which can be more complex to operate than other gravity systems. However, reservoir discharges to the downstream watercourse may be constrained by requirements for pollution and flood risk mitigation to downstream receptors. The paper outlines a range of options considered and employed across multiple sites to manage the discharge of outflows from permanent siphon systems. It discusses constraints encountered – actual and theoretical – and outlines potential lessons learned for future reservoir siphon schemes. The discussion considers methods of siphon priming; the philosophy of ‘wet’ test operations; how outflows can be terminated; and if and by what means outflows are conveyed to the downstream watercourse or are otherwise managed.
Craig Goch reservoir is the most upstream reservoir in the Elan Valley cascade and forms part of the water supply system serving Birmingham via Severn Trent Water and areas of mid-Wales via Dŵr Cymru Welsh Water. During a reservoir drawdown to facilitate diving in the reservoir as part of works for a ‘Measure to be Taken in the Interest of Safety’ under Section 10 of the Reservoirs Act 1975 the butterfly valve on the turbine bypass failed. The turbine bypass at Craig Goch is the most effective means of drawdown at the reservoir in normal operation and emergencies as it is considerably larger in diameter than the scour. Following the closing of the upstream guard valve it was found that the butterfly plate from the valve had sheared off and been ejected from the end of the pipe. The paper shares lessons learnt from a valve failure and covers the likely transient pressures experienced, the poor design of the valve that had been installed, the risk of using a valve for purposes for which it was not intended and the importance of an operable secondary means of isolation, especially where the asset is a critical water resource.
