Real-time water quality monitors are being deployed at various wild swimming locations and beaches throughout southern England to help people gauge their immediate risk of illness from polluted water. This initiative follows a successful pilot study at Warleigh Weir near Bath, where an AI-based system accurately predicted high bacterial levels 87% of the time.
Why Are AI-Based Water Quality Monitoring Systems Needed?
Wessex Water is installing these advanced sensors at three freshwater sites in Dorset, Somerset, and Hampshire, and two coastal locations in Bournemouth. This comes after the successful trial at Warleigh Weir, where the artificial intelligence system demonstrated its effectiveness in predicting water contamination.
Meanwhile, Southern Water is testing a different monitoring approach at Tankerton in Kent and Langstone Harbour in Hampshire, with plans to extend the system to Hayling Island soon.
Despite current regulations requiring water companies and environmental regulators to test for pollution markers, there is no mandate to test for harmful faecal bacteria like E. coli and intestinal enterococci, except at designated bathing sites. Moreover, water sample results, which are sent to laboratories, are typically delayed by a week or more, complicating the assessment of immediate risks from sewage or agricultural runoff.
How Will These Technologies Affect Future Water Quality Monitoring?
Wessex Water initiated the use of these systems at Warleigh Weir following local demand for more transparent data on the impact of storm overflows and treatment processes on the River Avon’s water quality. According to Ruth Barden, the director of environmental solutions at Wessex Water, the goal is to ensure that people can safely enjoy their watercourses and understand the impact of water quality issues.
One of the innovative systems tested is developed by UnifAI Technology, a UK-based startup. This AI-based approach does not measure bacteria directly but infers high levels of E. coli or enterococci by analysing data from real-time sensors placed upstream. These sensors monitor pH, temperature, turbidity, dissolved oxygen, and ammonia levels.
Under the Environment Act 2021, water companies are required to install such sensors at sites upstream and downstream of storm overflow and wastewater treatment discharges to enhance water quality and wildlife protection. During an initial six-month training period, the AI learns to correlate bacteria levels with sensor data patterns. Following this, an app provides public alerts on water quality every half-hour, signalling potential high bacteria levels.
Dan Byles, the chief commercial officer at UnifAI, emphasises that while the system does not declare water safe, it does issue alerts about potential issues.
Wessex Water plans to expand this technology to additional swimming spots including Farleigh Hungerford, Fordingbridge, and Poole Park lagoon, along with Bournemouth and Boscombe piers, with real-time alerts expected by 2025. The company is also negotiating with landowners and river users at 20 sites across south-west England for further installations.
The expansion of these technologies promises to offer valuable insights into how storm overflow discharges and other factors influence water quality. Dan Byles notes that this could help in creating a “digital twin” of river systems for better monitoring.
River Action UK has generally supported the introduction of real-time pollution monitoring but stresses that it should not divert attention from addressing the root causes of pollution. James Wallace, the chief executive of River Action, argues for stronger enforcement of pollution laws to compel investment in maintaining and upgrading sewage treatment facilities.
Southern Water is also exploring the application of UnifAI’s machine learning models to coastal water data, using algorithms to estimate bacteria levels based on light scattering. However, as the technology is still in the research phase, the data is not yet available to the public.
