View PDF Version
 
DOI: 10.1039/D4EW00693C (Paper) Environ. Sci.: Water Res. Technol., 2025, Advance Article

A comprehensive analysis of combined stormwater overflow (CSO) discharges to coastal waters and data communications by a UK water company between 2021 and 2023

Alex T. Ford *a, Richard Boakes b, James Richardson c and Mike Owens d
aInstitute of Marine Sciences, University of Portsmouth, Ferry Road, Portsmouth, PO4 9LY UK. E-mail: alex.ford@port.ac.uk
bSchool of Computing, University of Portsmouth, UK
cTop of the Poops, London, UK
dHayling Sewage Watch, Hayling Island, UK

Received 20th August 2024 , Accepted 4th September 2025

First published on 4th September 2025

Abstract

Sewage discharges to aquatic environments present a real danger to human and ecosystem health. Event duration monitors (EDMs) from combined stormwater overflows (CSOs) are now fitted to over 90% of storm overflows in England and Wales. These have transformed our understanding of consented and non-consented discharges of sewage and wastewater from UK water companies. In 2018, Southern Water Services Ltd launched “Beachbuoy” which is an online ‘near’ real-time platform to inform customers when EDMs have been activated at particular CSOs and bathing water sites. Since April 2022, this water company categorized CSO discharges as genuine, genuine but non-impacting, and not genuine (false alarms by EDMs). We analyzed Beachbuoy data to provide an overview of CSO discharges and EDM activity and performance in the region. Across all assets, between December 2020 and February 2023 there were 7[thin space (1/6-em)]164[thin space (1/6-em)]656 genuine (impacting and non-impacting) minutes of discharges of which 19% overall were regarded as non-impacting of bathing water locations. Non-impacting discharges from all assets often persisted beyond multiple tidal cycles suggesting the impacts on bathing waters may need to be reevaluated. Discharges classed as ‘not genuine’ (false alarms) were highly variable between CSO for which some recorded false discharges 100% of the time. There were very strong correlations between the triggering of genuine and not genuine discharges and time of the day. Overall, 39% of all total minutes discharged and 14% of discharge events were classified as not genuine. Sewage releases from CSOs were more likely to happen between 7–10 am indicating that earlier morning patterns in human behaviours are substantially impacting the infrastructure's ability to tackle increased capacity in the system through precipitation. We discuss the appropriateness of classifying sewage discharges as ‘non-impacted’ and whether data should also be obtained on false negative discharges (EDMs not activating) as well as false positive discharges (not genuine). We call for better transparency of the data and models used by the water industry and regulation of how this information is presented to the public.

image file: d4ew00693c-p1.tif

Alex Ford

Alex Ford is a Professor of Biology at the University of Portsmouth with expertise in aquatic pollution. He is an elected board member of the International Panel on Chemical Pollution (IPCP).

image file: d4ew00693c-p2.tif

Richard Boakes

Richard Boakes is an Associate Head of Department in the School of Computing at the University of Portsmouth. He has expertise in computer programming and the use of disruptive technologies which bring about systemic change.

image file: d4ew00693c-p3.tif

James Richardson

James Richardson is a computer programmer with expertise in data visualisation. He directs a website Top of the Poops which uses publicly available data on sewage discharges into rivers and coastal waters.

image file: d4ew00693c-p4.tif

Mike Owens

Mike Owens is a citizen scientist and founder of a campaign group called Hayling Sewage Watch which raises awareness of poor water quality in South East England.


Water impact

Event Duration Monitors (EDMs) have revolutionised our understanding of the extent to which wastewater discharges occur into aquatic environments. We reveal the frequency and duration of sewage discharges to coastal environments plus the variability in their performance from one UK privatised water company. We highlight how EDM data can be better communicated to the public.

Introduction

Sewage discharges to aquatic environments present a risk to human and ecosystem health.1–3 Not only do discharges contain human pathogens but also a vast array of toxic chemical contaminants including pharmaceuticals, plastic additives, flame retardants, surfactants, metals and PFAS compounds. Because many of the pathogens are only expected to last in seawater for relatively short periods of time (days/weeks) and are often linked with visible signs of discharge such as wet wipes and sanitary products, the impacts of longer-lasting but invisible chemical contaminants are often overlooked. For example, sewage is a major source of PFAS compounds4–6 and these are toxic, bioaccumulate and classed as the most environmentally persistent substances among organic chemicals.7 Classed as ‘forever chemicals’ some have put estimates of over 1000 years on their breakdown and we are already starting to see these compounds bioaccumulating in our offshore wildlife.8–10 Therefore, we must understand the amounts of sewage countries are discharging into coastal environments as their impacts like those of other persistent chemicals (e.g. PCBs) may be seen for generations after they are banned.

