Summary
The purpose of this paper is to provide additional information and insights gained on manuscript SPE-209980-MS, accepted for presentation at the 2022 Society of Petroleum Engineers Annual Technical Conference and Exhibition (Esparza et al. 2022).
The energy sector has been identified as one of the main contributors to emissions of anthropogenic greenhouse gases. Therefore, sustainability in the sector is mainly associated with the advancement in environmental and social performance across multiple industries. Individual firms, particularly those belonging to the oil and gas (O&G) industry, are now assessed for their environmental, social, and governance (ESG) performance and their impact on climate change. To meet the different key performance indicators (KPIs) for corporate social responsibility (CSR) and ESG, the planning, development, and operation of O&G infrastructure must be conducted in an environmentally responsible way.
Today, operators calculate their own emissions, which are typically self-reported annually, usually relying on emission factors to complement the lack of emission measurement data. This paper discusses how methane detection of O&G infrastructure using remote sensing technologies enables operators to detect, quantify, and minimize methane emissions while gaining insights and understanding of their operations via data analytics products. The remote sensing technologies accounted for in this paper are satellite and aerial platforms operating in tandem with data analytics, providing a scheme to support sustainability initiatives through the quantification of some ESG metrics associated with methane emissions. This paper presents examples of measurements at O&G sites taken with satellites and aircraft platforms, providing evidence of methane emissions at the facility level. A discussion of each platform and how they work together is also presented. Additionally, this paper discusses how these data insights can be used to achieve sustainability goals, functioning as a tool for ESG initiatives through the incorporation of analytical models.
Introduction
This paper provides additional information that supplements the original findings, model parameters, and conclusions submitted in manuscript SPE-209980-MS, which was presented at the 2022 Society of Petroleum Engineers Annual Technical Conference and Exhibition. This paper reflects the most recent changes to the model and examples discussed in Esparza et al. (2022).
The general concept of sustainability lies in the vision that incorporates the economic, social, and environmental dimensions (Purvis et al. 2019), making it a multigenerational overarching endeavor. Therefore, sustainable development entails the actions taken by each generation to define what needs to be sustained, what needs to be developed, and the tension between short-term business development imperatives and long-term impact on the environment (Parris and Kates 2003). Businesses in multiple industries have been prompted to demonstrate how each of them operate in a manner that protects the environment while enhancing society and profit (Windsor 2001). This led to the inclusion of CSR precepts into business models (Latapí Agudelo et al. 2019). By incorporating CSR initiatives, companies build their unique corporate image by pursuing projects on social and human rights, supply chain and sourcing, and environmental goals (Du et al. 2010).
The energy sector has been identified as one of the main contributors to emissions of anthropogenic greenhouse gases (Höök and Tang 2013). The O&G industry, due to the inherent impacts on the environment, community development, and global warming, has continuously engaged in CSR practices (Frynas 2005). The companies within this industry are expected, by society in general, to conduct self-assessment of their impact and potential risks associated with their processes, prompting them to take actions that exceed what is required by the governing law (Kirat 2015). Therefore, each company sets KPIs as metrics for sustainable development. These KPIs are measured, analyzed, and reported, which could have the potential to affect interoperability and strategic decisions (Mojarad et al. 2018).
Sustainability in the O&G industry is mainly associated with the advancement in environmental and social performance across the industry. Thus, the assessment of such advancement is currently being gauged by diverse metrics that may not conform to a standardized context across the industry (Schneider et al. 2013). These sustainability metrics typically provide nonfinancial information that can later be translated by financial analysts (in practice, after applying corresponding assumptions) into a commercial impact. Hence, the global metrics and policies that have defined competitiveness across nations are constantly changing as companies from all industries are embracing innovation that makes them more competitive and differentiated (Boons et al. 2013).
Due to increasing external pressures from governments, stakeholders, and potential investors, a proper understanding of the relationship between sustainability innovation and a firm’s competitiveness is necessary, as it can assist in the creation of operational strategies, leading to a firm’s future self-preservation (Hermundsdottir and Aspelund 2021). It has been noted that companies that invest early in sustainability innovation and products have the opportunity of developing a competitive advantage (Boons et al. 2013).
