Main

Global waste production has nearly tripled since 1965, reaching 2 billion tonnes per year in 2016 and, with a growing population6, is expected to further increase by 70% by 20507. Close to 70% of waste currently ends up in landfills or dumping sites7, in which anaerobic decomposition of organic material produces methane. Methane is a short-lived (with about a nine-year lifetime8) but potent greenhouse gas and its anthropogenic emissions make it the second most important contributor to human-induced climate change after anthropogenic carbon dioxide emissions, accounting for ~30% of current positive warming relative to pre-industrial temperatures (1850–1900 average)9. Deep and rapid reductions in global anthropogenic methane emissions are essential to keep net warming below 1.5 °C by 21001,10. Methane emissions from solid waste currently amount to 38 million tonnes per year, roughly 10% of total anthropogenic methane emissions2, and could reach 60 million tonnes annually by 205011. However, some mitigation options are available, for example, banning organic waste in landfills, source separation, reuse, recycling or treatment with an anaerobic digester11. If these are implemented to their fullest potential, 2050 methane emissions from solid waste could be as low as 11 million tonnes per year11.

Solid waste emission estimates are based on widely used first-order decay models12 that are used in country-level reporting of methane emissions13 as well as at facility scale. Different variants of such models exist and can yield very different results for similar facilities14. The parameters (for example, methane generation potential of the waste) that drive these models are also uncertain and specific to each facility15,16. Finally, waste disposal management practices can greatly impact methane emissions, from unmanaged dumping sites to managed sanitary landfills that include linings, covers and gas capture systems of variable efficiency17. Considering all of these uncertainties, independent observations of methane emitted from waste disposal sites are critical and can be obtained through various on-ground and/or aerial-measurement methods18 that are deployable at the site level and that can provide emission estimates at high continuous temporal resolution. Complementarily to these site-specific approaches, satellites offer extensive global coverage, providing consistent observation sets across a large number of sites. We present here a global-scale survey of methane emissions from waste disposal sites using 1,447 high-resolution satellite observations.

Satellite remote sensing of atmospheric methane can have an active role in methane emission mitigation by locating emission hotspots and identifying the super-emitting sources they contain19. Over the past decade, a range of spaceborne instruments have been transformative for methane imaging from space3,20,21,22,23,24. They provide spatial images of atmospheric methane concentrations that enable the detection of anthropogenic emission plumes. These consist of strong enhancements in methane concentration that extend downwind from localized emission sources, as illustrated in Fig. 1. Calibrated mass-balance approaches are employed to translate these instantaneous snapshots into emission rates (Methods), validated by single-blind controlled release25. Our study focuses on measurements from GHGSat’s high-resolution (~25 × 25 m2) methane imaging satellites, which capture targeted 12 × 15 km² scenes and detect facility-scale plumes arising from localized sources at emission rates as low as 100 kg h–1. They can be attributed to individual sources across oil and gas facilities (onshore and offshore), coal mines and waste disposal sites5,26,27,28. Many of these individual sources were first coarsely spatially identified5,29,30,31,32,33 with the Tropospheric Monitoring Instrument (TROPOMI) on board the Sentinel-5 Precursor satellite34. It maps the atmospheric methane concentration with daily global coverage and a resolution of up to 7 × 5.5 km2, enabling the detection4 of methane plumes that can be followed-up—in a so-called tip-and-cue strategy—by targeted GHGSat observations to pinpoint their exact sources. This approach has been demonstrated for four urban areas with strongly emitting waste disposal sites5. Here we present a global GHGSat-based survey of methane emissions from waste disposal sites across 130 urban areas in 47 countries during 2021–2022.

