Abstract
Africa and the Middle East contribute significantly to global fossil fuel production, yet the associated carbon emissions from these extraction and processing activities remain insufficiently reported. The current reporting is not only delayed but also heavily reliant on energy consumption statistics that are difficult to verify. Here, we introduce an alternative approach to indirectly infer and map anthropogenic carbon dioxide (CO2) emissions using the coemitted nitrogen oxides (NOx) emissions derived from TROPOMI satellite observations. We find that current global inventories underestimate total CO2 emissions by about 30% for Africa and the Middle East. About 35% of this underestimation is related to newly identified CO2 emission sources, which are mainly attributable to oil and gas infrastructure. The analysis reveals that these underestimated emissions could lead to a delay of ∼5 years in achieving the 2030 Nationally Determined Contribution targets in Africa and the Middle East. This study provides an effective tool for constructing an independent carbon emission accounting framework based on satellite observations and offers a scientific database for guiding climate action.
Satellite-based CO2 emission quantification technique reveals underestimation of CO2 emissions in Africa and the Middle East.
INTRODUCTION
Africa and the Middle East are not only core to the global energy economy and the focal point of transformation (1–2) but also represent key sources of uncertainty in global carbon cycling and climate models (3–5). Africa and the Middle East collectively account for ∼40% of global oil production (6). During crude oil extraction, substantial quantities of associated petroleum gas are routinely released and combusted through gas flaring (7). Globally, gas flaring in 2024 resulted in 389 million tonnes (Mt) of carbon dioxide (CO2) equivalent emissions, of which 52% originated from Africa and the Middle East (8). Beyond CO2, flaring emits hazardous pollutants such as benzene, naphthalene, and black carbon, which have been linked to elevated risks of cancer and respiratory diseases (9). Despite its substantial environmental and public health impacts, yet flaring-related emissions remain inadequately represented in official energy planning and greenhouse gas inventories (10). As a result, the true magnitude of carbon emissions in Africa and the Middle East remains highly uncertain, underscoring the need for more comprehensive monitoring and accounting.
Under the United Nations Framework Convention on Climate Change (UNFCCC) framework, each country’s Nationally Determined Contribution (NDC) targets and emission reduction commitments are based on their national emission inventories (11–12). Setting feasible and credible NDCs targets critically depends on accurate and comprehensive estimates of carbon emissions. Moreover, with only 5 years left for companies and governments to meet their commitments under for example the Zero Routine Flaring by 2030 initiative (6), clarifying the overall carbon emission landscape in Africa and the Middle East has become increasingly urgent. However, due to the delayed collection of energy consumption data, national inventories typically have a time lag of about 1.5 years (13) and are associated with considerable uncertainties, especially in developing countries. Additionally, while the global uncertainty in fossil fuel CO2 emissions is ∼±8%, the uncertainty in regional gridded CO2 emission inventories varies significantly due to differences in spatial allocation methods and the proxy data used. For instance, uncertainties in inventories are ±40% in Europe and ±150% in the United States (14–16) and are even greater in underdeveloped countries or regions.
Recent studies have attempted to use satellite data from the Orbiting Carbon Observatory-2 (OCO-2) and OCO-3 missions (17–19) to quantify CO2 emissions from isolated point sources (17–21). However, the sparse spatiotemporal sampling of XCO2 observation data provides only limited instantaneous emission information at specific locations. For example, only about 6% of the total soundings returned by OCO-2 each month passes quality screening (22). A limited number of studies have used TROPOspheric Monitoring Instrument (TROPOMI) and OCO-2 satellite data to quantify greenhouse gas emissions in African and Middle Eastern cities, revealing that bottom-up (BU) inventories significantly underestimate CO2 emissions from cities in the Middle East (23, 24). However, there are few reports (25, 26) on the total carbon emissions and spatial distribution of Africa and the Middle East based on satellite observations.
Here, we propose a method for independent accounting of anthropogenic CO2 emissions in Africa and the Middle East based on TROPOMI NO2 satellite observation data. We first subtract TROPOMI-based soil NOx emissions (27) from total top-down NOx emissions derived by the Daily Emissions Constrained by Satellite Observations (DECSO) inversion algorithm (28, 29) to obtain anthropogenic NOx emissions. We then quantify satellite-based CO2 emissions using the derived anthropogenic NOx emissions and CO2-to-NOx emission factor ratios () from BU inventories. The results of this framework reveal an underestimation of reported emissions in existing inventories for the Middle East and Africa. We further analyze the causes of the underestimation. Our analysis shows that existing BU inventories not only omit some emissions from oil and gas infrastructure in these regions but also significantly underestimate CO2 emissions from the transportation sector. This implies that the NDC targets now set for the region risk being delayed. This study lays a solid foundation for building an independent satellite-based top-down CO2 emission accounting framework and provides policy-makers with a scientific database for monitoring carbon emission dynamics and guiding climate action.
RESULTS
Gridded satellite–based CO2 emission map
Based on the analysis of TROPOMI NO2 observations and NOx-to-CO2 approach, we estimate the CO2 emissions in 2019 for Northern Hemispherical Africa (NHA), the Middle East, and Europe together to be 7.2 ± 0.6 billion tonnes (Gt) of CO2 year−1 (± 1σ of four CO2-to-NOx ratios for deriving CO2 Sat, shown in Fig. 1). The average BU CO2 emissions (light blue bars in Fig. 1A) from five widely used inventories—the Copernicus Atmosphere Monitoring Service (CAMS), the Emissions Database for Global Atmospheric Research (EDGAR), the Open-Data Inventory for Anthropogenic Carbon dioxide (ODIAC), the Global Infrastructure emissions Detector (GID), and the Global Gridded Daily CO2 Emissions Dataset (GRACED)—report the CO2 emissions to be 6.3 ± 0.2 Gt CO2 year−1 (± 1σ of the five inventories), about 13% lower than the satellite-based CO2 emissions. Figure 1A shows that our satellite-based CO2 emissions in NHA, the Middle East, and Europe are 0.8, 3.6, and 2.8 Gt CO2 year−1 (red lines and dark blue bars in Fig. 1A), respectively, differing 28, 30, and 15% compared to BU inventories, respectively. The total uncertainty of our estimates is lie between 12 and 33% (see Materials and Methods), propagated from three components: 25% from the uncertainty of NOx emission retrieval algorithm, 10% from the uncertainty of NOx-to-CO2 approach, and 33% from the uncertainty of the variation of CO2-to-NOx ratios. Due to Europe’s comprehensive, accurate, and timely anthropogenic activity data and BU inventory–calculated emission factors (30), the discrepancy between satellite estimates and BU inventories is the smallest (15%) among the three regions, which also indirectly indicates the reliability of our estimation results. It should be noted that, even in Europe where the aggregate discrepancy is small, the method still reveals substantial spatial discrepancies (fig. S2), due to the offsetting of regional positive and negative biases. However, the high uncertainty in the BU data from Africa (31) and the Middle East (24) leads to an underestimation of ∼30%.
Fig. 1. Satellite-based CO2 emission map and comparison of satellite estimation and BU inventories for three regions.
(A) Monthly variation of satellite-based CO2 emission (called CO2 Sat) in 2019 and the percentage difference between satellite estimation and BU inventories (called CO2 BU). (B) Satellite-based CO2 emission map in 2019, which shows clear ship CO2 emissions track over the sea captured by satellite. (C to E) The zoomed-in views of the regions marked by the red boxes in (B). The difference map between CO2 Sat and CO2 BU (fig. S1) are shown in fig. S2.
Figure 1B presents the spatial distributions of satellite-based CO2 emission in 2019 at a resolution of 0.2° × 0.2°, with the easily identified major emission regions, urbanized and industrialized areas. Line shape road emissions of satellite-based CO2 emissions are less prominent than those of BU inventories, especially in the Middle East (figs. S1 and S2). The satellite-based CO2 emissions indicate higher emissions in urban and industrial areas than BU inventories, which show higher background emissions in rural areas (blue area in fig. S2). Unlike BU inventories that rely on proxy data (population distribution, point sources such as power plants and industrial facilities, residential activities, and agricultural patterns) (32), the spatial distribution of satellite-based CO2 emission is directly derived from satellite NO2 observations and CO2-to-NOx ratios, thereby reducing uncertainties associated with proxy-based spatial allocation.
