Global observational coverage of oil and gas methane sources with TROPOMI

doi:10.21203/rs.3.rs-2681923/v1

Download PDF

Article

Global observational coverage of oil and gas methane sources with TROPOMI

https://doi.org/10.21203/rs.3.rs-2681923/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 05 Oct, 2023

Read the published version in Scientific Reports →

You are reading this latest preprint version

Satellite observations have been widely used to measure methane (CH₄) emissions from the oil and gas (O&G) industry, particularly by revealing previously undocumented, very large emission events and basin-level emission estimates. However, the performance of retrieving CH₄ mixing ratios from the short-wave infrared band of passive remote sensing instruments is spatially and temporally varied due to sunlight, earth surface properties, and atmospheric conditions. Thus, the reliability of satellites is also varied. To assess the reliability of passive remote sensing satellite instruments for routine monitoring of CH₄ emissions, we investigated the global observational coverage of the TROPOMI (onboard the Sentinel-5P), which has been providing a daily global CH₄ mixing ratio since 2017. A 0.1° x 0.1° gridded global map that indicates the average number of days with valid observations from TROPOMI from 2019 to 2021 was generated by following the measurement retrieval quality-assurance threshold (≥0.5). We found TROPOMI had promising observational coverage over dryland regions (maximum: 58.6%) but limited coverage over tropical regions and high latitudes (minimum: 0%). Cloud cover and solar zenith angle were the primary factors at high latitudes, while aerosol optical thickness was the primary factor over dryland regions. To further assess the country-level reliability of satellites for detecting and quantifying CH₄ emissions from the O&G sector, we extracted the average annual TROPOMI observational coverage (TOC) over O&G infrastructure for 160 countries. Seven of the top-10 O&G-producing countries had an average annual TOC < 10% (< 36 days per year), which indicates the limited ability to routinely identify large emissions events, track their duration, and quantify emissions rates using inverse modelling. We further assessed the potential performance of the latter by combining TOC and the uncertainties from the global O&G inventory. Results inverse modeling with CH₄ measurements from TROPOMI may only be accurate for quantifying emissions from O&G sectors in countries located in dryland and mid-latitude regions. Conclusively, the current passive-sensing satellites have low potential for frequent monitoring of large methane emissions from O&G sectors in countries located in tropical and high latitudes (e.g., Canada, Russia, Brazil, Norway, and Venezuela). Therefore, alternative methods should be considered for routine emissions monitoring in the corresponding regions.

Earth and environmental sciences/Environmental sciences

Earth and environmental sciences/Environmental social sciences

Methane (CH₄) is a potent greenhouse gas (GHG) with 84 times the global warming potential of CO₂ over a 20-year period [1]. Atmospheric CH₄ concentrations have increased 260% since the pre-industrial era; an increase that is primarily due to human activities [2, 3]. Reducing CH₄ emissions is internationally recognized as the most effective strategy to mitigate near-term global warming [4]. Following the United Nations Framework Convention on Climate Change (UNFCCC) Conference of the Parties 26 (COP26), more than 100 countries committed to the Global CH₄ Pledge – a voluntary international initiative targeting a 30% reduction in CH₄ emissions across all economic sectors in participating countries by 2030. Emissions reductions goals are becoming more ambitious and regulations more stringent. As such, it is necessary to evaluate annual progress towards CH₄ emissions reduction targets and monitoring emission sources at global scales.

Satellites are one of approaches that can provide global empirical CH₄ emissions measurement. Satellites can play an important role in detecting large emission events and estimating emissions on global and regional scales [5, 6] Most satellite instruments measure CH₄ passively in the shortwave infrared (SWIR), relying on the absorption characteristics of CH₄ at 1650 and 2300 nm [5]. With appropriate satellite retrieval methods, SWIR radiation can be used to infer column-integrated CH₄ mixing ratios [5, 6]. In the past two decades, many studies have estimated CH₄ emissions rates using various satellite measurements, such as Sentinel-2, Sentinel-5P, GOSAT, GHGSats, PRISMA, Gaofen5, Ziyuan1 [7–26]. Due to the growing interest in monitoring CH₄ emissions from space, such as initiatives like the UN’s Methane Alert and Response System (MARS), many satellites originally designed for different purposes but with effective ranges to detect CH₄ in the SWIR are now being used for this application. This includes systems such as PRISMA, Landsat-8, Worldview-3, and Sentinel-2 [17, 15, 27].

