Use a standardized framework to track air pollution

is founder and chief executive officer of Respirer Living Sciences.

Aglobal air pollution report, World Air Quality Report 2023: Region and City PM2.5 Ranking, featured 83 Indian cities in the top 100 most polluted cities of the world. While the reality of air pollution in India is undeniable, the ranking of cities requires an unbiased measurement framework. The data used in this report is a combination of readings from regulatory air-quality monitors and low-cost sensors. While this approach has merit, the weak link of the study is the ‘spatial and temporal’ extent of the monitoring. For example, the yearly average of Delhi, reported at 102 microgram per cubic metre (ug/m3), is based on readings from 40 locations across the city with a cumulative uptime of 93%, whereas that of Siwan in Bihar, ranked No. 7 with 90ug/m3, is based on a single location, and Gurugram’s 17th rank reading of 84ug/m3 is drawn from four locations. The number of monitoring locations and their uptime that go into calculating the city’s average (and ranking) need standardization.

For city-level air pollution rankings, there are three critical parameters: one, the spatial coverage of the measurements; two, the temporal coverage (‘uptime’ of monitors or the number of hourly data points reported every month); and three, what pollutants are measured (PM2.5, PM10, NO2, CO, SO2, etc). Ranking cities by PM2.5 is sufficient, but a more nuanced ranking would factor in levels of NO2, O3 and CO as well.

Scrutiny of the report’s 83 cities reveals that 15 of these had no public or government-backed air-quality data, so it is likely that data from privately run sensors was used to track their annual levels. The government uses Continuous Ambient Air Quality Monitoring Stations (CAAQMSs), whose readings are reported on the Central Pollution Control Board (CPCB) portal. Alternate technology is available and community-set-up sensors have scientificallyvalid readings too. The report’s data for 52 other cities is sourced from only one government-owned CAAQMS. Relying on a single monitor to check an entire city’s air quality has many challenges, the biggest being that the reading may not be sufficiently representative and could give us a biased impression. About 10 locations included in the list (such as Morar and Banposh) do not qualify as cities on the list of monitored cities on the CPCB portal. Since only 16 cities on this list have more than one monitor, should the others be part of such global rankings?

A standardized framework for reporting city-level pollution levels is essential not just for ranking reliability, but also from a policy perspective, as the National Clean Air Programme (NCAP) uses such data for decisions on where to disburse funds.

Launched on 10 January 2019, the NCAP identified 102 ‘non-attainment’ cities at the onset. By 2024, it had identified 131. Its entire ₹9.934.4 crore budget is allocated by data from manual air quality monitors. These offer a “minimum 104 measurements in a year, at a particular site, taken twice a week 24 hourly at uniform intervals.” Alongside these, NCAP cities run CAAQMSs, which cost around $20,000 to monitor each pollutant, thus cumulatively costing as much as $200,000 for all notified air pollutants.

While the lack of a standardized framework doesn’t necessarily mean single-monitor cities should not be included in rankings, it’s important to move towards a peer-reviewed approach that does not have spatial and temporal coverage gaps.

Any ranking needs to meet globally-accepted ‘FAIR’ data principles. In other words, it should have the attributes of: Findability, so that people at large can find the underlying data used to compute these rankings; Affordability, which means the technology used to generate the data should be easily affordable by cities that want to track their monthly air quality levels and join these rankings; Interoperability, so that the data is in a format that allows for use with other data systems and not provided in sealed documents (as is the case of manual monitoring data in India); and Reproducibility, which means that the findings should be entirely reproducible by another independent agency that

A robust system that offers us reliable data is crucial not just to be sure of how cities are faring against pollutants, but also to guide allocations under the National Clean Air Programme. undertakes the same exercise with equivalent technology. For national-level rankings to hold credibility, the data used should adhere to these FAIR data principles. Only then would the rankings form a valuable record in the context of understanding the current situation and checking which cities are becoming worse and which are getting better over time. Using FAIR data principles will also allow us to compare city-level air quality data independent of the size or population density of a city, and eventually help us scale these rankings to cover the country’s 7,000-plus census cities and towns.

To effectively compute city-level air quality rankings, the authorities in India need to release guidelines on what constitutes adequate spatial and temporal coverage of a city from an air-quality representation perspective. The three prevalent mechanisms in use—manual air samplers, CAAQMs and sensor-based monitors—all have their respective strengths and limitations. A prudent city administration would look at adopting a judicious mix of all three systems. This would allow it to create a comprehensive approach based on an affordable, accessible and irrefutable methodology to track air-quality data. It’s time we attained such clarity on the air around us.