Our methodology enables clear detection of linear ship emission tracks over the ocean surface (sea area in Fig. 1, B and D), although it does not rely on prior ship trajectory data or spatial distribution information from BU inventories. Using a domain-averaged emission factor (CO2-to-NOx ratio), we estimate total ship emissions at 0.6 Gt, compared to the 0.2 Gt reported in BU inventories. This discrepancy may be partially attributed to uncertainties in ship emission factors. However, our results also suggest that current BU inventories underestimate maritime shipping emissions, which aligns with previous findings (33).
Underestimated CO2 emission from oil and gas infrastructure
Our satellite-based CO2 emissions reveal systematic underestimations in current BU inventories across Africa and the Middle East. The underestimations are primarily attributable to (i) missing emission sources in BU datasets and (ii) underestimation of known sources such as urban and industrial areas (discussed in the following section). Figure S4 presents the spatial distribution of newly identified CO2 emissions, which are defined as grid cells that report finite satellite-based emissions but have near zero BU emissions. These satellite-detected emissions total 332 Tg CO2 year−1 across the study domain, accounting for 4.6% (35% of 13%) of the total CO2 Sat, with 98% concentrated in Africa and the Middle East. The newly identified emissions represent 7% of total satellite-based CO2 emissions (CO2 Sat) in the combined Africa and the Middle East regions, constituting 19 and 5% of Africa’s and the Middle East’s total CO2 Sat, respectively.
We find that the newly identified CO2 emissions that are missing or severely underestimated in existing inventories are closely associated with oil and gas infrastructure (Figs. 2 and 3). Spatially, the newly identified CO2 emissions exhibit high consistency with the distribution of gas flaring CO2 emissions (Fig. 2, A and B), showing colocation patterns observed in regions such as northeastern Algeria, western Libya, the Iraq-Iran border, Kuwait, Yemen, the United Arab Emirates, and Oman. In addition, in Egypt and Iran, the newly identified CO2 emissions align with the spatial distribution of major oil and gas pipelines (Fig. 2C and fig. S7). These facilities are also known as major contributors to anthropogenic methane emissions due to frequent methane leaks (34–36). Therefore, these spatial consistencies likely reflect coemitted emissions associated with oil and gas production and transport processes. Furthermore, at the national scale across Africa and the Middle East, the newly identified CO2 emissions are significantly correlated with oil and gas sector CH4 emissions [correlation coefficient (r) = 0.57] and gas flaring CO2 emissions (r = 0.69) (Fig. 3, A and B). This suggests that current BU inventories may systematically miss or underestimate the carbon emission intensity associated with oil and gas exploitation in this region.
Fig. 2. Spatial patterns and temporal variation of newly identified CO2 emissions.
Comparison of the spatial patterns of newly identified CO2 emissions (A), gas flaring CO2 emissions (B), and CH4 emissions from the oil and gas sector (C) in 2019. The gas flaring CO2 equivalent emissions are derived from the World Bank’s Global Gas Flaring Tracker (6), which estimates flare volumes using VIIRS satellite observations. The flared gas is assumed to consist of 81% methane and the remaining fraction of heavier hydrocarbons. Additionally, 98% of the flared gas is assumed to be combusted, whereas the remaining 2% is directly released into the atmosphere. CH4 emissions from the oil and gas sector are taken from the 2019 EDGAR (32) inventory (version 2025, fuel exploitation: oil and gas sectors). The linear CH4 emission features shown in (C) are spatially consistent with the distribution of oil and gas pipelines (fig. S7). The corresponding spatial distributions over the entire study region for (B) and (C) are shown in figs. S5 and S6. (D) Monthly variation of CO2 emission from newly identified emission grid cells spatially linked to industrial heat sources (34), grouped by four subcategories (nonmetal, metal, coal processing, and oil and gas).
Fig. 3. National-scale magnitude of newly identified CO2 emissions.
Comparison of newly identified CO2 emissions with gas flaring CO2 emissions and oil and gas CH4 emissions in terms of correlation (A and B) and magnitude (C and D). Each circle in (B) and (C) represents a country, with the size scaling with the magnitude of the national CO2 sat emission in 2019 (unit, teragrams of CO2 per year). The color of the circle is used to distinguish the different countries. The correlation coefficient (r) is based on the Spearman rank correlation, and the P values are derived from a permutation test with 10,000 samples.
To attribute the grid cells of the newly identified CO2 emissions, we overlay them with an industrial heat source dataset (34), which is derived from time series of the Visible Infrared Imaging Radiometer Suite (VIIRS) Nightfire product and classifies industrial point sources based on temperature characteristics of nighttime thermal anomalies. This dataset divides industrial heat source plants into four subcategories (nonmetal, metal, coal processing, and oil and gas) plants (fig. S8). Specifically, we found that the monthly variation curve of these newly identified CO2 emissions belonging to oil and gas plants showed distinct seasonal characteristics with winter peaks and summer lows (Fig. 2D). This feature is consistent with the variation in monthly demand of natural gas in 45 countries reporting data to the Joint Organizations Data Initiative database (35–37). This phenomenon likely stems from CO2 emissions associated with natural gas operations, such as the flaring of fugitive methane and combustion emissions from gas-powered equipment (38, 39), all of which intensify during winter heating seasons.
The underreported BU CO2 emission at country scale and city scale
In addition to the ∼4.6% of newly identified CO2 emissions missing from the BU inventories (fig. S4), the BU inventories underestimated ∼8.4% of CO2 emissions in the study area, which are mainly from urban and industrial areas (a total underestimate of 13%, fig. S10 and red area in fig. S2). At the national scale, substantial underestimate is observed for Algeria (114 Mt year−1, 45%), Iran (314 Mt year−1, 34%), Egypt (262 Mt year−1, 50%), and Iraq (190 Mt year−1, 51%) (Fig. 4). At the same time, considerable variations exist among different BU inventories (fig. S9), particularly for Bahrain (66%), Yemen (52%), Equatorial Guinea (52%), Ethiopia (34%), and Lebanon (32%). While total European satellite-based estimates show the closest agreement with BU inventories overall, Germany, France, and Poland are overestimated by 13, 10, and 57%, respectively, whereas Italy and Spain are underestimated by 20 and 9% (Fig. 4B).
Fig. 4. Comparison of country-scale satellite-based CO2 emissions and BU inventories in Africa, the Middle East, and Europe in 2019.
(A) Top five countries in CO2 emissions for each region. (B) The absolute differences and percentage differences between satellite-based CO2 emissions and BU inventories. The error bars in (A) and (B) denote the range of ±SD of four calculations that represent the uncertainty of our method . “KSA” represents the Kingdom of Saudi Arabia.
At the city scale, the satellite-based CO2 estimates of urban emissions are generally higher than the reported values from BU inventories (fig. S10). This conclusion is consistent with the conclusion mentioned above that the top-down CO2 emissions are systematically higher than BU estimates. Because our TROPOMI NOx-to-CO2 approach depends on satellite-derived NOx (Eq. 2), the higher satellite-based NOx observed in cities compared to BU NOx directly accounts for the correspondingly higher CO2 estimates (fig. S11).
To further understand the causes of discrepancies between CO2 BU and CO2 Sat at the national and city scales, we analyzed the sectoral structure of both CO2 BU and NOx BU inventories for countries (Algeria, Iraq, Iran, Egypt, Turkey, Saudi Arabia, and Nigeria) with large inventory discrepancies (yellow bars in Africa and the Middle East in Fig. 4B). Because anthropogenic CO2 and NOx are nearly coemitted, the proportional contributions of various sectors to both inventories are expected to be consistent. However, we found structural discrepancies in the sectoral contributions between the CO2 BU and NOx BU inventories for these countries. As shown in Fig. 5A, we identified a notable disparity between the transport sector’s contribution to national NOx BU and CO2 BU emissions in multiple countries (e.g., Algeria, Egypt, and Iran). Although transportation is a major subsection of Iran’s total greenhouse emissions (40), it accounts for 64% of NOx BU emissions but only 20% of CO2 BU emissions (red and red-hatched bars in Fig. 5A). Furthermore, the positive correlation between the discrepancy in the transport sector contribution and the difference between CO2 Sat and CO2 BU suggests that biases in the transport sector may play an important role in explaining the observed discrepancies (Fig. 5C).