The Tropospheric Monitoring Instrument (TROPOMI) onboard of Sentinel-5 Precursor outperforms other satellite instruments with the shortest revisit time (less than 24 hours) with global coverage up to 5.5 x 3.5 km² pixel resolution. Thus, it has become the most popular satellite in the past five years to quantify regional/basin-scale emissions, reconcile bottom-up CH₄ emission inventories, validate annual emission estimates from different sectors, and identify ultra-emitter events [11, 14, 20, 22, 23, 24, 26, 28, 29]. However, due to coarse spatial resolution, it often cannot locate the precise source that is emitting. Like other passive satellite instruments, the actual performance and observational coverage of TROPOMI are limited by the instrument parameters, artifacts of the atmospheric components, and land surface characteristics [30, 31].

While many more satellites will be launched into the space in the next decade, it remains unknown how they will advance global CH₄ emissions monitoring. Understanding the global observational coverage of TROPOMI reveals geographical challenges for monitoring CH₄ emissions and factors that are likely to affect the performance of similar satellites (i.e., polar-orbiting sun-synchronized, low-earth orbiting satellites with passive sensing systems). This global analysis addresses three questions: First, where and when does TROPOMI have the most (and least) valid observations? Second, which measurement-impeding factor is the coverage of TROPOMI most sensitive to? Third, how reliable is TROPOMI for different applications? Answers to these questions will also provide insight into the potential observational performance of planned CH₄ emissions monitoring satellites. To address these questions, we investigated the global observational coverage of the TROPOMI data products to understand the geographical distribution of valid observations for CH₄ emission detection and quantification.

Global TROPOMI observational coverage (TOC).

Figure 1 shows the global distribution of 3-year average TOC over land between 2019 and 2021. The global mean and median TOCs are 6.6% and 3.8% over land, respectively. The global pattern indicates the highest values are over the subtropics and mid-latitude dryland (arid to sub-humid) regions and the lowest values over high latitude regions and the equator (Fig. 1). These patterns are similar to number of valid observations found in past studies [6, 24]. There are almost no valid observations from TROPOMI over tropical regions along the equator throughout the year. Low TOC is also found over regions with high elevation terrain and orographic features, such as the Rocky Mountains in North America, the Andes in South America, and the Tibetan Plateau in Asia. Overall, Fig. 1 outlines the climate and topographic conditions that are favorable and less favorable for applying TROPOMI (and other similar satellites) to monitor CH₄ emissions.

Seasonal variability of TOC.

Seasonal variations in the average TOC over different regions are shown in Fig. 2. The average seasonal TOC is lowest in MAM (4.6%) and highest in SON (7.7%). In all four seasons, TOC is very low in a ~ 20° latitudinal zone centered at the Intertropical Convergence Zone (ITCZ) and at high latitudes except for Greenland in JJA and Antarctica in DJF when anticyclones establish over those regions. TOC is near zero above 50°N in DJF, including most of Canada, Europe, and Russia, and there is a very marginal increase in TOC over these regions in the other seasons. Globally, TOC is highest over Northern Africa (Sahel and Sahara) and Middle East in SON and DJF but decreases in MAM and remains low over Northern Africa and southern portions of the Middle East into JJA. Seasonal variations of TOC over Northern Africa coincide with shifts in the ITCZ and resulting effects from the African monsoons and increased dust loading in MAM and JJA [32, 33]. TOC over Southern Africa is highest in JJA and lowest in DJF, which coincides with changes in the position of the ITCZ and associated monsoons. Over Central Asia and Australia, TOC is highest in SON and lowest in DJF.

TOC-reducing factors.

The effects of different surface, atmospheric, and geometric factors on TOC are shown in Fig. 3. Cloud cover has the largest impact on TOC, with a global average annual reduction of 40.0% (146 days), followed by SZA (11.5% or 42 days), AOT (0.07% or 0.27 days), and surface albedo (0.04% or 0.13 days). Approximately 79.3% of land-based regions are impacted by persistent cloud cover for at least 90 days (24.7%), with several regions experiencing persistent cloud cover for more than 240 days (66%; e.g., high latitudes, Southeast Asia, Amazon rainforest, and Tibetan Plateau). The effects of both cloud cover and SZA are pronounced over broad swaths of Canada and Russia – two of the largest oil and gas producing countries in the world. In boreal winter, high SZA at high northern latitudes results in no valid observations for up to 3 months. Surface albedo has the smallest impact on TOC globally and is mostly influential only in Greenland. AOT has the most pronounced effects on TOC over major desert regions, such as the Sahara, Arabian, Taklimakan, and Gobi, which is expected as these are major source regions for the global dust cycle [32].

TOC for monitoring CH ₄ emissions from the O&G sector.

