the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Mar 2026
Methane Emissions Estimation from China's Leading Coal Production Hub: A Hybrid Hyperspectral Satellite Observations and Emission Inventory Framework
Abstract. Accurate estimation of coal mine methane (CMM) emissions in Shanxi Province, China's leading coal production hub, is essential for mitigating China's anthropogenic methane emissions. Hyperspectral remote sensing is an emerging method for real-time methane monitoring with significant potential for optimizing CMM emission factors. However, limited satellite revisit frequencies can introduce biases in CMM emission estimates. To address these issues, we developed a Hierarchical Bayesian Inversion Algorithm utilizing time-series observations from seven hyperspectral satellites in Shanxi (2019–2023), comprising 215 methane plumes from 26 coal mines, to estimate annual CMM emission rates with limited satellite revisit frequency. Subsequently, we integrated multi-source satellite observations with inventory data to estimate CMM emissions in Shanxi province. Our analysis yields a CMM emission factor of (7.9 ± 1.4)×10-3 Tg/Mt for Shanxi, with CMM emissions reaching 11 ± 2 Tg/yr in 2023. We demonstrate that CMM emissions follow a right-skewed distribution in Shanxi Province, where low-frequency extreme methane emission events (≥10000 kg/h) constitute approximately 25 % of all time-series observations. Additionally, our results reveal that capacity reduction policies initially decreased CMM emissions, but subsequent production recovery led to emission increases, with asymmetric responses to coal price fluctuations. Our findings establish a novel strategy for CMM accounting from hyperspectral satellite observations.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(11943 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (extended)
- RC1: 'Comment on egusphere-2025-4610', Anonymous Referee #4, 31 Mar 2026 reply
-
RC2: 'Comment on egusphere-2025-4610', Anonymous Referee #5, 26 Jun 2026
reply
Reviewer Report General assessment
The manuscript addresses an important and timely topic, namely the satellite-based quantification of methane emissions from coal mining regions. This subject is of broad international relevance, particularly in the context of the Global Methane Pledge and the rapidly growing capabilities of satellite observing systems. The focus on Shanxi province is well justified, as it represents one of the most significant coal-producing regions globally and a major source of anthropogenic methane emissions.
The general idea of the study is sound. Coal mining systems, especially ventilation shafts and drainage systems, represent relatively strong and spatially localized emission sources, which in principle makes them suitable for remote sensing–based plume detection and quantification. The use of hyperspectral satellite observations combined with plume detection and mass balance approaches is well established in the literature.
The manuscript claims novelty based on two aspects: first, the use of multiple satellite platforms, and second, the application of Bayesian inversion to reconcile observations and derive emission estimates. Both elements are potentially valuable and could constitute a meaningful contribution. However, in its current form, the manuscript does not adequately demonstrate the robustness, transparency, and physical interpretability required for publication in a journal such as Atmospheric Chemistry and Physics.
The core issue is that the study remains largely descriptive, while its most innovative components are underdeveloped. At the same time, key methodological and interpretative aspects are either insufficiently justified or entirely missing. In particular, the manuscript lacks a critical assessment of uncertainties, does not properly validate the multi-satellite approach, and does not incorporate the necessary geological and engineering context required to interpret methane emissions from coal mines.
As a result, although the study has clear potential, substantial revision is required before it can be considered for publication.
Multi-satellite approachThe use of seven satellite platforms is presented as a central strength of the manuscript. While this approach indeed offers the potential to increase observational coverage and improve statistical robustness, it also introduces significant methodological complexity that is not sufficiently addressed.
Combining observations from different satellite instruments is not straightforward. Each platform differs in spatial resolution, retrieval algorithm, detection threshold, and sensitivity to atmospheric conditions. As demonstrated in recent studies, emission estimates derived from different sensors for the same source can vary substantially, often by tens of percent and in some cases by an order of magnitude. These discrepancies arise from differences in plume resolution, wind field assumptions, and retrieval uncertainties.
In this context, the manuscript assumes, without sufficient justification, that observations from all seven platforms can be treated as consistent and combined into a single dataset. The validation presented (based on a controlled release experiment) is not representative of real-world conditions and cannot be used to support such a strong assumption. This is particularly problematic in the case of Shanxi, where terrain complexity, variable meteorological conditions, and optical challenges make retrieval significantly more uncertain than in controlled experiments.