Fig. 1: Examples of urban- and facility-scale satellite observations of methane plumes.
Fig. 1: Examples of urban- and facility-scale satellite observations of methane plumes.
Full size image

a–o, Examples of GHGSat facility-scale (b,d,f,g,h,j,k,m) and TROPOMI-detected urban-area (a,c,e,l) methane emission plumes for urban areas in Charlotte (USA, a,b), Bucharest (Romania, c,d), Hyderabad (India, e,f), Guadalajara (Mexico, g), Córdoba (Argentina, h), Hong Kong (China, j), Bangkok (Thailand, k) and Casablanca (Morocco, l,m). The spatio-temporal distributions of all GHGSat plume origins and Sentinel-2-detected surface activity (structural changes between visual Sentinel-2 images; Methods) for the Casablanca landfill are shown in n and o, respectively. Black crosses mark site locations, whereas white dots represent the GHGSat plume origins and thick black contours demarcate landfill site boundaries. White arrows (a–h, j–m) illustrate the wind direction sampled from the ERA5 reanalysis48, with the associated labels indicating the wind speed. Plumes overall follow the reanalysis wind direction, with some exceptions at low wind speeds. Background imagery relies on non-concurrent Sentinel-2 data (2022). Copernicus Sentinel-2 data in parts a–h,j–o are adapted from Google Earth Engine49,50. Scale bars, 50 km (a,c,e,l), 1 km (b,d,f,g,h,j,k,m,n,o).

One-third of methane emission plumes detected in TROPOMI data from the year 2021 is related to urban areas4. In 2021 and 2022, we detect 897 plumes with TROPOMI across 46 urban areas among the 130 covered by GHGSat (Extended Data Fig. 1 and Supplementary Notes 1 and 2). These detections—which depend on observational coverage and the magnitude of emissions (Extended Data Fig. 2 and Supplementary Note 3)—are located on six different continents, with the majority coming from Asia (Fig. 2). TROPOMI plumes illustrate the mitigation potential concentrated in urban areas, which harbour a range of sources such as wastewater treatment, natural gas distribution and incomplete combustion35. Waste disposal sites, however, are some of the most concentrated and mitigable sources in urban areas and are therefore the facilities that we focus on in our GHGSat analysis.

Fig. 2: Distribution of GHGSat-observed waste disposal sites and their urban areas.
Fig. 2: Distribution of GHGSat-observed waste disposal sites and their urban areas.
Full size image

Location of the 151 waste disposal sites observed by GHGSat satellites, and the 46 out of 130 corresponding urban areas for which methane emission plumes have been detected in TROPOMI data (grey). GHGSat methane emission rate distributions over logarithmically spaced bins are given for all sites (black line), and separately for managed landfills (orange) and dumping sites (purple). This site classification has been manually determined using satellite and aerial imagery from Google Earth (see main text). The red line shows cumulative emissions. The site-level and urban-area-level data supporting this figure are provided in the Supplementary Information.

A global facility-scale satellite survey

We use 1,447 clear-sky observations acquired by GHGSat’s C1–C5 satellites in 2021 and 2022. These were targeted at 151 different waste disposal sites located in 130 urban areas scattered over six continents, as shown in Fig. 2. Only sites for which at least one methane emission plume has been detected by GHGSat are included, meaning that our sample is on the upper end of the global waste disposal site emission rate distribution. The median number of GHGSat observations per site is 5, with 23 sites that have been observed at least 20 times (Extended Data Fig. 3). These are opportunistic observations that could be made in parallel to regular GHGSat activities—a substantial fraction (51%) of which intersect with TROPOMI-detected urban methane hotspots.

Out of the 1,447 observations, 1,013 show at least one emission plume above GHGSat’s detection threshold (1,085 plumes in total, Fig. 1; quantified as described in Methods). We conservatively consider the emission rate of the 434 site-level null detections to be zero even though we may miss (possibly diffuse) emissions that are lower than the GHGSat detection threshold. The positive plume detection rate per site ranges from 7% (two plumes among 30 observations at Icheon, South Korea) to 100% (which we find for 74 sites). The median of the plumes’ detected methane emission rates is 2.4 t h–1, with 5th and 95th percentiles of 0.5 t h–1 and 15.1 t h–1, respectively. The median relative uncertainty of these emission rates is ~45% (Supplementary Note 4).

