the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Jul 2025
High-resolution regional inversion reveals overestimation of anthropogenic methane emissions in China
Abstract. Methane (CH4), the second most important anthropogenic greenhouse gas, significantly impacts global warming. As the world's largest anthropogenic CH4 emitter, China faces challenges in accurately estimating its emissions. Top-down methods often suffer from coarse resolution, limited data constraints, and result discrepancies. Here, we developed the Regional Methane Assimilation System (RegGCAS-CH4) based on the WRF-CMAQ model and the EnKF algorithm. By assimilating extensive TROPOMI column-averaged dry CH4 mixing ratio (XCH4) retrievals, we conducted high-resolution nested inversions to quantify daily CH4 emissions across China and Shanxi Province in 2022. Nationally, posterior CH4 emissions were 45.1 TgCH4·yr⁻1, 36.5 % lower than the EDGAR estimates, with the largest reductions in the coal and waste sectors. In North China, emissions decreased most significantly, mainly attributed to the coal and enteric fermentation sectors. Posterior emissions in coal-reliant Shanxi Province decreased by 46.3 %. Sporadic emission increases were detected in major coal-producing cities but were missed by the coarse-resolution inversion. Monthly emissions exhibited a winter-low, summer-high pattern, with the rice cultivation and waste sectors showing higher seasonal increases than those in EDGAR. The inversion significantly improved XCH4 and surface CH4 concentration simulations, reducing emission uncertainty. Compared to other bottom-up/top-down estimates, our results were the lowest, primarily because the high-resolution inversion better captured local emission hotspots. Sensitivity tests underscored the importance of nested inversions in reducing the influence of boundary condition uncertainties on emission estimates. This study provides robust CH4 emission estimates for China, crucial for understanding the CH4 budget and informing climate mitigation strategies.
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
(4150 KB) - Metadata XML
-
Supplement
(751 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-2669', Anonymous Referee #1, 24 Jul 2025
Feng et al. performed a high-resolution inversion to quantify methane emissions from China. The authors have employed nested WRF-CMAQ simulations, with particular fine resolution (9 km) over Shanxi, the major coal mining province in China. They found that the prior inventory severely overestimates coal mining emissions in North China. The study is overall well executed and clearly presented. I am happy to recommend publication of the manuscript after the following comments are addressed, mainly regarding clarification of the methodology.
Main comments:
1. The study discussed methane emissions from different sectors, but how the sector partitioning is done is not described. I'd suggest authors to provide further methodological details.
2. The authors have applied TROPOMI XCH4 L2 data. An earlier version of the TROPOMI data have shown substantial regional biases over East China, which may cause errors in the inversion. It would be good if the authors can have some discussion or conduct evaluation on this issue, for instance, using TCCON sites in China.
3. I do appreciate that the authors have performed evaluation for meteorology parameters against independent data, which most of existing studies have not done. This is crucial for characterizing model transport errors and understanding the difference between inversion systems. However, the discussion is overly simple. I'd suggest the authors to expand the results on meteorology evaluation (especially wind). In particular, the evaluation over the D02 domain provides crucial information because of the complex terrain in Shanxi.
4. The authors used the optimized emissions from the D01 inversion as prior emissions for the D02 inversion. This implies that the observations over D02 are used twice in the optimization of emissions. From the Bayesian standpoint, this is problematic as it leads to over-confidence in observations.
5. The paper in general lacks uncertainty characterization for emission flux estimates. For regions with limited observation coverage (e.g., Southern China), it is unclear to what degree the posterior estimates depend on prior estimates.
Minor suggestions:
L80: key source-> "point source scale" or "local scale"?
L95: unclear -> uncertain
L103: To my knowledge, IMI is not an operational inversion system, but more like open-source software. So it may be improper to characterize it as a US system. Similar issues may exist for other listed systems.
L188-189: Any quantitative estimates how much error it will incur for D01 and for D02 respectively, by deactivating the chemical oxidation?