In its current form, the manuscript does not provide a quantitative assessment of inter-satellite variability, nor does it test the sensitivity of results to the inclusion or exclusion of specific datasets. Without such analysis, the combination of multiple platforms may introduce bias rather than reduce uncertainty.
Summary of required improvements:
- Quantify differences between individual satellite products for the same sources.
- Perform sensitivity analysis (e.g. removing individual datasets).
- Clearly justify the assumption of independence and consistency across platforms.
- Avoid overinterpretation of aggregated datasets without validation.
The Bayesian inversion is presented as a key methodological component of the study, yet its implementation is not sufficiently described to allow for proper evaluation.
The manuscript does not provide a clear mathematical formulation of the inversion. The definitions of prior distributions, likelihood functions, and posterior estimates are either missing or insufficiently explained. Furthermore, there is no quantitative evaluation of inversion performance, such as uncertainty reduction or goodness-of-fit metrics. The reader is left with a qualitative assessment based primarily on graphical outputs, which is not sufficient for a method of this complexity.
In addition, key variables used in the equations are not properly defined. For example, Equation 7 introduces a variable that is not described in the text, and it is unclear whether the estimator corresponds to a mean, median, or another statistical quantity. The relationship between prior and posterior uncertainties is also not explained.
As a result, the inversion behaves as a “black box”, preventing reproducibility and limiting the scientific value of the work.
Summary of required improvements:
- Provide a full mathematical description of the inversion.
- Clearly define all variables and assumptions.
- Present prior and posterior values for each mine.
- Quantify uncertainty reduction and inversion performance.
A fundamental weakness of the manuscript is the absence of a physical interpretation of the results. While the study identifies spatial and temporal variability in methane emissions, it treats this variability largely in statistical terms, without connecting it to the underlying processes governing emissions.
Methane emissions from coal mines are strongly controlled by geological and operational factors, including:
- coal rank and gas content,
- depth and structural conditions,
- hydrogeology,
- mining methods and sequencing,
- ventilation and methane drainage practices.
These factors are well documented in the literature and are essential for interpreting observed emission patterns. The omission of this context significantly limits the credibility of the analysis.
In particular, temporal variability should not be treated as purely stochastic. In reality, emission variability is largely driven by operational schedules, such as extraction from different seams with varying methane content. High-emission episodes are often associated with known “sweet spots” in coal seams, characterized by elevated methane content.
Without incorporating this knowledge, the study cannot adequately explain why the reported emission factors are among the highest found in the literature. This lack of interpretation is a major deficiency.
Summary of required improvements:
- Integrate geological and mining-related literature.
- Relate emissions to coal seam properties and structure.
- Interpret variability in terms of mining operations rather than pure statistics.
Closely related to the previous point, the manuscript lacks critical information about the mining systems themselves. In practice, methane emissions originate from specific infrastructure components, particularly ventilation shafts and drainage systems, rather than from the mine as a single entity.
It is common for a single mine to have multiple ventilation shafts, which may be spatially separated from the central facilities. Additionally, methane drainage systems can significantly alter the spatial distribution of emissions by extracting gas before it reaches ventilation systems, or by introducing new leakage points.
The manuscript does not consider these aspects, nor does it provide sufficient justification for the selection of analyzed mines. Without information on the location and number of emission points, the identification of methane plumes remains uncertain.
Summary of required improvements:
- Justify the selection of study sites.
- Include information on ventilation shafts and emission points.
- Discuss the role of methane drainage systems.
- Evaluate how operational practices affect emission estimates.
The treatment of uncertainty is insufficient and in several cases methodologically incorrect. The manuscript does not apply standard uncertainty propagation principles. In particular, independent uncertainties appear to be combined linearly instead of using a root-sum-of-squares approach, which is the correct method under the assumption of independence.
Although this approach may produce conservative estimates, it is not formally valid and should be corrected. In addition, the separation between different types of uncertainty (e.g. measurement error, model uncertainty, inter-satellite variability) is not clearly defined.