Recurrent observations allow us to investigate the potential drivers of the detected emission variability. We compared site-wise emission variability against several meteorological variables (10 m wind speed, 2 m temperature, surface pressure, surface pressure change and accumulated precipitation over two weeks) as well as the hemisphere-corrected day in the year, but we did not find any significant link between them (Extended Data Fig. 4 and Supplementary Note 5); however, past on-site studies have indicated that surface pressure change drives landfill methane emissions36,37,38. Our findings based on satellite observations of high-emitting active sites are consistent with recent airborne-based results39. This finding could be explained by meteorological driving producing too small emission changes compared to single observation uncertainty for the sites included in our dataset.

The median site-wise averaged emission rate is 1.2 t h–1 (including null detections), with 5th and 95th percentiles of 0.1 t h–1 and 6.8 t h–1, respectively (Methods, Extended Data Fig. 5 and Supplementary Note 4). The lowest three site-averaged detected emission rates are found at a Canadian landfill in British Columbia (0.03 ± 0.04 t h–1), an Italian landfill near Rome (0.04 ± 0.03 t h–1) and at a South-African landfill near Gqeberha (0.06 ± 0.04 t h–1). The highest three site-averaged detected emission rates are found at the Norte III landfill in Buenos Aires, Argentina (22.0 ± 1.9 t h–1), at a landfill near Hong Kong, China (10.0 ± 2.7 t h–1) and at a landfill near Tehran, Iran (9.4 ± 4.9 t h–1). Averaged emission rates have a median relative uncertainty of 45%, which accounts for both single observation and sampling uncertainties, calculated consistently across all sites (Methods). Using satellite and aerial imagery from Google Earth, we manually classify the 151 waste disposal sites into two categories: 108 managed landfills (sites with organized structures for burying waste, for example, featuring covers) and 43 dumping sites (with informal gathering of waste). Within this dataset, managed landfills and dumping sites do not show a statistically significant differennce in the total detected emission rate distributions. However, when normalized by the total site area, managed landfills show significantly lower area-normalized emission rates compared with dumping sites, thus demonstrating the expected effects of emissions mitigation by closing and covering modules of the landfill (Extended Data Fig. 6 and Supplementary Note 6). Overall, the distribution of site-wise averaged detected emissions is heavy-tailed, with the 60 (40%) strongest-emitting sites (47 managed landfills and 13 dumping sites) accounting for 80% of total emissions (Supplementary Note 4). This estimated skewness is probably conservative as the 100 kg h–1 detection threshold and selective targeting of GHGSat would limit the inclusion of low-emitting sites. This detection threshold enables to cover 54% of the sites (assuming sufficiently localized emissions sources) included in the facility-scale waste disposal site emission database compiled by Climate Tracking Real-Time Atmospheric Carbon Emissions (Climate TRACE)40, but these 54% of sites amount to 96% of total emissions (Extended Data Fig. 7 and Supplementary Note 4). Overall, the 151 waste disposal sites observed here represent a small fraction of the global total number of landfills (over 10,000 are included in the Climate TRACE datasets40) but, assuming constant emissions, their collective instantaneous emission rate scales up to a yearly total of 2.8 million tonnes. This corresponds to 7.4% of 2022 global solid waste emissions in version 8 of the Emissions Database for Global Atmospheric Research (EDGAR) inventory2.

