Preprints
https://doi.org/10.5194/egusphere-2025-1407
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/egusphere-2025-1407
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
15 May 2025
| 15 May 2025
Optimizing Ammonia Emissions for PM2.5 Mitigation: Environmental and Health Co-Benefits in Eastern China
Abstract. Ammonia (NH3) is a key precursor of PM2.5, contributing to the formation of secondary inorganic aerosols and playing a crucial role in haze events. However, current bottom-up emission inventories in China often underestimate NH3 emissions, particularly with significant uncertainties in urban areas. This study developed a "top-down" iterative algorithm that integrates the IASI satellite observations with the WRF-Chem model to optimize bottom-up NH3 emissions, and further quantified the impacts of source-specific emission reductions on PM2.5 pollution. The result reveals that the updated NH3 emissions in Eastern China for 2016 amounted to 4.2 Tg·yr-1, 27.3 % higher than prior estimations. The optimized NH3 emissions peak in summer at 463.1 Gg·mon-1, with agricultural sources accounting for 85 %, while winter emissions drop to 217 Gg·mon-1 when the contribution from non-agricultural sources (e.g., industry, vehicle) significantly increases. The optimized NH3 emission significantly improved the simulated results of NH3 concentration, both in terms of magnitude (31 %~42 %) and variations (17 %~55 %). Sensitivity simulations show that a 30 %~60 % reduction in NH3 emission led to decreases of 1.5~8.8 μg·m-3 in city-level PM2.5 concentrations and the potential effect of reducing non-agricultural emissions is comparable with that from agricultural sources. Furthermore, the NH3 reduction positively impacts public health, resulting in a 6.5 %~10.3 % decrease in premature deaths attributed to PM2.5 exposure. These findings provide strong data support for air quality research and offer valuable insights into the potential air quality and public health benefits of NH3 emission reduction.
How to cite. Tang, K., Zhang, H., Xu, G., Chang, F., Xu, Y., Miao, J., Cui, X., Jin, J., Li, B., Li, K., Liao, H., and Li, N.: Optimizing Ammonia Emissions for PM2.5 Mitigation: Environmental and Health Co-Benefits in Eastern China, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2025-1407, 2025.
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.
Download & links
- Preprint
(1691 KB) - Metadata XML
-
Supplement
(1441 KB) - BibTeX
- EndNote
Status: final response (author comments only)
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
-
RC1: 'Comment on egusphere-2025-1407', Anonymous Referee #1, 20 May 2025
The authors use IASI NH3 observations as well as forward WRF-Chem simulations to improve estimates of NH3 emissions in China, resolving both sectors and seasonality. Overall, the manuscript is well-argued, and the findings are of broad interest to the community. My main concern is the source attribution: I am skeptical of such a large increment to non-agricultural ammonia emissions for reasons outlined below. Remaining comments are mostly to clarify the methodology and its potential shortcomings. I support publication after these comments are addressed.
Major comments:
- The authors attribute the ammonia emissions underestimate in the model almost entirely to non-agricultural emissions (Figure 2). However, temporally, the authors put the largest posterior increase in ammonia emissions in spring and summer (Figure 5) when I would expect agricultural emissions to be most important (fertilization time + favorable meteorology). At the same time, your sources are close or overlap in space (Figure 2), especially considering the smoothness of the modeled and observational total column NH3 (Figure 6). Moreover, because of co-linearity, I am not sure how well the MLR (Eqn 4) can separately fit the alpha and beta parameters and thus separate source sectors. For these reasons, I am skeptical of the source attribution in this study. I am more confident in total ammonia emissions magnitudes.
- The MLR does not account for chemistry occurring between emission and observation. Could you comment on how this affects the results?
Minor comments:
- Line 75: It is not clear to me what the 1%–50% figure represents here. Is this reduction in PM2.5 per unit NH3 emissions reduced?
- Which version of the IASI NH3 data do you use?
- Lines 165–166: What do the index of agreement and mean fractional bias mean, intuitively?
- Line 167: What does the C mean in these equations? I presume ammonia column concentrations?
- How do you convert simulated NH3 to total column densities comparable to IASI (e.g. the SA in line 205)?
- Figure 3: What does the red box represent?
- Line 214: What is A_blank? Also, what are these A variables more generally? They are not defined.
- Equation 4: The SA variables are referring to simulated values, right? This is not clear.
- Line 219: How is D_i^j,k calculated, in terms of the variables already given?
- Line 221: What do you mean by “excessive residuals”? What is the judgement coefficient and how is it calculated?
- Line 222: What goodness of fit test/metric do you use? How did you pick this 0.3–1 acceptability range?
