India’s electricity demand nearly quadrupled between 2000 and 2020, met largely by coal, which supplied about 76% of power and drove a sharp rise in SO2 emissions, sulphate aerosols and urban PM2.5.
A Nature Cities study links this coal-heavy growth to higher aerosol loading and a 30% jump in PM2.5-related urban deaths, even as clean cooking cut household pollution.
India’s electricity demand nearly quadrupled between 2000 and 2020 as its GDP per capita more than tripled, but the expansion was met predominantly with coal, contributing to a sharp rise in urban air pollution and particulate matter-related deaths, according to a new study.
Published in Nature Cities, the study by Basudev Swain of the University of Oxford and colleagues found that coal supplied about 76 per cent of India’s electricity generation during the period. Urban aerosol loading more than doubled, while estimated urban deaths attributable to fine particulate matter (PM2.5) increased from about 0.67 million in 2000 to 0.87 million in 2020.
The researchers found that the expansion of coal-based electricity was associated with higher sulphur dioxide (SO2) emissions, increased sulphate aerosols and rising urban PM2.5 concentrations. At the same time, rising incomes and greater access to clean cooking reduced household air pollution and associated deaths, creating a shift in the pollution burden from indoor to outdoor environments.
The findings are relevant to India’s current energy trajectory even as the electricity mix begins to change. The International Energy Agency (IEA) said India’s coal-fired power generation declined by about 3 per cent in 2025, partly because of a strong monsoon that boosted hydropower and reduced cooling and agricultural electricity demand. However, the IEA expects India’s electricity demand to grow at an average 6.4 per cent a year between 2026 and 2030, with solar photovoltaic expected to meet about half of the additional demand and coal about a quarter.
India’s GDP per capita increased from about $600 to $700 in 2000 to more than $2,000 by 2020, while electricity demand rose from approximately 450 terawatt hours (TWh) to almost 2,000 TWh, the study said.
Access to electricity increased from about 60 per cent of the population to more than 95 per cent during the same period. However, coal continued to account for about 75 to 80 per cent of electricity generation. In 2020, coal accounted for 75.9 per cent, followed by hydropower at 7.7 per cent, wind at 6.6 per cent, solar at 4 per cent, nuclear energy at 2.8 per cent, natural gas at 2.7 per cent and oil at 0.2 per cent.
India had 71 coal-fired power plants before 2000. Another 182 were commissioned between 2001 and 2024, taking the total to 253 by 2024. The plants are concentrated particularly across the Indo-Gangetic Plain and the Horseshoe Belt, areas that also contain major urban agglomerations, industrial centres and economic corridors.
The study’s econometric analysis found a statistically significant positive relationship between GDP per capita and electricity demand. A 1 per cent increase in GDP per capita was associated with an approximately 1.14 per cent increase in electricity demand, with a 95 per cent confidence interval of 1.112 to 1.167 and P < 0.001.
The relationship between economic growth and electricity demand is likely to remain important. The IEA expects India to add more than 570 TWh to annual electricity consumption between 2026 and 2030. Industry is expected to contribute about one-third of this growth, while cooling is expected to account for more than 20 per cent as ownership and use of air conditioners increase.
The IEA also expects India’s variable renewable energy share, comprising solar photovoltaic and wind, to rise from 14 per cent of electricity generation in 2025 to 24 per cent in 2030. The total renewable energy share is projected to exceed one-third by 2030, compared with 24 per cent in 2025.
Urban aerosol optical depth (AOD), a measure of the number of aerosols in the atmosphere, increased from less than 0.2 in 2000 to more than 0.4 by 2020, the study found.
Sulphate accounted for 35.5 per cent of total urban AOD in 2000, rising to 46.4 per cent in 2020. Over the same period, the contribution of dust declined from 30.4 per cent to 21.5 per cent, while organic aerosols accounted for 19.8 per cent and 18.6 per cent, respectively. Black carbon accounted for 8.4 per cent in 2000 and 7.8 per cent in 2020.
The researchers linked the increase in sulphate to rising SO2 emissions from coal-based electricity generation. Energy-sector SO2 emissions increased from less than 1 million tonnes a year in the late 1990s to about 7 to 8 million tonnes a year by the early 2020s, becoming the dominant source of sulphur emissions. Industrial emissions also rose to about 3 to 4 million tonnes a year.
Using GEOS-Chem sectoral simulations, the researchers estimated that the power sector’s contribution to total urban AOD increased from about 16 per cent in 2000 to 36 per cent in 2020, making it the largest anthropogenic contributor. Industrial emissions accounted for about 32 per cent in 2020. Power sector and industrial AOD increased by about 125 per cent and 130 per cent, respectively.