Governments around the world have pledged action on addressing climate change (Berrang-Ford et al. 2019). Industries that have an impact on the environment are now observed by both the government and financial industry. Individual firms, particularly those belonging to the O&G industry, are now assessed for their ESG performance and their impact on climate change (Dye et al. 2021). For more than a decade, the investment community has favored the inclusion of nonfinancial metrics and intangibles. Financial frameworks such as the International Financial Reporting Standards in Europe, the Generally Accepted Accounting Principles in the United States, and the eXtensible Business Reporting Language have adopted these nonfinancial metrics (Boerner 2011).
To meet the different KPIs for CSR and ESG, the planning, development, and operation of O&G infrastructure must be conducted in an environmentally responsible way. For this, understanding the potential air quality impact and greenhouse gas budgets from emissions of organic pollutants and methane is mandatory (Brantley et al. 2014). Several studies have demonstrated the discrepancies among bottom-up and top-down approaches, in which methane emissions are higher than estimates based on inventories using top-down methodologies (Allen 2014; Zavala-Araiza et al. 2015; Johnson et al. 2017; Robertson et al. 2017). Hence, the use of remote sensing technology to accurately detect and quantify methane emissions is of utmost relevance.
The objective of this paper is to discuss how methane detection from O&G infrastructure using remote sensing technologies enables operators to detect, quantify, and minimize the emissions while gaining insights and understanding of their operations via data analytics products. The remote sensing technologies accounted for in this study are satellite and aerial platforms operating in tandem with data analytics, providing a scheme to support sustainability initiatives and ESG metrics. This paper presents examples of methane measurements at O&G sites taken with satellites and aircraft platforms. A discussion of each platform and how they work together is also presented. In addition, this paper proposes how these data can be used to achieve sustainability goals, functioning as a tool for ESG initiatives through the incorporation of analytical models.
Contextual Basis
The adoption of ESG reporting by the O&G industry is of increasing relevance for the financial sector, the government, and the public. As the impact of methane on the environment is understood, reducing O&G-related emissions represents an immediate activity toward lowering the short-term radiative forcing (Scoones 2023) and therefore, actual actions toward KPIs for ESG and CSR purposes. However, there is a problem of comparability across the different companies in the sector (Cardoni et al. 2019). This means that the quantitative and qualitative characteristics of the data vary between different sets of information, as no singular tool or framework is used.
The use of satellites enables persistent measurements across the globe. This remote sensing platform allows the use of the same instrument in the constellation, effectively minimizing the error attributable to differences in instrumentation hosted by different satellites, thereby decreasing the number of error biases. Therefore, satellites can be used in reconciliation, reporting, and verification of emission estimates (Cooper et al. 2022). Also, the use of targeted aerial platforms using the same sensor as that used for the satellite platform will increase the refinement of the data, which will only observe the same instrument bias.
GHGSat is the first organization to combine the same sensor at different altitudes to operate within a tiered monitoring system (Esparza et al. 2021). This provides unique advantages in combining the data, particularly by eliminating the need for conversion and equivalency. This same approach has been used in a study in the Permian Basin, in which not only does the combination show equivalent methane emissions reductions when compared to optical gas imaging practices alone but also proved to be more cost-effective in the analysis (Esparza et al. 2023). The application of data analytics modeling techniques to build a financial tool targeted at ESG and sustainability stakeholders represents an additional important utility.
Methodology
The platforms used in this study use spectrometry principles to perform measurements of methane across the atmospheric column to capture excess background methane concentration. These measurements are taken remotely (i.e., remote sensing) from either space or the air, by analyzing the amount of light absorbed by the gas at spectral frequencies specific to methane. The data set generated by the daily accumulation of data provides the opportunity to generate insights by performing analytical studies. Fig. 1 summarizes the multiple platforms and processes associated with the tiered monitoring approach. This approach provides unique advantages in linking the data, notably by eliminating the need for adaptation between different platforms and the compounding of uncertainties.