The water industry in the United Kingdom is publicly owned in Northern Ireland and Scotland, a private not-for-profit organization company in Wales and is made up of nine privately owned water and sewerage companies in England plus a smaller number which provide water only services.11 Much of the sewerage infrastructure pre-1960s is a combined system which takes foul and surface water runoff. Built into this infrastructure were combined stormwater overflows (CSOs) which, under license from the Environment Agency, allows water companies to legally discharge untreated effluent into water courses during periods of extreme weather (rain or snowmelt). Between 2020 and 2022, there were an estimated 7.5 million hours of sewage discharges into watercourses in England and Wales.12 These untreated discharges despite being diluted present an impact to the health of water courses and a human health risk to water users.1,13,14

There has been considerable controversy over the frequent discharges from combined stormwater overflows (CSOs) in England prompting critics to suggest a failure in environmental and financial regulation.15 Currently, there are ongoing environmental and financial investigations into these water companies by the regulators and investigations into the regulators themselves by a newly formed Office of Environment.15–18 Furthermore, there are several judicial inquiries into the government in the UK High Court by NGOs and businesses plus civil actions against water companies for rebates on water bills. In 2021 Southern Water pleaded guilty to 51 counts of discharging untreated sewage into coastal waters between 2020 and 2025 which was estimated to be 16–21 billion liters.19 As a consequence, they were fined £90 million and the financial regulator forced the company to discount bills for their poor performance.

There are estimated to be 15[thin space (1/6-em)]000 CSOs in the UK. In 2013 the Department of Fisheries, Agriculture and Rural Affairs (DEFRA) insisted that the water industry had a better understanding of the performance of their wastewater assets and insisted on a program of fitting Event Duration Monitors.12 These EDMs work by either measuring flow or using non-contact radar, ultrasonic, pressure devices and floats.20 The Environment Agency hopes that the data from EDMs will “inform water company management of storm overflows under current and future pressures; and when they aren't meeting the bar, we can take robust action based on sound evidence to secure improvements, using enforcement if necessary”. Details of where and how best to fit EDMs have been published by the Chartered Institute of Water and Environmental Management.20 They report that special consideration should be taken over the location of EDMs to ensure accuracy (with minimum noise and interference), repeatability (consistently return the same measurement data for comparable spill events), reliability (minimize the likelihood of damage, make maintenance or repairs easier) and safety. Environment Agency data highlights the rapid 14-fold increase in the coverage of monitors, from 862 in 2016 to 12[thin space (1/6-em)]393 in 2021;12 (Fig. S1). The take-up of these monitors varies considerably between companies with Anglian Water having 54.3% CSOs listed as having EDMs and Southern Water having 98%;12 (Table S1). Southern Water Services Ltd covers the South and Southeast coast of England, treating the wastewater and supplying water services to 2.6 million customers. The assets cover 700 miles of coastline, 84 bathing water sites where they have just under 200 CSOs which discharge directly to the sea.

In 2018, Southern Water launched “Beachbuoy” which is an online platform which is meant to inform customers when EDMs have been activated at particular CSOs and bathing water sites (https://web.archive.org/web/20230330145232/https://www.southernwater.co.uk/water-for-life/our-bathing-waters/beachbuoy). This initially focused on two locations Langstone Harbour close to Portsmouth (UK) and the Isle of Wight before being expanded across their region. How the data has been presented on Beachbuoy has evolved since its initiation launch from originally having EDMs displaying under review before being categorized as genuine or not genuine via human validation. Initially, all discharges were considered genuine and the public were alerted. In April 2022 Southern Water made changes to the way the data was presented by differentiating between ‘impacting’ and ‘non-impacting’ discharges through what they refer to as dynamic modelling (Fig. 1). Discharges which are not expected to impact the water quality at bathing water locations are categorized as non-impacting and are meant to consider unique location-based hydrographic models and the tides and weather coupled with the known degradation rates of faecal coliform bacteria. The time it takes to validate spills as genuine or not genuine can vary spatially and temporally.


image file: d4ew00693c-f1.tif
Fig. 1 Screenshots of Southern Water's Beachbuoy app utilizing Google Earth to display a and b) red squares and yellow triangles when a release had occurred in the past 24 and 72 hours respectively (2018–2022) and c and d) the more recent displays when icons would stay white circles until a verification exercise had been conducted and then change to red or yellow icons indicating releases in the past 24 and 72 hours respectively (2022 to date).

Southern Water point out that “Following feedback, we released an upgrade in September 2022 that enables Beachbuoy to now account for tidal conditions. This means that the tool also considers the impact a release has on a local bathing water based on the location of the outfall, the duration of the release and tidal conditions at the time. While previously the Beachbuoy alert cautiously took the worst-case scenario when it was flagged on the app, this could inadvertently cause unnecessary worrying for the public and the tourism industry alike”.

The Environment Agency generates summary data on the discharge frequencies from storm overflows and how they compare to the sector averages. The current data for Southern Water is presented in Table S1 however these do not look at EDM performance in terms of false alarms, temporal trends or categorised such as non-impacting discharges. Therefore, this study aimed to investigate the publicly available data from the Beachbuoy website and determine the extent of false alarms and proportions of discharges categorized as non-impacting discharges across bathing sites, season and time. The EDM data is totally new to the wider community of environmental scientists, and as yet there is little academic appraisal of its reliability. This is appraisal is important because the data is already being used by industry, regulators and even scientists in quantitative calculations about sewage impact on the environment.21 The objective therefore of this paper was to make the data available in a format that trends, patterns and general observations could be summarized and help inform those studying the impacts on human and ecosystem health to understand the temporal and geographic extent of spills.