The study’s findings are consistent with the broader health burden from ambient PM2.5 in India. The State of Global Air estimated that India had about 980,000 deaths attributable to PM2.5 in 2019, although this estimate covers the country as a whole and is based on a different methodology and year from the Nature Cities study.
Mean urban PM2.5 concentrations increased from approximately 35 to 38 micrograms per cubic metre in 2000 to more than 40 to 45 micrograms per cubic metre by 2020. Several years after 2015 exceeded 45 micrograms per cubic metre.
Across northern and eastern India, PM2.5 concentrations in many urban regions increased to more than 60 to 80 micrograms per cubic metre by 2020. Power-sector PM2.5 increased from generally less than 5 micrograms per cubic metre in 2000 to 10 to 20 micrograms per cubic metre across large parts of northern and eastern India in 2020, with some locations approaching 20 to 25 micrograms per cubic metre.
The researchers estimated that power-sector PM2.5 increased by about 400 per cent between 2000 and 2020, compared with about 90 per cent for industrial emissions. Transport-related PM2.5 declined by about 40 per cent.
The resulting health burden also increased. Total urban PM2.5-attributable deaths rose from about 0.67 million in 2000 to 0.87 million in 2020, an increase of about 30 per cent.
Deaths attributable to power-sector emissions increased from about 0.015 million to nearly 0.30 million during the same period, an increase of about 1,800 per cent. Industrial-attributable deaths increased from about 0.03 million to 0.14 million, or about 350 per cent.
The largest increases in power-sector-attributable deaths were concentrated across the Indo-Gangetic Plain and Horseshoe Belt. In several urban and industrial corridors, power-sector-attributable deaths rose from generally below 30 to 60 deaths per grid cell in 2000 to more than 90 to 150 deaths per grid cell by 2020.
The study also identified a contrasting trend in household air pollution. Access to clean fuels and technologies for cooking increased from about 25 per cent of the population to nearly 75 per cent between 2000 and 2023.
Household air pollution deaths declined from about 1.25 to 1.30 million a year around 2010 to approximately 1 million in 2023. In contrast, outdoor air pollution deaths increased from about 0.22 million in 1990 to nearly 0.95 million by 2023. Urban PM2.5-attributable deaths estimated by the study increased from about 0.7 million in 2000 to 0.9 million in 2020.
The expansion of liquefied petroleum gas (LPG) access has continued since the period covered by much of the study. Government data show that more than 10.57 crore Pradhan Mantri Ujjwala Yojana (PMUY) connections had been provided by May 2026, while total LPG connections increased from 14.52 crore in 2014 to 33.39 crore in 2026. The government said average annual LPG consumption among PMUY beneficiaries increased from 3.68 refills in 2021-22 to 4.71 in 2025-26.
The government has also continued targeted financial support for LPG use. For financial year 2025-26, PMUY beneficiaries were eligible for a subsidy of Rs 300 per 14.2 kilogram cylinder for up to nine refills a year. The government said average per capita consumption among PMUY consumers rose from about three refills in 2019-20 to 4.47 in 2024-25.
The study found that a 1 per cent increase in GDP per capita was associated with an approximately 1.12 per cent increase in clean cooking access. Electricity demand was positively associated with power sector SO2 emissions, while SO2 emissions were associated with higher urban AOD and PM2.5 concentrations.
The clean cooking transition, however, does not necessarily eliminate exposure to air pollution. The study’s central finding is that as households move away from polluting solid fuels, emissions associated with electricity generation can become more important where electrification is supplied predominantly by coal-fired power plants.
India’s recent electricity trajectory shows both continuity and change. According to the IEA, coal-fired generation declined by about 3 per cent in 2025, while electricity demand growth slowed to 1.4 per cent because of an early monsoon. Strong growth in solar and wind generation and higher hydropower output also contributed to the decline.
The decline was not necessarily a structural break in coal dependence. The IEA expects India’s coal demand to return to its historical growth trend in 2026, rising by 4.2 per cent to 1.353 billion tonnes, as electricity demand increases and industrial coal use remains strong.
At the same time, the IEA expects about half of India’s additional electricity demand through 2030 to be met by solar photovoltaic and about one-quarter by coal. It projects that India’s electricity-sector emissions will rise by an average 2.4 per cent annually between 2026 and 2030 as demand grows, despite expanding low-emissions generation.
The researchers cautioned that the absolute health estimates should be interpreted with care because the assessment used MERRA-2 PM2.5 data, which may underestimate absolute concentrations in densely populated urban areas. They said the resulting bias could lead to conservative estimates of PM2.5-attributable deaths, although the long-term trends and relative changes were supported by the analysis.