The bootstrap analysis is also problematic. The methodology is not clearly explained, and the selection of subsets appears arbitrary. The interpretation of results is not statistically justified, particularly in the context of seasonal variability, where the dataset is insufficient to draw robust conclusions.
Summary of required improvements:
- Apply proper uncertainty propagation methods (RSS).
- Clearly define uncertainty components.
- Reconsider or remove the bootstrap analysis.
- Avoid unsupported conclusions regarding seasonality.
The manuscript contains a number of technical inconsistencies that should be addressed:
- Temporal coverage is uneven and biased toward a single year.
- The criteria for selecting mines are not explained.
- Methane drainage systems are completely omitted.
- Equation 1 does not correctly represent a weighted emission factor.
- Equations 2–6 do not follow correct uncertainty propagation principles.
- Equation 7 introduces undefined variables.
- Units used for emission factors are unconventional and confusing.
- Several figures are unclear or misleading in their current form.
The manuscript addresses an important and relevant problem and includes potentially valuable methodological elements. However, in its current form, it does not meet the standards required for publication.
The main deficiencies relate to:
- insufficient validation of the multi-satellite approach,
- lack of transparency in the Bayesian inversion,
- absence of physical and operational interpretation,
- incorrect or unclear treatment of uncertainty.
These issues are fundamental and require substantial revision.
Recommendation: Major revision required.
Citation: https://doi.org/10.5194/egusphere-2025-4610-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 853 | 350 | 62 | 1,265 | 90 | 118 |
- HTML: 853
- PDF: 350
- XML: 62
- Total: 1,265
- BibTeX: 90
- EndNote: 118
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
This manuscript presents an approach to estimate coal mine methane emissions in Shanxi by combining multi-satellite hyperspectral observations with a hierarchical Bayesian inversion framework. The topic is timely and relevant, particularly in the context of increasing interest in facility-scale methane monitoring using emerging satellite observations. The use of multiple hyperspectral datasets and the attempt to address sparse temporal sampling through a hierarchical Bayesian framework are both valuable contributions. The integration of satellite observations with inventory data is also well motivated, and the results provide useful insights into emission factors as well as spatial and temporal patterns of methane emissions.
Overall, the manuscript is clearly written and addresses an important problem. The methodology is promising, and the results are of potential interest to both the remote sensing and atmospheric science communities. While a few aspects could benefit from further clarification and discussion, I believe these can be addressed with relatively minor revisions.
Major comment
One point that may deserve a bit more careful discussion concerns the use of satellite observations to characterize the distribution of methane emissions. While the general approach is well motivated and has been explored in previous studies, it is important to consider the detection limits of hyperspectral instruments. For example, the detection threshold for instruments such as EnMAP is on the order of ~500 kg/h under favorable conditions, and this threshold may be higher in regions with complex or heterogeneous surface conditions, such as Shanxi.
In this context, it would be helpful for the authors to elaborate on how such detection limits might influence the inferred emission distribution. In particular, there may be a tendency for satellite observations to preferentially capture stronger emission events, while smaller but more frequent sources remain undetected. This could potentially affect the characterization of the emission distribution and the relative importance of extreme events. In related studies that focus specifically on emission distributions, airborne campaigns are often used, partly because they can achieve substantially lower detection limits and therefore provide a more complete sampling across the full range of emission strengths. A brief comparison or discussion along these lines would help place the current results in context.
Related to this, the choice of Shanxi as the primary study region is understandable given its importance in coal production, but its heterogeneous surface conditions may pose additional challenges for plume detection. A brief discussion on how representative the observed emission distribution is, given these observational constraints, would help strengthen the interpretation of the results.
Minor comments
1. The description of the hierarchical Bayesian framework could be slightly clarified to improve readability, particularly for readers less familiar with this type of method.
2. It would be helpful to briefly clarify the extent to which plume identification relies on manual interpretation. In addition, in Figure 2 (lower right panel, EMIT), part of the detected plume appears visually correlated with underlying terrain features, which may suggest a potential false positive. A brief clarification on how such cases are handled would improve confidence in the plume identification procedure.
3. The discussion on detection limits could be expanded slightly to note potential implications for smaller emission sources.