Changes in air pollutant emissions in China during two clean-air action periods derived from the newly developed Inversed Emission Inventory for Chinese Air Quality (CAQIEI)

Kong, Lei; Tang, Xiao; Wang, Zifa; Zhu, Jiang; Li, Jianjun; Wu, Huangjian; Wu, Qizhong; Chen, Huansheng; Zhu, Lili; Wang, Wei; Liu, Bing; Wang, Qian; Chen, Duohong; Pan, Yuepeng; Li, Jie; Wu, Lin; Carmichael, Gregory R.

doi:https://doi.org/10.5194/essd-16-4351-2024

Articles | Volume 16, issue 9

https://doi.org/10.5194/essd-16-4351-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/essd-16-4351-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 9

Data description paper

|

26 Sep 2024

Data description paper |

| 26 Sep 2024

Changes in air pollutant emissions in China during two clean-air action periods derived from the newly developed Inversed Emission Inventory for Chinese Air Quality (CAQIEI)

Lei Kong, Xiao Tang, Zifa Wang, Jiang Zhu, Jianjun Li, Huangjian Wu, Qizhong Wu, Huansheng Chen, Lili Zhu, Wei Wang, Bing Liu, Qian Wang, Duohong Chen, Yuepeng Pan, Jie Li, Lin Wu, and Gregory R. Carmichael

Download

Final revised paper (published on 26 Sep 2024)
Supplement to the final revised paper
Preprint (discussion started on 21 Dec 2023)
Supplement to the preprint

Interactive discussion

Status: closed

RC1:
'Comment on essd-2023-477', Anonymous Referee #1, 01 Feb 2024
This is a valuable high-resolution emission dataset for China, and the paper is well-structured.
However, I have some questions and suggestions:

In Figure 4, compared to other pollutants, NO_x and VOC emissions still show distributions in the Tibet region of China . Could you provide more information about the sources of NOx and VOC in this region?

Could you display the temporal trends in emissions from 2013 to 2020 for China and its sub-regions, comparing different inventories (multiple lines) to better illustrate the changes in emission patterns over time?

As a high-resolution grid product, is it possible to include a comparison at the grid scale with other inventories?

In Figure 12, the profiles presented in this study differ from previous research, especially for SO₂ and PM₁₀. Are the monthly simulation results for SO₂ and PM₁₀ superior to the simulation results using other inventories?

Additionally, it would be interesting to see further validation results from more models (e.g., WRF-CMAQ) using this dataset as input in the future.
Citation: https://doi.org/10.5194/essd-2023-477-RC1
- AC1: 'Reply on RC1', Lei Kong, 10 Mar 2024
  
  We appreciate the reviewer for his/her constructive and up-to-point comments. We have carefully considered the comments and revised the manuscript accordingly. Please refer to the attached file for more details.
  
  Citation: https://doi.org/10.5194/essd-2023-477-AC1
RC2:
'Comment on essd-2023-477', Anonymous Referee #2, 02 Feb 2024

The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2023-477/essd-2023-477-RC2-supplement.pdf

Citation: https://doi.org/10.5194/essd-2023-477-RC2
- AC2: 'Reply on RC2', Lei Kong, 10 Mar 2024
  
  We appreciate the reviewer for his/her constructive and insightful comments. We have carefully considered the comments and revised the manuscript accordingly. Please refer to the attached file for more details.
  
  Citation: https://doi.org/10.5194/essd-2023-477-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Lei Kong on behalf of the Authors (10 Mar 2024) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (18 Mar 2024) by Kuishuang Feng

RR by Anonymous Referee #3 (18 Apr 2024)

Suggestions for revision or reasons for rejection

General comments:
The authors effectively answered the reviewer's questions and improved the quality of data visualization and the coherence of the writing. More specifically, the revised manuscript presents more evidence supporting the use of O3 on constraining VOC emissions. The discussion of the uncertainty and potential limitations of top-down emissions derived from this study further enables readers to fully comprehend the data's strengths and weaknesses. There are only a few minor and specific comments that need to be addressed before publication in ESSD.

