the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Oct 2025
Driving Factors of Oxalic Acid and Enhanced Role of Gas-Phase Oxidation under Cleaner Conditions: Insights from 2007–2018 Field Observations in the Pearl River Delta
Abstract. Secondary organic aerosol (SOA) is a dominant constituent of fine particulate matter, exerting significant impacts on both climate and human health. Oxalic acid (C2), a key end-product formed from the oxidation of volatile organic compounds, can provide insights into the formation mechanism of SOA. Thus, long-term measurements of C2 and related compounds help understand the changes in SOA formation with decreasing pollutant levels. In this study, C2 and its homologs, along with five primary anthropogenic source markers and three SOA markers, were measured in the Pearl River Delta (PRD) during 2007–2018. The concentrations of C2 and its homologs did not exhibit significant downward trends, despite substantial reductions in anthropogenic emissions, for example, biomass burning (−11 % yr−1), vehicle emissions (−17 % yr−1), and cooking emissions (−7 % yr−1). Correlation analysis revealed that aerosol liquid water content (ALWC) and Ox (O3 + NO2) were the main drivers of C2 variation. Moreover, the relative contribution of biogenic SOA increased under cleaner conditions. A machine learning model was applied to quantify the contributions of anthropogenic precursors emission, biogenic precursors emission, aqueous-phase oxidation processes, and gas-phase oxidation processes to C2 variability. As pollution levels declined, the contribution of gas-phase oxidation increased from 24 % to 48 %, whereas that of aqueous-phase oxidation declined from 35 % to 20 %. This shift indicated a transition from aqueous-phase to gas-phase pathways in C2 and SOA formation. Our findings highlight the increasing importance of gas-phase oxidation and underscore the need for effective ozone control strategies to further reduce SOA in the future.
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
(1575 KB) - Metadata XML
-
Supplement
(571 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2025-4624', Anonymous Referee #1, 18 Nov 2025
-
AC1: 'Reply on RC1', Yunfeng He, 16 Dec 2025
We sincerely thank you for your time and valuable comments. We have added more information about machine learning to enhance the robustness of our results, and carefully revised the manuscript to improve its clarity and enhance the readers' understanding. Our point-by-point responses are marked in blue and the corresponding changes to the original text are shown below each response. We have attached the response letter below, please check it. We hope that these revisions adequately address the comments and concerns.
-
AC1: 'Reply on RC1', Yunfeng He, 16 Dec 2025
-
RC2: 'Comment on egusphere-2025-4624', Anonymous Referee #2, 26 Nov 2025
This work analyzed a long-term variation of oxalic acid in atmospheric aerosols in the Pearl River Delta from 2007 to 2018. Aerosol liquid water and Ox are the driving factors of C2 formation, and gas-phase oxidation would play a more important role than aqueous-phase oxidation as air pollution decreases. The long-term data is valuable and informative. However, the data interpretation needs to be more rigorous. I may suggest a major revision before publication.
Major comments:
The ALWC and Ox are identified as the drivers of C2 variation. The authors also highlight the increased contribution of gas-phase oxidation and decreased contribution of aqueous-phase oxidation to C2 formation as pollution levels declined. However, high air pollution level is usually accompanied by high humidity and low Ox.
How did the authors exclude the impacts of different emission levels when addressing the influence of RH and Ox on C2 variation?
In Fig. 5, the correlation between C2 and Ox is always higher than that between C2 and ALWC under IT1, 2, 3, and 4. The correlation between C2 and ALWC is weak under any IT condition.
Section 3.4: For the machine learning analysis, the authors quantify the contributions of different sources using some input parameters. The rationality of this approach needs to be elaborated. For example, air temperature, solar radiation, and relative humidity are used to represent the emission of biogenic precursors (lines 318-319). Do all these meteorological factors promote the emission of biogenic precursors? For biogenic emissions, is there a synergistic or antagonistic mechanism between these factors? Please explain in detail. The reasonability of using the input parameters to represent other sources also needs to be elaborated.
For the contribution of gas-phase oxidation versus aqueous-phase oxidation, is the result here obtained based on machine learning comparable to those reported in published literature?
Figures 2 and 3: I am curious about the high levels since 2013. Please explain the reasons.
Lines 250-252: “Meanwhile, the correlations between C2 and ASOA markers became weaker. These results suggested ….” Please explain this statement. In Table 1, I did not find an obvious decreasing trend for the correlation coefficients between C2 and Phthalic acid (changing from 0.28 under IT1 to 0.31 under IT4) or DHOPA (changing from 0.49 under IT1 to 0.32 under IT4) from IT1 to IT4.
Lines 287-290: I did not see an obvious difference in the correlation efficiency between C2 and Ox or between C2 and ALWC from IT1 to IT4. The change of Pearson r values seems small.
