The air quality index was invented for one useful property: to collapse a dozen different pollutants into a single number and a single colour. For a morning check on your phone, that is perfect. But every compression has a price — and for AQI the price is that the same number describes two completely different situations: a city with a flat, moderate background, and a city where clean days alternate with toxic ones.
PM2.5: the usual culprit
Of all the components of the index, the one that most often drives the result is fine particulate matter, PM2.5 — the particles small enough to reach deep into the lungs. Their concentration is not constant: in winter, when heating comes on and air stagnates under an inversion, it rises sharply.
Monthly PM2.5 series. The dashed line marks the WHO annual guideline of 15 µg/m³.
Notice two things at once. First, the pronounced seasonality: winter months are several times dirtier than summer ones. Second, the slow year-over-year decline in the mean — the result of switching to cleaner fuels and moving dirty industry out. The annual average may look acceptable, but that average hides the winter peaks, the times when breathing is genuinely harmful.
Not the mean, but the distribution of days
It is far more honest to show not the average AQI but what a year is made of — how many days it spent in each category.
Each year's bar is split into days by category, from Good (green) to Severe (dark red).
Here real progress is visible: the green zone widens while the red one shrinks. It is this chart, not a running average, that answers the question a resident cares about: “How many days this year could I freely take the kids for a walk?” The average AQI does not answer it — two cities with the same average can have wildly different numbers of hazardous days.
Air that breathes with the traffic
Nitrogen dioxide, NO₂, has an almost direct link to transport. Its daily curve traces the road-traffic profile: two peaks, morning and evening.
Average NO₂ by hour, normalised to the peak. The two peaks coincide with traffic hours.
The match with traffic is no coincidence — it is the proof of source. When a pollutant’s daily curve traces the traffic curve, there is no point arguing about where it comes from: it is exhaust. And so the remedies are obvious too — they lie in the domain of transport, not of abstract “environment”.
What a reader should check
When you are shown an “annual average index”, ask two questions. How many days this year fell into the hazardous categories — since those are what create the harm? And which pollutant dominates — if it is NO₂ with two daily peaks, the conversation is about transport; if it is winter PM2.5, it is about heating and fuel. A single averaged number answers neither.