Fig. 5. Structural biases in sectoral emissions of BU CO2 inventories.
(A) The percentage contribution of four sectors (transport, industry, energy, and residential) to national CO2 (colored bars, left y axis) and NOx (hatched bars, right y axis) emissions in BU inventories. “KSA” represents the Kingdom of Saudi Arabia. The BU inventories are from CAMS. Panels (A) and (B) share the same legend, with bar colors representing the same emission sectors. (B) The difference between NOx BU and BU CO2 BU contribution percentages (NOx BU% − CO2 BU%, left y axis) for each sector, with positive values indicating disproportionately high NOx contribution relative to CO2. The magenta line with circles shows the ratio of NOx Sat to NOx BU emissions (right y axis), where values of >1 indicate NOx Sat > NOx BU. (C) Relationship between the disparity in the percentage contributions of the transport sector to NOx and CO2 emissions (NOx BU% − CO2 BU%, x axis) and the difference between CO2 Sat and CO2 BU (CO2 Sat − CO2 BU, y axis). Each circle represents a country, with the size scaling with the magnitude of the national CO2 sat emission in 2019. The color of the circle is used to distinguish the different countries. Relationships for the other three sectors with (CO2 Sat − CO2 BU) are shown in fig. S12. The correlation coefficient (r) is based on the Spearman rank correlation, and the P value is derived from a permutation test with 10,000 samples.
To further illustrate the implications of the CO2 emissions underestimation in the Middle East and Africa, we quantify its potential impact on meeting the NDC targets by 2030 (Fig. 6). Using the satellite-derived underestimation ratios to scenario-based future emission pathways, we find that the underreporting of emissions could cause delays in achieving these targets. As shown in Fig. 6, the time delay in meeting the NDC targets for the region could be ∼5.2 years. More specifically, Algeria, Nigeria, and Egypt are particularly vulnerable, with projected delays of 7.6, 6.8, and 6.3 years, respectively. These findings underline the urgent need for countries in the Middle East and Africa to enhance their emission inventory accounting systems. In regions where statistical data is outdated or incomplete, satellite-based estimates should be integrated as a key reference when setting NDC targets. Furthermore, these results can be used by international bodies including the UNFCCC to prioritize capacity-building and support in these regions to ensure that climate commitments are both credible and achievable.
Fig. 6. Impact of underestimated carbon emissions on delayed achievement of NDC targets in the Middle East and Africa.
Solid lines represent emission pathways based on the EN_INDCi2030_900 scenario from the TIAM-ECN 1.1 model (65) for the Middle East and Africa region and three countries (Algeria, Nigeria, and Egypt). The EN_INDCi2030_900 scenario builds on the implementation of current nationally determined contributions by 2030 within a 900-Gt CO2 cumulative emission constraint. The colored shaded areas represent the potential delays (in years) to achieving these 2030 targets, resulting from applying the 2019 scale of emission underestimates to the projected pathways. Emission pathway data are from the IPCC’s Sixth Assessment Report (AR6) Scenario Database. We selected emission scenario data for the Middle East and Africa region and for three countries, Algeria, Nigeria, and Egypt, as these countries represent cases where emissions in the study area are underestimated and national-scale scenario data are available.
DISCUSSION
At any given time, more than 10,000 gas flares are burning worldwide (7), more than half of which are located in Africa and the Middle East. Yet these emissions are underreported by oil companies in their official disclosures, leaving the true scale of carbon emissions in Africa and the Middle East largely unknown. This study provides a complementary method for carbon emission accounting driven by satellite observation data, using inversion techniques and the TROPOMI NOx-to-CO2 approach. It enriches the scientific database essential for clarifying global carbon budgets and enhancing transparency in climate governance. This approach expands the existing emission monitoring framework and establishes an independent verification tool that does not rely on spatial distribution information of BU inventories or proxy data. It enhances our ability to accurately track emissions, especially in regions where BU inventories are incomplete or outdated, playing a vital role in refining emission inventories and filling data gaps.
Specifically, this study shows that the CO2 emissions reported in existing BU inventories are systematically underestimated in the Middle East and the NHA regions. The spatial distribution and quantitative results of satellite-based CO2 emissions indicate that this underestimation is reflected in two aspects: (i) Satellite observations identify emission signals from sources missing in the current inventories. These newly identified sources show spatial colocation and correlation with oil and gas infrastructure across the Middle East and Africa. (ii) The CO2 Sat emissions at both national and city scales are higher than those reported in the CO2 BU inventories. Further analysis reveals notable structural biases in the sectoral contributions of BU CO2 and NOx inventories, which introduce uncertainties into the CO2-to-NOx emission factors.
This study finds that the underestimation of CO2 emissions in existing inventories for the Middle East and NHA should be given due attention. By applying the emission underestimation ratios derived from satellite data to scenario-based future emission pathways, we find that these underestimated emissions could lead to a delay of ∼5.2 years in meeting the 2030 NDC targets for the Middle East and Africa. This finding suggests that, in regions with outdated inventories, satellite-based estimates should be used as a key reference when setting NDC targets. Furthermore, because underestimations are mostly found in developing countries (Algeria, Iran, Egypt, and Iraq) with less capacities in compiling their emission inventory, global leaders from private and public sectors need to exert additional efforts to enhance the inventory-building capacity in these regions to ensure that climate commitments are both credible and achievable.
However, it is crucial to acknowledge the limitations of this study. First, although the spatial patterns and temporal resolution of CO2 Sat are based on satellite observations, the method is still limited by the reliability of existing inventories for emission factors, especially in developing countries and regions. However, sensitivity analyses indicate that the main conclusions of this study remain consistent across a range of emission factor assumptions. Future research could focus on reducing estimation uncertainties by directly obtaining emission factors from satellite observations. Specifically, the use of upcoming satellites, such as Copernicus Anthropogenic Carbon Dioxide Monitoring constellation (CO2M) and Twin Anthropogenic Greenhouse Gas Observers (TANGO), which can simultaneously observe both CO2 and NO2, will be crucial for refining emission factors and improving the accuracy of CO2 estimates. Through structural analysis of the inventories, we found substantial underestimation of CO2 emissions from the transport sector in the BU CO2 inventories of Middle Eastern and African countries, which results in a lower CO2-to-NOx ratio. Nevertheless, the fact that CO2 Sat for these countries still significantly exceed BU estimates suggests that actual CO2 emissions may be even higher than those estimated in this study. In addition, while recent studies have reported substantial biases in the absolute magnitude of the Global Fire Assimilation System (GFAS) fire emissions (41), the high spatial correlation between observations and GFAS fire emissions ensures that fire locations are accurately identified. Therefore, using GFAS as a spatial mask avoids the impact of these intensity biases on our results. Similarly, the TROPOMI overpass at about 13:30 local time may introduce representativeness uncertainty by missing diurnal variability in emissions. Applying the method to observations from geostationary satellites or instruments with varying overpass times could help mitigate this limitation in future studies.
MATERIALS AND METHODS
TROPOMI NO2 satellite observations
We use tropospheric NO2 observations with a resolution of 5.5 km by 7 km (5.5 km by 3.5 km after 6 August 2019) from the TROPOMI on the Sentinel-5P satellite, which has an overpass time of about 13:30 local time (42). Sentinel-5p provides daily global coverage, but observations need to be filtered to exclude cloudy pixels. In this study, we use the version 2.4 reprocessed Level 2 TROPOMI NO2 product for the year 2019 (43).