Based on the global variation of TOC in Fig. 1, we investigated the potential of TROPOMI for monitoring CH₄ emissions from O&G infrastructure in 160 countries across the world. The boxplots in Fig. 4 show distributions of average annual TOC over O&G infrastructure in partner (n = 89) and non-partner (n = 71) countries of the Global Methane Pledge (after COP26). Some countries have large regional differences in TOC, as expressed by wide interquartile ranges and whiskers (e.g., Oman), while others have very small differences (e.g., Norway). Most of the 160 countries have relatively low median average annual TOC (< 15%) over their O&G infrastructure. In aggregate, non-partner countries have a higher median average annual TOC (4.7%) compared to partner countries (1.9%).

For the O&G infrastructure in each country in Fig. 4, we calculated the median 3-year average number of consecutive days and gap days with valid observations (see Supplementary Table S1). This provides an indication of the ability to monitor large emissions events and determine when they begin and end, which helps investigate the source and root cause, determine the total mass or volume of the emission, and potentially determine penalties or fees. As an example, Kuwait's O&G infrastructure has the highest median average TOC (36.7%), highest median consecutive number days with valid observations (3 days), and the fewest gap days (4 days). Except Kuwait, results for the rest top-10 O&G producing countries [34] are as follows: USA (9.5%, 2 days, 15 days), Russia (3.6%, 2 days, 37 days), Saudi Arabia (20.8%, 3 days, 8 days), Canada (4%, 2 days, 31 days), Iraq (22.3%, 3 days, 8 days), China (4.6%, 1 day, 42 days), UAE (9%, 2 days, 21 days), Iran (6% 2 days, 35 days), and Brazil (2%, 1 day, 90 days). Some of these countries represent reasonable targets for routine monitoring, particularly those located in dryland regions (Kuwait, Saudi Arabia, Iraq, and UAE), while monitoring capabilities in other tropical or high latitude countries are very limited (Russia and Canada). It is worth noting that Iran is in a dryland region but has low TOC because most of its O&G infrastructure is in regions with undulating topography (see Supplementary Fig. S1).

The median of average annual TOCs for the O&G sectors in the USA and China were relatively low because the O&G infrastructure in these countries are spatially distributed across different latitudes and climate zones. Low TOC in Russia and Canada is mainly a result of high cloud cover and high SZA in boreal winter. The differences in TOCs among the top-10 producing countries indicates not all countries receive equal coverage of potential sources, which could create an observational bias where countries with high TOC receiver more attention in the scientific literature and associated media. For example, large emissions in Algeria (23.2%, 3 days, 7 days) and Turkmenistan (20%, 3 days, 9 days) have been featured in many peer-reviewed publications since 2017 [9, 12, 13, 15, 17, 25, 35], whereas large emissions from countries like Venezuela (0%, 0 day, 304 day), which holds the largest crude oil reserves in the world, are largely unknown based on satellite data due to persistent cloud cover. The ability to apply passive satellite remote sensing to detect large emissions quickly is therefore unequal among major O&G producing countries.

Potential effect of TOC on regional CH ₄ emission quantification.

Figure 5 shows the suitability scores of atmospheric inverse modeling for global O&G infrastructure. Most O&G producing regions have moderate suitability scores (> 0.5), such as the Permian Basin in the U.S., the Tarim Basin, China, and Hassi R’Mel in Algeria. Therefore, emissions estimates from atmospheric inverse modelling over these regions can be used to infer actual emission rates, although uncertainties may exist. Considerable grid cells in O&G producing regions at high latitudes, such as the Western Canadian Sedimentary Basin (Canada) and the West Siberian Petroleum Basin (Russia), also have moderate suitability scores, which is primarily because of the low GFEI relative uncertainties (see Supplementary Fig. S2). Nonetheless, there are still many grid cells in these two regions with low suitability scores (< 0.3) due to the low TOC. Following the global TOC distribution, consistently low suitability scores (< 0.3) were found for several O&G producing countries near equator such as Venezuela, Kenya, and Nigeria. This demonstrates the limits of TROPOMI in quantifying emissions with inverse methods for some large O&G basins. High suitability scores (> 0.7) were found along pipelines and over multiple O&G producing regions in Turkmenistan. This is due to the combined effect of the low uncertainties of BU inventory and high TOCs. Overall, our analysis indicates methods like Integrated Methane Inversion 1.0 [29] is not a universal solution to estimating regional emissions. As such, reconciling emissions estimates to BU inventories should not rely exclusively on TROPOMI in O&G sectors with low suitability such as Canada, Russia, and some countries in South America, Africa and Europe.