Instrument
The primary instrument on the satellite platform is an imaging spectrometer operating under a wide-angle fixed-cavity Fabry-Pérot (WAF-P) design (Germain et al. 2016; Jervis et al. 2021; Esparza et al. 2021). This sensor measures solar radiation in the shortwave infrared (SWIR) spectrum at 1635 to 1670 nm (spectral resolution 0.1 nm) (McKeever et al. 2017). Besides the WAF-P, the instrument’s optical design is of high relevance, allowing the configuration of three lens groups and beam folding mirrors (Jervis et al. 2021). The same sensor technology was later incorporated into an instrument designed to fly on a light aircraft (Deglint et al. 2021; Esparza and Gauthier 2021). At this lower altitude, the detection threshold is one order of magnitude lower than the satellite version, along with higher resolution imagery. These two platforms at different altitudes provide different spatial and temporal resolutions. Understanding that no single solution can provide complete coverage of O&G facilities affordably, a combination of technologies and instruments working together in a tiered observation system is increasingly recognized as an efficient and cost-effective way to monitor assets for methane emissions and assist sustainability efforts (Esparza and Gauthier 2021).
Satellite Platform
The company GHGSat has one demonstrator satellite D (“Claire,” launched in 2016) (McKeever et al. 2017) and eight commercial satellites in the constellation. The commercial satellites C1 (“Iris,” launched in 2020), C2, (“Hugo,” launched in 2021), C3 through C5 (“Luca,” “Penny,” and “Diako,” launched in 2022), and C6 through C8 (“Mey-Lin,” “Gaspard,” and “Océane,” launched in 2023) share the same specifications. The nominal detection threshold is 100 kg∙h−1, with a spatial resolution of less than 30 m and a standard field of view (FOV) of approximately 12×12 km (Jervis et al. 2021). The altitude for each of these satellites is around 500 km from the ground. In this platform, measurements of methane concentration are performed along the atmospheric column, from the ground to the top of the atmosphere (Varon et al. 2018). The vertical sensitivity depends on atmospheric absorption and scattering, making clear-sky conditions optimal for detections in the SWIR spectrum. The instrument pointing capability on this platform allows for fine spatial and spectral resolution (Jacob et al. 2022).
A recent example of a detection made with the satellite platform is shown in Fig. 2 . This image depicts a methane concentration map from an O&G infrastructure emitting at a calculated rate of 1009 kg∙h−1 ± 40% on 19 June 2022.
Methane concentration map for an emission detected with the satellite platform in Algeria on 19 June 2022. The methane concentration is overlaid over background imagery provided by Mapbox®.
Methane concentration map for an emission detected with the satellite platform in Algeria on 19 June 2022. The methane concentration is overlaid over background imagery provided by Mapbox®.
Aircraft Platform
The same satellite company has two aircraft variant instruments (AV1 and AV2), which share the same specifications. The nominal detection threshold is 10 kg∙h−1, with a spatial resolution of less than 1 m flying at 3 km above ground level (Deglint et al. 2021). At that altitude, it provides an across-track swath width of approximately 0.75 km (Esparza and Gauthier 2021; Esparza et al. 2023).
A recent example of a detection made with the aircraft platform is shown in Fig. 3 . This image depicts a methane concentration map from an O&G facility emitting at a calculated rate of 278 kg∙h−1 ± 39% on 5 October 2022.
Methane concentration map for an emission detected with the aircraft platform in Louisiana, United States, on 5 October 2022. The methane concentration is overlaid over background imagery taken during the observation by a mounted auxiliary camera.
Methane concentration map for an emission detected with the aircraft platform in Louisiana, United States, on 5 October 2022. The methane concentration is overlaid over background imagery taken during the observation by a mounted auxiliary camera.
Detection and Quantification
The retrieval of methane concentration in a column is typically denoted in parts per billion, as it represents the column-average dry molar mixing ratio, X (Qin et al. 2017). A plume is detected when the value of X is greater than the background concentration of methane, leaving the difference between these values (∆X) as the enhancement detected in parts per billion. Mid-latitude and desert-like areas provide greater opportunities for successful observations, whereas the poles present low sun angles, cloudiness, and seasonal darkness (Jacob et al. 2022). Fig. 4 depicts the impact of clouds on the surface reflectance (left image) on a partially cloud-covered FOV observation. The image on the right depicts the methane layer, in which the darker areas indicate suboptimal methane values where the clouds were present. Fig. 5 shows the surface reflectance of a fully cloud-covered FOV.
Satellite observation showing the surface reflectance layer (left) and the methane enhancement layer (right) and the impact of clouds.
Satellite observation showing the surface reflectance layer (left) and the methane enhancement layer (right) and the impact of clouds.
Satellite observation exhibiting the surface reflectance layer saturated with clouds.