The dataset collected during 2007-2018 is valuable and informative. My concern is the uncertainty caused by long-term storage. How long after sampling were these samples analyzed? How much of the C2 organic acid could change during storage?
In addition, in lines 115-116, I may suggest adding a table in the supplementary to detail the sample information.
Specific comments:
Please specify the data source of solar radiation in the method section.
In Figure 2, 3, or other similar figures, modify the name of the y-axis to the corresponding species. It would be easier for readers.
Figure captions need to be revised. For example, “The concentrations decreased from 864 ± 283 ng m-3 (2007) to 307 ± 122 ng m-3 (2018), ….”, “Pearson’s r values between C2 and ALWC decreased from 0.43 to 0.15, while those between C2 and Ox increased from 0.28 to 0.68.” or similar statements should not be described in the figure caption.
Line 225 and Table 1: Change “IT4 (25 ug/m3 >PM2.5)” to “PM2.5 < 25 ug/m3”.
Lines 244-245: Please show the data or other evidence on the higher temperature, solar radiation, or humidity in PRD.
Lines 330-332: Please show evidence on the statement that lower ALWC favors the C2 compounds from the particle-phase to the gas-phase.
Citation: https://doi.org/10.5194/egusphere-2025-4624-RC2 -
AC2: 'Reply on RC2', Yunfeng He, 16 Dec 2025
We sincerely appreciate your time and constructive comments. In response, we conducted a more detailed data interpretation and incorporated relevant supporting evidence to strengthen the reliability of our conclusions. We have also carefully revised the manuscript to improve clarity and facilitate better understanding for readers. Our point-by-point responses are marked in blue and the corresponding changes to the original text are shown below each response. We hope that these revisions adequately address the comments and concerns. We have attached the response letter below, please check it.
-
AC2: 'Reply on RC2', Yunfeng He, 16 Dec 2025
Status: closed
-
RC1: 'Comment on egusphere-2025-4624', Anonymous Referee #1, 18 Nov 2025
General Comments:
The manuscript reports long-term field observations of di-acids and its related primary and secondary markers from anthropogenic and biogenic sources at a site in the PRD region, China. It also combines these observations with machine-learning methods to investigate and quantify potential contributions of major drivers to the variation of oxalic acid. Their major findings highlight the increasing importance of gas-phase oxidation in forming SOA. Overall, the topic is valuable with good-quality datasets, but the manuscript needs clearer methodological descriptions, stronger validation of the machine-learning attribution, and more mechanistic and systematic support before publication. For the machine learning methodology part, the attribution is potentially interesting, but I am concerned about robustness given the relatively small dataset (~400 observations) and 11 features. With this sample size there is a substantial risk of overfitting and unstable feature attributions, especially if the data are temporally autocorrelated. I would suggest a major revision.
Specific Comments:
- Attribution to gas- vs aqueous-phase pathways is mechanistically simplified; aqueous production depends on pH, transition metals, oxidant availability, and organic composition, the author may consider adding more feature variables in the machine learning model;
- line 84: you should spell out an abbreviation (ALWC) the first time it appears in the main text even if you already defined it in the abstract.
- Figure 1: there’s almost no exact content in the figure. The author may consider adding back-trajectories or removing this figure to the SI.
- line 209: Malic acid is a plausible product of biogenic VOC photooxidation, but it is not a unique tracer. Given the winter, urban-influenced atmosphere, anthropogenic VOCs and combustion sources could contribute substantially.
- line 226: The authors normalize oxalic acid and related species by PM2.5 to reduce dilution effects. I would rather recommend using primary and inertia tracers such as ΔCO as a more appropriate normalizer for removing dilution.
- Figure 3: add oxalic acid data in this figure.
- line 255-258: I do not find enough evidence supporting the two sentences claiming the limited contribution of anthropogenic VOCs and meteorology.
- Table 1: how may data points are in each category?
- Figure 6: The author should consider using the same features to predict other di-acids to see if these features can well capture the variation of other di-acids.
Citation: https://doi.org/10.5194/egusphere-2025-4624-RC1 -
AC1: 'Reply on RC1', Yunfeng He, 16 Dec 2025
We sincerely thank you for your time and valuable comments. We have added more information about machine learning to enhance the robustness of our results, and carefully revised the manuscript to improve its clarity and enhance the readers' understanding. Our point-by-point responses are marked in blue and the corresponding changes to the original text are shown below each response. We have attached the response letter below, please check it. We hope that these revisions adequately address the comments and concerns.
-
RC2: 'Comment on egusphere-2025-4624', Anonymous Referee #2, 26 Nov 2025
This work analyzed a long-term variation of oxalic acid in atmospheric aerosols in the Pearl River Delta from 2007 to 2018. Aerosol liquid water and Ox are the driving factors of C2 formation, and gas-phase oxidation would play a more important role than aqueous-phase oxidation as air pollution decreases. The long-term data is valuable and informative. However, the data interpretation needs to be more rigorous. I may suggest a major revision before publication.