Satellite systems have enhanced the ability to detect and quantify high magnitude (“super” or “ultra”) CH₄ emissions events globally from the O&G sector, as well as other large anthropogenic sources like landfills [36], and are helping shape strategies and initiatives to respond and ultimately reduce those emissions [37]. Until this study, a detailed analysis connecting limiting conditions to the reliability of satellites for routine monitoring of CH₄ emissions from global O&G targets had not been reported. Our analyses indicate that it is difficult to reliably monitor CH₄ emissions from O&G infrastructure in countries outside the subtropics and dryland regions with TROPOMI and likely with other LEO sun-synchronized passive satellite systems (see Supplementary Table S2) on a routine or on-demand basis. If large CH₄ releases occur in Venezuela, for example, our results indicate that TROPOMI and similar passive satellite systems are unlikely to detect them. This could have several ramifications beyond the specific emissions events. First, our understanding of the spatial distribution of large emissions events based on satellite observations is incomplete [24]. CH₄ emitted from undetected events may contribute substantially to the global CH₄ budget. However, they will continue to escape satellite detection and reporting, causing uncertainties in emission estimates as undocumented sources. Second, as outlined earlier, some countries may become targeted by research and monitoring programs, including corollary media reports, owing to high observational coverage (e.g., Turkmenistan and Algeria), while others with low observational coverage are overlooked. These geographic disparities reflect the fact that more research and monitoring attention is likely to be directed at regions where the technology is more effective and where results can be published in the peer-review literature or media.

Low TOC over tropical and high latitude regions indicates these settings are less reliable for deriving satellite detections and quantifications. This includes broad swaths of Canada and Russia – two of the largest O&G producers globally. Many O&G producing countries in these regions experience prolonged periods (months) when there are no valid observations owing to persistent cloud cover and high SZA, and large emissions events, if present, can go undetected (see Supplementary Table S1). For emissions quantification, our results show that the majority of O&G producing regions have low to moderate suitability scores. Hence, using the inversion method with satellite observations to accurately infer regional emissions may only be feasible in some O&G producing regions.

Our analyses show that there is an abundance of regions globally with O&G targets and seasons when emissions, if present, will be hidden from passive satellite systems for extended periods of time. This presents a challenge for monitoring programs and resulting policies and actions to reduce emissions. With knowledge of limiting conditions, some O&G producers could strategically schedule activities so that large emissions resulting from intentional releases (e.g., blowdowns) are undetectable by satellites, like in winter at high latitudes or during monsoon seasons. This may be more challenging in the future with planned active satellite systems such as MERLIN, which will use a LiDAR sensor [27], but other planned passive satellites are likely to encounter similar limitations.

On an annual basis, this study demonstrates that passive satellite monitoring of CH₄ emissions from O&G infrastructure should not be viewed as a form of global-scale continuous monitoring with the ability to acquire regular snapshots for detecting large emissions events or regional quantifications. At best, the analysis of TROPOMI data reveals that passive satellite monitoring is periodic and based on a global patchwork of high and low observational coverage. This is unlikely to change substantially with the addition of new passive satellite systems. In some regions, passive satellites like TROPOMI can provide high temporal coverage for detecting large emissions and quantifying rates from O&G sources, but in many regions, they are occasionally useful or not at all. Understanding the current spatial and temporal gaps of TROPOMI and the variables affecting its performance provides insight on similar challenges that forthcoming satellite systems may face.

Satellite remote sensing is often promoted as a leading component of a comprehensive CH₄ emissions monitoring system because of their global coverage, such as the UN’s Methane Alert and Response System (MARS) launched at COP27. However, the quality of the observations for detecting and quantifying CH₄ emissions from O&G targets and other sources is not equal across the globe. Comprehensive monitoring programs must therefore incorporate a mix of other technologies and methods to fill gaps in satellite observational coverage. In many regions, satellites using passing remote sensing will serve a very limited role in monitoring. With considerable reliance on satellites in international initiatives like MARS, it is imperative that the scope of monitoring approaches is expanded to include other leading technologies and methods.

Overview.

To assess TROPOMI observational coverage (TOC), we calculated the mean proportion of the year with valid CH4 observations for TROPOMI. The valid observations of TROPOMI spanning three years were determined using the quality assurance value (qa_value) in TROPOMI level-2 data product. In this analysis, a TROPOMI day was defined as one day in which a given observation from TROPOMI had qa_value ≥ 0.5. 3-year average TOC was analyzed globally using calculated TROPOMI days of geographical grids from 2019 to 2021. Additionally, we investigated the global impacts of four TOC-reducing factors: cloud cover fraction, solar zenith angle, surface albedo (SWIR), and AOT (NIR). Moreover, we assessed the reliability of TROPOMI for monitoring emissions from global O&G sectors by extracting TOC of grids associated with O&G sectors. Finally, we conducted a suitability analysis to evaluate the potential suitability of using atmospheric inverse modeling to quantify the regional emission of global O&G producing regions.