Satellite observation exhibiting the surface reflectance layer saturated with clouds.
The quantification method used is integrated mass enhancement (IME) (Varon et al. 2020). The computation of the source rate (Q) requires the plume enhancements in IME, wind speed (Ueff) representative of the site (e.g., operational meteorological database of a local anemometer), and its relationship with the characteristic plume size (L) (Cusworth et al. 2019).
where
A comprehensive explanation of the IME method and quantification used by this platform is described in Varon et al. (2018). It is important to note that this method infers source rates with an error of 5% to 12% depending on instrument precision, which could range between 1% and 5%. Also, additional errors are introduced when the local wind speed data are not available, expanding the error range. Nonetheless, Sherwin et al. (2023) showed that this platform’s calculated emission rate was highly correlated with the true source in a blind test. Additional details of this test are described in the Validation of Measurements section.
Third-Party Satellite Data Analytics
Many Earth-observing satellites in existence are deemed public satellites. Some of these satellites can be used to detect and measure methane emissions globally. The most notable of these satellites are those of the Copernicus system, which is the European Union’s largest Earth observation program, managed by the European Space Agency. Geospatial expertise and computational power are required to process the data effectively (Edwards et al. 2020). This proprietary third-party satellite data analytics tool effectively uses these data sets, which are leveraged to complement targeted satellite observation data.
Description of the Analytical Suite for ESG
To aggregate the heterogeneous data sets to calculate ESG company-level emissions, an ESG pipeline was developed. The analytical suite for ESG relies on a set of integrated analytical components to process observation data sets into meaningful insights for operators. At a high level, the analytical suite for the ESG product follows four steps, which are depicted in Fig. 6 .
Schematic to reflect the flow of data within the analytic suite for ESG.
Step 0: Assumptions
Analytical suites built to estimate company-level emissions based on direct measurements from satellite data are novel. Therefore, the product is underpinned by a list of assumptions related to the technical limitations of methane monitoring as well as to bespoke or proprietary algorithms:
The targeted satellites have a minimum detection limit (MDL) of 100 kg∙h−1 at 50% probability of detection (McKeever and Jervis 2022).
Sentinel 2 has an MDL of 3000 kg∙h−1. Jacob et al. (2022) provide a range of 1400 to 4000 kg∙h−1.
Plumes follow a power law distribution. This distribution allows the estimation of the percentage of emissions below the MDL as well as the volume of methane emitted.
The average emissions rate (including nulls) that is observed at a site is representative of the average emissions rate for that site overall.
Emissions should be attributed to the operator of a facility rather than the owner.
Step 1: Data Import and Cleaning
The analytic suite for ESG uses satellite and plans to use aircraft data integrated with data from the new analytics product, which processes public satellite data from Sentinel 2 (with additional public satellites to be incorporated in a later stage), allowing for greater observation coverage.
The targeted satellites performed observations of specific sites. As a result, they cannot observe large regions without support from public satellites. Currently, emissions data collected from Sentinel 2 are used to complement the targeted satellites’ observations, but data from other existing and new public satellites will be incorporated. These public satellites have higher detection thresholds and lower spatial resolution when compared to the targeted satellites used in this study but provide higher temporal resolution (Jacob et al. 2022). The latter characteristic allows for increased coverage as well as highlighting points of interest for targeted satellites to observe (i.e., tip and cue of satellite observations).
A facilities data set is created by ingesting and processing data from sources like the British Columbia Oil and Gas Commission and the US Environmental Protection Agency’s public data sets. Additional facilities are identified using a custom-built facility detector algorithm that uses computer vision to identify facilities, which are then validated by subject matter experts.
The first element of complexity in this process is transforming the data from multiple sources into a standardized output and deduplicating the data. Duplicate records are commonplace in the combined data set due to data coming from multiple sources. First, the data are standardized so that the attribute values are consistent. Then, the data are deduplicated using a combination of purpose-built proprietary algorithms and an unsupervised learning model.
Step 2: Identification of Observed Facilities and Matching of Plumes to Emitting Facilities
The analytical tool combines satellite observations with the deduplicated facility list. Each satellite observation has an FOV (the on-the-ground area that the observation covers); facilities within the FOV are said to be observed. Public satellites take observations across large areas of the globe at once. The third-party data analytics tool splits these observations into smaller regions, which are 0.1° latitude by 0.1° longitude in size. A facility is said to be observed by a public satellite if the region that contains it has been processed. Plumes are matched to facilities based on plume source location, facility location, and status (e.g., active or plugged wells).