Methods

An Environmental Information Regulation (EIR) request was put to Southern Water Ltd on the 6th March 2023 which is a UK Statutory Instrument (2004) right of access to environmental information held by UK public authorities (ERI 1751). The data requested were for all the event ID's, site numbers, bathing sites, outfalls, activation start and end times plus activity fields from 2021-01-01 to 2023-03-06 as a csv file. On the 3rd April 2023 the water company reported they required more time to provide a comprehensive response and this would take a further 20–40 days. This EIR (1751) request was refused on the 12th April 2023 on the basis that the data was publicly available on the BEACHBUOY website (previously available via https://www.southernwater.co.uk/water-for-life/beachbuoy). Southern Water response stated that “Under the EIR Regulations and guidance issued by the Information Commissioner's Office, we are not required to create information in order to answer a request and as we do not already create, process or record this information, it cannot be said that we “hold” this information for the purposes of Regulation 12(4)(a) EIR. To clarify, Southern Water do not have a downloadable excel file for Beachbuoy, however, you are able to do this yourself. Just select an outfall you are interested in, and you can the copy and paste the data from the Historical Release Table into an excel spreadsheet”. The team were directed to Southern Water's official flow and spill reporting data however this was not suitable as it no longer includes discharges categorized as not-impacting or not genuine (https://web.archive.org/web/20231212162144/https://www.southernwater.co.uk/our-performance/environmental-performance/flow-and-spill-reporting). Southern Water are known to use a variety of different EDM sensors (optical, radar and float) across their assets although the types use at a particular location or not public knowledge.

The Beachbuoy website provides a list of all storm release activity relating to coastal bathing waters. Information that appears on the site is displayed in chunks of 25 spills/discharges. Anyone wishing to get a copy of the complete dataset of spills would need to spend many hours manually requesting every one of more than 800 pages, then copying and pasting 25 rows of data from each. Furthermore, because the data presented is dynamic with discharges constantly changing category from unverified to verified, and assigned genuine or not, this means the data is in constant flux making overviews of the data logistically challenging. As a result citizen science software developers wrote programs to automate this labour-intensive process which were optimised, and the results of each program were cross-checked for accuracy (https://github.com/csodata/tools). These yielded over 21[thin space (1/6-em)]000 rows of data (one row per spill event) and over half a million data items. We have made these programs available via GitHub to enable others to replicate our analysis.

Results

Between December 2020 and March 2023 there were 25[thin space (1/6-em)]294 discharges categorized on Beachbuoy (Fig. 2; Tables 1 and 2) across 77 classified bathing water areas and 186 outfalls. Presented alphabetically by CSO name and on a log[thin space (1/6-em)]10 scale, the median, interquartile ranges and minimum and maximum for each CSO are presented in Fig. 2. The median discharge duration for a typical CSO is between 10 and 1000 minutes with some extending beyond 100[thin space (1/6-em)]000 minutes.
image file: d4ew00693c-f2.tif
Fig. 2 Box plot with median, interquartile ranges and minimum/maximum discharges of genuine and genuine non-impacting discharges from Southern Water outfalls (plotted alphabetically by discharge location).
Table 1 Frequency of discharges events between December 2022 and March 2023 categorised as genuine, genuine not-impacting and not genuine from 176 Southern Water outfalls
Year Row labels Genuine Genuine – non impacting Not genuine Under review Under review – non impacting Grand total
2020 2020 1         1
  Qtr4 1         1
2021 2021 10[thin space (1/6-em)]439         10[thin space (1/6-em)]439
  Qtr1 3236         3236
  Qtr2 1679         1679
  Qtr3 1974         1974
  Qtr4 3550         3550
2022 2022 5708 3765 1502 147 270 11[thin space (1/6-em)]392
  Qtr1 1746 1 39     1786
  Qtr2 451   50     501
  Qtr3 637 390 168     1195
  Qtr4 2874 3374 1245 147 270 7910
2023 2023 1152 1511 472 178 149 3462
  Qtr1 1152 1511 472 178 149 3462
Grand total   17[thin space (1/6-em)]300 5276 1974 325 419 25[thin space (1/6-em)]294