TROPOMI level-2 data product and data quality assurance flag.

Sentinel-5P (TROPOMI) – a low earth orbit near polar sun-synchronous earth observation satellite – relies on passive remote sensing to map atmospheric methane columns on daily basis. It produces 13 or 14 netCDF data files each day for the level-2 data product (i.e., one data file per orbit). The pixel size of images is 7 km × 7 km for data processed before 06 August 2019 and 7 km × 5.5 km for data collected after 06 August 2019. In this study, we downloaded Sentinel-5P TROPOMI CH₄ Level 2 data products from 01 January 2019 to 31 December 2021 from NASA's Goddard Earth Sciences Data and Information Services Center.

The TROPOMI level-2 data product uses the retrieval quality assurance flag (qa_value) to indicate whether pixels (XCH₄ – column averaged methane concentration) can be used to detect and quantify emissions. The qa_value is a continuous quality descriptor, varying between 0 and 1, depending on the instrumental viewing angle, content of atmospheric components, surface albedo, topography, etc. [38]. As suggested by the data providers and previous studies, a minimum qa_value for a valid XCH₄ measurement is 0.5 [11, 14, 22, 24, 30, 38]. This represents a XCH₄ retrieved under cloud-free conditions and filtered for conditions with sufficient atmospheric reflectance (solar zenith angle ≤ 70°), viewing zenith angle (< 60°), smooth topography (1 standard deviation of surface elevation, ≤ 80 m within a 5-km radius), low absorption and scatter radiation from aerosols in the atmosphere (aerosol optical thickness in the near-infrared ≤ 0.3) and dark surface (surface albedo in shortwave infrared ≥ 0.02).

TROPOMI day and TROPOMI observational coverage (TOC).

In this study, we define a TROPOMI day as a pixel-covered area (hereinafter a "pixel") with at least one valid observation from TROPOMI (i.e., a XCH₄ measurement with a qa_value ≥ 0.5) on a given day. The proportion of TROPOMI days per year for each pixel was then used to calculate the global TROPOMI observational coverage (TOC). To count TROPOMI days per year for each pixel, we constructed a workflow to sequentially determine if a given pixel and timestep combination satisfied the qa_value threshold (Fig. 6). We created two 3D global arrays, A₁ and A₂, to store the number of valid observations per day and TROPOMI days per year, respectively. Thus, the depth of A₁ is either 13 or 14 (depending on the number of files per date-stamp), and the depth of A₂ is 365 (366 for the year 2020). The column and row of both 3D arrays are longitude and latitude, respectively, and each pixel has a spatial resolution of 0.1° × 0.1°.

The daily workflow (Fig. 1) starts by finding all data files corresponding to the date of the day ($d$). A quality assurance (QA) filter was used to find all valid pixels (${QA}_{ij}$) for each file. Next, we found the nearest neighbor grid cells (${C}_{ij}$) in A₁ for all valid pixels by searching for the minimum Haversine distance between geographic coordinates of valid pixels and ${C}_{ij}$. Then, we assigned 1 to ${C}_{ij}$ to indicate one valid observation. This workflow was done recursively for all files. Following our definition of a TROPOMI day, we converted A₁ to a 2D binary array (M). The grid cell of M with at least one valid observation indicates one TROPOMI day. At the end of the daily workflow, we assigned M to A₂ based on the date-stamp. This daily workflow was repeated from 01 January 2019 to 31 December 2021. Finally, we summed A₂ and divided it by the number of years of analysis ($n$ = 3) to find the average annual TROPOMI days per pixel. Then, we divided the average annual TROPOMI days by 365 and multiplied by 100 to calculate the average annual TOC (${TOC}_{ij}$). The general formulas for calculating ${TOC}_{ij}$ are shown in Eqs. 1 and 2. In addition to average annual TOC, TROPOMI days were examined seasonally for all land-based grid cells. We also assessed average seasonal TOC for four seasons (01 September to 30 November, 01 December to 28/29 Feb, 01 March to 31 May, and 01 June to 31 August) to show the temporal variation in TROPOMI's data availability. This analysis was performed for 3 years from 2019 to 2021.

$${TOC}_{ij}=\left(\frac{\sum _{1}^{n}\sum _{1}^{365}{C}_{ij}}{n}\right)÷365\times 100$$

where

$${C}_{ij}=\left\{\begin{array}{c}0, {QA}_{ij}<0.5\\ 1, {QA}_{ij}\ge 0.5\end{array}\right.$$

Assessment of factors that reduce TOC.

