On Sunday 26 April 2026, a record 59,830 people streamed through the streets of London for the TCS London Marathon.
Most of them had one goal in mind: to finish. One of them had another: Sabastian Sawe crossed the line in 1:59:30, a time that rewrote the London Marathon record books and marked the first sub-two hour marathon in a major race setting.
As the running world applauded Sabastian, a quieter question was being asked: what was air quality like along the route when history was being made?
Air Aware Labs, the London based startup and developer of AirTrack, provided route air quality data, communications and reporting support for the 2026 TCS London Marathon. This included pre-event participant education, hourly air quality forecast screens for elite athletes, a specific event day web app containing live information accessible by QR code, and a post-event monitoring report.
The data showed that air quality during a marathon is not one thing. It changes hour by hour which means that a two-hour elite runner and a six-hour recreational participant can experience very different pollution conditions.
Air quality changed through the race
Nitrogen dioxide (NO2), a combustion and traffic related pollutant, had fallen sharply by the time the event was underway. Overnight concentrations around the route were elevated, but as road closures progressively came into effect, route adjacent NO2 dropped to around 17-19 µg/m³ by 09:00 BST.
Across the event window, the route adjacent average was 18 µg/m³ – around a quarter lower than the preceding week’s daytime average, pointing to the clear air quality benefit of reduced traffic on marathon day. However, interestingly marathon day NO2 was actually higher than the previous Sunday (19th April), in large part due to changes in weather systems. The 19th April involved clean northerly air and a deep, well mixed atmosphere, which dispersed emissions efficiently. By event day (26 April) a slightly more stagnant easterly regime had set in under high pressure: winds were light and the mixing layer had shrunk, so the air held on to pollution rather than clearing it. The same low traffic Sunday therefore recorded higher NO₂ because the atmosphere was less able to disperse it.
Fine particulate matter, PM2.5, followed a different daily pattern. Concentrations were elevated at the start of the event, at around 18 µg/m³, before declining steadily through the morning to around 9-11 µg/m³ by midday. The event hour mean was 12 µg/m³, higher than the equivalent period in the preceding week and substantially higher than the low pollution 2023 race.
However, this was influenced by a regional rise in PM2.5 across southern England in late April. In other words, even with local traffic reduced, wider regional air pollution still shaped the conditions participants experienced, outside of the organisers control.
Ozone moved in the opposite direction. It rose from around 60 µg/m³ at 09:00 BST to more than 91 µg/m³ by 15:00 BST. Unlike NO2, ozone is not emitted directly from exhaust pipes. It forms in the atmosphere through sunlight driven chemistry, and on warm, bright days it can build during the afternoon. It can also irritate the lungs, making it particularly relevant for people exercising for long periods. For faster participants finishing before midday, ozone levels were moderate. For those still on the course into the afternoon, ozone became the main air quality concern.
The elite runner and the rest of us
One of the most striking findings came from comparing two participants at opposite ends of the finishing time spectrum. Elite runner Sabastian Sawe ran from 09:35 to 11:34 BST, completing the marathon in just under two hours. During that time, his elite cardiovascular system was ventilating at a very high rate. Patrick Hicks, a 70 year old recreational runner whose AirTrack data we were able to match to official split timing, was on course from 11:07 to 17:39 BST. He ran for more than six and a half hours at a lower, sustained breathing rate.
For PM2.5, Sawe ran through the elevated morning period whereas Patrick ran later in the day, when concentrations had fallen. The result was surprising: their estimated total inhaled PM2.5 doses were broadly comparable.
Sawe’s high ventilation rate in dirtier morning air gave an estimated inhaled PM2.5 dose of 221-289 µg. Patrick’s lower ventilation rate, sustained over a much longer period in cleaner afternoon air, gave an estimated dose of 253-368 µg. These are modelled dose estimates rather than direct biological measurements, but they show why timing, duration and breathing rate all matter.
The ozone story points the other way. By the time Patrick was completing the final third of the event, ozone had risen above 85 µg/m³. Sawe was long finished by then.
This matters because elite athletes are often assumed to carry the greatest air pollution burden at mass participation events simply because they breathe at such high rates. The data suggest a more nuanced picture. Slower, longer, duration participants – who are typically less studied – may accumulate comparable air pollution doses while also encountering higher afternoon ozone.
For older participants, and for those with asthma or other respiratory conditions, that combination deserves attention.
What this means for major events
The 2026 TCS London Marathon will be remembered for one number: 1:59:30. The air quality data offer a quieter companion to that figure: a precise account of the conditions in which that record fell, and a foundation for doing better. For event organisers, this kind of analysis can feed directly into participant communications, sustainability reporting, medical planning and future event design. The goal is not just to measure what participants breathe. It is to reduce it.
Air Aware Labs builds personal air quality exposure tools and event air quality intelligence products. Find out more at airawarelabs.com.
Leave a Reply