Table 2 Duration of discharges (minutes) between December 2022 and March 2023 categorised as genuine, genuine not-impacting and not genuine from 176 Southern Water outfalls
Year Row labels Genuine Genuine – non-impacting Not genuine Under review Under review – non-impacting Grand Total
2020 2020 339         339
  Qtr4 339         339
2021 2021 2[thin space (1/6-em)]612[thin space (1/6-em)]472         2[thin space (1/6-em)]612[thin space (1/6-em)]472
  Qtr1 1[thin space (1/6-em)]085[thin space (1/6-em)]528         1[thin space (1/6-em)]085[thin space (1/6-em)]528
  Qtr2 298[thin space (1/6-em)]399         298[thin space (1/6-em)]399
  Qtr3 400[thin space (1/6-em)]613         400[thin space (1/6-em)]613
  Qtr4 827[thin space (1/6-em)]932         827[thin space (1/6-em)]932
2022 2022 2[thin space (1/6-em)]315[thin space (1/6-em)]123 950[thin space (1/6-em)]533 2[thin space (1/6-em)]571[thin space (1/6-em)]231 14[thin space (1/6-em)]368 22[thin space (1/6-em)]508 5[thin space (1/6-em)]873[thin space (1/6-em)]763
  Qtr1 286[thin space (1/6-em)]098 68 363     286[thin space (1/6-em)]529
  Qtr2 48[thin space (1/6-em)]954   66[thin space (1/6-em)]249     115[thin space (1/6-em)]203
  Qtr3 100[thin space (1/6-em)]699 39[thin space (1/6-em)]923 30[thin space (1/6-em)]269     170[thin space (1/6-em)]891
  Qtr4 1[thin space (1/6-em)]879[thin space (1/6-em)]372 910[thin space (1/6-em)]542 2[thin space (1/6-em)]474[thin space (1/6-em)]350 14[thin space (1/6-em)]368 22[thin space (1/6-em)]508 5[thin space (1/6-em)]301[thin space (1/6-em)]140
2023 2023 883[thin space (1/6-em)]788 402[thin space (1/6-em)]401 359[thin space (1/6-em)]939 12[thin space (1/6-em)]176 15[thin space (1/6-em)]423 1[thin space (1/6-em)]673[thin space (1/6-em)]727
  Qtr1 883[thin space (1/6-em)]788 402[thin space (1/6-em)]401 359[thin space (1/6-em)]939 12[thin space (1/6-em)]176 15[thin space (1/6-em)]423 1[thin space (1/6-em)]673[thin space (1/6-em)]727
Grand total   5[thin space (1/6-em)]811[thin space (1/6-em)]722 1[thin space (1/6-em)]352[thin space (1/6-em)]934 2[thin space (1/6-em)]931[thin space (1/6-em)]170 26[thin space (1/6-em)]544 37[thin space (1/6-em)]931 10[thin space (1/6-em)]160[thin space (1/6-em)]301

Data entries scrutinized during this study found there were many incidences where one event ID had multiple entries. This sometimes could be explained by instances whereby an event needed to be verified and subsequently had its category changed. For some event IDs there were over 1000 logs which probably highlights faulty monitoring devices which were allocated the same ID. Other incidents occurred where one event ID number was attributed to several locations.

In total, these data represented over 10 million minutes of discharges (10[thin space (1/6-em)]095[thin space (1/6-em)]826 minutes or 7010 days) and are presented in yearly quarters in Tables 1 (event counts) and 2 (event duration) and in descending order in Fig. 3. The top 20 outfalls which between them discharged 5[thin space (1/6-em)]389[thin space (1/6-em)]965 minutes of genuine (including non-impacting) are presented in Table S2.


image file: d4ew00693c-f3.tif
Fig. 3 Sum (minutes − log[thin space (1/6-em)]10) of genuine and genuine non-impacting discharges from Southern Water outfalls.

The largest proportion of discharges would take place in the wetter months October to February with the amount discharging in the winter periods varying over the winter months of 2020/2021, 2021/2022 and 2022/2023 (Fig. 4). The winter period of 2022/2023 highlights visually the proportion of discharge categories as not impacting or not genuine since the new classification came into consideration in April 2022.


image file: d4ew00693c-f4.tif
Fig. 4 The (a) number and (b) sum duration (minutes) of discharges from Southern Water outfalls between January 2021 and January 2023 categorized by genuine, genuine non-impacting and not genuine.

Excluding the discharges categorized as under review the total number of discharges was broadly similar in 2021 (10[thin space (1/6-em)]439 events) compared to 2022 (10[thin space (1/6-em)]975 events; Table 1). However, the total number of minutes reported discharged was substantially more in 2022 (5[thin space (1/6-em)]836[thin space (1/6-em)]887 minutes) compared to in 2021 (2[thin space (1/6-em)]612[thin space (1/6-em)]472 minutes; Table 2). Discharges were further categorized as non-impacting or not genuine from April 2022 onwards. The total number of minutes discharged in 2022 was broken down into 40% classed as genuine (2[thin space (1/6-em)]315[thin space (1/6-em)]123 min), 16% as genuine non-impacting (950[thin space (1/6-em)]533 min) and 44% as not genuine (2[thin space (1/6-em)]571[thin space (1/6-em)]231 min). For comparison from the 1st quarter of 2023 for which data was captured 54% classed as genuine (883[thin space (1/6-em)]788 min), 24% as genuine non-impacting (402[thin space (1/6-em)]401 min) and 22% as not genuine (359[thin space (1/6-em)]939 min). In 2022, based on the number of discharge events (counts), 52% were considered genuine, 34% were classed as genuine non-impacting and 14% not genuine based on the counts of discharge events (Table 1). Therefore, ‘not genuine’ discharges or false alarms make up 14% of discharges as classified by counts, but 44% of the total minutes discharged. The average duration (min) of genuine, genuine non-impacting and non-genuine discharges were 336, 256 and 1485 minutes respectively. Those classified as not genuine would, on average, last 3 times longer in duration than genuine discharges.