L230: How do you specify the R matrix? Also explain specifically that R is an error covariance matrix for what.
L232: Ep: Power plant sources? Seems something copied from a CO2 study.
L234: No need to capitalize O in oil
L251: Would 1 day be too short for adequate observation constraint, if you assume that prior errors are independent from one day to the next (L272-273)?
Table 1: What do the last two columns (building, mature) stand for?
Table 1 and related discussion (e.g., L360): EDGAR v8.0 is used as prior information. Recent studies have shown that EDGAR has large errors in the spatial and seasonal distribution in rice emissions (Chen et al., 2025; Liang et al., 2024). I'd suggest the authors to briefly discuss the impact on emission quantification and sector attribution in Northeast and East China.Chen et al.: Global Rice Paddy Inventory (GRPI): a high-resolution inventory of methane emissions from rice agriculture based on Landsat satellite inundation data, Earth's Future, 2025.
Liang et al.: Satellite-based Monitoring of Methane Emissions from China's Rice Hub, Environmental Science & Technology, 2024.Table 2: Just a comment: The comparison with local observations, which are sensitive to emission adjustment, is valuable.
Citation: https://doi.org/10.5194/egusphere-2025-2669-RC1 -
RC2: 'Comment on egusphere-2025-2669', Anonymous Referee #2, 19 Aug 2025
This study presents a high-resolution CH4 emission inversion system using the WRF-CMAQ model coupled with the ensemble Kalman filter assimilation method. It assimilated high-resolution XCH4 concentrations from TROPOMI to estimate CH4 emissions over China and Shanxi Province. The research highlights that the prior inventory significantly overestimates CH4 emissions from coal and waste sectors in China, and reveals seasonal characteristics of sectoral CH4 emissions. Additionally, the research innovatively adjusts the model's boundary conditions based on TROPOMI XCH4 retrievals, and discusses the impact of boundary condition uncertainties on regional CH4 emission inversion through sensitivity experiments. The scope and quality of this work are suitable for publication in ACP , and I recommend the manuscript be accepted after minor revisions based on the following comments.
1. Lines 33–35 and 548–550: Regarding the conclusion that this study's top-down emission inversion is lower compared to others, the authors mainly attribute this to the higher resolution used here, which captures finer emission details—a point supported by comparing results from the inner and outer nested domains. While this explanation is reasonable, it may not be entirely sufficient. Other factors like differences in methods, models, and observations—including how chemical processes, soil sinks, and boundary conditions are represented—could also contribute significantly. It would be good if the authors could briefly discuss these aspects in the Discussion.
2. Figure 1b: It shows that TROPOMI data has a high missing rate across the country, especially in the south, with some areas having coverage for only 10% of the dates. For regions with long periods of no observation, how are the posterior emissions represented? Also, what is the impact of this representation on daily，monthly and yearly CH4 emission estimates? It would be helpful if the authors could discuss this as well.
3. Figure 3 on page 20 shows the differences in prior and posterior emissions for different sectors. How are the sectors distinguished in the posterior emissions? It is indeed challenging to differentiate sectoral emissions in top-down emission inversion. Typically, sectoral emissions in the posterior are calculated based on the proportional grid emissions from the prior inventory. Does this study use the same method?
4. The title of Figure 4 needs to specify the time period covered by the data. Is it the average for the entire year of 2022?
5. On line 509, P24, 'Except for LF site...', this description is incorrect. Not only is the LF site underestimated, but the TY site is also significantly underestimated, as well as the JC site. It is recommended to correct the description."
Citation: https://doi.org/10.5194/egusphere-2025-2669-RC2
Data sets
Anthropogenic CH4 Emissions over China in 2022 Inverted Using TROPOMI XCH4 Retrievals Shuzhuang Feng https://zenodo.org/records/15602944
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 438 | 77 | 17 | 532 | 25 | 8 | 19 |
- HTML: 438
- PDF: 77
- XML: 17
- Total: 532
- Supplement: 25
- BibTeX: 8
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1