Delhi’s average AQI fell to 102 in September, its lowest September average in five years. Yet the city recorded slightly higher average concentrations of PM2.5 and PM10 than in September 2025. The apparent contradiction reveals a central problem in reading urban air-quality data: the headline index is not a direct average of the pollutants most closely associated with health risks.
The Central Pollution Control Board data analysed by The Indian Express shows that Delhi’s average PM10 concentration rose from 104 micrograms per cubic metre in September 2025 to 106 µg/m³ this year. Average PM2.5 increased from 35 µg/m³ to 39 µg/m³. At the same time, monthly averages of nitrogen dioxide and ozone were lower than they were in September 2025.
The Commission for Air Quality Management reported Delhi’s September average AQI at 102, compared with 105 in both 2025 and 2024, 108 in 2023 and 104 in 2022. Under the National Air Quality Index classification, an AQI of 102 falls within the ‘Moderately polluted’ category, which covers readings from 101 to 200.
That combination of numbers is not a statistical error. It reflects what the AQI is designed to do, and also what it cannot communicate on its own.
The AQI is a composite measure based on individual pollutants, including PM2.5, PM10, nitrogen dioxide, ozone, sulphur dioxide, carbon monoxide and ammonia. A sub-index is calculated for each pollutant. The overall daily AQI is then determined by the pollutant with the highest sub-index rather than by adding or averaging all pollutant concentrations.
This means that a rise in PM2.5 does not automatically produce a corresponding rise in the overall AQI. If PM2.5 increases but remains below the sub-index of another pollutant, the final AQI can remain unchanged. It can also decline if the pollutant that was previously driving the index falls significantly.
Sunil Dahiya, lead analyst and founder at Delhi-based think tank Envirocatalysts, explained that the same logic applies to PM10. If the sub-index for ozone or another pollutant remains higher than the PM2.5 sub-index, the overall AQI will be controlled by that pollutant. Conversely, if the dominant pollutant falls, the AQI can decline even while PM2.5 or PM10 rises.
For residents, this distinction matters because the AQI is often treated as a single shorthand for the city’s pollution burden. It is useful for communicating the worst-performing pollutant on a given day, but it does not describe every pollutant equally. A lower monthly AQI therefore cannot be read as proof that all harmful pollutants have declined.
The difference becomes more important when monthly and cumulative comparisons are placed alongside one another. CAQM said Delhi recorded 194 days with an average AQI below 200 between January 1 and September 28. That was the highest count for the corresponding period in the last five years. The figure stood at 187 days in 2025, 188 in 2024, 193 in 2023 and 146 in 2022.
The number of days in the Poor to Severe+ categories also fell to 79 during the period, compared with 86 each in 2025 and 2024, 80 in 2023 and 127 in 2022. These figures indicate an improvement in the distribution of daily AQI categories compared with several previous years, particularly 2022.
However, the cumulative PM2.5 average from January 1 to September 28 was 69 µg/m³, the same as in 2025. It was lower than 81 µg/m³ in 2024, 73 µg/m³ in 2023 and 80 µg/m³ in 2022. The data therefore supports a more limited conclusion than a broad declaration of clean-air improvement: several indicators have improved over earlier years, but the most closely watched fine-particle measure has not improved compared with 2025.
CAQM told The Indian Express that its evaluation uses cumulative trends from January through the month under review. The commission considers monthly averages as well as the number of days in different AQI categories. This approach gives greater weight to the duration and distribution of air-quality conditions than a single monthly pollutant reading would.
That institutional method also explains why different summaries of the same season can appear to point in different directions. One measure may track the number of days below a threshold, another may track the pollutant with the highest sub-index, and a third may report the average concentration of a specific particulate matter fraction. Each answers a different question.
PM2.5 and PM10 are not interchangeable either. The first refers to finer particles, while PM10 covers a larger particulate fraction. The supplied data identifies both as key pollutants associated with harmful air pollution, but their concentrations can respond differently to local conditions. A city’s air-quality assessment therefore depends not only on the index selected but also on which pollutant is being examined.
September adds another layer of uncertainty to short-term comparisons because it is a transition period as Delhi moves towards the post-monsoon phase. Pollution concentrations can change sharply with weather conditions, including wind speed, rainfall and humidity. A monthly average captures the conditions that occurred during that period, but it does not by itself establish how the same trend will continue in subsequent months.
This is important for governance because the index used to communicate air quality also influences how authorities and the public interpret the urgency of pollution-control measures. A lower average AQI may suggest fewer severe days, while rising PM2.5 may indicate that exposure to fine particles remains a concern. Both can be true at the same time.
The data also shows why the city’s air cannot be assessed through one number alone. The September AQI comparison presents a positive trend against recent September averages, and the January-to-September count of days below an AQI of 200 is the strongest in five years. But the PM2.5 comparison with 2025 is flat on a cumulative basis and higher for September, while PM10 also rose during the month.
The central urban question is therefore not whether Delhi’s air improved or worsened in a simple, citywide sense. It is which measure improved, over what period, and what pollutant was driving the reported change. Without that distinction, a composite index can obscure deterioration in a specific pollutant even when the overall number moves in the opposite direction.
The evidence supplied by CAQM and CPCB confirms that Delhi’s September AQI was lower than in the previous five Septembers listed, while PM2.5 and PM10 were higher than in September 2025. The figures are not contradictory because the AQI is controlled by the highest pollutant sub-index rather than by an average of pollutant concentrations. What remains essential in tracking the coming months is whether the improvement in AQI categories is accompanied by sustained reductions in particulate matter as post-monsoon conditions develop.