The same city can show a different AQI on different services

Two services can show a different colour, a different AQI and even a different PM2.5 concentration for the same city at the same apparent moment. At first glance one of them seems to have to be wrong, but the difference often begins much earlier, with the choice of monitoring station, averaging period or the rules used to convert a measurement into an index.

A comparison only becomes meaningful once it is clear that the services are actually comparing the same thing.

The same concentration can receive a different category

AirQuality.city uses U.S. EPA concentration breakpoints for PM2.5 and PM10, while Croatia's official portal applies the European Air Quality Index. These are not simply differently coloured versions of the same table.

The U.S. EPA uses a numerical AQI from 0 to 500 and categories from Good to Hazardous. The European index uses Good, Fair, Moderate, Poor, Very poor and Extremely poor, with different concentration thresholds. For PM2.5, for example, the American Good band ends at 9.0 µg/m³ while the European Good band ends at 5 µg/m³.^[1][2]

Two services can therefore display the same 8 µg/m³ PM2.5 value and still use a different colour and description. In that case the measurement is the same; the interpretation is different.

Services do not necessarily use the same monitoring station

A city name can create the misleading impression that there is one air-quality value for the entire urban area. In reality, a roadside station, an urban-background site and a residential monitor can record different concentrations at the same time.

One service may automatically select the station nearest the city centre, another the station closest to the user, a third an average across several locations, and a fourth the worst current reading. Without checking the source, the city name alone is not enough for a valid comparison.

Timing and averaging can change the result without any error

A concentration representing the last few minutes is not the same as a one-hour or 24-hour average. Croatia's official system uses 24-hour moving averages for the PM component of its displayed European index, while AirQuality.city currently calculates its PM index from the latest processed value available to the portal.^[3]

AirQuality.city can therefore react more quickly to a sudden local spike, but the result should not be treated as a direct copy of the official Croatian or US AQI. For current particulate pollution, the official U.S. AirNow system uses NowCast, which combines several previous hours and changes their weighting when concentrations are changing rapidly.^[4]

The pollutant set may be different as well

The overall AirQuality.city AQI is currently driven by PM2.5 and PM10. NO₂, O₃, SO₂ and CO can be displayed separately for official stations but do not enter that overall result.

The European index can include PM2.5, PM10, NO₂, O₃ and SO₂, with the overall category determined by the least favourable pollutant level available.^[2] An official station can therefore have relatively low particulate matter but elevated ozone, producing a worse European overall index than the PM-based result on AirQuality.city.

Operational data is not the same as final validated data

Near-real-time measurements are valuable because they are fast, but Croatia's official system explicitly notes that current values are not yet final and may later change during validation.^[3]

One service may ingest a new value immediately, another refresh later, and a third may wait for an additional processing step. During a rapidly changing episode, even a small timing difference can make the displays look unrelated.

A low-cost sensor and a reference instrument are not the same device

SensorBox and other low-cost optical sensors are useful for local trends, but their performance depends on technology, calibration and environmental conditions. The US EPA notes that high humidity, temperature and other factors can affect sensor readings and that low-cost devices should not automatically be treated as equivalent to regulatory monitors.^[5]

If one service uses a low-cost sensor on a balcony and another uses a reference station several kilometres away, a difference between the numbers does not by itself tell us which is "correct". They may describe a different place, instrument and time window.

Compare raw concentrations before comparing colours

The most useful sequence is straightforward: first identify the same station, then check the same time and concentration in µg/m³, and only after that compare AQI values and colours.

If the concentrations agree but the categories do not, the index methodology is almost certainly the reason. If the concentrations differ too, look for differences in station, timing, averaging, instrument or source.

A different result is not evidence of an error. Without the methodology, the number alone does not carry enough context to support that conclusion.