To minimize biases linked to free-tropospheric contributions to the column, the partial column from the land surface up to the 500-hPa level is calculated from the TROPOMI NO2 retrieval (29, 44). The area-weighted averages of tropospheric NO2 observations are processed into super-observations of 0.2° × 0.2° grid cells to reduce the representativeness errors compared to individual observations (45–48).
NOx emissions from DECSO inversion algorithm
DECSO is a state-of-the-art inversion algorithm developed at the Royal Netherlands Meteorological Institute (KNMI) with the aim of daily updates of short-lived atmospheric composition emissions (28, 29, 49). DECSO assimilates emissions with an extended Kalman filter based on the comparison of satellite observed TROPOMI NO2 and simulated concentrations driven by the chemical transport model CHIMERE v2020r3 (50). DECSO can detect newly emerged emission sources without relying on a priori emissions. The sensitivity of concentrations to emissions is determined by using multiple forward two-dimensional (2D) trajectories to calculate the transport of the species (29). DECSO has been successfully applied to provide NOx and NH3 emissions for different regions (46, 51–56). We use DECSO-derived monthly total NOx emissions (NOx total DECSO), version 6.4, for 2019 over Europe, the Middle East, and Africa (fig. S13) at a resolution of 0.2° × 0.2°. In a postprocessing step, the total NOx emissions are separated into soil NOx and anthropogenic NOx emissions (27, 29). Our former proposed method derives soil NOx emissions (NOx soil DECSO) of each biogenic sector by fitting monthly average emissions of selected grid cells, which are dominated by forest, cropland, and shrubland land cover type according to the land-use dataset obtained from the Copernicus Global Land Service (57). The gridded monthly soil NOx emissions are determined by multiplying the land use fraction by the corresponding biogenic sector soil NOx emissions. After subtracting the soil NOx emissions from the total NOx emissions, the remaining NOx are anthropogenic NOx emissions (NOx Sat). Both soil and anthropogenic NOx emissions are validated in previous studies (27, 29).
In Central Africa (subregions D and E in fig. S13), NOx Sat are affected by biomass burning emissions (BBEs). We apply CAMS global monthly BBE based on fire radiative power (GFAS, version 1.2) (58) as a mask to exclude biomass burning influencing in our analysis. We assume that there are almost no other anthropogenic emissions in a single grid cell where the fires occurred. Biomass burning–influenced NOx emissions are identified by marking grid cells where monthly GFAS is greater than 0, shown in fig. S15C.
Across the entire study region, the DECSO-derived total NOx is decomposed into three components: ∼40% NOx soil DECSO, 49% anthropogenic NOx Sat emissions, and 11% BBE-influenced NOx Sat (fig. S14). Specifically, the contribution of NOx soil DECSO varies across subregions, accounting for 55% in NHA, 23% in the Middle East, and 26% in Europe. This indicates that soil NOx and BBE-influenced NOx Sat constitute a substantial fraction of total NOx and, if not removed, would significantly interfere with the estimates of anthropogenic NOx Sat. In our previous study (27), the uncertainty of NOx soil DECSO was quantified to be about 37%. In terms of magnitude, NOx soil DECSO is ∼2.7 times larger than CAMS soil NOx, whereas NOx Sat is comparable to CAMS anthropogenic NOx emissions (about 1.1 times) (figs. S15 and S16). Spatially, NOx soil DECSO exhibits higher values over cropland and forested regions, whereas anthropogenic NOx is concentrated in urban and industrial areas, showing clearly distinct spatial patterns (fig. S14, B and D). After removing BBE-influenced NOx Sat and NOx soil DECSO, the resulting anthropogenic NOx Sat shows improved agreement with NOx BU in both spatial distribution and magnitude (figs. S14D and S15D; solid and dashed red lines in fig. S16). This improvement indicates that the applied source separation method effectively removes nonanthropogenic contributions and improves the accuracy of NOx Sat. Therefore, the NOx Sat used in this study to infer CO2 emissions has been filtered to exclude contributions from wildfire and soil sources (e.g., soil microbial processes, agricultural fertilization, and nitrogen deposition).
BU emission inventories for CO2-to-NOx ratios
We use global monthly BU anthropogenic NOx (NOx BU) and CO2 (CO2 BU) emissions in 2019 to calculate CO2-to-NOx ratios both from CAMS (CAMS-GLOB-ANT version 6.2 for CAMS-NOx and CAMS-CO2) (59) and Emissions Database for Global Atmospheric Research (EDGAR; version 8.1 for EDGAR-NOx and version 8.0 for EDGAR-CO2) (32). CAMS-GLOB-ANT are used in forecasts and reanalysis to provide emissions for the most recent years (59). EDGAR emissions at a resolution of 0.1° × 0.1° are used as a basis for the CAMS emissions and then are extrapolated to the current year using trends derived from the CEDS totals per country and sector (59, 60). CAMS-GLOB-ANT v6.2 is derived from EDGAR v6 and CEDS emissions. We choose CAMS and EDGAR inventories to calculate CO2-to-NOx ratios because both provide CO2 and NOx emissions for the year of 2019.
Other global gridded CO2 emissions datasets from ODIAC (61), GID (62), and GRACED (63, 64) are used for comparison. All emission inventories are regridded on a 0.2° × 0.2° grid the same as NOx total DECSO. Unless otherwise specified, all numerical comparisons between satellite-based CO2 emissions and BU CO2 inventories in this study refer exclusively to anthropogenic CO2 emissions over land. Shipping CO2 emissions over ocean are only presented in Fig. 1B.
The CO2-to-NOx ratio () follows the equation below. We calculate two types of : grid ratio and country ratio. The grid ratio refers to the ratio of CO2 to NOx within each single grid cell (fig. S17, A and C). The country ratio refers to the average ratio of a single country by dividing the national total by the national total (fig. S17, B and D). Two types of CO2-to-NOx ratios (grid ratio and country ratio) from CAMS and EDGAR generate in total four ratios to infer CO2 Sat
| (1) |
Note that all BU CO2 emission inventories referenced in this study represent fossil fuel combustion emissions, which are used for both CO2-to-NOx ratio calculations and comparisons with satellite-based CO2 estimates.
Satellite-based CO2 emissions inferred from TROPOMI NOx-to-CO2 approach
We infer anthropogenic CO2 emissions (CO2 Sat) based on NOx Sat and emission factor ratio and the following equation
| (2) |
In this study, we apply the grid ratio and country ratio of to test the variability of our method. The estimates of national CO2 emissions would be identical if we multiply BU NOx by grid CO2-to-NOx ratio or country ratio of the BU inventory.
The flowchart of our approach is shown in fig. S18. The derived CO2 Sat are compared with BU CO2 inventories on a city scale and a country scale. In our study area, we selected the top 60 cities with an urban area greater than 500 km2. The urban area shapefile was accessed from www.naturalearthdata.com/downloads/10m-cultural-vectors/10m-urban-area/. The countries are divided into three regions (Europe, the Middle East, and NHA) for analysis. Note that all references to “Africa” in this work denote the Northern Hemispheric Africa region as delineated in fig. S13.
The newly identified CO2 emissions from satellite-based CO2 emission
In Eq. 2, extremely low BU CO2 or BU NOx cause anomalous , which subsequently lead to unrealistic satellite-based CO2 emission. Here, we replace the grid ratio with BU NOx lower than 0.02 Gg NO2 year−1 (fig. S20B) or with a ratio lower than 40 (fig. S20C) by the country ratio, because these values are probably caused by incomplete or very uncertain information on either the BU CO2 or the BU NOx emissions, which only occur in remote uninhabited locations in the desert or high mountains (fig. S20A). The minimum ratio threshold is determined by histogram statistics of in Europe (fig. S19). To avoid unrealistic CO2 emissions caused by NOx Sat noise, the with NOx Sat lower than 0.05 Gg NO2 year−1 are set to zero (fig. S21). The noise threshold of DECSO is determined by the theoretical minimum of the emission covariance following the “observations minus analysis” statistics of the assimilation results (28, 49). In general, the grid cells with replaced ratio (fig. S21B) generate about 4.6% CO2 emissions (bright red bars in Fig. 3A), which we define as newly identified sources. These satellite-derived CO2 emissions corresponding to grid cells where country ratios replaced grid-specific ratios are designated as the newly identified CO2 emissions (CO2 Sat new) in this study.
