Defra's national air-quality monitors, every hour from 2009 to 2025 · United Kingdom
Bonfire Night: the night UK air monitors most often record their dirtiest hour of the year
Every national fine-particle (PM2.5) monitor in the UK has one worst hour a year. It could fall on any day. Again and again it falls on the night of 5 November: 282 times out of 1,206 since 2009, against 27 for the next-worst night and 2.2 for an average winter night. At Liverpool Speke it did in 12 of 14 years.
What happens to the air on Bonfire Night, hour by hour?
Line up every monitor's readings for 5 November and the hours either side. Until early evening, Bonfire Night air is ordinary November air. Then, from about five in the evening, fine particles climb for three hours. At 20:00 the median reading, across all monitors and years, is 34 micrograms per cubic metre, against 9 at the same hour on other November nights: 3.7 times as much.
Fine particles are called PM2.5: specks smaller than 2.5 thousandths of a millimetre, small enough to pass deep into the lungs. On this night they come from bonfires and from the fireworks themselves, whose colours come from metal salts. They linger: at dawn on 6 November the air still holds about twice its usual load.
5–6 NovemberOther November nights
Where did the smoke go each Bonfire Night?
Each light is a monitor; its glow and its plume show the reading that hour, and the smoke leans the way that night's modelled wind was blowing (there is no wind for 2009 and 2022, so there it drifts at random). Pick a year, then play the night from five in the afternoon to six the next morning. Tap a monitor for its story. Only the lights are measurements; the smoke between them is drawn to show the wind, not measured.
Fine particles in that hour, micrograms per cubic metre.
Which monitors have their worst hour on Bonfire Night most often?
Counting only monitors with at least 8 complete years. A monitor near a big public display or among houses with back-garden fireworks catches the smoke; one in a windy place or by a road that is dirty all year often does not.
Is it just cold, still weather?
Partly. Smoke stays where it is made on a calm, cold night, and November has many such nights. The weather helps explain why some Bonfire Nights are far worse than others: in 2014, with a median wind of 2.9 metres a second, the typical monitor's highest hour reached 124 µg/m³; in 2013, with 6.9 metres a second, it reached 28 µg/m³.
So compare like with like. For each monitor and each winter, take the other November to February nights with the same wind (within 1 metre a second) and temperature (within 3 degrees), and set Bonfire Night against them. Across 922 cases (one monitor in one winter), Bonfire Night air was typically 2.8 times as dirty as on the matched nights. The factor changes from year to year, from 0.9 in 2015 to 4.3 in 2011: in 9 of 15 years it was at least double.
The wind and temperature come from a weather model (WRF, run by WSP for Defra) calculated for each monitor, not from instruments at the site, and start in 2010.
How bad does it get on the government's own air quality index?
For past days, Defra's Daily Air Quality Index rates fine particles by their daily mean, from 1 (low) to 10 (very high). Smoke after dark lifts the average for 5 November, and for 6 November when it lingers past midnight. In 2014, 7 of 64 monitors reached band 10, the highest; in 2023, none reached even 'moderate'.
| Year | Monitors | Moderate or worse | High or worse | Very high |
|---|---|---|---|---|
| 2009 | 57 | 12 | 2 | 1 |
| 2010 | 55 | 20 | 4 | 1 |
| 2011 | 64 | 23 | 9 | 3 |
| 2012 | 61 | 20 | 6 | 2 |
| 2013 | 62 | 1 | 0 | 0 |
| 2014 | 64 | 24 | 15 | 7 |
| 2015 | 65 | 3 | 0 | 0 |
| 2016 | 65 | 2 | 1 | 0 |
| 2017 | 69 | 11 | 5 | 2 |
| 2018 | 77 | 30 | 6 | 3 |
| 2019 | 77 | 3 | 0 | 0 |
| 2020 | 77 | 36 | 11 | 1 |
| 2021 | 76 | 2 | 1 | 0 |
| 2022 | 95 | 3 | 0 | 0 |
| 2023 | 99 | 0 | 0 | 0 |
| 2024 | 111 | 16 | 2 | 0 |
| 2025 | 153 | 3 | 0 | 0 |
Worst of the two daily means, 5 and 6 November, at each monitor with at least 18 hourly readings that day. Bands: moderate from 36, high from 54, very high from 71 µg/m³. Levels before and after about 2019 are not directly comparable: the instruments changed.