Cloud cover, solar zenith angle (SZA), viewing zenith angle, topography, aerosol optical thickness (AOT), and surface albedo in SWIR are six factors that can affect the quality of TROPOMI's observations and measurement of XCH₄ [30, 31]. Except for topography, the rest factors are included in the TROPOMI level-2 data product and are measured by Sentinel-5P simultaneously when measuring XCH₄ or from observations with co-located satellites (e.g., some cloud cover fraction from the L2 data product was measured from the Visible Infrared Imaging Radiometer Suite on the U.S. S-NPP). Since the current L2 TROPOMI data product already removed observations with viewing zenith angles that greater than 60 degrees [38], we only quantified the effect of the rest four factors on TOC according to the workflow described in Fig. 6. The goal was to determine how global variations in TOC are affected by each factor. First, we created filters to separately find pixels with cloud cover fraction > 0, SZA > 70°, surface albedo (SWIR) < 0.02, and AOT (NIR) > 0.3. When any one of these conditions is met, the observations are invalid because the qa_value < 0.5, so TOC cannot be derived according to Eqs. 1 and 2. Second, we calculated the proportion of times per year for each factor that failed to meet its quality assurance requirement. This was performed separately for each factor, averaged for three years, and converted to a percentage. It is worth nothing that the qa_value of a given grid cell and day combination can be affected by multiple factors. These factors also affect XCH₄ retrieval for other passive remote sensing satellites. Therefore, investigating the impacts of different factors on TROPOMI also reveal potential limitations for other similar satellites.

Impact of TOC on monitoring CH ₄ emissions from the O&G sector.

We evaluated how TOC affects the ability to monitor CH₄ emissions from the O&G sector. We used an inventory (0.1° × 0.1°) of O&G sector CH₄ emissions (Fuel Exploitation Gas, Fuel Exploitation Oil, and Oil Refineries and Transformation Industry) from the Emissions Database for Global Atmospheric Research v6.0 [39] and excluded TOC of non-O&G grid cells in our analysis. We created a binary grid that denoted the presence or absence of O&G activity to mask areas that had a high probability of containing O&G derived CH₄ emissions. We also calculated the average consecutive TROPOMI days and average gap (i.e., number of days between two TROPOMI days) to further assess the continuity and observational gaps of using TROPOMI (or similar satellites) to monitor O&G CH₄ emissions in different countries. We focused our analysis on the top-10 O&G producing countries: USA, Russia, Saudi Arabia, Canada, Iraq, China, United Arab Emirates, Iran, Brazil, and Kuwait [34].

Next, we evaluated the affect of TOC on regional CH₄ emission quantification using TROPOMI observations. Atmospheric inverse modelling combined with Bayesian inference is the most widely used technique to estimate regional CH₄ emissions with satellite observations [14, 22, 23, 29]. The number of valid satellite observations, the accuracy of the bottom-up (BU) inventory data (also known as the prior), and uncertainty are three factors that can impact the confidence of a posterior estimation [6, 14, 20, 40, 41]. Accurate posterior estimates require an adequate number of valid observations to perform Bayesian inference. Therefore, error in the posterior estimate depends on errors in observations and the prior. Observational error has an inverse relationship with the number of observations; error is reduced as the number of observations increases. This indicates that TROPOMI (or similar satellites) are less suitable for monitoring CH₄ emissions in regions with limited valid observations and large prior uncertainty. To identify these regions, we estimated the suitability score ($SS$) of using the inversion method to estimate CH₄ emissions by integrating global TOC of O&G sector grid cells and the relative uncertainties ($RU$) of CH₄ emissions in the BU inventory. The $SS$ was calculated using the following equation:

$$SS= 1-(RU- RU\times TOC)$$

Where $RU$ is the relative standard deviation of the O&G prior (between 0 and 100%) from the Global Fuel Exploitation Inventory v2 [42], which was created to improve the accuracy of inversion method and was evaluated by comparing to other BU inventories in the previous studies [42]. As such, the high $RU$ identifies regions with large uncertainty with respect to inventory of CH₄ emission. The value of $SS$ also lies between 0 and 1. Regions with $SS$ close to 1 are regions suitable for inversion method, which happens when $RU$ close to 0 (low relative uncertainty), TOC close to 1 (high observational coverage from TROPOMI), or both.

Data availability

The datasets analyzed during the study, if not in the Supplementary Information, are available from the corresponding author on reasonable request.

