Estimating Methane Emissions by Integrating Satellite Regional Emissions Mapping and Point-Source Observations: Case Study in the Permian Basin

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I have carefully reviewed the manuscript titled "Estimating Methane Emissions by Integrating Satellite Regional Emissions Mapping and Point-Source Observations: Case Study in the Permian Basin." This study estimates methane emissions in the Permian Basin by combining regional monitoring from TROPOMI with point-source detections from CARBON MAPPER. This is a valuable effort, as exploring how point-source satellites and regional monitoring satellites can work together is an important issue.

Due to time constraints, I did not go through the calculation process in detail, and my comments focus on the experimental results section. First, I find the ratio between the event-based results and the divergence method results to be too high. The detection thresholds for point-source satellites typically range between 100 kg/h (TANGO) and 450 kg/h (EMIT), which means many emissions will go undetected. According to prior studies by Omara et al., this ratio in the Permian Basin is generally around 10–15%. I suspect that the main reason for the discrepancy lies in an underestimation from the divergence method. I recommend the authors carefully recheck the relevant calculations.

Second, the event-based method itself is a very interesting topic. However, the cited work by GAO (2025) does not appear in the reference list, so I was unable to access its details. I suggest the authors verify the references and, if possible, include a detailed description of the method in an appendix.

 

Author Response

Comment 1: Due to time constraints, I did not go through the calculation process in detail, and my comments focus on the experimental results section. First, I find the ratio between the event-based results and the divergence method results to be too high. The detection thresholds for point-source satellites typically range between 100 kg/h (TANGO) and 450 kg/h (EMIT), which means many emissions will go undetected. According to prior studies by Omara et al., this ratio in the Permian Basin is generally around 10–15%. I suspect that the main reason for the discrepancy lies in an underestimation from the divergence method. I recommend the authors carefully recheck the relevant calculations.

Response 1: Thank you for your thoughtful comments. We agree that the discrepancy between the two methods need to be careful discussed, and we actually see two possible contributing factors.

First, as you rightly noted, many plumes are likely missed by TROPOMI given its detection threshold. At the same time, the divergence method does not depend on individual plume detection but instead estimates emissions from the transport of methane column mixing ratios above and below the planetary boundary layer. While smaller plumes (e.g., <100 kg h⁻¹) may be undetected, their methane will still mix into the atmosphere on annual timescales. We therefore think that part of the underestimation may mainly from how the divergence method averages emissions and from horizontal transport of methane outside the study region.

Second, the event-based method is designed not only to sum detected plumes but also to account for potentially undetected ones using a model calibrated from historical plume distributions. This feature helps capture emissions that satellites may miss. Therefore, the ratio will be higher than just aggregate all detected plumes.

Comment 2: Second, the event-based method itself is a very interesting topic. However, the cited work by GAO (2025) does not appear in the reference list, so I was unable to access its details. I suggest the authors verify the references and, if possible, include a detailed description of the method in an appendix.

Response 2: Thanks for reminding us. We’ve verified and updated the reference list. We are sorry for missing the very important reference – Gao et al., 2025. This work is currently still under review. The preprint version of this paper can be accessed at https://eartharxiv.org/repository/view/8906/ or https://doi.org/10.31223/X59M72. The paper includes details of the methodology and an example of how to use it to estimate annual emissions.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a robust and well-structured study on the integration of satellite-based methods for estimating methane emissions in the Permian Basin. The methodology is described adequately and supported by detailed quantitative and comparative analysis. The datasets and references are up-to-date and credible, and the results are consistent with previous findings. The manuscript provides highly valuable information. I recommend the publication of this article, subject to some minor revisions.

 

• Lines 54-66: Consider adding an indication of the detection limits for point-source detectors and regional mappers to clarify the operational differences between the two approaches.

 

• Line 57: The phrase “at the site-level on the ground” might read more smoothly as “at the site level”, since “on the ground” is implied and may be redundant.

 

• Lines 70-71: It may be helpful to include a quantitative definition or range for "super-emitter" to better contextualise the detection capabilities of regional mappers.

