Windsor sewage treatment works, run by Thames Water, has 1 monitored storm overflow discharging to Thames. In 2025 it recorded 34 spills totalling 333 hours, against 133 spills and 1,735 hours in 2024. By hours spilling, it ranks 95 of 233 Thames Water works, where 1 spilled longest.
Works size: 39,096 population equivalent (WINDSOR, HAM ISLAND, OLD WINDSO STW", UWWTD 2018). Used for the spill volume estimate.
Every spill in 2025
Each blue mark is one spill, placed by when it started and sized by how long it lasted. Hover over a mark for the date and duration.
Estimated spill volume by month, 2025
Hours spilling each month × the estimated spill rate for a works of this size. The bars are central estimates and the whiskers show the low–high range. Method.
Cubic metres (m³), estimate. Hours are measured by the event duration monitor.
Longest single spills
Started (GMT)
Overflow
Duration
Estimated volume
26 Jan 2025, 14:08
Windsor
3.5 days
22,672 – 142,895 m³
5 Jan 2025, 02:46
Windsor
2.4 days
15,467 – 97,487 m³
31 Jan 2025, 06:52
Windsor
21.7 h
5,799 – 36,551 m³
1 Feb 2025, 08:44
Windsor
18.0 h
4,803 – 30,272 m³
30 Jan 2025, 13:36
Windsor
14.7 h
3,914 – 24,666 m³
Estimated 2025 spill volume and pollution
Estimate, not a measurement. Monitors record how long an overflow spills, not how much. This calculator back-calculates a spill rate from the size of the works, using the Environment Agency's permit formulae, and multiplies it by published storm-sewage concentrations. How it works.
Volume, central estimate223,341 m³range 88,961 – 560,705 m³
In litres223.3 millionrange 89.0 million – 560.7 million L
Olympic swimming pools89at 2,500 m³ each
Organic load, as people's raw sewage335,011person-days of untreated BOD
Assumed spill rate while spilling: 74.1 – 467.2 L/s (central 186.1 L/s), from a dry-weather flow of 74.1 L/s.
1.9 / 6.3 / 9.3 mg/L Low: Germany, Brombach 2005; typical/high: Paris range 3.3–9.3 midpoint and top, Gasperi 2012 (via botturi); no UK value found
Total phosphorus
107 kg
581 kg
3,028 kg
1.2 / 2.6 / 5.4 mg/L Low/high: Paris, Gasperi 2012; typical: Slovakia (via botturi). UK figure of 10 mg/L treated as an outlier
E. coli
8.9 trillion organisms
223.3 trillion organisms
5.6 × 10¹⁵ organisms
10,000 / 100,000 / 1.0 million per 100 mL Median range at CSO and retention-tank outlets, Stott et al. 2018 (via botturi); typical = log midpoint
Low combines the low spill rate with low concentrations, and high combines high with high, so the true figure is very likely inside the range. Long spills become more diluted as they go on, so the central figure tends to overstate the load of long events. The calculator does not estimate the effect on the receiving water, which depends on river flow or tide at the time of the spill.
Dry-weather flow (DWF) = population × water use per person × (1 + infiltration). Water use is 136.5 L per person per day, England's per-capita consumption for 2024–25 [EA]. Infiltration of groundwater into sewers is taken as 0–40% of that flow [Escritt, via HBF]. The Environment Agency defines DWF as PG + I + E [EA].
Population. The size of each works comes from the load entering it as reported under the Urban Waste Water Treatment Directive, in population equivalents (1 p.e. = 60 g of BOD a day) [UWWTD]. We treat population equivalent as population. That overstates flow where trade effluent adds load, and no official source links the two. Works serving fewer than 2,000 p.e. are not reported, so their size has to be entered by hand.
Low spill rate = DWF. This is the method used by Giakoumis & Voulvoulis (2026), which they note likely underestimates storm loads [G&V].
High spill rate = Formula A − flow to full treatment = (DWF + 1360P) − 3DWF. This is the largest flow the permit expects the storm tanks to handle before it passes on to treatment or overflows [EA]. The central rate is the geometric mean of the low and high rates.
Volume = spill rate × hours spilling, with hours taken from the event duration monitor.
Pollution = volume × published storm-sewage concentrations [Botturi et al.][US EPA]. Organic load is also shown as person-days of untreated sewage at 60 g of BOD per person per day [UWWTD].
What this cannot tell you. It gives no real flow for any single spill, and no effect on the river or sea, which depends on the flow, tide and the other pressures on the water at the time. It gives no figure for spills from overflows on the sewer network, which serve catchments of unknown size. It is a way to put hours into rough physical terms, not a substitute for flow monitoring.