A side-by-side ledger of what people in both cities actually breathed over the last 24 months, where the pollution comes from, and what the published projections say about the years ahead.
Lahore sits in the Indo-Gangetic Plain smog basin; Islamabad sits 500 m higher against the Margalla Hills. Over the same 24 months Lahore averaged roughly twice Islamabad's fine-particle load, and its smog season runs longer and peaks harder.
Both cities are moderate-to-poor from March to September, then step up together in October. Lahore's step is two to three times taller: November 2024 averaged 286 µg/m³, a typical-day AQI of 336, the worst month either city recorded. The one month Islamabad came close was December 2025, when it hit 122 µg/m³ against Lahore's 171.
Hover or tap a month for both values and the AQI category. The WHO annual guideline is 5 µg/m³, which would sit on the baseline of this chart.
US AQI is computed from the monthly mean PM2.5 using the EPA breakpoints, so it reads as "the AQI of a typical day that month". Daily peaks run far higher: IQAir logged Lahore days above 500 AQI in November 2024.
Ground networks in Pakistan mostly measure particulates, so the gas and dust picture below comes from the Copernicus CAMS atmospheric reanalysis. Treat the levels as indicative and the seasonal shape as robust. Values are 24-month means for winter (Oct–Feb) and the rest of the year.
Source apportionment for Lahore has been studied repeatedly; Islamabad has no dedicated apportionment study, so its picture is assembled from emissions inventories, Pak-EPA statements and regional studies. Percentages differ by study because they measure different things (emission tonnage vs. ambient PM2.5 contribution) in different seasons.
The same inventory's headline, quoted everywhere as "83% transport", counts all pollutants by mass and is dominated by carbon monoxide. Restricted to fine particles, burning of crop residue and waste is as large as traffic.
Sectoral Emissions Inventory of Lahore, Table 3 · reproduced in the CM Punjab Smog Mitigation Plan 2024
Diesel trucks and buses are a fifth of the fleet but over 65% of transport PM2.5. The only study to separate brick kilns. Punjab EPA publicly rejected its "overwhelmingly local" framing and pointed to cross-border sources.
Unveiling Pakistan's Air Pollution, Exhibit 3.3 · EPA response via Dawn
The only estimate of what people actually breathe rather than what is emitted. It puts household solid-fuel burning first, a sector both Urban Unit inventories score at 0.1%. Punjab-wide the same model gives 53% of PM2.5 from Punjab itself, 9% from other provinces, 13% from neighbouring countries.
Cited in the World Bank Punjab Clean Air Program PID and ESSA; the advisory itself is unpublished
The most-cited breakdown in Pakistani policy documents. FAO's satellite analysis of aerosol point sources affecting Lahore in winter attributes 65% to Pakistan, 17% to India, 11% to Afghanistan, 5% to Iran.
FAO, Emission inventory for Punjab, pp. xvi, 21, 60
Total 12,040 t of PM2.5 a year for the twin cities. The capital's problem is mostly traffic: over a million vehicles registered in a decade plus roughly 500,000 daily entries, per the Excise Department.
Unveiling Pakistan's Air Pollution, pp. 45, 73–75
Vehicles and road dust; non-zigzag brick kilns on the outskirts (demolished in sectors H-16 and H-17 after an October 2025 deadline); steel units; round-the-clock construction machinery on Blue Area high-rises; pollution drifting in from Rawalpindi; CDA burning garbage in public bins. PM2.5 makes up about 90% of the pollution load. December 2024 was the worst month on record, with a daily AQI peak of 250.
Dawn, 16 Nov 2024 · Dawn, 30 Dec 2024 · The Nation, 30 Oct 2025
Source fingerprinting in the industrial area (2023 samples, EPA-PMF) finds steel mills, marble processing, e-waste incineration, battery processing and vehicles, with higher PM in cooler months from thermal inversions in the Margalla basin. No study quantifies Margalla forest fires or cross-border transport for the capital.
Journal of Trace Elements and Minerals, 2025 · PIXE study, Nilore, 2012
From late October a temperature inversion sits at 700–900 hPa over the Indo-Gangetic Plain, the mixing layer collapses to about 300 m, winds drop to 0–3 mph and humidity climbs past 75%, so smoke nucleates fog. Rice-residue burning in both Punjabs is confined to October–November, which is why it matters more to smog than its annual tonnage suggests. Islamabad's basin, ringed by the Margalla Hills, traps its own air in autumn and spring.