Uncertainties of estimated CO2 emissions
We determine the uncertainty of CO2 Sat by three uncorrelated components , , and . represents the uncertainty from the DECSO algorithm and is estimated to be about 25% for monthly emissions based on model errors, observation errors, and representation errors (29). refers to the SD of four CO2-to-NOx ratios for deriving CO2 Sat emissions and is about 10% (fig. S22). represents the uncertainty of CO2-to-NOx ratios and is calculated by the average SD of sector ratios from transport (17%), residential (42%), industry (35%), and energy (41%) sectors, which is 33% (fig. S23). The total uncertainty of CO2 Sat is estimated following error propagation as
| (3) |
We estimate the total uncertainty of the CO2 Sat monthly emissions to be ∼43% for an individual grid cell. If all three error components are assumed to be random, the uncertainty of annual emissions for a grid cell would be 12% due to temporal averaging. However, if is assumed as a systematic error that remains constant the year, then the annual uncertainty is instead estimated to be 33%. In practice, may exhibit some seasonal variability due to changes in fuel composition across different months, and, therefore, the annual uncertainty is expected to lie between these two bounds. Regardless of these grid-level assumptions, the uncertainty in country-level or regional total emissions is much smaller due to the spatial cancellation of independent errors.
Our base estimates are derived using the CAMS grid ratio. For the newly identified CO2 emissions where specific ratios are unknown, we applied the CAMS country ratio for the conversion. However, for those CO2 Sat new that are likely associated with oil and gas infrastructure, the actual ratio may deviate substantially from the country-average values. To assess the sensitivity of CO2 Sat new to these ratio variations, we perform estimates using sector-specific ratios from CAMS (energy, industry, residential, and transport), as well as country ratio from the EDGAR inventory (table S2). These ratios span a range from low emission intensity sectors (e.g., transport) to high emission intensity sectors (e.g., industry) and partly represent the possible range associated with different emission sources, including oil and gas activities. The results show that the estimated CO2 Sat new varies from 0.9 to 3.3 times the base estimates. The highest values are obtained using the industry-sector ratio, whereas the lowest values are derived from the transport-sector ratio. For Africa and the Middle East, the estimated underestimation remains within the range of 29 to 40%, and the contribution of CO2 Sat new to the total underestimated emissions ranges from 33 to 69%. Overall, although the absolute magnitude of emissions depends on the choice of ratio, the conclusion that BU inventories underestimate emissions remain consistent across different ratio assumptions, indicating the robustness of our results.
Quantification of emission underestimation impact on NDC targets
To quantify the impact of CO2 emission underestimation on achieving NDC targets, we developed a scenario-based approach to quantify the delay risk in NDC targets. First, the country-specific or specific-region emission underestimation ratios for 2019 were calculated as the percentage difference between our satellite-based CO2 emission estimates CO2 Sat and the corresponding BU inventories (CAMS). The underestimation ratio Ui for country (or region) i is defined as
| (4) |
Then, we selected a benchmark future emission pathway reflective of current climate policy commitments. The “EN_INDCi2030_900” scenario from the TIAM-ECN 1.1 integrated assessment model from IPCC Sixth Assessment Report (AR6) Scenario Database was chosen as our policy-reference pathway. This pathway represents a future where current NDCs are implemented by 2030 under a global carbon budget of 900 Gt CO2. We extracted the projected reference emissions for the Middle East and Africa region, as well as for Algeria, Nigeria, and Egypt, which were selected due to their substantial emission underestimation and the availability of national-scale scenario data. The calculated underestimation ratios Ui applied to these pathways were 44.7% for the Middle East and Africa region and 81.0, 34.6, and 100.0% for Algeria, Nigeria, and Egypt, respectively.
Third, we calculated an adjusted emission pathway to account for systematic reporting biases. For each year in the projection (2030–2050), we applied the Ui to the model’s projected reference emissions ENDC900, generating an adjusted emission pathway ESatNDC900 that incorporates the persistent underestimation
| (5) |
This calculation assumes that the relative underestimation bias in the inventory reporting remains constant without corrective intervention. Last, the time delay in achieving the NDC target was quantified. The target emission level for 2030, ENDC900(2030), is defined by the model’s policy-reference pathway. We then identified the calendar year in which the adjusted pathway ESatNDC900 is projected to lastly decrease to ENDC900(2030). The difference between this calendar year and the target year of 2030 represents the projected delay (in years) attributable to the emission underestimation.
Acknowledgments
We thank the team that realized the TROPOMI instrument and its operational data products, consisting of the partnership between Airbus Defense and Space Netherlands, KNMI, SRON, and TNO, commissioned by NSO and ESA. We thank Z. Deng and K. He from the University of Hong Kong for the revision of the figures and structure of the paper. We thank P. Ciais from Laboratoire des Sciences du Climat et de l’Environnement for providing support on data analysis. We thank M. Zhou from Institute of Atmospheric Physics, Chinese Academy of Sciences, for suggestions on improving the quality of the Figures. We thank Q. Zhong from Peking University for providing advice about data processing.
Funding:
This work was supported by the National Key R&D Program of China, grant no. 2023YFE0113000 to Z.L.; National Natural Science Foundation of China, grant no. T2525015 to Z.L.; and Carbon Neutrality and Energy System Transformation (CNEST) program to Z.L.
Author contributions:
Conceptualization: X.L., R.v.d.A., J.d.L., and Z.L. Data curation: X.L. Formal analysis: X.L. Funding acquisition: Z.L. Investigation: X.L., R.v.d.A., J.D., and Z.L. Methodology: X.L., R.v.d.A., and J.d.L. Project administration: R.v.d.A. and Z.L. Resources: X.L., R.v.d.A., J.D., H.E., and Z.L. Software: X.L., R.v.d.A., and J.D. Supervision: R.v.d.A., J.d.L., and Z.L. Validation: X.L. Visualization: X.L. Writing—original draft: X.L., J.d.L., H.E., and Z.L. Writing—review and editing: X.L., R.v.d.A., J.d.L., J.D., H.E., and Z.L.
Competing interests:
The authors declare that they have no competing interests.
Data, code, and materials availability:
TROPOMI data are available at www.tropomi.eu/data-products/nitrogen-dioxide. NOx emissions derived by CAMS soil NOx emissions are available at https://permalink.aeris-data.fr/CAMS-GLOB-SOIL. CAMS anthropogenic NOx and CO2 emissions are obtained from ECCAD: https://permalink.aeris-data.fr/CAMS-GLOB-ANT. EDGAR anthropogenic NOx and CO2 emissions are available at https://edgar.jrc.ec.europa.eu/emissions_data_and_maps. DECSO emissions are available via www.temis.nl/emissions/decso_nox.php. All data and code needed to evaluate and reproduce the results in the paper are present in the paper and/or the Supplementary Materials. This study did not generate new materials.
Supplementary Materials
This PDF file includes:
Figs. S1 to S23
Tables S1 and S2
REFERENCES
- 1.International Energy Agency (IEA), “Africa energy outlook 2022” (IEA, 2022); www.iea.org/reports/africa-energy-outlook-2022.
- 2.Organization of the Petroleum Exporting Countries (OPEC), “Annual statistical bulletin” (OPEC, 2024); https://publications.opec.org/asb/Home.