Step 3: Estimate Emissions at the Facility Level
A key challenge in estimating emissions for a facility that has been observed is that no satellite can view a site constantly. Therefore, an algorithm must be used to estimate what is happening between observations to generate an accurate total.
Fig. 7 illustrates three hypothetical sites that are emitting 0% of the time, 60% of the time, and 20% of the time, over the course of 10 days. Five observations of the region where the three sites are located were taken by satellite. For Site A, no emission was detected in these five observations, and no emission occurred over the course of 10 days, meaning the site is found to emit 0% of the time. For Site B, three emissions were detected in these five observations, and six emissions actually occurred over the course of 10 days, meaning the site was found to emit 60% of the time. For Site C, two emissions were detected in these five observations, and four emissions actually occurred over the course of 10 days, meaning the site is found to emit 20% of the time. The intuition and assumption of this method is that the average emissions rate (including nulls) that is observed at a site is representative of the average emissions rate for that site overall.
Chart describing the stochastic nature of emissions from three hypothetical O&G sites emitting 0% of the time, 60% of the time, and 20% of the time, respectively.
Chart describing the stochastic nature of emissions from three hypothetical O&G sites emitting 0% of the time, 60% of the time, and 20% of the time, respectively.
This calculation follows that the emission rate estimate () for a site is given by summing the total emissions within a period and dividing that by the total number of observations within the same period for a given site. This method allows the estimation of emissions totals without making decisions about the persistence or intermittency of plumes, as this method works well for both.
To estimate total emissions for a period, the emission rate estimate, , measured in kg∙h−1, is multiplied by the number of hours in the period. Where there are no observations of a facility for a given period, estimates are made based on the emissions profiles of similar facilities belonging to the same company, where the assumption is that unobserved facilities emit at the same rate as observed.
Aggregating data from distinct satellites poses a challenge. One of the most important of these challenges is the differences in MDLs across different instruments. The targeted satellites used in this study can detect emissions as small as 100 kg∙h−1, whereas ESA’s Sentinel 5P TROPOMI instrument and their Sentinel 2 instrument have lower detection limits of 10 t∙h−1 and between 1 and 3 t∙h−1, respectively (Jacob et al. 2022; Sherwin et al. 2023). In addition to detecting and measuring emissions, the targeted satellites also estimate unobserved emissions for their clients, and when doing so, the different lower limits of detection for each satellite must be considered and modeled accordingly.
Step 4: Estimate Emissions at a Company Level
To generate total emissions for a single company, each facility is mapped to an operator, and these operators are mapped to the ultimate owning company. The emissions of a company are the sum of all the facilities they operate.
Limitations
Like any model, there are limitations to the emissions estimation algorithm. These limitations occur from data processing to algorithm development. One limitation to using satellite data for estimating emissions is associated with MDLs, as some emissions go undetected due to satellite MDLs. The aforementioned method for estimating emissions below the MDL assumes a single MDL for each satellite (GHGSat and Sentinel 2, respectively). However, the MDL is also dependent on surface reflectance. Future algorithm versions will assume a series of different MDLs per satellite type by accounting for different surface reflectance levels. Another limitation is associated with estimating company-level emissions, which is dependent on having a complete facilities database. However, facilities data are sparse and continuously changing. For example, in highly competitive markets such as the US, wellheads can be traded frequently, changing their ownership. Although the targeted satellite data set is growing daily, there is a continuous effort to improve the data and keep a record of historical facilities ownership. To improve the data set, there is an aim to continue building upon the in-house facility detector algorithms and partnering with more institutions to obtain timely and accurate operator and ownership data. It is acknowledged that to provide a whole-rounded emissions product, offshore emissions need to be accounted for. Offshore emissions are more challenging to detect with satellite technology due to their surface reflectance behavior over water bodies. For this, a research project to detect offshore emissions using Glint mode is under way. On 30 September 2022, this modality was tested during the explosions of Nord Stream II over the Baltic Sea (GHGSat Inc 2022). The expectation is to continue gathering offshore data to eventually incorporate it into the ESG model when sufficient.