Frontiers in Sustainable Cities, 2023 · Environ. Sci. Pollut. Res., 2026
No published model projects a monthly or annual AQI number for Lahore or Islamabad specifically. What exists are policy targets, regional scenario models, and trend readings. They are listed with their horizon and what they actually promise.
The only model-based, city-level projection that exists. The World Bank labels it provisional, to be revised once Punjab's emissions inventory is built. The program funds 600 electric buses, 5,000 super-seeders for rice stubble, kiln conversions and monitoring, and closes in June 2030. The "35% cut in a decade" figure in press coverage is a Pakistan-wide World Bank target, not a Lahore projection.
Punjab Clean Air Policy, notified April 2023. Against IQAir's 2022 value of 97.4 µg/m³ that implies about 68 µg/m³ in 2030, still 13 times the WHO guideline. It is a target with no modelling behind it. The same policy promises all brick kilns on zigzag technology by 2030 and a 25% cut in NOx, SOx, ozone and CO from 2021 levels.
World Bank and IIASA GAINS modelling for the "Punjab (Pakistan)" region, the closest scenario set for both cities. About 40% of the achievable reduction in the Indo-Gangetic Plain depends on controls outside one's own region, which is the case for treating the two Punjabs, Haryana and Delhi as one airshed. The Bank's later program document puts Punjab on a rising path instead: 52 today to 58 by 2030 without action.
Striving for Clean Air, 2023, Tables A.4.1 and A.4.3 · Purohit et al., 2024
Adopted in the World Bank Country Partnership Framework for 2026–35 and echoed in the National Clean Air Policy, which targets PM2.5 emissions 38% below baseline in 2030 and 81% below in 2040. SEI's modelling of that policy's 18 measures finds a 36–39% emissions cut by 2030 if implemented. A September 2026 review estimates full implementation would avert about 129,500 deaths a year by 2040.
National Clean Air Policy 2023 · Slater et al., SEI 2024 · Dawn, 8 Sep 2026
IIASA GAINS modelling for Pakistan as a whole. A separate 2025 disease-burden model finds PM2.5 rising 0.2–0.5 µg/m³ a year over four decades and PM2.5-attributable deaths passing 290,000 a year by 2030 on the current path, versus about 29,000 under mitigation.
Mir et al., Environmental Science & Policy 2022 · Khan et al., Environmental Pollution 2025
A 2026 five-model CMIP6 ensemble focused on Lahore finds increasingly persistent 700–900 hPa inversions and longer stagnation windows through 2049 under SSP5-8.5. Regional work agrees: no emissions pathway improves South Asian air by 2050 (Kumar et al. 2018); SSP3-7.0 adds about 8 µg/m³ to South Asian PM2.5 by 2050 (Xu et al. 2022); each extra winter stagnation day in the plain adds roughly 1 µg/m³ (Zhou et al. 2024). None of these give a Lahore or Islamabad number.
Ghani et al., 2026 · Kumar et al., 2018 · Xu et al., 2022 · Zhou et al., 2024
Pak-EPA's Vehicular Emission Control Action Plan (Oct 2025) sets Euro-5 fuel by 2027 and Euro-6 by 2030 with no PM2.5 or AQI goal. The CDA and Asian Development Bank scoped an Air Quality Improvement programme in December 2025, with no figures yet. The nearest proxies are the Punjab scenarios above and the Air Quality Life Index reading: at 51 µg/m³ (2023) Islamabad residents lose 4.5 years of life expectancy against the WHO guideline, Lahore Division residents 5.8.
APP, 19 Oct 2025 · Dawn, Dec 2025 · AQLI Pakistan fact sheet 2025
Punjab launched a pre-emptive anti-smog drive on 3 September 2026 without a severity outlook; CAMS issues only five-day forecasts. The 2025 season was markedly better than 2024 (Lahore's November average fell 37%, the daily peak from 539 to 237 µg/m³, and no day above AQI 300), but PAQI attributes much of that to favourable weather and warns it can reverse. The direction of the 24-month series is consistent with that: Lahore's Jan–Aug 2026 mean of 70 µg/m³ sits between its 2025 (68) and 2024 (75) figures for the same months.
The World Bank rows are for Lahore Division (13 million people), a wider and slightly cleaner area than the IQAir city reading. Even the policy target leaves Lahore where Islamabad was in December 2024 on an annual basis.