Eighteen percent of the specific outlet locations have discharges categorized as not genuine for more than 50+% of the occasions (Fig. S2a). Twenty two percent of the specific outlet locations have discharges categorized as not genuine for more than 50+% of the overall minutes (Fig. S2b). A number of the outlets had 100% of their discharges classed as non-genuine. These were Albion Groyne Brighton, Black Rock Marine Drive Brighton, Cinque Ports Way Hastings No. 1, Foads Hill Cliffs End, Hurst Point View Totland, Roedean Close Bexhill, Shoreham No. 2, Sunnyfield Road Barton On Sea, Village Road Alverstoke, Military Road Alverstoke, Military Road Ramsgate No. 2 and Roedean Close Bexhill (Fig. S2). Results of the Spearman correlation indicated that there is a significant positive relationship (r(185) = 0.885, p < 0.001) between the proportion of the spills classified as not genuine and the proportion of the total overall minutes from outfalls (Fig. S3). However, there were many exceptions to the relationship. For example, many outlets had less than 5% of the discharges categorized as not genuine based on the counts but over 90% of the discharges based on the total minutes.

There was a clear and strong positive correlation between the outfalls which discharge most frequently (counts) and the ones which discharged for the longest duration (min) (Spearman rank, r(185) = 0.875, p < 0.001; Fig. 5).


image file: d4ew00693c-f5.tif
Fig. 5 Correlation between discharges categorised as either genuine (including non-impacting) based on the number of discharge events (log[thin space (1/6-em)]10 counts) and duration (log[thin space (1/6-em)]10 minutes) between Dec 2020 and March 2023.

Discharges by time of day

Discharges irrespective of classification would occur significantly more frequently during the morning (6–10 am) compared to other hours of the day with the exception of 2–3 am (Kruskal–Wallis X = 146.1, treatments = 24, values = 12[thin space (1/6-em)]206, p < 0.0001; Fig. 6).
image file: d4ew00693c-f6.tif
Fig. 6 Median number of discharges (minutes) categorized by hours of the day between December 2020 and February 2023. Those coloured red had significance differences from those coloured blue. Error bars were removed to aid interpretation.

When broken down into 5-minute time bins and plotted as sum duration minutes over 24 hours over there was a clear correlating daily patterns between those discharges classified as genuine (impacting and non-impacting) and those deemed to be not genuine (false alarms) (Fig. 7). The lowest point for minutes discharged was between 3–6 am but from 7 am onwards through to 11 am there would be increasingly higher numbers of wastewater discharges. Plotting the two variables (genuine and not genuine) together revealed clear and significant positive relationships between the times of the day when genuine and not genuine spills occurred (linear regression (F = 1996; DFn, DFd 1, 286; p < 0.0001) Y = 10.57 × X − 17[thin space (1/6-em)]435 R2 = 0.8747; Fig. 8).


image file: d4ew00693c-f7.tif
Fig. 7 Sum of minutes binned by 5-minute intervals across a 24-hour period discharged across all CSOs between Dec 2020 and March 2023 categorised as either genuine (including non-impacting) and not genuine.

image file: d4ew00693c-f8.tif
Fig. 8 Correlation between discharges categorised as either sum genuine (including non-impacting) minutes and not genuine minutes the between Dec 2020 and March 2023.

In 2022 Southern Water decided to categorize discharges not only on whether they were genuine or not but whether they were ‘impacting’ or not. Between Jan 2022 and Jan 2023, there were 5276 discharges categorized as non-impacting representing over 1.3 million minutes. The rank order of those CSOs reported as discharging the most non-impacting discharges in terms of events and duration are presented in Fig. S4.

Fig. 9 is a box plot of log[thin space (1/6-em)]10 median, interquartile range and minimum/maximum duration of discharges (minutes) discharged to every bathing water area across the South East region covered by Southern Water. Most bathing water sites have discharges which last a median time of between 100 and 1000 minutes (1.6 h to 16.7 h). However, we determined that many bathing water areas have discharges categorized as non-impacting that proceed the length of 1 and often many 4+ full tidal cycles. Some locations had discharges lasting a maximum of 100[thin space (1/6-em)]000 minutes (∼70 days) and were classed as non-impacting.


image file: d4ew00693c-f9.tif
Fig. 9 Box plot of log[thin space (1/6-em)]10 median, interquartile range and minimum/maximum duration of discharges (minutes) categorised as non-impacting to every bathing water area across the South East region covered by Southern Water between April 2022 and March 2023. The red dashed line indicates one complete tidal cycle (1 × low and 1 × high tide).