- 3.Friedlingstein P., O’Sullivan M., Jones M. W., Andrew R. M., Hauck J., Landschützer P., Le Quéré C., Li H., Luijkx I. T., Olsen A., Peters G. P., Peters W., Pongratz J., Schwingshackl C., Sitch S., Canadell J. G., Ciais P., Jackson R. B., Alin S. R., Arneth A., Arora V., Bates N. R., Becker M., Bellouin N., Berghoff C. F., Bittig H. C., Bopp L., Cadule P., Campbell K., Chamberlain M. A., Chandra N., Chevallier F., Chini L. P., Colligan T., Decayeux J., Djeutchouang L. M., Dou X., Duran Rojas C., Enyo K., Evans W., Fay A. R., Feely R. A., Ford D. J., Foster A., Gasser T., Gehlen M., Gkritzalis T., Grassi G., Gregor L., Gruber N., Gürses Ö., Harris I., Hefner M., Heinke J., Hurtt G. C., Iida Y., Ilyina T., Jacobson A. R., Jain A. K., Jarníková T., Jersild A., Jiang F., Jin Z., Kato E., Keeling R. F., Klein Goldewijk K., Knauer J., Korsbakken J. I., Lan X., Lauvset S. K., Lefèvre N., Liu Z., Liu J., Ma L., Maksyutov S., Marland G., Mayot N., McGuire P. C., Metzl N., Monacci N. M., Morgan E. J., Nakaoka S. I., Neill C., Niwa Y., Nützel T., Olivier L., Ono T., Palmer P. I., Pierrot D., Qin Z., Resplandy L., Roobaert A., Rosan T. M., Rödenbeck C., Schwinger J., Smallman T. L., Smith S. M., Sospedra-Alfonso R., Steinhoff T., Sun Q., Sutton A. J., Séférian R., Takao S., Tatebe H., Tian H., Tilbrook B., Torres O., Tourigny E., Tsujino H., Tubiello F., van der Werf G., Wanninkhof R., Wang X., Yang D., Yang X., Yu Z., Yuan W., Yue X., Zaehle S., Zeng N., Zeng J., Global carbon budget 2024. Earth Syst. Sci. Data 17, 965–1039 (2025). [Google Scholar]
- 4.Palmer P. I., Feng L., Baker D., Chevallier F., Bösch H., Somkuti P., Net carbon emissions from African biosphere dominate pan-tropical atmospheric CO2 signal. Nat. Commun. 10, 3344 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hubau W., Lewis S. L., Phillips O. L., Affum-Baffoe K., Beeckman H., Cuní-Sanchez A., Daniels A. K., Ewango C. E. N., Fauset S., Mukinzi J. M., Sheil D., Sonké B., Sullivan M. J. P., Sunderland T. C. H., Taedoumg H., Thomas S. C., White L. J. T., Abernethy K. A., Adu-Bredu S., Amani C. A., Baker T. R., Banin L. F., Baya F., Begne S. K., Bennett A. C., Benedet F., Bitariho R., Bocko Y. E., Boeckx P., Boundja P., Brienen R. J. W., Brncic T., Chezeaux E., Chuyong G. B., Clark C. J., Collins M., Comiskey J. A., Coomes D. A., Dargie G. C., de Haulleville T., Kamdem M. N. D., Doucet J.-L., Esquivel-Muelbert A., Feldpausch T. R., Fofanah A., Foli E. G., Gilpin M., Gloor E., Gonmadje C., Gourlet-Fleury S., Hall J. S., Hamilton A. C., Harris D. J., Hart T. B., Hockemba M. B. N., Hladik A., Ifo S. A., Jeffery K. J., Jucker T., Yakusu E. K., Kearsley E., Kenfack D., Koch A., Leal M. E., Levesley A., Lindsell J. A., Lisingo J., Lopez-Gonzalez G., Lovett J. C., Makana J.-R., Malhi Y., Marshall A. R., Martin J., Martin E. H., Mbayu F. M., Medjibe V. P., Mihindou V., Mitchard E. T. A., Moore S., Munishi P. K. T., Bengone N. N., Ojo L., Ondo F. E., Peh K. S. H., Pickavance G. C., Poulsen A. D., Poulsen J. R., Qie L., Reitsma J., Rovero F., Swaine M. D., Talbot J., Taplin J., Taylor D. M., Thomas D. W., Toirambe B., Mukendi J. T., Tuagben D., Umunay P. M., van der Heijden G. M. F., Verbeeck H., Vleminckx J., Willcock S., Wöll H., Woods J. T., Zemagho L., Asynchronous carbon sink saturation in African and Amazonian tropical forests. Nature 579, 80–87 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.The World Bank, “Global gas flaring data” (The World Bank, 2025); www.worldbank.org/en/programs/gasflaringreduction/global-flaring-data.
- 7.“Gas flaring: What is it and why is it a problem?,” BBC, 29 September 2022; www.bbc.com/news/science-environment-63051458.
- 8.The World Bank, “Global gas flaring tracker report 2025” (The World Bank, 2025): www.worldbank.org/en/programs/gasflaringreduction/publication/2025-global-gas-flaring-tracker-report.
- 9.“Toxic gas putting millions at risk in Middle East, BBC finds,” BBC, 28 November 2023; www.bbc.com/news/science-environment-67522413.
- 10.Zhizhin M., Matveev A., Ghosh T., Hsu F.-C., Howells M., Elvidge C., Measuring gas flaring in Russia with multispectral VIIRS Nightfire. Remote Sens. 13, 3078 (2021). [Google Scholar]
- 11.Intergovernmental Panel on Climate Change (IPCC), in 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC National Greenhouse Gas Inventories Programme, Institute for Global Environmental Strategies, 2006), pp. 2108–2111.
- 12.UNFCCC, “Decision 18/CMA.1. Modalities, procedures and guidelines for the transparency framework for action and support referred to in Article 13 of the Paris Agreement” (FCCC/PA/CMA/2018/Add.2, 2018).
- 13.Deng Z., Ciais P., Tzompa-Sosa Z. A., Saunois M., Qiu C., Tan C., Sun T., Ke P., Cui Y., Tanaka K., Lin X., Thompson R. L., Tian H., Yao Y., Huang Y., Lauerwald R., Jain A. K., Xu X., Bastos A., Sitch S., Palmer P. I., Lauvaux T., d’Aspremont A., Giron C., Benoit A., Poulter B., Chang J., Petrescu A. M. R., Davis S. J., Liu Z., Grassi G., Albergel C., Tubiello F. N., Perugini L., Peters W., Chevallier F., Comparing national greenhouse gas budgets reported in UNFCCC inventories against atmospheric inversions. Earth Syst. Sci. Data 14, 1639–1675 (2022). [Google Scholar]
- 14.IPCC, “Emissions trends and drivers,” in IPCC 2022: Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, A. R. Shukla, J. Skea, R. Slade, A. Al Khourdajie, R. van Diemen, D. McCollum, M. Pathak, S. Some, P. Vyas, R. Fradera, M. Belkacemi, A. Hasija, G. Lisboa, S. Luz, J. Malley, Eds. (Cambridge Univ. Press, 2022). chap. 2, pp. 215–294.