Modelled and GHGSat-based rates disagree

Figure 3 compares facility-level GHGSat-detected methane emission rates against national site-level reporting programs41,42,43 (which are based on process-based models or gas capture efficiency assumptions) and emissions obtained from data-driven models developed by the non-profit coalition, Climate TRACE40. National reporting data exclusively cover managed landfills in the United States, Canada and some EU countries, whereas Climate TRACE has more global coverage and includes dumping sites (Extended Data Fig. 8 and Supplementary Note 7). Overall, we find no correlation between satellite-based and reported or modelled estimates (r = 0.03 for reported emissions, and r = 0.18 for Climate TRACE), with differences showing an insignificant bias and a large scatter, exceeding the averaged emission rates (Supplementary Note 8). Analysing managed landfills and dumping sites separately does not change this conclusion. Although no overall bias is found between reported and GHGSat-based estimates, emissions from 14 (out of 37) landfills are at least twice as large as what is reported to national programmes. As the US Greenhouse Gas Reporting Program includes reports based on two different methodologies—one based on gas capture efficiency and the other based on waste decay modelling—we can compare our results for the US to both (Fig. 3 separates US sites depending on which reporting method was chosen by the facilities). Comparing both estimates for all US landfills to GHGSat-detected emissions (Extended Data Fig. 9 and Supplementary Note 9), we observe that the approach based on gas capture efficiency tends to underestimate emissions (by a factor two), whereas the approach based on waste decay modelling tends to overestimate them (by a factor 1.5). These US results are consistent with a recent investigation of landfill emission models used for reporting44 and with aerial-based observations39. Neither method shows a strong correlation with our results. These findings highlight the critical importance of coordinating bottom-up modelling efforts with independent observations of landfill emissions to improve the understanding of facility-scale waste emissions.

Fig. 3: Discrepancy between satellite-based and bottom-up emission estimates.
Fig. 3: Discrepancy between satellite-based and bottom-up emission estimates.
Full size image

Comparison of site-wise methane emission rates observed by GHGSat against data included in reporting programmes (left) and emissions calculated by the Climate TRACE non-profit (right), both averaged over the corresponding GHGSat observation years. Reported and Climate TRACE data are provided as annual totals and have been converted to hourly rates assuming constant emission. Error bars show the site-wise-averaged GHGSat emission uncertainty. The one-to-one line is shown in black. The sites marked by dashed-circles in both panels drive the high correlations for EU site reports (r = 0.76) and Climate TRACE dumping sites (r = 0.54). If removed, these correlations drop to r = −0.62 and r = −0.15, respectively.

Plume sources relate to site activity

Figure 1 shows that high-resolution observations also allow us to pinpoint where detected emissions originate from within a solid waste disposal facility. To understand these origins, we compare manually verified GHGSat emission plume origins with surface activity detected from clear-sky Sentinel-2 10-m resolution RGB observations (Methods and Supplementary Note 10). A landfill near Casablanca (Morocco; Fig. 1l–o) is a clear example, as both GHGSat plume sources and landfill surface activity show north-to-south migration as time progresses and a new section of the landfill is developed in the southwest. Across 107 facilities that have a sufficient number of clear-sky Sentinel-2 images and high-quality surface activity detection results, we find that 44 (41%) show a statistically significant proximity (P-value < 0.05) between surface landfill activity and GHGSat plume source location. When considering only the 21 sites with at least 16 identifiable plume origins in GHGSat observations, we find statistically significant proximity for 18 (86%) of them (Extended Data Fig. 10). Upon revisiting our dataset showing that total site area-normalized emission rates are significantly lower for managed landfills compared with dumping sites (Extended Data Fig. 6 and above), we conclude that the small fraction of open active areas in managed landfill accounts for almost all of the emissions detected by GHGSat. This highlights the predominant role open modules have in managed landfill emissions. This result is consistent with reports of methane emissions being observed originating from landfill work faces in on-ground and satellite-based studies for a limited number of sites5,45, and with an extensive aerial survey covering the United States that showed the prevalence of work faces in total landfill emissions39,46. Our dataset shows the active surface is the dominant emission source across management and economic development levels. This emphasizes the need to quantify emissions from the active surface, underscoring the importance of repeated observations to both reliably estimate mean emissions and to narrow down on (potentially migrating) source locations within a landfill. This spatial information can for example help site operators focus mitigation efforts more effectively. Our dataset also includes two example plumes originating from adjacent facilities: a biogas plant near the Las Dehesas landfill near Madrid and from a wastewater treatment plant near Shanghai (both filtered from the analysis; Supplementary Note 11). They illustrate the mitigation potential that satellite observation can detect in facilities related to and neighbouring waste disposal sites.