 

• Line 137: Consider rephrasing “comprising 13 or 14 orbits of data” to “comprising approximately 13–14 orbits per day” for clarity.

 

• Line 164: The acronym "PBL" may not be immediately clear to some readers. As done in Appendix A, it would be preferable to use the full definition the first time it is mentioned in the text.

 

• Lines 235-248: The event-based (PRE) method relies on certain assumptions regarding event duration and the availability of detections/non-detections. Clarifying how the method performs under limited observation frequency or data gaps (e.g. due to cloud cover or revisit times) would be helpful in assessing its robustness.

 

• Lines 255-262: It would be interesting to understand whether the absence of resolved events had a significant impact on the results. In other words, whether this term represents an essential component of Equation (5), or if its contribution is relatively minor in the overall emissions estimate.

 

• Lines 459-461: The R² value is described as indicating a "strong correlation", though in some contexts it might be considered moderate. It would be helpful either to further support this interpretation or to consider rephrasing it as "moderate".

 

• The manuscript presents a detailed and valuable analysis of methane monitoring using satellites and aerial campaigns. To further enrich the discussion on multi-scale monitoring capabilities, it might be beneficial to briefly incorporate the emerging role of drones (UAVs). These platforms offer exceptional operational flexibility and extremely high spatial resolution, proving useful for characterising individual sources and providing fine-scale validation data (e.g., Fosco et al., 2025; Corbett et al., 2023; Golston et al., 2018), thereby completing the comprehensive overview of observation strategies.

Author Response

1. Lines 54-66: Consider adding an indication of the detection limits for point-source detectors and regional mappers to clarify the operational differences between the two approaches.

R1: Thank you for suggestion. We’ve added the minimum detection limit for both point-source detectors (~100 kg hr^-1) and regional mappers (~25000 kg hr^-1) in the revised manuscript. Please see lines 57 and 64.

2. Line 57: The phrase “at the site-level on the ground” might read more smoothly as “at the site level”, since “on the ground” is implied and may be redundant.

R2: Thank you for the suggestion. We’ve removed “on the ground” please see line 54.

3. Lines 70-71: It may be helpful to include a quantitative definition or range for "super-emitter" to better contextualise the detection capabilities of regional mappers.

R3: Thank you for the suggestion. We are following the definition from Lauvaux et al. (reference 24), which the super-emitter’s quantitative detection limit is 25000 kg hr^-1. We’ve added it in the revised manuscript.

4. Line 137: Consider rephrasing “comprising 13 or 14 orbits of data” to “comprising approximately 13–14 orbits per day” for clarity.

R4: Thank you for the suggestion. We’ve made change. Please see line 135.

5. Line 164: The acronym "PBL" may not be immediately clear to some readers. As done in Appendix A, it would be preferable to use the full definition the first time it is mentioned in the text.

R5: Thank you for the suggestion. The definition of PBL has been added at the first appearance. We also checked the uses of other abbreviations throughout the manuscript.

6. Lines 235-248: The event-based (PRE) method relies on certain assumptions regarding event duration and the availability of detections/non-detections. Clarifying how the method performs under limited observation frequency or data gaps (e.g. due to cloud cover or revisit times) would be helpful in assessing its robustness.

R6: Thanks for your comments; you are absolutely correct. The accuracy of the method is closely tied to the availability of detections and non-detections, or in other words, the number of events. Unfortunately, we are unable to assess this quantitatively, as no true emissions values are available for verification. This will definitely be one of our focuses for future work. We have added clarification in Section 3.2 (lines 378–389).

7. Lines 255-262: It would be interesting to understand whether the absence of resolved events had a significant impact on the results. In other words, whether this term represents an essential component of Equation (5), or if its contribution is relatively minor in the overall emissions estimate.

R7:  Thank you for your comments. Missing resolved events (RE) will impact the overall annual emissions estimation; however, this term is not an essential component of the framework. Before expanding on this point, the general equation for calculating emissions from any type of event is: rate × duration.