Discussion

Water companies in England have come under intense public and political scrutiny after it has been revealed that untreated wastewater regularly discharges into rivers and coasts, often during periods of no or low rainfall.15 As a result of this pressure, the government mandated the need for EDMs to be fitted into all WWTPs and CSO discharge points. Leading the early rollout of the EDM has been Southern Water whose area covers the South East coast of England. Here we highlight that over 2 years between Dec 2020 and March 2023 there were 25[thin space (1/6-em)]294 untreated discharges including those still being reviewed representing 10[thin space (1/6-em)]160[thin space (1/6-em)]301 minutes (7076 days) registered on the Beachbuoy database.

Are all EDM false alarms really false alarms?

The company changed the way it displayed and represented the data during that period introducing categories such as non-impacting discharges and not genuine discharges for those considered false alarms (or false positives). We highlight that “not genuine” discharges are quite frequent occurrences with EDM devices and 14% discharges by event counts and 44% duration in minutes. Several CSOs have only been presented as having ‘not genuine’ discharges and have never purportedly had a real stormwater discharge event. What we know is devices are reporting false readings over 14% of the time by event count, what we don't know is the percentage of the time they do not detect discharges when real events are occurring (false negatives). Neither the water companies nor the government agencies provide data on the performance of EDM technologies. During this study, we also highlight very strong correlations in daily readings of not genuine discharges to those occurring that are considered genuine which brings into question whether some false alarms are indeed false but true discharges. The occurrence of genuine and not genuine increasing and decreasing at the same periods throughout the day leaves two plausible explanations. Firstly, whatever causes the genuine discharges also triggered false alarms or secondly, some of the alarms reported as false are indeed real.

Also interesting is the timing of the genuine and not genuine discharges which are most likely to occur during early morning 6 onwards through to about 11 am. This indicates that the triggering of EDMs is not also just a component of excess rainfall but also the daily patterns of the general public such as showering, setting washing machines/dishwashers and flushing toilets before leaving for school or work. This supports Giakoumis and Voulvoulis21 who recently analysed EDM data and concluded that their findings reveal the chronic under capacity of the English wastewater systems as a fundamental cause behind the increased frequency and duration of CSO spills.

Are non-impacting discharges really ‘non-impacting’?

Since spring 2022 when discharges were categorised as ‘genuine impacting’ or ‘non-impacting’, there was approximately 35% reported as non-impacting in 2022 and over 50% for the 1st quarter of 2023. In total, there had been 1[thin space (1/6-em)]352[thin space (1/6-em)]934 minutes (940 days) of non-impacting discharges over the study period. We have established that the median discharge duration for a typical CSO to the coastal environment is between 100 and 1000 minutes however many of these outfalls have discharges which exceed several tidal cycles and some lasted over 70 days.

In 2022, Southern Water decided to change the way it displayed and presented its discharge data. The iconography on the Beachbuoy portal was changed so that alerts didn't immediately appear as red icons but white icons until discharged had been characterized as genuine or not or impacting or non-impacting. At the time the water company received criticism for the approach with the UK broadsheet newspaper running with the headline “Southern Water alters pollution alert tool to curb automatic red alerts”22 and the BBC “Southern Water's sewage discharge app changes provokes row”23 in September 2022. A spokesperson for Southern Water at the time said “This is an important step for the tool, ensuring we provide accurate, fair and clear information to wild swimmers, kayakers, paddleboarders and all users of the beautiful beaches across our region. Beachbuoy is leading the way in providing near real-time data about storm releases, but we must ensure it goes further to inform the public about the impacts to the watercourse and not limit their enjoyment of their local bathing water. We'll never hide data though, with all releases still available on the website”.

From April 2022 discharges considered to be genuine were further categorized as to whether they were genuine or genuine non-impacting. This also received criticism for several reasons. Firstly, the algorithms and oceanographic models from which the data was based were not available for scrutiny plus the term ‘non-impacting’ if the models are correct only provides data for designated bathing water areas and estimates ‘risk’ of human pathogens in seawater. However, these do not consider chemical contaminants in wastewater and ecological impacts, or discharges modelled currents flowing in the opposite direction to designated bathing water beaches.24 In July 2023, on the back of this criticism Southern Water changed the terminology from “non-impacting” to “non-impacted” to better represent the condition of the specific bathing water location. However, Ford et al.15 pointed to the fact that these ‘non-impacting’ discharges or indeed ‘non-impacted’ discharges could well be ‘impacting’ water quality for human and ecological health in areas not close to designated bathing water locations. We therefore suggest that if algorithms are based on the tidal cycles, wind and ocean currents this data be made available and fully scrutinized to determine the validity of those statements.