Validation of Measurements
Controlled releases of natural gas are planned, and coordinated tests are planned to independently validate the performance and accuracy of the measuring instrument. An independent, single-blinded validation study of different methane-detecting satellites and their measurement capabilities was performed in 2021 (Sherwin et al. 2023). In the study, several different data processing teams and five different satellites (GHGSat-C2, Sentinel 2, PRISMA, Worldview 3, and Landsat 8) were assessed. Sherwin et al. showed that GHGSat-C2 achieved a quantification accuracy better than ±20% for each emission studied, the best accuracy recorded in the research. Also, it highlights that it was the only satellite that detected an emission less than 1000 kg∙h−1, with the smallest emission detected at 197 kg∙h−1 (Sherwin et al. 2023). The team analyzing Sentinel 2 measurements was able to detect most of the emissions due to the daily revisit time with an average accuracy of ±45.5%.
Within the analytic suite for ESG product presented here, rather than merely complementing targeted satellite measurements with Sentinel 2 or Sentinel 5P observations, the public data are more often used to guide the more targeted observations (i.e., tip and cue). This approach ensures that the valuable characteristics of each satellite are harnessed, that is, the coverage and frequent revisits of public satellite data are enriched with the accuracy of the targeted satellite data.
In the US, O&G operators currently estimate their emissions using a method based on emissions factors, but studies have shown that those estimates are understated by approximately 60% (Alvarez et al. 2018). Emissions estimates based on direct measurements, such as those intended to be produced by the analytic suite for ESG product, are considered to be more accurate than engineering calculations and emissions factors due to the level of granularity in the data (Cooper et al. 2022).
The analytic suite for ESG product builds on the accuracy of its own emission measurements and those from third-party data sources to provide operators and other stakeholders with measurement-based and timely emissions estimations for O&G operators. There are many tools or initiatives that exist today to assist operators in calculating their emissions for their annual disclosures such as the Science Based Target initiative or the Sustainable Accounting Standard Board reporting framework. Although they are well defined, they have some pitfalls that diminish their reliability or attractiveness as a sufficient metric or benchmark for operators to use to differentiate themselves from one another. Emissions are often reported annually and are therefore not current; they are also estimated without using actual measurements and without any third-party involvement. The proposed solution, by its nature, involves third parties and furthermore aspires to provide timely and unbiased emissions estimates based on measurements.
Results
When aggregating total emissions for an operator, emissions totals are calculated on a per quarter, monthly, and annual basis. Table 1 shows a summary of a sample data set of annual emissions for three companies. An expanded data set is included in Table S-1 of the Supplementary Materials. The first emissions total is the total source emission rate, showing the sum of all emission rates observed in metric tons of methane per hour (tCH4∙h−1). Next, the total methane emissions for all company facilities are estimated via the method described in the Methodology section. This metric is calculated from observed facilities and plumes, includes estimates for methane emissions from unobserved facilities, and then is adjusted to take into account plumes too small to be detected by satellites. Data aggregated over quarters and months are adjusted using annual emissions calculations to improve consistency. The final emissions total builds on the previous total by converting it to carbon dioxide equivalents (CO2e) using the adjustment factor of 25, which is in line with UNFCCC reporting guidelines (UNFCCC 2014).
Subset of the synthetic output data set from the analytic suite for ESG product.