Not all operational data provide rate estimates, but most provide duration, since it is typically known in oil and gas operations (e.g., scheduled or intentional events). The primary difference between RE and PRE (partially resolved events) lies in duration estimation. For RE, duration is directly available, and its uncertainty is almost negligible. In contrast, PRE relies on non-detections to estimate event duration, making it strongly dependent on data availability or survey/measurement frequency (e.g., satellite revisit time; related to the previous comment).

With the availability of RE, the temporal resolution of events at a given site is finer. As a result, the period requiring simulation of unresolved events becomes more constrained, and the overall simulation results become more accurate. Subsequently, the annual site-level emissions estimation will become more accurate as well. More details of the event-based framework can be found in Gao et al. (2025, preprint available at https://eartharxiv.org/repository/view/8906/ or https://doi.org/10.31223/X59M72). This work is currently under review.  

8. Lines 459-461: The R² value is described as indicating a "strong correlation", though in some contexts it might be considered moderate. It would be helpful either to further support this interpretation or to consider rephrasing it as "moderate".

R8: Thank you for the suggestion. We’ve replaced the “strong” with “moderate” and revised the sentence. Please see line 469.

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript focuses on methane emissions in the Permian Basin. The study proposes an innovative methodological framework that integrates regional divergence methods (TROPOMI ) and point-source event-based approaches (Carbon Mapper) to quantify methane emissions in the Permian Basin, demonstrating clear methodological contributions and empirical value. The research design is sound, the data coverage is comprehensive, and the results align with historical studies, revealing a downward trend in emissions. However, the following critical revisions are required before the manuscript can be accepted:

 

1. The font type and size in figure captions need to be modified to differentiate them from the main text. 
2. In Figure 7, the caption "The error bars indicate the" is incomplete and lacks a period.
3. Professional terminology should be standardized throughout the manuscript. For example, after the first mention of "Methane (CH₄)," subsequent references should consistently use either "Methane" or "CH₄."
4. Line 362 references "Figures 6d, 6e, and 6f," but Figure 6 in the manuscript only contains subfigures (a) and (b). This likely refers to Figure 5d–f, as the content (fitting parameters a and b, mean, and standard deviation) matches that section. If these figures belong to supplementary material, please clarify this explicitly (e.g., state "Figures S6d–f in the Supporting Information"). 
5. The manuscript does not explain why the divergence method underestimates instantaneous high-emission events in overlapping grid cells or quantify the contribution of the event-based method to the integrated results.
6. The uncertainty of the integrated results is presented numerically but lacks an explanation of how it was calculated.
7. Recent key studies on emission reduction mechanisms are not discussed. The conclusion attributing the decline solely to "regulatory efforts, operational improvements, and increased public and industry scrutiny has been evident over the past six years. " lacks credibility without supporting evidence.
8. The resolution of Figures 4–6 is insufficient; hotspot regions appear blurry and lack clarity for accurate interpretation.
9. The introduction fails to distinguish this study from similar research. Add a paragraph at the end of the introduction emphasizing that this integrated method addresses the core issue of "underestimated instantaneous high-emission events" for the first time.
10. The term "PBL" (Planetary Boundary Layer) must be defined upon its first appearance in Appendix A.
11. Supplement Appendix A with the bilinear interpolation formula and error analysis.
12. The units in the manuscript are inconsistent. In Appendix B, "MT yr⁻¹" should likely be replaced with "Tg yr⁻¹" even though Tg and MT represent the same order of magnitude. This change is necessary to ensure standardized notation and align with conventional scientific reporting practices.

Comments for author File: Comments.pdf

Author Response

1. The font type and size in figure captions need to be modified to differentiate them from the main text. 

R1: Thanks for the suggestion. We have modified the font size to meet the formatting requirements of Remote Sensing. All captions are now set to font size 9.

2. In Figure 7, the caption "The error bars indicate the" is incomplete and lacks a period.

R2: Thank you for pointing this out. We’ve revised the typo, and the complete sentence should be “The error bars indicate the estimates uncertainties.” Please see line 400.

3. Professional terminology should be standardized throughout the manuscript. For example, after the first mention of "Methane (CH₄)," subsequent references should consistently use either "Methane" or "CH₄."