Summary

We have highlighted that the CSOs served by Southern Water discharged over 7000 days of untreated effluent into coastal environments in just over 2 years between December 2020 and March 2023. This is the equivalent of over 19 years of continuous discharges to the South Coast of England of untreated wastewater containing chemical contaminants and pathogens to the English Channel and Southern North Sea. Investigations by the BBC25 and Unearthed26 found that these discharges often occur during dry weather and directly or close to Special Areas of Conservation (SACs), Sites of Special Scientific Interest (SSSIs), RAMSAR sites for the protection of migratory seabirds, UNESCO World Heritage Sites and protected shellfish grounds. Therefore, the information presented by the water companies must be accurate and in a form that is easily understood to fully understand the impacts on human health and the environment. The data presented within was denied to the researchers via a statutory Environmental Information Request (EIR) because the data was publicly available online. However, in this instance, the data was not available in a way that is or was easily scrutinized. Water Companies in England have been criticised for not making environmental data easily available to researchers and environmental organizations. For example, the BBC during their dry spills investigation was denied data from 6 or 9 major water companies in England because they said “they couldn't provide information because they were already being investigated for potential illegal spilling by industry regulator Ofwat and the EA”.25 However, OFWAT the financial regulator of the water industry has recently warned water companies they are endangering public trust by withholding data.27 In response, a senior lawyer for South West Water is reported as saying in an article for the Times newspaper that “it is the regulators and not the press or public” who should be able to view information regarding “complex technical and regulatory matters”.27

We have highlighted that many EDMs on CSOs are faulty and provide many false alarms which at some specific locations can be 100% percent of the time. We have also highlighted an worrying correlation between discharges classified as genuine and not genuine which warrants further investigation as they collectively occur at similar times of the day. Water Companies in England self-report their discharges to the environmental regulator. It is important therefore that the regulators and public get accurate information on discharges. We have highlighted that discharges formerly classified as “non-impacting” before recently being renamed “not impacted” occur with considerable frequency and duration. The water companies may have modelled the risks posed based on death rates of bacteria in seawater and tidal currents. These frequent and lengthy discharges are likely to have an impact on the aquatic environment. It is expected that more water companies will make real-time or near real-time monitoring available to the public. We encourage all water companies to make their data and modelling open source. We also suggest that regulators serving the environment (EA), finances (OFWAT) and information commissioner to provide guidance on how EDM data by water companies are presented to the public.

Conflicts of interest

ATF has in the past received funding from water companies. JR and MO work for organizations which promote the impacts of sewage into aquatic environments.

Data availability

Supplementary information: Supplementary material contains data on a) numbers of event duration monitors commissioned b) Percentages of sewage discharges considered not genuine by outfall c) Relationship between not genuine counts and not genuine minutes d) Percentages of non-impacting discharges by outfall e) summary data on event duration performance across the sector and f) worst performing CSOs by minutes discharged. See DOI: https://doi.org/10.1039/D4EW00693C.

All data is publicly available and/or available on request.

