Start with the conclusion, because it reverses the usual framing: in almost every city on earth, exercising outdoors is better for you than not exercising, and it is not close.
A comparative risk assessment by Tainio and colleagues modelled the trade-off across background PM2.5 concentrations from 5 to 200 µg/m³. At the global average urban background of 22 µg/m³, you would need to cycle seven hours a day, every day, before the pollution harm caught up with the physical activity benefit. Walking, sixteen hours. Even at 100 µg/m³, a level that would look like a bad smog day in most of the world, harms only exceeded benefits after ninety minutes of daily cycling. And if the alternative to cycling is driving rather than sitting at home, the benefit holds out to three and a half hours.
So this article is about optimisation, not risk avoidance. You are going to train. The question is whether you can shift the hour and pick up a cleaner lungful for free.
Why the hour matters at all
Dose is concentration multiplied by how much air you move multiplied by how long you move it. Exercise attacks the middle term hard.
At rest you breathe somewhere around 6 to 8 litres of air per minute. In a study of interval training, minute ventilation averaged 77 L/min during moderate intervals and 107 L/min during high-intensity intervals. That is roughly a tenfold to fifteenfold increase in the volume of air passing through your lungs.
Two things compound it. Somewhere around 35 L/min, most people switch from breathing through the nose to breathing through the mouth. Your nose is a surprisingly good filter and humidifier, and above that threshold you stop using it. Deeper, faster breathing also drives particles and reactive gases further into the small airways where clearance is slowest.
The practical upshot: an hour of hard running in moderate air can deliver more inhaled pollutant than a whole day of sitting in the same air. Which hour you choose is worth something.
Two clocks, running opposite directions
Here is where the standard advice goes wrong. Air pollution does not have one daily curve. It has at least two, and they peak at opposite ends of the day.
Particles peak morning and night. An analysis of more than seventeen million hourly measurements across 3,110 monitoring sites worldwide found a consistent shape: PM2.5 peaks between 7:00 and 10:00 local solar time, peaks again between 21:00 and 23:00, and reaches its daily minimum between 15:00 and 17:00. The driver is atmospheric mixing rather than emissions alone. Overnight and early morning, a cool shallow layer of air traps everything close to the ground. As the sun warms the surface, that layer lifts and the same pollution disperses through a much taller column.
Ozone peaks mid-to-late afternoon. Ground-level ozone is not emitted by anything. It is manufactured in the air when sunlight drives reactions between nitrogen oxides and volatile organic compounds. That takes hours, so ozone bottoms out in the morning, when nitric oxide from traffic actively destroys it, and builds to a maximum typically between 1 PM and 5 PM. Ozone is a strong airway irritant, and the effect is worse during exercise specifically because heavy breathing delivers more of it deeper into the lung.
Put those together and mid-afternoon is simultaneously your cleanest hour for particles and your worst hour for ozone. There is no time of day that is clean on both counts every day of the year.
So the early-morning run is safest?
Not automatically, and this is the claim most fitness articles repeat without checking.
Early morning is genuinely the low point for ozone. On a hot, sunny, still summer day, a 6 AM run avoids the photochemical peak entirely, and that is a real win.
But on a cold, still winter morning, the same hour can be the worst of the day for particles. A nocturnal temperature inversion traps overnight emissions in a thin layer at street level, and in regions where households heat with wood or coal, that layer is dense with combustion particles by dawn. The EEA’s 2026 air quality assessment attributes PM2.5 exceedances in Türkiye and several Western Balkan countries mainly to solid fuel burning for domestic heating, combined with older vehicle fleets. If you live somewhere like that, your winter dawn run may be the single dirtiest hour you breathe all day.
The honest rule is seasonal:
- Hot, sunny, still, summer. Ozone limits you. Train early morning or after sunset.
- Cold, still, winter, especially with wood or coal heating nearby. Particles limit you. Train late morning through afternoon, once the inversion has broken up.
- Windy or rainy, any season. Both clocks flatten. Almost any hour works.
- Wildfire smoke. Neither clock applies. Smoke plumes are driven by fire behaviour and wind, so check the actual hourly number rather than the time.