Company . | Period . | Geographic Area . | Start Date . | End Date . | Duration . | Total Observations . | Total Source Rate (tCH4/h) . | Total Methane Emissions for All Co. Facilities (tCH4) . | Total Methane Emissions for All Co. Facilities (tCO2e) . |
---|---|---|---|---|---|---|---|---|---|
Co. 1 | 2022 | Anadarko | 01/01/2022 | 31/12/2022 | 365 | 4 | – | – | – |
Co. 1 | 2022 | Denver | 01/01/2022 | 31/12/2022 | 365 | 2 | – | – | – |
Co. 1 | 2022 | Greater Green River | 01/01/2022 | 31/12/2022 | 365 | 27 | – | – | – |
Co. 1 | 2022 | Marietta | 01/01/2022 | 31/12/2022 | 365 | 396 | – | – | – |
Co. 1 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 4,668 | 4 | 33 371 | 834 264 |
Co. 1 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 4 | – | – | – |
Co. 1 | 2022 | Western Gulf | 01/01/2022 | 31/12/2022 | 365 | 17 | – | – | – |
Co. 2 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 8,271 | 11 | 12 247 | 306 178 |
Co. 3 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 767 | 1 | 1074 | 26 842 |
Co. 3 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 1,818 | 1 | 9912 | 247 797 |
Co. 4 | 2022 | Anadarko | 01/01/2022 | 31/12/2022 | 365 | 67 | – | – | – |
Co. 4 | 2022 | Canada | 01/01/2022 | 31/12/2022 | 365 | 49 | – | – | – |
Co. 4 | 2022 | Denver | 01/01/2022 | 31/12/2022 | 365 | 5 | – | – | – |
Co. 4 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 24 | – | – | – |
Co. 4 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 11,371 | 3 | 18 113 | 452 823 |
Co. 4 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 10 | – | – | – |
Co. 4 | 2022 | Williston | 01/01/2022 | 31/12/2022 | 365 | 6 | – | – | – |
Co. 5 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 1,766 | – | – | – |
Co. 5 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 3,925 | 1 | 2575 | 64 379 |
Co. 5 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 156 | – | – | – |
Company . | Period . | Geographic Area . | Start Date . | End Date . | Duration . | Total Observations . | Total Source Rate (tCH4/h) . | Total Methane Emissions for All Co. Facilities (tCH4) . | Total Methane Emissions for All Co. Facilities (tCO2e) . |
---|---|---|---|---|---|---|---|---|---|
Co. 1 | 2022 | Anadarko | 01/01/2022 | 31/12/2022 | 365 | 4 | – | – | – |
Co. 1 | 2022 | Denver | 01/01/2022 | 31/12/2022 | 365 | 2 | – | – | – |
Co. 1 | 2022 | Greater Green River | 01/01/2022 | 31/12/2022 | 365 | 27 | – | – | – |
Co. 1 | 2022 | Marietta | 01/01/2022 | 31/12/2022 | 365 | 396 | – | – | – |
Co. 1 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 4,668 | 4 | 33 371 | 834 264 |
Co. 1 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 4 | – | – | – |
Co. 1 | 2022 | Western Gulf | 01/01/2022 | 31/12/2022 | 365 | 17 | – | – | – |
Co. 2 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 8,271 | 11 | 12 247 | 306 178 |
Co. 3 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 767 | 1 | 1074 | 26 842 |
Co. 3 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 1,818 | 1 | 9912 | 247 797 |
Co. 4 | 2022 | Anadarko | 01/01/2022 | 31/12/2022 | 365 | 67 | – | – | – |
Co. 4 | 2022 | Canada | 01/01/2022 | 31/12/2022 | 365 | 49 | – | – | – |
Co. 4 | 2022 | Denver | 01/01/2022 | 31/12/2022 | 365 | 5 | – | – | – |
Co. 4 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 24 | – | – | – |
Co. 4 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 11,371 | 3 | 18 113 | 452 823 |
Co. 4 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 10 | – | – | – |
Co. 4 | 2022 | Williston | 01/01/2022 | 31/12/2022 | 365 | 6 | – | – | – |
Co. 5 | 2022 | Fort Worth | 01/01/2022 | 31/12/2022 | 365 | 1,766 | – | – | – |
Co. 5 | 2022 | Permian | 01/01/2022 | 31/12/2022 | 365 | 3,925 | 1 | 2575 | 64 379 |
Co. 5 | 2022 | Tx-La-Ms Salt | 01/01/2022 | 31/12/2022 | 365 | 156 | – | – | – |
Utilizing the data set shown in Table 1 , Fig. 8 can be generated to depict a graphical analysis of the number of observations and estimated total methane emissions per region.
Observations and emissions by region using a synthetic output data set from the analytic suite for ESG.
Observations and emissions by region using a synthetic output data set from the analytic suite for ESG.
Processing and presenting the data in this way provides an operator with impartial emissions estimates for its entire operations based on real measurements. The unbiased nature of this methodology allows an operator to benchmark itself in local and global geographies. A useful benchmark metric is the detection intensity metric. It tells the reader the percentage of emission detections per site observation, regardless of the size of the site or the scale of the operator. Using the data set in Table S-1 in the Supplementary Materials, a comparison of detection intensity between 10 US companies is shown in Fig. 9 .