Is Bonfire Night getting cleaner?
The readings on ordinary winter nights are lower than they were. But between about 2018 and 2020 the network replaced its fine-particle instruments, and Defra found the new ones read lower than the old, which had over-read. So these numbers cannot say how much cleaner the air itself became.
Comparing each monitor with its own other nights in the same winter should cancel much of the instrument change, though not necessarily all: Defra notes that one of the new instruments, the Fidas, can under-read air heavy with soot. On that comparison, Bonfire Night's multiple shows no steady change across the years (rank correlation with the year: 0.00); the weather moves it far more.
Why 5 November?
On 5 November 1605 a plot to blow up the House of Lords during the state opening of Parliament was discovered, and Guy Fawkes was found guarding the gunpowder. An Act of Parliament the following year made the day one of thanksgiving; it stayed in force until 1859. Bonfires, and later fireworks, have marked it ever since.
FAQ
Is Bonfire Night the most polluted night of the year in the UK?
At the national monitors it is the night that most often holds a monitor's single worst hour of the year: 23% of 1,206 monitor-years since 2009, against 2.2% for the next night, 6 November.
How much worse is the air on Bonfire Night?
At 20:00 the typical monitor reads 34 µg/m³ of fine particles, 3.7 times the same hour on other November nights. Against nights with the same wind and temperature, the whole night is typically 2.8 times as dirty.
Where is Bonfire Night smoke worst?
It depends on the year and the wind. The monitors where it most often brings the worst hour of the year include Liverpool Speke (12 of 14), Leeds Centre (12 of 16) and Glasgow Townhead (8 of 12).
Does New Year's Eve compare?
Not at these monitors: New Year's Eve held the worst hour in 10 of 1,206 monitor-years, against 282 for Bonfire Night.
How this was measured
Hourly PM2.5 from the Automatic Urban and Rural Network (AURN), the UK's national monitoring network run for Defra and the devolved governments, 2009–2025, ratified data as published in the openair files (one monitor-year, Canterbury 2024, is ratified only to 4 November), where each time stamp marks the start of the hour. A monitor-year (one monitor in one year) counts when at least 75% of its hours have a reading: 1,206 monitor-years at 165 monitors. Times are GMT all year, the clock time in November; summer worst hours on the monitor cards are an hour behind British Summer Time.
Bonfire Night is the 13 hours from 17:00 on 5 November to 06:00 on 6 November. A worst hour between 06:00 and 17:00 is a daytime worst hour and is not assigned to a night. If two hours tie for a monitor's worst, the earlier one counts. For the weather comparison, each night is the mean from 18:00 to 06:00.
The tests were written down before they were run, with their pass marks. Two claims failed them and are worded here as the data allows: that the weather match leaves Bonfire Night at least twice as dirty in nearly every year, and that ordinary winter-night readings fell by at least 40% (they fell 39% between 2010–12 and 2023–25, a span that includes the change of instruments).
- Hourly AURN data and site list (openair files), Defra uk-air.defra.gov.uk
- Daily Air Quality Index bands, Defra uk-air.defra.gov.uk
- The openair R package (data import and modelled weather) openair-project.github.io
- PM2.5 instrument equivalence: evidence synthesis report (2024), Defra uk-air.defra.gov.uk
- Air quality on Bonfire Night 2023, Ricardo for Defra ricardo.com
- The Gunpowder Plot and the Thanksgiving Act, UK Parliament parliament.uk