R3: Thanks for the suggestion. CH₄ are now being used throughout the manuscript.  

4. Line 362 references "Figures 6d, 6e, and 6f," but Figure 6 in the manuscript only contains subfigures (a) and (b). This likely refers to Figure 5d–f, as the content (fitting parameters a and b, mean, and standard deviation) matches that section. If these figures belong to supplementary material, please clarify this explicitly (e.g., state "Figures S6d–f in the Supporting Information"). 

R4: Thank your for pointing out this typo. You are correct. All referenced figures are subplots of Figure 5. We’ve made the change. Please see lines 351 and 356.   

5. The manuscript does not explain why the divergence method underestimates instantaneous high-emission events in overlapping grid cells or quantify the contribution of the event-based method to the integrated results.

R5: Thanks for your comments. The key limitation of the divergence method is that it relies on calculating the change in methane mixing ratios above and below the PBL. Therefore, when averaging the daily divergence to quantify regional emissions for each cell, short-duration events may not be well captured. This is briefly mentioned in Lines 414–420.

6. The uncertainty of the integrated results is presented numerically but lacks an explanation of how it was calculated.

R6: The overall uncertainty was estimated as the root mean square error (RMSE) across all grid cells. In other words, we recalculated the RMSE using the uncertainties of each grid cell derived from the divergence method, together with the uncertainties of the event-based substituted grid cells. We’ve added a sentence to clarify that (Line 427).

7. Recent key studies on emission reduction mechanisms are not discussed. The conclusion attributing the decline solely to "regulatory efforts, operational improvements, and increased public and industry scrutiny has been evident over the past six years. " lacks credibility without supporting evidence.

R7: The investigations on the emission reduction mechanisms are rare for now. The findings derived from airborne measurements suggest methane emissions in Permian basin trended down in recent years. But the exact causes of the reduction have not been fully unpacked. Moreover, our work focuses on the application of two different methods in methane emissions from O&G production regions. The deep discussion of the emission reduction mechanisms is not in scope of this work. To make rigorous conclusions, we’ve revised the statement (see lines 456 - 461) as follows:

“This downward trend suggests that the impact of regulatory efforts, operational improvements, increased efficiency of leak detection and repair program with advanced detection techniques, and increased public and industry scrutiny on reducing CH₄ emission from O&G sector is noteworthy over the past six years (Sherwin et al., 2025; Leblanc 2025). Further investigations into emission reduction mechanisms are needed to unpack the precise causes.”

8. The resolution of Figures 4–6 is insufficient; hotspot regions appear blurry and lack clarity for accurate interpretation.

R8: Thank you for the suggestion. We have attached high-resolution versions of all figures. The reduced quality in the PDF file is likely due to automatic resolution compression.

9. The introduction fails to distinguish this study from similar research. Add a paragraph at the end of the introduction emphasizing that this integrated method addresses the core issue of "underestimated instantaneous high-emission events" for the first time.

R9: Thanks for your suggestions. We’ve added the paragraph at the end of the introduction. Please see lines 100 -106.

10. The term "PBL" (Planetary Boundary Layer) must be defined upon its first appearance in Appendix A.

R10: We’ve defined abbreviation at its first appearance in both manuscript and appendix. We also checked uses of other abbreviations to ensure the full name of the abbreviations were provided at the first appearance.

11. Supplement Appendix A with the bilinear interpolation formula and error analysis.

R11: Thank you for your comments. We’ve added equations of bilinear interpolation and calculating root mean square errors to Appendix A. Please see lines 546 – 548 and lines 560 – 564.

12. The units in the manuscript are inconsistent. In Appendix B, "MT yr⁻¹" should likely be replaced with "Tg yr⁻¹" even though Tg and MT represent the same order of magnitude. This change is necessary to ensure standardized notation and align with conventional scientific reporting practices.

R12: Thank you for the suggestion. We’ve changed the unit in our revised submission. However, we kept the unit of t y^-1 for the results from Omara et al., 2024 to acknowledge the potent bias when upscale the originally reported hourly emission to annual emission.  