Benchmarking O&G companies in the US using the detection intensity metric based on the synthetic output data set.
Benchmarking O&G companies in the US using the detection intensity metric based on the synthetic output data set.
In cases where they are performing better than competitors, this can be used as a differentiator to attract potential customers seeking methane-conscious options, or alternatively, it can be used as an indicator that improvements need to be made if performance against competitors reflects attention to these different metrics. The data can also be used to unlock capital as more investors seek to meet their ESG KPIs. The output data set shows the emissions measured directly by the targeted satellite constellation, which then gets processed by proprietary algorithms, providing emissions estimates for the known infrastructure at the basin and country levels. Hence, this analytic suite for ESG product attempts to provide further insight into the aggregated methane emissions data.
Discussion
A tiered monitoring system that incorporates satellite and aircraft measurements with emissions analytics allows for the detection, quantification, and analysis of the emission events and provides the potential for more frequent methane emissions monitoring than the current practices (Deglint et al. 2021; Esparza and Gauthier 2021; Esparza et al. 2023). de Gouw et al. (2020) showed that the use of Sentinel 5P to monitor for methane emissions from O&G infrastructure in active areas in the US can provide the case for enabling timely monitoring for the very large leaks or ultra emitters (e.g., greater than 3000 kg∙h−1; de Gouw et al. 2020). The aircraft platform alone is capable of detecting lower emissions rates with a fine spatial resolution over onshore (Conrad et al. 2023) and offshore (Ayasse et al. 2022) facilities. However, the platform requires constant deployment to accurately characterize an area, which in turn may not be cost-effective. Nonetheless, Esparza et al. discuss this combination of technologies in a section of the Permian Basin, reporting that the tiered system could be equivalent or superior in methane reductions at a lower cost when compared with optical gas imaging surveys alone (Esparza et al. 2023). This system can not only find the very large leaks but incorporates surveys for the smaller and more frequent leak rates.
Fig. 10 shows a detection performed by the satellite platform (left) on 18 February 2022. On 22 February 2022, the same site was surveyed with the aircraft and was found that the leak persisted.
Concentration maps for emissions detected at the same site in the Permian Basin. Left image: Methane concentration plume observed with satellite overlaid on Mapbox® map background data. Right image: Methane concentration plume observed with an airborne instrument overlaid on a visible light reflectance image acquired at the same time with the onboard auxiliary camera.
Concentration maps for emissions detected at the same site in the Permian Basin. Left image: Methane concentration plume observed with satellite overlaid on Mapbox® map background data. Right image: Methane concentration plume observed with an airborne instrument overlaid on a visible light reflectance image acquired at the same time with the onboard auxiliary camera.
The detections shown in Fig. 10 demonstrate the effectiveness of a system that incorporates satellite and aircraft platforms. Tiered systems that incorporate different remote sensing platforms have the potential of decreasing the time between surveys while gaining additional insight when the data are processed by an analytic suite. In turn, results from the suite can serve to demonstrate progress in different ESG metrics.
Conclusions
A data analytics system aiming to assist the O&G industry was presented. The advantage of this system lies in the combination of publicly available satellite data and high-resolution data from targeted satellites and aircraft into a purpose-built data analytics suite. Analytical results can be determined for any geographical region, over any period with sufficient observations. These analytical results, in conjunction with frequent satellite methane monitoring and scheduled aircraft monitoring campaigns, can provide a top-down characterization of the equipment and infrastructure’s performance of the O&G facilities.
The satellite and aircraft platforms are scalable. With the increase in the satellite constellation, more capacity becomes available to target O&G infrastructure more frequently. GHGSat is currently in the process of increasing its constellation to a total of 10 commercial satellites dedicated to methane detection plus one for carbon dioxide detection(GHGSat Inc 2023). Also, the company is increasing the number of aircraft sensors with the prospect of improved performance. The data analytics suite for ESG is constantly evolving and is flexible enough to add any relevant data sets that become available, including satellite data from current and existing satellites.
Article History
Original SPE manuscript received for review 14 March 2023. Revised manuscript received for review 9 May 2023. Paper (SPE 215818) peer approved 11 May 2023.
Supplementary materials are available in support of this paper and have been published online under Supplementary Data at https://doi.org/10.2118/215818-PA. SPE is not responsible for the content or functionality of supplementary materials supplied by the authors.