Code availability

The TROPOMI observational coverage processing and analysis were programmed in Python with standard packages. The scripts can be found in https://github.com/MozhouGao/Global-TROPOMI-Observatioanl-Coverage. Data required to perform TROPOMI observational coverage calculation can be downloaded from either Copernicus Open Access Hub or NASA Goddard Earth Science Data and Information Services Center. The results can be reproduced by employing the equations, explanation, and parameters provided in the main text. Additional code and data will be made available upon request.

Acknowledgements

The authors thank the University of Calgary’s Global Research Initiative for funding support to MG.

Author contributions

M.G., Z.X., and C.H.H. designed the research. M.G. directed and performed the analyses. M.G., Z.X., C.V., C.H.H and T.E.B. wrote and edited the paper.

Competing interests

The authors declare no competing interests.

Forster, P. et al. The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. 923–1054 (2021).
Van Dingenen, R., Crippa, M., Maenhout, G., Guizzardi, D. & Denener, F. Global trends of methane emissions and their impacts on ozone concentrations, European Commission Publications Office. JRC Science For Policy Report. https://data.europa.eu/doi/10.2760/820175 (2018).
Saunois, M. et al. The Global Methane Budget 2000–2017. Earth System Science Data 12, 1561–1623 (2020).
Nature. Control Methane to Slow Global Warming — Fast. Nature 596, 461–461 (2021).
Jacob, D. J. et al. Satellite observations of atmospheric methane and their value for quantifying methane emissions. Atmospheric Chemistry and Physics 16, 14371–14396 (2016).
Jacob, D. J. et al. Quantifying methane emissions from the global scale down to point sources using satellite observations of atmospheric methane. Atmospheric Chemistry and Physics 22, 9617–9646 (2022).
Frankenberg, C. et al. Tropical methane emissions: A revised view from SCIAMACHY onboard ENVISAT. Geophysical Research Letters 35, (2008).
Wecht, K. J., Jacob, D. J., Frankenberg, C., Jiang, Z. & Blake, D. R. Mapping of North American methane emissions with high spatial resolution by inversion of SCIAMACHY satellite data. Journal of Geophysical Research: Atmospheres 119, 7741–7756 (2014).
Buchwitz, M. et al. Satellite-derived methane hotspot emission estimates using a fast data-driven method. Atmospheric Chemistry and Physics 17, 5751–5774 (2017).
Maasakkers, J. D. et al. Global distribution of methane emissions, emission trends, and OH concentrations and trends inferred from an inversion of GOSAT satellite data for 2010–2015. Atmospheric Chemistry and Physics 19, 7859–7881 (2019).
Pandey, S. et al. Satellite observations reveal extreme methane leakage from a natural gas well blowout. Proceedings of the National Academy of Sciences 116, 26376–26381 (2019).
Varon, D. J. et al. Satellite Discovery of Anomalously Large Methane Point Sources From Oil/Gas Production. Geophysical Research Letters 46, 13507–13516 (2019).
Varon, D. J. et al. High-frequency monitoring of anomalous methane point sources with multispectral Sentinel-2 satellite observations. Atmospheric Measurement Techniques 14, 2771–2785 (2021).
Zhang, Y. et al. Quantifying methane emissions from the largest oil-producing basin in the United States from space. Science Advances 6, eaaz5120 (2020).
Guanter, L. et al. Mapping methane point emissions with the PRISMA spaceborne imaging spectrometer. Remote Sensing of Environment 265, 112671 (2021).
Jervis, D. et al. The GHGSat-D imaging spectrometer. Atmospheric Measurement Techniques 14, 2127–2140 (2021).
Irakulis-Loitxate, I. et al. Satellite-based survey of extreme methane emissions in the Permian basin. Science Advances 7, (2021).
Lu, X. et al. Global methane budget and trend, 2010–2017: complementarity of inverse analyses using in situ (GLOBALVIEWplus CH₄ ObsPack) and satellite (GOSAT) observations. Atmospheric Chemistry and Physics 21, 4637–4657 (2021).
Ma, S. et al. Satellite Constraints on the Latitudinal Distribution and Temperature Sensitivity of Wetland Methane Emissions. AGU Advances 2, (2021).
Qu, Z. et al. Global distribution of methane emissions: a comparative inverse analysis of observations from the TROPOMI and GOSAT satellite instruments. Atmospheric Chemistry and Physics 21, 14159–14175 (2021).
Sadavarte, P. et al. Methane Emissions from Superemitting Coal Mines in Australia Quantified Using TROPOMI Satellite Observations. Environmental Science & Technology 55, 16573–16580 (2021).