Major comments:
The ALWC and Ox are identified as the drivers of C2 variation. The authors also highlight the increased contribution of gas-phase oxidation and decreased contribution of aqueous-phase oxidation to C2 formation as pollution levels declined. However, high air pollution level is usually accompanied by high humidity and low Ox.
How did the authors exclude the impacts of different emission levels when addressing the influence of RH and Ox on C2 variation?
In Fig. 5, the correlation between C2 and Ox is always higher than that between C2 and ALWC under IT1, 2, 3, and 4. The correlation between C2 and ALWC is weak under any IT condition.
Section 3.4: For the machine learning analysis, the authors quantify the contributions of different sources using some input parameters. The rationality of this approach needs to be elaborated. For example, air temperature, solar radiation, and relative humidity are used to represent the emission of biogenic precursors (lines 318-319). Do all these meteorological factors promote the emission of biogenic precursors? For biogenic emissions, is there a synergistic or antagonistic mechanism between these factors? Please explain in detail. The reasonability of using the input parameters to represent other sources also needs to be elaborated.
For the contribution of gas-phase oxidation versus aqueous-phase oxidation, is the result here obtained based on machine learning comparable to those reported in published literature?
Figures 2 and 3: I am curious about the high levels since 2013. Please explain the reasons.
Lines 250-252: “Meanwhile, the correlations between C2 and ASOA markers became weaker. These results suggested ….” Please explain this statement. In Table 1, I did not find an obvious decreasing trend for the correlation coefficients between C2 and Phthalic acid (changing from 0.28 under IT1 to 0.31 under IT4) or DHOPA (changing from 0.49 under IT1 to 0.32 under IT4) from IT1 to IT4.
Lines 287-290: I did not see an obvious difference in the correlation efficiency between C2 and Ox or between C2 and ALWC from IT1 to IT4. The change of Pearson r values seems small.
The dataset collected during 2007-2018 is valuable and informative. My concern is the uncertainty caused by long-term storage. How long after sampling were these samples analyzed? How much of the C2 organic acid could change during storage?
In addition, in lines 115-116, I may suggest adding a table in the supplementary to detail the sample information.
Specific comments:
Please specify the data source of solar radiation in the method section.
In Figure 2, 3, or other similar figures, modify the name of the y-axis to the corresponding species. It would be easier for readers.
Figure captions need to be revised. For example, “The concentrations decreased from 864 ± 283 ng m-3 (2007) to 307 ± 122 ng m-3 (2018), ….”, “Pearson’s r values between C2 and ALWC decreased from 0.43 to 0.15, while those between C2 and Ox increased from 0.28 to 0.68.” or similar statements should not be described in the figure caption.
Line 225 and Table 1: Change “IT4 (25 ug/m3 >PM2.5)” to “PM2.5 < 25 ug/m3”.
Lines 244-245: Please show the data or other evidence on the higher temperature, solar radiation, or humidity in PRD.
Lines 330-332: Please show evidence on the statement that lower ALWC favors the C2 compounds from the particle-phase to the gas-phase.
Citation: https://doi.org/10.5194/egusphere-2025-4624-RC2 -
AC2: 'Reply on RC2', Yunfeng He, 16 Dec 2025
We sincerely appreciate your time and constructive comments. In response, we conducted a more detailed data interpretation and incorporated relevant supporting evidence to strengthen the reliability of our conclusions. We have also carefully revised the manuscript to improve clarity and facilitate better understanding for readers. Our point-by-point responses are marked in blue and the corresponding changes to the original text are shown below each response. We hope that these revisions adequately address the comments and concerns. We have attached the response letter below, please check it.
-
AC2: 'Reply on RC2', Yunfeng He, 16 Dec 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 275 | 94 | 33 | 402 | 39 | 19 | 24 |
- HTML: 275
- PDF: 94
- XML: 33
- Total: 402
- Supplement: 39
- BibTeX: 19
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General Comments:
The manuscript reports long-term field observations of di-acids and its related primary and secondary markers from anthropogenic and biogenic sources at a site in the PRD region, China. It also combines these observations with machine-learning methods to investigate and quantify potential contributions of major drivers to the variation of oxalic acid. Their major findings highlight the increasing importance of gas-phase oxidation in forming SOA. Overall, the topic is valuable with good-quality datasets, but the manuscript needs clearer methodological descriptions, stronger validation of the machine-learning attribution, and more mechanistic and systematic support before publication. For the machine learning methodology part, the attribution is potentially interesting, but I am concerned about robustness given the relatively small dataset (~400 observations) and 11 features. With this sample size there is a substantial risk of overfitting and unstable feature attributions, especially if the data are temporally autocorrelated. I would suggest a major revision.
Specific Comments: