Every year a new set of statistics attempts to establish exactly how many million hours water companies have spent spilling raw sewage into Britain’s rivers and coastal waters. These figures are met with horror, but the truth is worse still: we do not actually know how much sewage is being spilled. New research shared exclusively with the New Statesman suggests that most of the monitors used by the water industry’s leading water company to track spills have generated inconsistent results. This implies that significant sewage spills into rivers such as the Derwent in Derbyshire and the River Severn in Shropshire have been missed, and thousands more have been misreported.
This data is used by the government to regulate the companies that provide water and sewage services, which are at the core of our critical national infrastructure. It is used to ensure that bathing water is safe to swim in, and to direct billions of pounds’ worth of investment.If the findings are correct then it is not fit for purpose.
Spills happen because each sewage treatment works has a limited capacity for the amount of untreated sewage (in litres per second) it can handle. During a period of heavy rain the flow can exceed this level and the extra rainwater, mixed with raw sewage from homes and businesses, will overflow into a storm tank. Once the tank is full (which can take a few hours) the sewage begins spilling from the tank into the environment.
The sewage that flows into treatment works (and the treated water that leaves them) are monitored by “flow meters”. These monitors receive Environment Agency-approved certification and have been relied upon for decades. In recent years, however, the government has sought to address the growing problem of sewage pollution by requiring the installation of a different kind of monitor – “event duration monitors”, or EDMs – on storm overflows. In December last year the Environment Secretary, Steve Barclay, announced that all of the UK’s nearly 15,000 outflows are now fitted with EDMs, which he said would collect a “wealth of data” to “ensure that we know the full extent to the problem – increasing transparency, revealing the worst-offending overflows, and enabling regulators to hold polluters to account”.
But independent analysis of the EDM data has found it to be highly questionable. Peter Hammond, a computational biologist who works with the campaign group Windrush Against Sewage Pollution, has analysed data from 200 sewage treatment works over a two-year period, and found that EDMs at more than half of these sites generated results that were inconsistent with the flow of sewage and local rainfall.
Hammond focused his analysis on data collected by Severn Trent, which is the water industry leader in Environmental Performance Assessments, having been awarded the highest four-star status every year for the past four years, and was the first water company in the UK to ensure all of its storm outflows were monitored by EDMs. Hammond believes that if there are questions to be asked about Severn Trent’s data, these questions are likely to apply to all the UK’s water and sewage companies.
A sewage spill at a wastewater plant has a clear pattern, Hammond explains: the untreated sewage flowing into the works rises to a certain rate (the “storm overflow rate”) and then typically levels off, as it overflows into the storm tank (which, when full, overflows into a watercourse such as a river). At this point the EDM – which measures the depth of fluid in the storm tank, usually with an ultrasonic sensor – should begin reporting a spill. However, Hammond found many examples in which the flow pattern clearly illustrates a spill, but the EDM has not recorded one.
These inconsistencies suggest that sewage is regularly being spilled into rivers, but not being recorded. “EDMs are the way that spills of untreated sewage are going to be monitored in the future,” Hammond told me, “so it’s worrying that there’s already signs of huge unreliability.”
Hammond’ ‘s findings include evidence that untreated sewage spilled into the River Derwent for hundreds of hours with no spills being recorded by the EDM.
These spills occurred from a sewage treatment works at Matlock in Derbyshire, in November 2022, a month in which heavy rain caused flooding across the UK. As the rainfall increased in Matlock, flow meter recorded sewage flowing into the works rising to around 100 per cent of the storm overflow rate, then flattening off – the clear signature of sewage overflowing into the storm tank. At one point this continued for around two weeks, but the EDM did not record a single spill.
At other times, spills were recorded by monitoring equipment when it was highly unlikely that they could be taking place, suggesting the monitors also routinely generate false positives.
In April 2021, for example, there was almost no rain for four weeks, and the flow of sewage into the Rainworth sewage works in Nottinghamshire never reached more than 70 per cent of the rate at which the storm overflow begins to be used – but the EDM recorded spills on 16 different days.
Water companies must tell the Environment Agency if their EDMs are not operational, and the industry has been criticised in the past for failing to ensure all EDMs are working. However, the data Severn Trent submitted to the regulator for these periods states that the EDMs at Matlock and Rainworth were operational for more than 99 per cent of the reporting period. There is no suggestion Severn Trent has acted improperly or broken the law in the way it has handled its data – the question is whether EDMs in general are sufficiently reliable.
Hammond has compiled a detailed report of his findings, which also shows that EDMs which appear reliable can also miss spills if they stop working. At Itchen Bank sewage works, on a stretch of the River Severn that Severn Trent has marked for improvement as a “bathing river”, an EDM appears to have missed at least eight spills while not operational. At Fleckney sewage works, which discharges into the Grand Union Canal, more than 200 hours of spills appear to have missed.
In all, Hammond’s analysis of 400 annual data series found that 55 per cent of EDMs generated results that appear inconsistent with the data on sewage flow and rainfall.
When the New Statesman shared the data and analysis from Rainworth and Matlock with Severn Trent, the company responded that it would not be able to comment without further investigation that included accounting for the complexities of each site, but that it does conduct retrospective reviews of EDM data. Rainfall and sewer catchment patterns, it noted, are complex and should be factored into the data from each individual monitor.
On 10 November 2021, Boris Johnson's government announced that its “world-leading” Environment Act had become law. Thanks to 22 Conservative MPs who voted against the government to include an amendment on sewage, the Act imposed a duty on water companies to reduce sewage discharges. However, it also enshrined the event duration monitor in UK law. The Act specifies that it is the “frequency and duration” of sewage spills that is important; the volume of sewage spilled is accounted for “where the information is available”.
Despite their legal status, almost none of the EDMs used on the almost 15,000 sewage outflows in England and Wales have been certified by the Environment Agency. The certification process was only introduced in March 2023, after the majority of EDMs had been installed. A guide compiled by the water industry in 2021 acknowledges that for some EDM sensors, “additional information is required to determine whether the sensor is reading correctly”. Some types of sensors “can be left in the incorrect position after high levels and provide unreliable data”, the guide warns, while even the more advanced sensors can suffer faults including a build-up of dirt or condensation that can cause them to “struggle to read properly” or to generate “noisy data”.
The Environment Agency describes EDM data as “key” to its regulation of water companies and says that the “robust and consistent” monitoring they provide helps it to direct around £1bn per year of investment in infrastructure. Environmental performance forms a key part of Ofwat’s (the water services regulation authority) considerations on the extent to which water companies are allowed to raise their bills.
There are signs that water companies themselves are aware that EDMs can be unreliable. At least 2,500 remote cameras have been installed at storm overflows around the UK to allow companies to check if spills are really happening, but the companies are not compelled to share the camera data with the Environment Agency.
This confusion appears convenient for water companies. Because EDMs only measure the duration of a sewage spill – and not the amount spilled – it is much harder for the government to impose fines or charges. The Liberal Democrat manifesto promises a “sewage tax”, while the Labour manifesto commits to “automatic and severe fines” and “independent monitoring of every outlet”. Such measures will be impossible without reliable data.
If flow meters were used to measure spills, said Hammond, “you might be charged per litre of what you’ve spilled. And if it turns out it was illegal as well, then maybe you could then add a fine on top. You wouldn't need to go to court, you would just introduce it as a charge. Most companies definitely don't want that.”
Dodgy data also helps to shape the public's perception of sewage spills. A small run-off of mostly rainwater and a billion-litre spill of raw sewage become the same data point, if they last for the same amount of time, and headlines can only report the dispassionate statistics of hours of spillage. Thanks to the government's failure to establish reliable facts about sewage spills, the true extent of water companies’ failure to protect the environment remains obscure.
[See also: The Great Stink: Britain’s pollution crisis]
Content from our partners
Related