Minor comments:
Comment 1: It is encouraging to see that the trend and the amount for most top-down emissions are consistent with previous studies or inventories. The primary concern is the posterior CO emissions, which are substantially higher than values reported in other research and increase by 2-3 times compared to the prior emissions. It is recommended to obtain independent measurements (e.g., data from field campaign) to further evaluate the reliability of the posterior CO emissions. Alternatively, the TROPOMI CO data may be useful for evaluating top-down CO emissions between 2018 and 2020 because the instrument is sensitive to the integrated amount of CO (Landgraf et al., 2016), including the contribution of the planetary boundary layer, making it particularly suitable for detecting surface sources of CO. The authors are not required to do this at this stage, but please consider conducting independent cross-validation for the top-down emissions that shown larger discrepancies to other studies in the future.

Comment 2: While the authors use the NOx-O3-VOC chemistry to support the use of O3 to constrain VOC emissions, the role of CO in O3 formation should be first investigated. I am concerned that the substantially greater posterior CO emissions may bring errors into the O3 simulation and impair the quality of VOC emissions inversion. It may be better to first constrain NOx and CO emissions and ensure that O3 can be properly simulated before optimizing VOC emissions with O3 data.

Comment 3: If both biogenic and anthropogenic VOC emissions are perturbed independently, they can be optimized separately. I'm curious why the authors did not explore doing this, which could aid in the intercomparison of top-down VOC emissions with results from other studies.

Specific comments:
Line 243-245: Underestimating the background error covariance results in an underestimating of the emissions adjustment rather than an overcorrection.

Figure 3: Since bias is typically defined as simulation minus observations and this figure displays prior and posterior OmF, please consider replacing "bias" with "OmF" in the figure legend and caption to prevent confusion and consist with the description in the manuscript.

Line 381-382: Shouldn't the NO2 OmF over the NCP and SE show negative values during the winter and positive values during the summer?

Line 883-884: As the posterior CO emissions derived in this work are substantially greater than those derived in other studies and inventories, it may be an exaggeration to state that the assimilation of CO surface observations is superior to the assimilation of satellite CO measurements. This statement requires further verification (e.g., comparison to independent CO measurement), and I recommend using a more neutral description here.

Figure S17: Since each plot displays different emissions species, please remove the regions (e.g., NW, SE) listed in the figures.

Reference:
Landgraf, J., aan de Brugh, J., Scheepmaker, R., Borsdorff, T., Hu, H., Houweling, S., et al. (2016). Carbon monoxide total column retrievals from TROPOMI shortwaveinfrared measurements. Atmospheric Measurement Techniques, 9(10), 4955–4975. https://doi.org/10.5194/amt-9-4955-2016

Hide

RR by Anonymous Referee #4 (17 May 2024)

Suggestions for revision or reasons for rejection

The authors present a high-resolution inversed air pollutant emission dataset over China for 2013–2020 by assimilating surface observations using EnKF method. Changes of air pollutant emissions in China were estimated based on this dataset with detailed explanations provided. They also compared the dataset with previous emission inventories and proposed some interesting points. This is a valuable dataset that can be useful for many applications. The revised paper is well-structed and has been improved from the original version with the comments raised during the first round of peer review well addressed. Therefore, I recommend publication of this manuscript after minor revisions in this stage:

1. Line 47-48. There is no definition of NCP and SE region.
2. Line 87. Change “chemical species” to “air pollutants”.
3. Line 189 – 191. This sentence is convoluted and difficult to understand, please rewrite this sentence.
4. Figure 14. Please clarify whether the natural sources are included in the calculation of the emission changes from different emission inventories.
5. Line 988. I agree that the limited observations can not fully constrain the emissions in China. Simultaneous assimilation of surface and satellite observation may help alleviate this problem. The authors could add some discussions on this point.
6. Section 6. The authors have proposed some interesting points regarding the air pollutant emissions in China during 2013 – 2020. However, as the authors stated, there are uncertainties in the estimated emission inventory due to the errors in meteorological simulation, CTM and settings of data assimilation, therefore I suggest that the authors need to clarify in the manuscript that further validation is required for the inversion inventory and its associated conclusions.