Reviewer 4 Report

Comments and Suggestions for Authors

Overview:

The methane emissions from the Permian Basin in the USA in 2024 were estimated by the divergence method using TROPOMI data and the point-source event-based method using Carbon Mapper’s data separately, then the method of integrating the event-based estimates with the divergence-based regional estimates was proposed to capture the short-duration super-emitter events. The manuscript provides a valuable method for monitoring regional methane emissions using satellite data. The manuscript is well written, and the structure is well organized. The manuscript is appropriate for the Journal, but it needs minor revision.

 

Specific comments:

1. Please set “4” in CH4 as a subscript throughout the manuscript.
2. Line 51: “Therefore, several launched satellites equipped with instruments” may be “Therefore, several launched satellites are equipped with instruments”.
3. Please put the right panel of Figure 1b to the bottom, as the font in Figure 1b in the right panel is too small.
4. It’s better to separate the content of Figure 2 into two panels, one panel with one method, as the font in Figure 2 is too small.
5. Line 233: “non-detect” -> “non-detection”.
6. Line 261: “Since operational data are not available. We only considered emissions from PREs” ->“Since operational data are not available, we only considered emissions from PREs”.
7. Line 289: “can be calculated” should be deleted.
8. Line 334-335: please describe simply the sources of the uncertainties of emissions.
9. Line 357: “Figures 6a to 6c” -> “Figures 5a to 5c”.
10. Line 362: “(Figures 6d, 6e, and 6f)” -> “(Figures 5d, 5e, and 5f)”.
11. Line 370: “event-based” -> “Event-based”.
12. Line 413: “…was not highlighted using the divergence method alone.” may be “…was not highlighted using the event-based method alone.”?

Author Response

1. Please set “4” in CH4 as a subscript throughout the manuscript.

R1: Thank you for the suggestion. Changes made throughout the manuscript and Appendix.

2. Line 51: “Therefore, several launched satellites equipped with instruments” may be “Therefore, several launched satellites are equipped with instruments”.

R2: Thank you for the suggestion. We’ve revised the sentence.

3. Please put the right panel of Figure 1b to the bottom, as the font in Figure 1b in the right panel is too small.

R3: Thank you for the suggestion. We have attached high-resolution versions of all figures. The reduced quality in the PDF file is likely due to automatic resolution compression.

4. It’s better to separate the content of Figure 2 into two panels, one panel with one method, as the font in Figure 2 is too small.

R4: Thank you for the suggestion. We have attached high-resolution versions of all figures. The reduced quality in the PDF file is likely due to automatic resolution compression.

5. Line 233: “non-detect” -> “non-detection”.

R5: Thank you for pointing out. Change made.

6. Line 261: “Since operational data are not available. We only considered emissions from PREs” ->“Since operational data are not available, we only considered emissions from PREs”.

R6: Thank you for pointing out. Change made.

7. Line 289: “can be calculated” should be deleted.

R7: The redundant words have been deleted in the revised manuscript.

8. Line 334-335: please describe simply the sources of the uncertainties of emissions.

R8: Thank you for your suggestions. We have added a paragraph clarifying the sources of uncertainty in event-based emissions estimation. Please see lines 378–389.

9. Line 357: “Figures 6a to 6c” -> “Figures 5a to 5c”.

R9: Thank you for catching this typo. Changes made.

10. Line 362: “(Figures 6d, 6e, and 6f)” -> “(Figures 5d, 5e, and 5f)”.

R10: Thank you for catching this typo. Changes made.

11. Line 370: “event-based” -> “Event-based”.

R11: Thanks, change made.

12. Line 413: “…was not highlighted using the divergence method alone.” may be “…was not highlighted using the event-based method alone.”?

R12: Thank you for catching this. Here we meant to state that not the entire Delaware basin was highlight from the divergence method. As indicated by Figure 4e, only the southern part (i.e., Texas side) of the Delaware basin was identified as emission hotspot from divergence method. The revised sentences are as follows:

“For example, as shown in Figure 4e, the New Mexico portion (norther side) of the Delaware Basin was not highlighted using the divergence method alone, but was identified with notable emissions using the event-based method (Figure 6b).”