- How do you perform the iterations (lines 226-229)? I presume you increment agricultural and non-agricultural emissions for each grid cell by following the fitted alpha and beta parameters, but how exactly and by what magnitude? What do you do for the next increment in the case where you reject the MLR results?
- What does Figure 5 look like if you split the prior and posterior bar plots up into AGR and nonAGR emissions? I am curious about how the source attribution varies with season.
- Line 291: What are the units of the RMSE? Is this referring to surface observations or satellite columns?
Typographical comments:
- Line 52: Figure should read “55.0%”
- Throughout: the en dash (–) should be used for numerical ranges instead of the tilde (~).
- Lines 179–180: this sentence ends abruptly (what is the RMSE of 10 referring to; what are its units).
- Line 226: I presume you mean alpha and beta as in Equation 4, not a and b here?
Citation: https://doi.org/10.5194/egusphere-2025-1407-RC1 -
AC1: 'Reply on RC1', Nan Li, 20 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1407/egusphere-2025-1407-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2025-1407', Anonymous Referee #2, 06 Jun 2025
This manuscript presents a top-down estimate of ammonia emissions in China using IASI satellite observations and WRF-Chem simulations. The study finds a notable contribution from non-agricultural sources and highlights their implications for PM₂.₅-related health burdens. The topic is of broad interest to the community. However, the manuscript requires major revision before it can be considered for publication. I have several concerns about the MLR-based top-down estimation approach, and the associated uncertainties need to be more clearly explained and discussed to evaluate the robustness of the results. Additionally, the interpretation of the top-down emission estimates should be strengthened by providing deeper insights and more comprehensive comparisons with existing literature.
Major comments:
- Although the authors do not mention, it seems to me that the MLR-based approach assumes a linear relationship between NH₃ emissions and column concentrations and attributes all discrepancies between satellite observed and simulated NH₃ columns to local NH₃ emissions. It is important for the authors to clarify these assumptions and discuss the associated uncertainties. Specifically, (1) How might the nonlinear response of NH3 concentrations to emissions affect the results? For example, do the summed contributions from individual sources (SA_agriculture + SA_non-agriculture + SA_transport) approximate the simulated total NH₃ column from the prior? (2) To what extent could transport of emissions from nearby grid cells influence the posterior results and cause spatial misattribution of emissions? (3) How might uncertainties in emissions of other pollutants (e.g., SO₂, NOₓ) or in the model’s representation of inorganic aerosol formation, impact the posterior estimates of NH₃ emissions? Could the large discrepancies between prior simulated and observed NH3 column concentrations due to these factors? Some quantitative discussion or sensitivity analysis on these points would help strengthen the credibility of the posterior estimates.
- It is not very clear to me how the current MLR framework can separate AGR and non-AGR NH₃ emissions. Some clarification would be helpful, as noted in the minor comments below. Generally, in each grid cell j, I would expect the temporal variations in SA_agriculture and SA_non-agriculture to be perfectly correlated and differ only in magnitude, thus they could not be separated in the regression. While WRF-Chem may simulate different day-to-day variations for SA_agriculture and SA_non-agriculture, that reflects the effects of transport from surrounding grid cells, which contradicts the assumption that transport effects are negligible. Also, it would be useful to explain why the emission corrections primarily affect the non-AGR sector. Given that non-AGR emissions are relatively small in the prior, one would expect SA_non-agriculture to be much smaller than SA_agriculture in Eqn (4). Is the regression coefficient b significantly larger than a, and if so, what is the reason for that? Specifically, in northern Henan, the posterior results show decreased AGR but increased non-AGR emissions compared to the prior, which seems hard to understand and needs further explanation.
- The finding of substantially higher non-agricultural (non-AGR) NH₃ emissions compared to prior estimates is certainly interesting and important. However, the discussion of the posterior results in Section 4 currently focuses mainly on reporting emission magnitudes and sectoral contributions, with limited interpretation of the underlying causes or contextualization within existing literature. I would encourage the authors to expand this discussion by addressing the following points: (1) What are the potential reasons for the apparent underestimation of non-AGR NH₃ emissions in current bottom-up inventories? (2) What types of non-agricultural sources (e.g., industrial processes, transportation) are most likely responsible, based on current understanding? (3) How do your findings about non-AGR emissions compare with previous top-down estimates? The discussion in Lines 259–281 is helpful, but it could be further strengthened by emphasizing on observation-based or model-based studies that have investigated non-AGR NH₃ sources. It would also be valuable to highlight how your results build upon or differ from those studies, and what novel insights your analysis contributes to this topic.
Minor comments:
- Line 26/28 and elsewhere: please remove the “·” between Tg and yr-1. Also, replace Gg mon-1 with either Gg month-1 or Gg mo-1 to follow standard unit concentions.