Shen, L. et al. Unravelling a large methane emission discrepancy in Mexico using satellite observations. Remote Sensing of Environment 260, 112461 (2021).
Chen, Z. et al. Methane emissions from China: a high-resolution inversion of TROPOMI satellite observations. Atmospheric Chemistry and Physics 22, 10809–10826 (2022).
Lauvaux, T. et al. Global assessment of oil and gas methane ultra-emitters. Science 375, 557–561 (2022).
Sánchez-García, E., Gorroño, J., Irakulis-Loitxate, I., Varon, D. J. & Guanter, L. Mapping methane plumes at very high spatial resolution with the WorldView-3 satellite. Atmospheric Measurement Techniques 15, 1657–1674 (2022).
Shen, L. et al. Satellite quantification of oil and natural gas methane emissions in the US and Canada including contributions from individual basins. Atmospheric Chemistry and Physics 22, 11203–11215 (2022).
Ehret, T. et al. Global Tracking and Quantification of Oil and Gas Methane Emissions from Recurrent Sentinel-2 Imagery. Environmental Science & Technology 56, 10517–10529 (2022).
Zhang, Y. et al. Observed changes in China’s methane emissions linked to policy drivers. Proceedings of the National Academy of Sciences 119, (2022).
Varon, D. J. et al. Integrated Methane Inversion (IMI 1.0): a user-friendly, cloud-based facility for inferring high-resolution methane emissions from TROPOMI satellite observations. Geoscientific Model Development 15, 5787–5805 (2022).
Hu, H. et al. Toward Global Mapping of Methane With TROPOMI: First Results and Intersatellite Comparison to GOSAT. Geophysical Research Letters 45, 3682–3689 (2018).
Cooper, J., Balcombe, P. & Hawkes, A. The quantification of methane emissions and assessment of emissions data for the largest natural gas supply chains. Journal of Cleaner Production 320, 128856 (2021).
An, Z. et al. Global Monsoon Dynamics and Climate Change. Annual Review of Earth and Planetary Sciences 43, 29–77 (2015).
Kok, J. F. et al. Contribution of the world’s main dust source regions to the global cycle of desert dust. Atmospheric Chemistry and Physics 21, 8169–8193 (2021).
EIA. Total energy production from petroleum and other liquids 2021 (2021). https://www.eia.gov/international/rankings/world?pa=288&u=2&f=A&v=none&y=01%2F01%2F2021 (Accessed 07 March 2023).
Schneising, O. et al. Remote sensing of methane leakage from natural gas and petroleum systems revisited. Atmospheric Chemistry and Physics 20, 9169–9182 (2020).
Maasakkers, J. D. et al. Using satellites to uncover large methane emissions from landfills. Science Advances 8, (2022).
Vogel, F. Chasing after methane’s ultra-emitters. Science 375, 490–491 (2022).
Landgraf, J., Lorente, A., Langerock, B. & Kumar, M. S5P Mission Performance Centre Methane [L2__CH4___] Readme (2022). https://sentinel.esa.int/documents/247904/3541451/Sentinel-5P-Methane-Product-Readme-File (Accessed 07 March 2023)
Crippa, M. et al. GHG emissions of all world countries - 2021 Report (2021). https://edgar.jrc.ec.europa.eu/dataset_ghg60 (Accessed 07 March 2023).
Raju, A. et al. Regional estimation of methane emissions over the peninsular India using atmospheric inverse modelling. Environmental Monitoring and Assessment 194, (2022).
Chen, Z. et al. Methane emissions from China: a high-resolution inversion of TROPOMI satellite observations. Atmospheric Chemistry and Physics 22, 10809–10826 (2022).
Scarpelli, T. R. et al. Updated Global Fuel Exploitation Inventory (GFEI) for methane emissions from the oil, gas, and coal sectors: evaluation with inversions of atmospheric methane observations. Atmospheric Chemistry and Physics 22, 3235–3249 (2022).

No competing interests reported.

TOCSupDocumentScientificReport.docx

Download PDF

Journal Publication

published 05 Oct, 2023

Read the published version in Scientific Reports →

Editorial decision: Major revision
02 Jul, 2023
Reviews received at journal
28 Jun, 2023
Reviews received at journal
30 Apr, 2023
Reviewers agreed at journal
13 Apr, 2023
Reviewers agreed at journal
12 Apr, 2023
Reviewers invited by journal
11 Apr, 2023
Editor assigned by journal
11 Apr, 2023
Editor invited by journal
21 Mar, 2023
Submission checks completed at journal
21 Mar, 2023
First submitted to journal
11 Mar, 2023

You are reading this latest preprint version

Global observational coverage of oil and gas methane sources with TROPOMI

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1