References

  1. C. Tuholske, B. S. Halpern, G. Blasco, J. C. Villasenor, M. Frazier and K. Caylor, Mapping global inputs and impacts from of human sewage in coastal ecosystems, PLoS One, 2021, 16(11), e0258898,  DOI:10.1371/journal.pone.0258898.
  2. S. L. Wear, V. Acuña, R. McDonald and C. Font, Sewage pollution, declining ecosystem health, and cross-sector collaboration, Biol. Conserv., 2021, 255(109010), 109010,  DOI:10.1016/j.biocon.2021.109010.
  3. A. Younger, S. Kershaw and C. J. A. Campos, Performance of storm overflows impacting on shellfish waters in England, Land, 2022, 11(9), 1576,  DOI:10.3390/land11091576.
  4. S. P. Lenka, M. Kah and L. P. Padhye, A review of the occurrence, transformation, and removal of poly-and perfluoroalkyl substances (PFAS) in wastewater treatment plants, Water Res., 2021, 199, 117187 CrossRef PubMed , Available from: https://www.sciencedirect.com/science/article/pii/S0043135421003857.
  5. S. Kurwadkar, J. Dane, S. R. Kanel, M. N. Nadagouda, R. W. Cawdrey and B. Ambade, et al., Per-and polyfluoroalkyl substances in water and wastewater: A critical review of their global occurrence and distribution, Sci. Total Environ., 2022, 809, 151003,  DOI:10.1016/j.scitotenv.2021.151003.
  6. A. T. Ford and F. Ginley, Insights into PFAS contaminants before and after sewage discharges into a marine protected harbour, Chemosphere, 2024, 366(143526), 143526,  DOI:10.1016/j.chemosphere.2024.143526.
  7. I. Cousins, J. C. DeWitt, R. Lohmann, J. Glüge, G. Goldenman and D. Herzke, et al., The High Persistence of PFAS is Sufficient for their Management as a Chemical Class [Internet], engrXiv, 2020, preprint,  DOI:10.31224/osf.io/exszc.
  8. C. Andvik, E. Jourdain, J. L. Lyche, R. Karoliussen and K. Borgå, High levels of legacy and emerging contaminants in killer whales (Orcinus orca) from Norway, 2015 to 2017, Environ. Toxicol. Chem., 2021, 40(7), 1850–1860,  DOI:10.1002/etc.5064.
  9. C. Wood, G. H. Balazs, M. Rice, T. M. Work, T. T. Jones and E. Sterling, et al., Sea turtles across the North Pacific are exposed to perfluoroalkyl substances, Environ. Pollut., 2021, 279(116875), 116875,  DOI:10.1016/j.envpol.2021.116875.
  10. G. Boldrocchi, D. Spanu, S. Polesello, S. Valsecchi, F. Garibaldi and L. Lanteri, et al., Legacy and emerging contaminants in the endangered filter feeder basking shark Cetorhinus maximus, Mar. Pollut. Bull., 2022, 176(113466), 113466,  DOI:10.1016/j.marpolbul.2022.113466.
  11. K. Buse and K. Bayliss, England's privatised water: profits over people and planet, BMJ, 2022, 378, o2076,  DOI:10.1136/bmj.o2076.
  12. Environment Agency, Event Duration Monitoring - Storm Overflows - Annual Returns [Internet], 2023, Available from: https://www.data.gov.uk/dataset/19f6064d-7356-466f-844e-d20ea10ae9fd/event-duration-monitoring-storm-overflows-annual-returns.
  13. F. Hassard, J. H. Sharp, H. Taft, L. LeVay, J. P. Harris and J. E. McDonald, et al., Critical review on the public health impact of Norovirus contamination in shellfish and the environment: A UK perspective, Food Environ. Virol., 2017, 9(2), 123–141,  DOI:10.1007/s12560-017-9279-3.
  14. A. O. Sojobi and T. Zayed, Impact of sewer overflow on public health: A comprehensive scientometric analysis and systematic review, Environ. Res., 2022, 203(111609), 111609,  DOI:10.1016/j.envres.2021.111609.
  15. A. T. Ford, A. C. Singer, P. Hammond and J. Woodward, Water industry strategies to manufacture doubt and deflect blame for sewage pollution in England, Nat. Water, 2025, 3(2), 231–243,  DOI:10.1038/s44221-024-00370-y.
  16. Office of Environmental Protection, Investigation into possible failures by Defra and the Environment Agency to comply with key laws to protect and improve water quality [Internet], 2025, Available from: https://www.theoep.org.uk/investigation/investigation-possible-failures-defra-and-environment-agency-comply-key-laws-protect.
  17. Independant Water Commission, Independent Water Commission: review of the water sector [Internet], 2025, Available from: https://www.gov.uk/government/publications/independent-water-commission-review-of-the-water-sector.
  18. Environment Agency, Environment Agency investigation into sewage treatment works, 2022, Nov 17, Available from: https://www.gov.uk/government/collections/environment-agency-investigation-into-sewage-treatment-works.
  19. Judiciary of England and Wales. Environment Agency vs Southern Water Ltd [Internet], 2021, [cited 2025 Jun 12], Available from: https://www.judiciary.uk/wp-content/uploads/2022/07/Southern-Water-Sentencing-Remarks.pdf.
  20. CIWEM. Event Duration Monitoring Good Practice Guide Version 3.0 [Internet], 2021, Available from: https://www.ciwem.org/assets/uploads/CIWEM_UDG_EDM_Good_Practice_Guide_2021_final.pdf.
  21. T. Giakoumis and N. Voulvoulis, Combined sewer overflows: relating event duration monitoring data to wastewater systems’ capacity in England, Environ. Sci., 2023, 9(3), 707–722,  10.1039/d2ew00637e.
  22. S. Laville, Southern Water alters pollution alert tool to curb automatic red alerts [Internet]. The Guardian, 2022, Available from: https://www.theguardian.com/environment/2022/sep/14/southern-water-pollution-alert-tool-automatic-red-alerts.
  23. BBC, Southern Water's sewage discharge app changes provokes row [Internet]. BBC, 2022, Available from: https://www.bbc.co.uk/news/uk-england-hampshire-62924995.
  24. C. Poopipattana, M. Suzuki and H. Furumai, Impact of long-duration CSO events under different tidal change conditions on distribution of microbial indicators and PPCPs in Sumida river estuary of Tokyo Bay, Japan, Environ. Sci. Pollut. Res., 2021, 28(6), 7212–7225,  DOI:10.1007/s11356-020-11046-x.
  25. BBC, Water firms illegally spilled sewage on dry days - data suggests [Internet]. BBC, 2023, Available from: https://www.bbc.co.uk/news/science-environment-66670132.
  26. E. O'Donnell, Interactive map: sewage spills in protected areas [Internet]. Unearthed, 2023, [cited 2025 Jun 12], Available from: https://unearthed.greenpeace.org/2023/08/04/interactive-map-sewage-spills-pollution-protected-areas/.
  27. A. Vaughan, Water firms told to stop withholding secret data on spills [Internet], The Times Newspaper, 2024, Available from: https://www.thetimes.com/uk/environment/article/water-companies-told-to-stop-withholding-secret-data-clean-it-up-l6pbd2lrm Search PubMed.
This journal is © The Royal Society of Chemistry 2025
Click here to see how this site uses Cookies. View our privacy policy here.