- 15.Hogue S., Marland E., Andres R. J., Marland G., Woodard D., Uncertainty in gridded CO2 emissions estimates. Earth’s Future 4, 225–239 (2016). [Google Scholar]
- 16.Super I., Dellaert S. N. C., Visschedijk A. J. H., Denier van der Gon H. A. C., Uncertainty analysis of a European high-resolution emission inventory of CO2 and CO to support inverse modelling and network design. Atmos. Chem. Phys. 20, 1795–1816 (2020). [Google Scholar]
- 17.Lin X., van der A R., de Laat J., Eskes H., Chevallier F., Ciais P., Deng Z., Geng Y., Song X., Ni X., Huo D., Dou X., Liu Z., Monitoring and quantifying CO2 emissions of isolated power plants from space. Atmos. Chem. Phys. 23, 6599–6611 (2023). [Google Scholar]
- 18.Nassar R., Hill T. G., McLinden C. A., Wunch D., Jones D. B. A., Crisp D., Quantifying CO2 emissions from individual power plants from space. Geophys. Res. Lett. 44, 10045–10053 (2017). [Google Scholar]
- 19.Schwandner F. M., Gunson M. R., Miller C. E., Carn S. A., Eldering A., Krings T., Verhulst K. R., Schimel D. S., Nguyen H. M., Crisp D., O’Dell C. W., Osterman G. B., Iraci L. T., Podolske J. R., Spaceborne detection of localized carbon dioxide sources. Science 358, eaam5782 (2017). [DOI] [PubMed] [Google Scholar]
- 20.Nassar R., Mastrogiacomo J.-P., Bateman-Hemphill W., McCracken C., MacDonald C. G., Hill T., O’Dell C. W., Kiel M., Crisp D., Advances in quantifying power plant CO2 emissions with OCO-2. Remote Sens. Environ. 264, 112579 (2021). [Google Scholar]
- 21.He C., Lu X., Zhang Y., Liu Z., Jiang F., Sun Y., Gao M., Liu Y., Lin H., Yang J., Lin X., Wang Y., Hu C., Fan S., Revisiting the quantification of power plant CO2 emissions in the United States and China from satellite: A comparative study using three top-down approaches. Remote Sens. Environ. 308, 114192 (2024). [Google Scholar]
- 22.Eldering A., Wennberg P. O., Crisp D., Schimel D. S., Gunson M. R., Chatterjee A., Liu J., Schwandner F. M., Sun Y., O’Dell C. W., Frankenberg C., Taylor T., Fisher B., Osterman G. B., Wunch D., Hakkarainen J., Tamminen J., Weir B., The Orbiting Carbon Observatory-2 early science investigations of regional carbon dioxide fluxes. Science 358, eaam5745 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Leguijt G., Maasakkers J. D., Denier van der Gon H. A. C., Segers A. J., Borsdorff T., Aben I., Quantification of carbon monoxide emissions from African cities using TROPOMI. Atmos. Chem. Phys. 23, 8899–8919 (2023). [Google Scholar]
- 24.Yang E. G., Kort E. A., Wu D., Lin J. C., Oda T., Ye X., Lauvaux T., Using space-based observations and lagrangian modeling to evaluate urban carbon dioxide emissions in the middle east. J. Geophys. Res. Atmos. 125, e2019JD031922 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Feng S., Jiang F., Wang H., Liu Y., He W., Wang H., Shen Y., Zhang L., Jia M., Ju W., Chen J. M., China’s fossil fuel CO2 emissions estimated using surface observations of coemitted NO2. Environ. Sci. Technol. 58, e2019JD031922 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhang Y., Du S., Guan L., Chen X., Lei L., Liu L., Estimating global 0.1° scale gridded anthropogenic CO2 emissions using TROPOMI NO2 and a data-driven method. Sci. Total Environ. 949, 175177 (2024). [DOI] [PubMed] [Google Scholar]
- 27.Lin X., van der A R., de Laat J., Huijnen V., Mijling B., Ding J., Eskes H., Douros J., Liu M., Zhang X., Liu Z., European soil NOx emissions derived from satellite NO2 observations. J. Geophys. Res. Atmos. 129, e2024JD041492 (2024). [Google Scholar]
- 28.Mijling B., van der A R. J., Using daily satellite observations to estimate emissions of short-lived air pollutants on a mesoscopic scale. J. Geophys. Res. Atmos. 117, 10.1029/2012JD017817 (2012). [Google Scholar]
- 29.van der A R. J., Ding J., Eskes H., Monitoring European anthropogenic NOx emissions from space. Atmos. Chem. Phys. 24, 7523–7534 (2024). [Google Scholar]
- 30.Kuenen J., Dellaert S., Visschedijk A., Jalkanen J. P., Super I., Denier van der Gon H., CAMS-REG-v4: A state-of-the-art high-resolution European emission inventory for air quality modelling. Earth Syst. Sci. Data 14, 491–515 (2022). [Google Scholar]
- 31.Valentini R., Arneth A., Bombelli A., Castaldi S., Cazzolla Gatti R., Chevallier F., Ciais P., Grieco E., Hartmann J., Henry M., Houghton R. A., Jung M., Kutsch W. L., Malhi Y., Mayorga E., Merbold L., Murray-Tortarolo G., Papale D., Peylin P., Poulter B., Raymond P. A., Santini M., Sitch S., Vaglio Laurin G., van der Werf G. R., Williams C. A., Scholes R. J., A full greenhouse gases budget of Africa: Synthesis, uncertainties, and vulnerabilities. Biogeosciences 11, 381–407 (2014). [Google Scholar]
- 32.Crippa M., Guizzardi D., Pagani F., Schiavina M., Melchiorri M., Pisoni E., Graziosi F., Muntean M., Maes J., Dijkstra L., Van Damme M., Clarisse L., Coheur P., Insights into the spatial distribution of global, national, and subnational greenhouse gas emissions in the Emissions Database for Global Atmospheric Research (EDGAR v8.0). Earth Syst. Sci. Data 16, 2811–2830 (2024). [Google Scholar]
- 33.Riess T. C. V. W., Boersma K. F., Prummel A., van Stratum B. J. H., de Laat J., van Vliet J., Estimating NOx emissions of individual ships from TROPOMI NO2 plumes. Remote Sens. Environ. 324, 114734 (2025). [Google Scholar]
- 34.Liu Y., Hu C., Zhan W., Sun C., Murch B., Ma L., Identifying industrial heat sources using time-series of the VIIRS Nightfire product with an object-oriented approach. Remote Sens. Environ. 204, 347–365 (2018). [Google Scholar]
- 35.International Energy Forum (IEF), “Monthly oil & gas data review” (IEF, 2024); www.ief.org/_resources/files/news/monthly-oil--gas-review/07-24.pdf.
- 36.US Energy Infromation Administration, Global natural gas market may experience a tighter supply-demand balance this winter (2024); www.eia.gov/todayinenergy/detail.php?id=63804.
- 37.Eurostat, Energy statistics - Latest trends from monthly data (2024); https://web.archive.org/web/20250606223859/https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Energy_statistics_-_latest_trends_from_monthly_data.
- 38.Gorchov Negron A. M., Kort E. A., Chen Y., Brandt A. R., Smith M. L., Plant G., Ayasse A. K., Schwietzke S., Zavala-Araiza D., Hausman C., Adames-Corraliza Á. F., Excess methane emissions from shallow water platforms elevate the carbon intensity of US Gulf of Mexico oil and gas production. Proc. Natl. Acad. Sci. U.S.A. 120, e2215275120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Du Y., Wang Y., He S., Yuan H., Xu W., Song Y., Liu B., Xu Y., Collaborative control of pollution and carbon reduction in the process of onshore oil and gas exploitation. J. Environ. Eng. Technol. 14, 371–378 (2024). [Google Scholar]
- 40.Heidari E., Bikdeli S., Mansouri Daneshvar M. R., A dynamic model for CO2 emissions induced by urban transportation during 2005–2030, a case study of Mashhad, Iran. Environ. Develop. Sustain. 25, 4217–4236 (2023). [Google Scholar]
- 41.de Laat A. T. J., Andela N., Forkel M., Huijnen V., Kinalczyk D., van Wees D., Sentinel-5p reveals unexplained large wildfire carbon emissions in the Amazon in 2024. Geophys. Res. Lett. 53, e2025GL115123 (2026). [Google Scholar]
- 42.Veefkind J. P., Aben I., McMullan K., Förster H., de Vries J., Otter G., Claas J., Eskes H. J., de Haan J. F., Kleipool Q., van Weele M., Hasekamp O., Hoogeveen R., Landgraf J., Snel R., Tol P., Ingmann P., Voors R., Kruizinga B., Vink R., Visser H., Levelt P. F., TROPOMI on the ESA Sentinel-5 Precursor: A GMES mission for global observations of the atmospheric composition for climate, air quality and ozone layer applications. Remote Sens. Environ. 120, 70–83 (2012). [Google Scholar]
- 43.J. H. G. M. V. Geffen, H. J. Eskes, K. F. Boersma, J. P. Veefkind, “TROPOMI ATBD of the total and tropospheric NO2 data products” (S5P-KNMI-L2–0005-RP, Royal Netherlands Meteorological Institute, 2022).