- Line 37-40. The summary statement is too general. It would be more informative to highlight the insights into non-agricultural ammonia emissions and their implications.
- Line 127: Is biomass burning emission also treated online? Just checking, as this is not commonly the case.
- Line 134: The “last accessed” date should reflect the actual date when the data were downloaded.
- In section 2.2, what’s the overpass time of IASI data? Do you use level-2 satellite data?
- Line 140-141: The description is unclear. How is the neural network applied to improve the data quality, was it developed by the authors or sourced elsewhere?
- Line 125/156: Which version of MEIC is used?
- Line 162-163 and Section 4.1: Please clarify how the model and observations are sampled for comparison.
- Line 179: The term “errors” is vague. Consider using clearer language such as “underestimated by 30%” or “biased low by 30%.”
- Line 181-189: Figure 6 can be described with the text here.
- Section 3: Please be consistent in the use of statistical metrics. RMSE is used for IASI comparisons, while IOA and MFB are used for surface observations. A brief explanation of why different metrics are applied, and what each evaluates, would be helpful.
- Line 195: You mention deriving posterior emissions for four months—how are prior/posterior simulations compared with observations across the seasons? Are the same scale factors applied to all three months in each season? Please clarify. Given that WRF-Chem simulations are available for the full year, it would be more consistent to derive monthly emissions for all 12 months, which should follow the same procedure and would not require much additional effort.
- Line 200: In Line 138, you mentioned that IASI data were regridded to the model resolution, but here you refer to single-pixel comparisons, which is somewhat confusing. Please clarify how the satellite data were matched to model outputs.
- Line 202: What does “regression factor” refer to? Is it the same as the emission scale factor?
- Line 206: Is TA_satellite the monthly average or the daily average of NH3 concentrations?
- Line 208: Should be “outside transportation, AGR emissions, non-AGR emissions, respectively’
- Line 211: The term “control emissions” is unclear. Do you mean emissions were zeroed out? Also, please replace “cycle” with “experiment.”
- Line 214: What is A_blank used for?
- Line 203/216: Earlier you use k for month and j for region, but later this is reversed. Please ensure consistency throughout. Also, using “grid cell j” is clearer than “region j” or “area j.”
- Line 217: It is unclear why does the regression is derived mathematically imply it needs to be corrected? Please clarify the motivation for adjusting the regression coefficients.
- Line 218-231: The description of the correction process is not very clear. It is unclear what is meant by “goodness of fit,” how the “invalid” regression coefficients are defined, and what fraction of them are removed. The phrase “make a trade-off” in Line 225 is vague and would benefit from clarification. Additionally, it is not explained how the adjustment factors aₙ and bₙ are derived or what their physical meaning is. The choice of a 30% threshold in Line 229 also seems arbitrary—particularly in high-NH₃ regions, where it could allow larger discrepancies between observations and simulations, but the physical basis for this threshold is not clearly explained.
-
AC2: 'Reply on RC2', Nan Li, 20 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1407/egusphere-2025-1407-AC2-supplement.pdf
-
RC3: 'Comment on egusphere-2025-1407', Anonymous Referee #3, 22 Jun 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1407/egusphere-2025-1407-RC3-supplement.pdf
-
AC3: 'Reply on RC3', Nan Li, 20 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1407/egusphere-2025-1407-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Nan Li, 20 Aug 2025
Viewed
Total article views: 800 (including HTML, PDF, and XML)
Cumulative views and downloads
(calculated since 15 May 2025)
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 716 | 64 | 20 | 800 | 28 | 10 | 19 |
- HTML: 716
- PDF: 64
- XML: 20
- Total: 800
- Supplement: 28
- BibTeX: 10
- EndNote: 19
Viewed (geographical distribution)
Total article views: 800 (including HTML, PDF, and XML)
Thereof 800 with geography defined
and 0 with unknown origin.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1
Latest update: 14 Sep 2025
Keqin Tang
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Haoran Zhang
School of Atmospheric Sciences, Nanjing University, Nanjing, China
Ge Xu
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Fengyi Chang
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Yang Xu
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Ji Miao
Department of Colorectal Surgery, Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University, Nanjing, China
Xian Cui
Rugao Hospital of Chinese Medicine, Nantong, China
Jianbin Jin
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Baojie Li
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Hong Liao
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Nan Li
CORRESPONDING AUTHOR
Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing, 210044, China
Short summary
Our research combined satellite observations with air quality modeling to establish a high-resolution ammonia emission inventory for Eastern China. Reducing ammonia emissions could lower particulate pollution levels by 1.5~8.8 micrograms per cubic meter and reduce related health risks. Meanwhile, sensitivity simulations highlight the critical need to non-agricultural emission controls for effective particulate mitigation.
Our research combined satellite observations with air quality modeling to establish a...