Our survey has extensive spatial coverage that brings top-down observation-based estimates of methane emissions for 151 waste disposal sites across six continents. It sheds new light on the mitigation potential of urban methane emissions and on the ability of high-resolution satellites to monitor methane emissions from waste disposal sites and support mitigation activities by pinpointing emission sources within the facility, highlighting the importance of the active surface. The availability of such high-resolution methane-sensitive satellite observations is currently increasing, with the expanding GHGSat constellation and new initiatives such as Carbon Mapper’s Tanager-1 satellite, as well as public hyperspectral satellite missions3. Across the 151 surveyed sites—assuming constant emissions to scale up the snapshot averages provided by satellites—we find that bottom-up and top-down satellite-based solid waste emission estimates cannot currently be reconciled at facility scale. This disagreement is consistent with past facility-scale studies using aerial measurements39 and country-scale studies using TROPOMI data47. These discrepancies highlight the importance of site-level data and practices, and call for further efforts that focus on both managed landfills and dumping sites, aiming to close the gap between current bottom-up and top-down understandings of methane emissions from solid waste. Ideally, such studies would involve partners operating waste disposal sites, bottom-up modellers, aerial and satellite-based methane observations augmented by complementary on-ground observations that can provide continuous measurements, including at night. An improved understanding of site-level solid waste methane emissions can support more effective emission mitigation strategies contributing to the worldwide efforts against climate change.

Methods

Automatic methane plume detection in TROPOMI data

TROPOMI51 on board the European satellite Sentinel-5 Precursor was launched in 2017. It observes backscattered sunlight in the near- and shortwave infrared around the 0.76 µm O2 and 2.3 µm methane bands, at approximately 1:30 pm local time. Total columns (vertically integrated concentrations) of methane with near vertically uniform sensitivity down to the surface are retrieved from these observations using a full-physics approach that accounts for the interfering impact of surface reflectance, aerosol and other geophysical variables on the shortwave infrared signal (v.2.6.0)34. TROPOMI is a methane flux mapper that offers daily global coverage with a 7 × 5.5 km2 spatial resolution at nadir. In addition to being used in long-term inverse analyses, its imaging capabilities enable the detection of anthropogenic methane emission plumes that arise from the world’s largest emitters29. We employ a two-step machine learning approach to explore TROPOMI data for methane emission plumes automatically4. We analyse and manually verify all plumes detected in 2021 and 2022 with estimated sources within 50 km from any of the landfills targeted by GHGSat. We apply the Integrated Mass Enhancement (IME) method20, calibrated specifically for TROPOMI using atmospheric transport simulations, to quantify the methane emission rate and its uncertainty for each TROPOMI-detected plume4.

Given TROPOMI’s spatial resolution (7 × 5.5 km2) compared with GHGSat’s (25 × 25 m2), we cluster the 151 landfills observed by GHGSat into 130 TROPOMI-relevant urban areas. For each urban area, we first apply a 2σ filter to remove outlier estimates that can be hampered by an unrepresentative plume mask due to variable meteorology or surface effects (for example, plume masks truncated by clouds for lower estimates). Then, relying on the remaining TROPOMI plume detections, we report their mean detected urban-scale methane emissions and their standard deviation. These averages only cover emissions detected as strong plumes and are not representative of mean urban emissions but do provide an indication of urban mitigation potential. Not detecting a plume does not imply that there are no emissions: it means that concentrated emissions are lower than the ~8 t h–1 TROPOMI plume detection threshold, or that the observational or geographical conditions did not allow for TROPOMI detection4. The discrepancy with mean emissions is verified for four different (above-average emitting and often detected) cities (Buenos Aires, Delhi, Mumbai and Lahore) where IME-based rates show a 7–47% overestimation (while agreeing within uncertainties) compared with urban-level methane emission estimates based on atmospheric inversions and TROPOMI data5 (Supplementary Note 1).