Hide

RR by Anonymous Referee #5 (22 May 2024)

Suggestions for revision or reasons for rejection

This study employs data assimilation techniques, national air quality monitoring networks, and an air quality model to provide inverse emission inventories for China during 2013-2020. The authors conduct a comprehensive and robust analysis, effectively addressing most concerns from previous reviewers. However, further clarification is required before this manuscript can be accepted for publication.

General comments:

1. Could the authors clarify the rationale behind selecting HTAP v2.2 2010 as the a priori emissions? Given the dataset is specific to China and ranges from 2013 to 2020, why not use more recent inventories (such as CEDS, updated until 2019) and/or China-specific inventories (such as MEIC)?
2. In Table 2, the results show much better performance for the a posteriori model simulation. However, the comparison might be unfair since the authors are comparing their inverse inventory with an outdated global inventory. I am curious about how simulations driven by the inverse inventories compare with simulations driven by more recent bottom-up inventories.
3. The authors mention the exclusion of meteorological and model errors in the ensemble simulation (lines 244 and 367) and briefly discuss future work on model biases (lines 1012-1018). More discussion is necessary. Could the authors provide results on the meteorological model performance against observations and discuss the potential impact of meteorological errors on the inverse emission inventory? For example:
a) If the WRF model systematically overestimates near-surface wind speed, what is the impact on the inventory?
b) If the WRF model systematically underestimates nighttime boundary layer height or mixing (e.g., Du et al., 2020, https://acp.copernicus.org/articles/20/2839/2020/), what is the impact on the inverse inventory?
c) Discussion of meteorological conditions should not be limited to these examples.
4. Why are the natural emission inventories used in the a priori emissions different from those used when comparing CAQIEI with other inventories? If the authors believe the CAMS and GFAS emission inventories are state-of-the-art, why not use them in the a priori emissions? Please clarify.
5. The authors admit in lines 364-366 that they only provide emissions of PM2.5 (PMF+BC+OC) due to a lack of speciated PM2.5 observations. In light of this, what is the rationale behind using PM2.5 observations to simultaneously constrain BC, OC, and primary unspeciated PM2.5 emissions, instead of using PM2.5 observations to constrain primary PM2.5 emissions? It seems that the authors try to provide emission inventories for BC, OC, and primary unspeciated PM2.5, while only presenting total primary PM2.5 emissions due to large uncertainties in its components.

Minor comments:

1. Lines 248-251 are confusing. If the authors do not consider coarse dust emission in the inversion, do they use PM10 concentrations driven by other sources to constrain PM10 emissions? Why assume large errors in simulated coarse dust concentration will impact the inversion of PM10 emissions? Please clarify.
2. Can the authors clarify whether all emissions are updated simultaneously at each iteration?
3. Do the figures in Figure 5 show the national total for China? Please specify in the caption.
4. In line 364, should "PM10 (PM2.5+PMF)" be written as "PM10 (PM2.5+PMC)"?
5. In Table 2, how is R estimated? Do the authors average all stations across China and calculate R for the averaged observation versus averaged simulation, or do they calculate R for each station and average the R values across China?
6. Please specify in line 318 the potential impact of nighttime O3 chemistry on the inversion to better illustrate the rationale for using MDA8 O3.
7. For Figure S16, consider adjusting the min and max values of the Y-axis to better illustrate annual variations. For example, consider not starting the Y-axis from 0.

Hide

ED: Publish subject to minor revisions (review by editor) (14 Jun 2024) by Kuishuang Feng

AR by Lei Kong on behalf of the Authors (25 Jun 2024) Author's response Author's tracked changes Manuscript

ED: Publish as is (12 Jul 2024) by Kuishuang Feng

AR by Lei Kong on behalf of the Authors (18 Jul 2024)

Download

Article (7759 KB)
Full-text XML

Short summary

A new long-term inversed emission inventory for Chinese air quality (CAQIEI) is developed in this study, which contains constrained monthly emissions of NO_x, SO₂, CO, PM_2.5, PM₁₀, and NMVOCs in China from 2013 to 2020 with a horizontal resolution of 15 km. Emissions of different air pollutants and their changes during 2013–2020 were investigated and compared with previous emission inventories, which sheds new light on the complex variations of air pollutant emissions in China.