Satellite measurements of methane over London, Los Angeles and New York suggest conventional emissions inventories may be substantially underestimating the amount of the greenhouse gas released by major cities, according to a new study.
Researchers at the University of Manchester developed a method using observations from the TROPOspheric Monitoring Instrument (TROPOMI) aboard Europe’s Sentinel-5 Precursor satellite to estimate methane emissions at city scale.1
Applying the technique to the three cities between 2021 and 2023, the researchers calculated emissions ranging from 5.99 to 11.90 tonnes per hour in London, 26.21 to 62.77 tonnes per hour in Los Angeles, and 30.85 to 44.77 tonnes per hour in New York.
The satellite-derived estimates were generally higher than those in conventional “bottom-up” emissions inventories, which calculate emissions using information on activities and associated emission factors.
Compared with the global EDGAR emissions inventory, the satellite estimates were approximately 1.5–3 times higher for London, 1.3–3.1 times higher for Los Angeles and 7–10.2 times higher for New York.
Comparisons with national inventories produced a less pronounced discrepancy for London. The researchers found that the UK’s National Atmospheric Emissions Inventory (NAEI) generally underestimated London’s emissions during the study period, except in 2021, with the difference reaching around 45% in 2022. For the two US cities, comparison with an EPA inventory indicated underestimates of approximately 56% for Los Angeles and 84% for New York.
The findings are published in the journal Atmospheric Measurement Techniques.
Improving satellite estimates
Methane emissions can be particularly difficult to quantify at city scale because they arise from numerous sources, potentially including waste management, wastewater treatment and leaks from natural gas infrastructure.
The Manchester researchers sought to improve estimates from satellite observations by refining a “source pixel” approach, which determines the enhancement in atmospheric methane over an urban area relative to surrounding background concentrations.
Their method incorporates hourly wind data within the planetary boundary layer – the lowest part of the atmosphere – rather than relying on surface wind measurements. It also determines background methane concentrations using areas located upwind of the city.
The researchers said these refinements provide a more physically representative assessment of how methane is transported through the atmosphere.
For London, the study found that methane enhancements over the surrounding background were relatively modest, making accurate determination of background concentrations particularly important. The authors calculated that around 80% of the uncertainty in London’s emissions estimates arose from variability in the measured methane enhancement, compared with around 20% attributable to wind uncertainty.
According to the EDGAR inventory used in the study, waste management accounted for approximately 63% of London’s methane emissions over 2021–23, while fossil-fuel-related sources such as leakage from natural gas transmission and distribution represented around 32%. Agriculture accounted for nearly 5%.
Monitoring cities from space
The researchers said previous measurement-led studies have also identified discrepancies between observed urban methane emissions and bottom-up inventories. Their own estimates were generally consistent with previous “top-down” studies of the three cities, with most falling within the calculated uncertainty ranges.
Differences between studies can arise from factors including the definition of the urban area being measured, the period covered, atmospheric conditions and the method used to determine background methane concentrations.
The team said the satellite approach could provide a repeatable means of monitoring emissions over cities without the need for extensive ground-based measurement networks.
This could allow satellite observations to be used to check and improve emissions inventories, monitor changes over time and assess whether methane-reduction targets are being achieved.
The researchers cautioned, however, that satellite coverage is affected by factors including cloud, resulting in uneven observations through the year. The relatively weak methane enhancement detected over London also means estimates there are particularly sensitive to the way background concentrations are determined.
Notes
[1] Long, H., Tsivlidou, M., Ricketts, H. and Allen, G., “Satellite-based global monitoring of urban-scale methane emissions”, Atmospheric Measurement Techniques, 19, 6171–6191 (2026), published 29 September 2026.