- 44.Douros J., Eskes H., van Geffen J., Boersma K. F., Compernolle S., Pinardi G., Blechschmidt A.-M., Peuch V.-H., Colette A., Veefkind P., Comparing Sentinel-5P TROPOMI NO2 column observations with the CAMS regional air quality ensemble. Geosci. Model Dev. 16, 509–534 (2023). [Google Scholar]
- 45.Boersma K. F., Vinken G. C. M., Eskes H. J., Representativeness errors in comparing chemistry transport and chemistry climate models with satellite UV-Vis tropospheric column retrievals. Geosci. Model Dev. 9, 875–898 (2016). [Google Scholar]
- 46.Ding J., van der A R. J., Eskes H. J., Mijling B., Stavrakou T., van Geffen J. H. G. M., Veefkind J. P., NOx emissions reduction and rebound in China due to the COVID-19 crisis. Geophys. Res. Lett. 47, e2020GL089912 (2020). [Google Scholar]
- 47.Miyazaki K., Eskes H. J., Sudo K., Global NOx emission estimates derived from an assimilation of OMI tropospheric NO2 columns. Atmos. Chem. Phys. 12, 2263–2288 (2012). [Google Scholar]
- 48.Rijsdijk P., Eskes H., Dingemans A., Boersma F., Sekiya T., Miyazaki K., Houweling S., Quantifying uncertainties of satellite NO2 superobservations for data assimilation and model evaluation. Geosci. Model Dev. 18, 483–509 (2025). [Google Scholar]
- 49.Ding J., van der A R. J., Mijling B., Levelt P. F., Space-based NOx emission estimates over remote regions improved in DECSO. Atmos. Meas. Tech. 10, 925–938 (2017). [Google Scholar]
- 50.Menut L., Bessagnet B., Briant R., Cholakian A., Couvidat F., Mailler S., Pennel R., Siour G., Tuccella P., Turquety S., Valari M., The CHIMERE v2020r1 online chemistry-transport model. Geosci. Model Dev. 14, 6781–6811 (2021). [Google Scholar]
- 51.Ding J., Miyazaki K., van der A R. J., Mijling B., Kurokawa J.-I., Cho S., Janssens-Maenhout G., Zhang Q., Liu F., Levelt P. F., Intercomparison of NOx emission inventories over East Asia. Atmos. Chem. Phys. 17, 10125–10141 (2017). [Google Scholar]
- 52.van der A R. J., de Laat A. T. J., Ding J., Eskes H. J., Connecting the dots: NOx emissions along a West Siberian natural gas pipeline. Npj Clim. Atmos. Sci. 3, 16 (2020). [Google Scholar]
- 53.Ding J., van der A R. J., Mijling B., Jalkanen J. P., Johansson L., Levelt P. F., Maritime NOx emissions over Chinese seas derived from satellite observations. Geophys. Res. Lett. 45, 2031–2037 (2018). [Google Scholar]
- 54.Ding J., van der A R., Mijling B., de Laat J., Eskes H., Boersma K. F., NOx emissions in India derived from OMI satellite observations. Atmos. Environ. 14, 100174 (2022). [Google Scholar]
- 55.Ding J., van der A R. J., Mijling B., Levelt P. F., Hao N., NOx emission estimates during the 2014 Youth Olympic Games in Nanjing. Atmos. Chem. Phys. 15, 9399–9412 (2015). [Google Scholar]
- 56.Ding J., van der A R., Eskes H., Dammers E., Shephard M., Wichink Kruit R., Guevara M., Tarrason L., Ammonia emission estimates using CrIS satellite observations over Europe. EGUsphere 24, 10583–10599 (2024). [Google Scholar]
- 57.M. Buchhorn, B. Smets, L. Bertels, B. De Roo, M. Lesiv, N. E. Tsendbazar, L. Linlin, A. Tarko, Copernicus Global Land Service: Land Cover 100m: version 3 Globe 2015–2019: Product User Manual, Zenodo (2020).
- 58.Kaiser J. W., Heil A., Andreae M. O., Benedetti A., Chubarova N., Jones L., Morcrette J. J., Razinger M., Schultz M. G., Suttie M., van der Werf G. R., Biomass burning emissions estimated with a global fire assimilation system based on observed fire radiative power. Biogeosciences 9, 527–554 (2012). [Google Scholar]
- 59.Soulie A., Granier C., Darras S., Zilbermann N., Doumbia T., Guevara M., Jalkanen J. P., Keita S., Liousse C., Crippa M., Guizzardi D., Hoesly R., Smith S. J., Global anthropogenic emissions (CAMS-GLOB-ANT) for the Copernicus Atmosphere Monitoring Service simulations of air quality forecasts and reanalyses. Earth Syst. Sci. Data. 16, 2261–2279 (2024). [Google Scholar]
- 60.C. Granier, C. Liousse, B. McDonald, P. Middleton, “Anthropogenic emissions inventories of air pollutants,” Handbook of Air Quality and Climate Change, H. Akimoto, H. Tanimoto, Eds. (Springer Nature Singapore, 2023), pp. 3–52. [Google Scholar]
- 61.Oda T., Maksyutov S., Andres R. J., The Open-source Data Inventory for Anthropogenic CO2, version 2016 (ODIAC2016): A global monthly fossil fuel CO2 gridded emissions data product for tracer transport simulations and surface flux inversions. Earth Syst. Sci. Data. 10, 87–107 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Tong D., Zhang Q., Davis S. J., Liu F., Zheng B., Geng G., Xue T., Li M., Hong C., Lu Z., Streets D. G., Guan D., He K., Targeted emission reductions from global super-polluting power plant units. Nat. Sustainability 1, 59–68 (2018). [Google Scholar]
- 63.Dou X., Hong J., Ciais P., Chevallier F., Yan F., Yu Y., Hu Y., Huo D., Sun Y., Wang Y., Davis S. J. J., Crippa M., Janssens-Maenhout G., Guizzardi D., Solazzo E., Lin X., Song X., Zhu B., Cui D., Ke P., Wang H., Zhou W., Huang X., Deng Z., Liu Z., Near-real-time global gridded daily CO2 emissions 2021. Sci. Data 10, 69 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Dou X., Wang Y., Ciais P., Chevallier F., Davis S., Crippa M., Janssens-Maenhout G., Guizzardi D., Solazzo E., Yan F., Huo D., Zheng B., Zhu B., Cui D., Ke P., Sun T., Wang H., Zhang Q., Gentine P., Liu Z., Near-real-time global gridded daily CO2 emissions. Innovation 3, 100182 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Riahi K., Bertram C., Huppmann D., Rogelj J., Bosetti V., Cabardos A.-M., Deppermann A., Drouet L., Frank S., Fricko O., Fujimori S., Harmsen M., Hasegawa T., Krey V., Luderer G., Paroussos L., Schaeffer R., Weitzel M., van der Zwaan B., Vrontisi Z., Longa F. D., Després J., Fosse F., Fragkiadakis K., Gusti M., Humpenöder F., Keramidas K., Kishimoto P., Kriegler E., Meinshausen M., Nogueira L. P., Oshiro K., Popp A., Rochedo P. R. R., Ünlü G., van Ruijven B., Takakura J., Tavoni M., van Vuuren D., Zakeri B., Cost and attainability of meeting stringent climate targets without overshoot. Nat. Clim. Change. 11, 1063–1069 (2021). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figs. S1 to S23
Tables S1 and S2
Data Availability Statement
TROPOMI data are available at www.tropomi.eu/data-products/nitrogen-dioxide. NOx emissions derived by CAMS soil NOx emissions are available at https://permalink.aeris-data.fr/CAMS-GLOB-SOIL. CAMS anthropogenic NOx and CO2 emissions are obtained from ECCAD: https://permalink.aeris-data.fr/CAMS-GLOB-ANT. EDGAR anthropogenic NOx and CO2 emissions are available at https://edgar.jrc.ec.europa.eu/emissions_data_and_maps. DECSO emissions are available via www.temis.nl/emissions/decso_nox.php. All data and code needed to evaluate and reproduce the results in the paper are present in the paper and/or the Supplementary Materials. This study did not generate new materials.