Where you run may matter more than when
Traffic pollutants fall off steeply with distance from the road. Ultrafine particles, black carbon, nitrogen dioxide and carbon monoxide typically decay toward background within roughly 100 to 200 metres of a busy road, with most of the drop happening in the first hundred.
There is a catch worth knowing. PM2.5 mass barely shows a near-road gradient at all. Several near-road studies found ultrafine particles and black carbon dropping sharply across the same distance where PM2.5 was elevated only modestly, on the order of twenty percent, because fine particle mass is regionally mixed rather than locally generated.
That means moving your route two streets away from the arterial makes a large difference to the traffic-specific pollutants and a modest one to the number your air quality app reports. It is still worth doing. It just will not show up in the figure you are checking.
Read the concentration, not the colour
Air quality indices are national conventions, not measurements. A US AQI of 100 and an Indian AQI of 100 describe very different air. The European index uses six bands anchored to WHO guideline levels, the US index uses six anchored to its own standards, and the numbers do not translate. The categories roughly translate, the numbers do not.
For anything involving travel or comparison, read raw PM2.5 in µg/m³. Useful anchors: the WHO 24-hour guideline is 15 µg/m³, and the WHO annual guideline is 5 µg/m³. Over nine in ten Europeans live above the annual guideline, so the guideline is a target rather than a threshold for cancelling your run.
Rough intensity guidance for a healthy adult, using PM2.5 concentration rather than index colour:
- Under about 15 µg/m³. Train normally.
- 15 to 35 µg/m³. Train normally. If you have asthma or a cardiovascular condition, prefer a park route over a roadside one and note how you feel afterwards.
- 35 to 55 µg/m³. Keep the session, reduce the intensity or the duration. Intensity is the more effective lever, because ventilation scales faster than time.
- Above 55 µg/m³. Move it indoors if you reasonably can. If you cannot, go easy and short. Sensitive groups should move indoors well before this point.
On ozone, the trigger is different: a hot, sunny, windless afternoon with a visible haze is the pattern to avoid, particularly if you get chest tightness or a dry cough after summer afternoon sessions.
Using SharedSky for this
Generic advice can tell you which clock probably governs your city. It cannot tell you which one governs you. Some people are noticeably ozone-sensitive and fine in particle haze. Others are the reverse. That difference only shows up in your own record.
SharedSky shows the air, pollen, pressure and heat where you are, and lets you log how you feel in one tap.
A workable routine:
- Check before you head out, not the daily average. The daily average smooths away exactly the peaks you are trying to dodge. Look at the current conditions and pick your window.
- Log within an hour of finishing. Post-workout is when airway irritation shows itself: cough, throat scratch, chest tightness, a headache, unusual fatigue. Log the sessions that felt fine too, using the “I’m fine now” button. A record of only bad sessions cannot show a pattern, because it has nothing to compare against.
- Note the context. The optional profile fields cover outdoor activity level, time spent outdoors, location type and any mitigation you took. Those are what let you separate “the air was bad” from “I went too hard”.
- Give it a few weeks across different weather. You need hot still days and windy days in the same record before the comparison means anything.
- Look for which pollutant tracks your symptoms. If your worst sessions cluster on hot sunny afternoons, ozone is your limiter and you should move to mornings. If they cluster on cold still mornings, particles are your limiter and you should move later.
One limitation worth stating plainly. SharedSky’s air data comes from the Copernicus Atmosphere Monitoring Service via Open-Meteo, which is a modelled regional product. It gives you your city’s curve accurately and it will not know that your usual route runs along a four-lane road. That gap is exactly what your own logs fill.
If you have asthma or a heart condition
Everything above describes optimisation for healthy adults. If you have asthma, COPD, angina or another cardiopulmonary condition, the thresholds are lower and the stakes are different. Exercise remains strongly beneficial for you, and the specifics of when, where and how hard belong in a conversation with your doctor rather than a blog post. A few weeks of logged sessions with the conditions attached makes that a much more productive conversation.
Want to optimize your outdoor training schedule? Check real-time hourly pollution cycles in SharedSky before heading out for your next workout.