A new satellite assessment by the Center for Energy and Clean Air Research (CREA), an international independent research organization, reports widespread PM2.5 pollution across India’s districts, states and airfields and recommends expanding clean air planning beyond urban boundaries.
The report, titled “Beyond City Limits: Satellite Assessment of PM2.5 Across Indian Airspaces, States and Districts,” by Manoj Kumar, Monish Raj, Panda Rushwood and Rosa Gierens, uses population-weighted estimates of PM2.5 using satellite data and ground-based measurements to assess annual and seasonal exposure.
PM2.5 refers to fine particles with a diameter of 2.5 micrometers or less, a key indicator of air pollution.
CREA’s district-level analysis found that 60% of Indian districts (447 out of 749) exceed the National Annual Air Quality Standard (NAAQS) for PM2.5 of 40 micrograms per cubic meter (µg/m³) and none meet the World Health Organization (WHO) guideline of 5 µg/m³. This unit, µg/m³, refers to micrograms of particulate matter per cubic meter of air, a standard measure of air pollution concentration.
The data show wide variations in exposure, with annual averages ranging from 21 µg/m³ in Tiruppur, Tamil Nadu to 112 µg/m³ in the north-west districts of Delhi. Several states and Union Territories are seeing widespread non-compliance, including Delhi, Assam, Punjab, Haryana, Chandigarh, Himachal Pradesh, Meghalaya, Tripura and Jammu and Kashmir.
The top 50 most polluted districts are concentrated in Delhi (11) and Assam (11), followed by Haryana (seven) and Bihar (seven). Other contributors include Uttar Pradesh (four), Tripura (three), Rajasthan (two), West Bengal (two) and one district each from Chandigarh, Meghalaya and Nagaland.
At the state level, 28 of the 33 assessed states and Union territories have at least one district exceeding the NAAQS and none meeting the WHO guideline.
Delhi ranks among the most polluted, recording an annual population-weighted PM2.5 of 101 µg/m³, which is 2.5 times the NAAQS and 20 times the WHO guideline.
States and Union Territories located in the Indo-Gangetic Plain, including Chandigarh, Haryana, Punjab, Uttar Pradesh, Himachal Pradesh, Uttarakhand and Jammu and Kashmir, show national averages above the NAAQS. The eastern and northeastern states – Bihar, West Bengal, Jharkhand, Odisha, Assam, Meghalaya, Nagaland, Tripura, Arunachal Pradesh – also exceed the national standard.
Apart from Sikkim in the northeast, South India records the lowest averages with Puducherry at 25 µg/m³, followed by Tamil Nadu, Karnataka, Telangana, Kerala and Andhra Pradesh, all below the NAAQS but still above the WHO guideline.
Seasonal analysis shows significant differences throughout the year. In winter (December through February), 616 of 749 counties (82%) exceed the NAAQS, which is associated with increased emissions and stagnant weather. During the summer (March to May), the number of exceedances decreases to 405 districts (54%). The biggest improvement comes in the monsoon (June to September), when 74 districts (10%) exceed the norm, mainly due to rainfall induced by atmospheric cleansing. Post-monsoon (October to November) there is a recovery, with 566 districts (75%) crossing the limit.
At the state level, 22 of 33 states report 100% of counties above the NAAQS in winter, and nine states report 100% exceedances in summer.
Assam (21 districts), Delhi (nine), Punjab (15) and Tripura (six) exceed the standards even during the monsoon, when much of India experiences cleaner air.
The study analyzes air ducts, defined as areas where air moves and mixes in regular patterns influenced by geography and weather. Airspace boundaries change seasonally with changes in wind and temperature. The Indo-Gangetic airship remains consistently unsatisfactory in winter, summer and post-monsoon.
The northeast airship emerges as a year-round problem, with Assam and Tripura maintaining elevated PM2.5 even during the monsoon. According to the report, most of the airwaves were below the NAAQS during the monsoon, except for the Assam-Tripura airwaves. The rapid return to high PM2.5 levels after the monsoon is presented as evidence that underlying emissions, not meteorology, are the underlying problem.
“Without airshed-based governance frameworks, integration of satellite monitoring into NCAP (National Clean Air Programme), sectoral emission targets and accountability mechanisms, millions in non-metropolitan India will remain excluded from clean air policies and locked into chronic pollution,” the report said.
CREA also says it is building near-real-time tools for public use. “To make air quality information more accessible and usable, CREA is developing daily PM2.5 concentration maps for India. These maps will soon be available for public use, helping policymakers, researchers and citizens track pollution patterns and supporting data-driven air quality management,” Kumar added.
The report highlights potential monitoring bias where station locations may underrepresent industrial corridors or high-traffic microenvironments in sparsely networked states. Dense ground monitoring in Delhi is cited as being in good agreement with satellite-derived PM2.5, indicating a better capture of local spatial variation compared to most states with fewer stations. The study suggests that expansion and better placement of ground-based monitors can improve accuracy and representativeness nationwide, while satellite observations provide consistent coverage to fill spatial gaps.
The PM2.5 data was generated using a machine learning model based on Kawano et al (2025). Daily estimates at 10 km resolution were produced for all of India using satellite observations of aerosols, NO₂ (nitrogen dioxide) and CO (carbon monoxide) combined with weather data. The model covered the period from September 2018 to March 2025.
CREA presented policy recommendations aimed at aligning air quality management with regional and sectoral realities. Recommendations include developing district-level Clean Air Action Plans beyond the current list of non-attainment cities; formal integration of satellite-derived PM2.5 data into NCAP for comprehensive assessment, validation and public dissemination; targeting regional and sectoral sources of emissions, including power generation, industrial activity, biomass burning and transport, particularly in the Indo-Gangetic Plain and North-Eastern States; incentivizing state-level accountability by linking clean air performance to fiscal mechanisms using standardized indicators from both satellite and ground-based data; and the adoption of airshed-based governance frameworks to coordinate emissions reductions and joint monitoring and accountability across neighboring jurisdictions.
The report identifies opportunities for differentiated targeting. Southern states with annual PM2.5 levels below the NAAQS can move toward WHO’s interim targets by adopting more stringent region-specific limits, allowing for incremental improvements while recognizing differences in baseline emissions between states.
In Maharashtra, CREA points to differences between the overall state average near the NAAQS threshold and district-level hotspots. Fourteen of the 36 districts exceed the annual PM2.5 standard in Maharashtra. Reported annual averages (µg/m³) include Ratnagiri (47.77), Raigarh (47.03), Gadchiroli (46.81), Dhule (42.83), Chandrapur (42.43), Jalgaon (42.38), Thane (41.63), Sindhudurg (40.40), Agarak (40.40), and Agra (40.40). (formerly Ahmadnagar, 40.34), Aurangabad (40.41), Nandurbar (41.16), Palghar (41.13) and Pune (40.03). Seasonal data show Mumbai averages 37.07 µg/m³ annually, but 63.17 µg/m³ in winter; Raigarh at 47.03 µg/m³ annually, with 66.08 µg/m³ in winter and 60.17 µg/m³ in monsoon; Ratnagiri at 47.77 µg/m³ annually, with 68.86 µg/m³ in winter and 58.85 µg/m³ in monsoon. The report describes Chandrapur as hosting a dense cluster of coal-fired industries and power plants with consistent exceedances of the annual standard.
Due to limitations in ground monitoring during the study period, Ladakh, Andaman and Nicobar Islands and Lakshadweep were excluded from the assessment. The remaining 33 states and Union territories covering 749 districts were included. Seasonal air waves ranging between nine and 11 in number were analyzed to explore regional differences, Mr Kumar said.