GHGSat observations and emission quantification

GHGSat-C1 to -C5 instruments were launched between 2020 and 2022. Satellites C1-2-3 perform measurements in the morning around 10 am (local time), while satellites C4-5 perform measurements in the afternoon around 2 pm (local time). These instruments estimate the total column (vertically integrated content) of methane at ~25 × 25 m2 resolution over targeted 12 ×15 km2 domains52 from backscattered sunlight measurements in the shortwave infrared near 1.65 µm, that provide near-surface sensitivity. The GHGSat instruments have an empirically measured methane column precision range of 1.4–2.9%53, which allows them to observe emission plumes from point (for example, a gas pipeline leak) or very localized sources (for example, active faces of landfills) emitting more than ~100 kg h–1 (this detection threshold increases with wind speed)28. Pixels exhibiting local spatially correlated methane column enhancements above background are clustered together and considered to belong to a plume26. We apply the IME method20 to estimate an emission rate Q based on a delineated plume and the local wind speed sampled from a meteorological model. We have:

$$Q=\frac{{U}_{\mathrm{eff}}}{L}\sum _{i}{\rm{\Delta }}{{X}_{\mathrm{CH}}}_{4,i}{a}_{i}$$

where Ueff is the effective wind speed, calibrated against the 10-m wind speed based on a set of large Eddy simulations (LES)5; \(L=\sqrt{\sum _{i}{a}_{i}}\) is the plume length computed as the square-root of the plume total area, where ai is the area of the ith pixel included in the plume; and \({\rm{\Delta }}{{X}_{\mathrm{CH}}}_{4,i}\) is the local enhancement above the background of the methane total column for this ith pixel. Here we use an effective wind speed calibration specific to landfills, based on LES of area sources: \({U}_{{\rm{e}}{\rm{f}}{\rm{f}}}=0.34\times {U}_{10\,{\rm{m}}}+0.66\) (ref. 5), where \({U}_{10\,{\rm{m}}}\) is the 10-m wind speed sampled from the GEOS-FP meteorological reanalysis54. The emission rate uncertainty calculation includes contributions from (1) wind speed error; (2) methane column retrieval error; and (3) IME calibration error26.

The calibration of this mass-balance approach against LES of known synthetic emission rates ensures that the estimated rates correctly account for the different advective transport conditions explored within the set of LES. Beyond this calibration on simulations, numerous real-life validation efforts have been organized, including controlled-release experiments, which are the validation gold standard. Notably, GHGSat participated and showed excellent agreement with metered emission rates in internal controlled releases, as well as in two single-blind controlled-release campaigns, where the true emission rates (and wind speeds) are not known to the satellite data providers and the comparisons are done by a third party (in this case a research group from Stanford University)25,55. Beyond controlled releases, landfill emission rates obtained through aerial methane imagery with an instrument that can detect plumes down to 10 kg h–1 have been validated against traditional aerial mass-balance results18,39. Besides, an in-depth study of two landfills near Madrid that included both similar airborne observations and GHGSat satellite observations showed that GHGSat satellite-based estimates match the total of airborne-detected plumes for same day observations within uncertainties56. Combined, these results show that GHGSat satellite-based observations can provide accurate estimates of methane emissions from waste disposal sites.

Estimating site-level GHGSat averages

Three outcomes are possible for any individual waste disposal site observation during a single overpass by GHGSat: (1) no plume is detected; (2) only one plume is detected; and (3) several plumes (arising from the same site) are detected. In the first case, we conservatively consider the emission rate to be equal to zero, with no uncertainty. In the second and third cases, we apply the IME method to each plume separately to quantify its emission rate and uncertainty. In the third case, we sum together all of the detected plume emission rates (and sum their respective uncertainties quadratically) to obtain an emission rate for the whole site.

Given a set of observations for a waste disposal site, we employ a two-step random sampling approach to evaluate the site-level averaged emission rate and its uncertainty, accounting for both single-observation and sampling uncertainties. First, in a bootstrapping approach, we randomly (N = 100,000) resample our set of observations by randomly picking single observations with replacement. This enables us to generate an ensemble of averaged emission rates for which we also compute corresponding uncertainties assuming that single observations are independent Gaussian variables. We then sample a Gaussian distribution (N = 1,000) for all these ensemble elements relying on their respective rates and uncertainties. Finally, we report the mean and standard deviation across this two-step random sampling approach as averaged emission rate and its uncertainty. This method accounts for the single-observation uncertainties and is especially useful to handle bi- or multi-modal site-wise emission rate distributions that can have a peak at zero (all of the observations without any detection) and one or several peaks for positive emission rate values (all of the observations with detected plumes).

Comparison of GHGSat and reported or calculated emissions

For site-wise GHGSat-based methane emission rate comparison against site-wise reported values included within national reporting programs, we manually match sites based on addresses (no distance threshold is used). To compare site-wise GHGSat-based methane emission rates against values modelled by Climate TRACE, we only select sites for which we find matches within a 2 km distance of GHGSat targets, and then only consider the facilities within these 2 km that show the minimum distance from GHGSat targets (Supplementary Note 7 details the other data selection criteria specific to each dataset we compare with). Reported and Climate TRACE data are provided as annual totals and have been converted to hourly rates assuming constant emissions.

Landfill surface activity detection from Sentinel-2 imagery

Managed landfills and dumping sites are active and constantly evolving as they accept new waste: they expand and their active surface(s) move(s) to accommodate the incoming waste. High-resolution visual imagery can be used to track the surface activities at waste disposal sites. To compare the spatio-temporal distributions of GHGSat-detected methane plumes origins and landfill activities, we devise an image analysis scheme to automatically detect surface activity from time series of clear-sky 10-m resolution Sentinel-2 satellite images of waste disposal sites.

For each of the 151 waste disposal sites observed by GHGSat, we convert the time series of Sentinel-2 clear-sky visual RGB images to greyscale by using the National Television Standard Committee’s formula57:

$$\mathrm{Greyscale}=0.299\times R+0.587\times G+0.114\times B$$

We then apply a three-image moving filter (over time) based on local structural analysis58,59, which determines surface activity in a given image by detecting overlapping structural changes that occur between this image and the previous one, and between this image and the next one. Using manually outlined landfill masks based on the latest Google Earth imagery, we only consider surface activity that is detected within landfill boundaries, and use filters to ignore pixels associated with water, clouds or cloud shadows. We smooth the raw activity map with a median filter to remove spatially inconsistent noise and only keep spatially consistent activity clusters. Individual activity clusters are then identified, outlined with convex hulls and stored as surface activity results. For each landfill, surface activity results are manually verified before being included in the analysis (Supplementary Methods 1; illustrations are provided in Supplementary Note 10).

Comparison of landfill surface activity results and GHGSat methane plume origins

The wind direction allows to estimate the plume origin as the most upwind highly enhanced pixel included in the plume mask. We also manually verify this result and pinpoint the approximate source(s) of all GHGSat plumes, allowing one to select multiple sources for overlapping plumes originating from the disposal site where appropriate. We use these source locations to compare to the Sentinel-2 based surface activity analysis.

For a given plume, we use the minimum distance between the manually determined plume origin and the nearest outline of a surface activity cluster detected in the closest-in-time Sentinel-2 image as the proximity metric. We set the metric to zero if a plume origin falls inside a detected activity cluster. Consequently, the lower the metric value, the closer the source is to a detected surface activity cluster. We also compute the same metric for N = 10,000 points randomly drawn within the landfill boundaries. This comparison is conservative because it is possible that GHGSat plumes show sources outside of landfill boundaries (their metric values have no upper boundary) whereas these random points can only be located inside (their metric values have an upper boundary).

For each site, we compute the averaged metric across all GHGSat-detected methane emission plumes and compare this result with the distribution of averaged metric values obtained for the N = 10,000 randomly drawn points. We then evaluate the P-value probability of randomly obtaining averaged metric values that are lower than the GHGSat-based result. We consider that GHGSat plume origins show a statistically significant proximity with detected landfill surface activity if we obtain a P-value lower than 0.05 (Supplementary Methods 2).

Supplementary Note 10 showcases examples from different landfills and present an overview of P-value results for all landfills where surface activity could be detected.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.