Mass-constrained source apportionment of nitrate-containing particles in eastern China using a SPAMS-NMF framework

Zhu, Wenfei; Wang, Qinghong; Wang, Qingsong; Shi, Jialin; Sun, Yu; Huo, Juntao; Sun, Yi; Chen, Jia; Zhao, Yue; Li, Mei; Chen, Jun; Lou, Shengrong; Chen, Hui

doi:10.5194/egusphere-2026-2447

Preprints

https://doi.org/10.5194/egusphere-2026-2447

Preprints

03 Jul 2026

| 03 Jul 2026

Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).

Mass-constrained source apportionment of nitrate-containing particles in eastern China using a SPAMS-NMF framework

Wenfei Zhu, Qinghong Wang, Qingsong Wang, Jialin Shi, Yu Sun, Juntao Huo, Yi Sun, Jia Chen, Yue Zhao, Mei Li, Jun Chen, Shengrong Lou, and Hui Chen

Abstract. Winter haze in eastern China is increasingly dominated by particulate nitrate, but its compositions and source contributions remain poorly constrained. We conducted intensive winter observations at a regional site in suburban Shanghai and developed a semi-quantitative single-particle framework combining SPAMS with factor analysis to attribute nitrate to specific particle types and sources in near real time. Nitrate occurred mainly in the accumulation mode (~0.5–0.7 μm), implying high regional persistence. Four nitrate-containing particle classes (NO₃Lv1–Lv4) were identified. A highly aged class (NO₃Lv4) accounted for the largest fraction of both particle number and nitrate mass loadings. This class showed extensive internal mixing with elemental carbon, ammonium, potassium-rich combustion markers, and chloride-depleted sea salt, indicating coupling between secondary inorganic nitrate, combustion emissions, and processed marine aerosol. Episode analysis and factorization further revealed that severe nitrate build-up reflects the co-occurrence of stagnant boundary layers, humid nocturnal heterogeneous formation of both ammonium and non-ammonium nitrate, long-range transport from northern industrial and urban corridors, and marine influence. These results indicate that winter nitrate haze in the Yangtze River Delta is governed by joint NO_x–NH₃ chemistry, primary combustion, and regional transport, requiring coordinated multipollutant control.

Received: 28 Apr 2026 – Discussion started: 03 Jul 2026

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 (PDF, 12830 KB)

Supplement (7707 KB)

Download & links

Preprint (12830 KB)
Metadata XML
Supplement (7707 KB)
BibTeX
EndNote

Wenfei Zhu, Qinghong Wang, Qingsong Wang, Jialin Shi, Yu Sun, Juntao Huo, Yi Sun, Jia Chen, Yue Zhao, Mei Li, Jun Chen, Shengrong Lou, and Hui Chen

Status: open (until 14 Aug 2026)

Post a comment Subscribe to comment alert

CC1:
'Comment on egusphere-2026-2447', Nima Zafarmomen, 05 Jul 2026 reply
The manuscript presents a study on winter nitrate-containing particles in eastern China using a semi-quantitative SPAMS–NMF framework. The authors conducted autumn–winter observations at the Dianshan Lake Atmospheric Superstation near Shanghai and combined single-particle aerosol mass spectrometry with bulk measurements, meteorology, and non-negative matrix factorization to identify nitrate-containing particle classes and source contributions. The study finds that nitrate-containing particles were mainly in the accumulation mode around 0.5–0.7 μm, with the aged NO₃Lv4 class dominating both particle number and nitrate mass contribution. The manuscript is relevant to winter haze formation, nitrate aerosol chemistry, and source apportionment in the Yangtze River Delta.
The manuscript should more clearly explain the novelty of the proposed SPAMS–NMF framework. The authors state that the method is mass-constrained and semi-quantitative, but it would be helpful to clarify how this approach improves upon previous SPAMS-based source apportionment studies and what the main methodological advancement is.

The regression-based conversion from SPAMS particle information to nitrate mass concentration needs further justification. The reported correlation coefficient of the MLR model is moderate, and the authors should discuss the uncertainty this introduces into the subsequent NMF source apportionment and particle-class mass estimates.

The interpretation of the four nitrate-containing particle classes should be made more concise and transparent. The transition from NO₃Lv1 to NO₃Lv4 is described as increasing nitrate enrichment and aging, but the criteria used to distinguish these classes should be summarized more clearly for readers who are less familiar with SPAMS spectral classification.

The discussion of nocturnal heterogeneous nitrate formation should be expressed more cautiously. Since NO₃, N₂O₅, and ClNO₂ were not directly measured, the authors should clearly distinguish between direct observational evidence and inferred mechanisms based on nighttime enhancement, high relative humidity, and stagnant boundary-layer conditions.

Several figures are information-rich but difficult to read, especially Figures 1, 2, 5, 6, and 7. The authors should improve figure readability by increasing font sizes, simplifying legends where possible, and ensuring consistent notation for NO₃ particle classes and source factors.

The authors are strongly recommended to cite recent work on high-spatiotemporal-resolution monitoring of traffic-related particles in complex urban environments. In particular, Yeganeh, B., Shakerdonyavi, A., Zafarmomen, N., and Taheri, A.: Comprehensive spatiotemporal analysis of long-term mobile monitoring for traffic-related particles in a complex urban environment, Atmospheric Pollution Research, 102870, 2025, is highly relevant. That study provides long-term mobile-monitoring evidence for PM2.5 and black carbon variability and can strengthen the discussion of traffic-related particulate pollution, spatial heterogeneity, combustion-related particles, and urban aerosol source characterization.

The source interpretations from NMF would benefit from additional uncertainty analysis. Since NMF solutions are not unique, the authors should report how factor number selection was justified and whether sensitivity tests were performed using different factor numbers or initialization settings.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2447-CC1
- CC3:
  'Reply on CC1', Zhiheng Liao, 06 Jul 2026 reply
  
  Dear Editor and Authors:
  I have serious doubts about the true purpose of Nima Zafarmomen's eponymous community comments on multiple ACPD articles (including this article). Nima Zafarmomen has recently posted numerous comments on papers covering very different topics (see attached comments), and in all of them, he strongly recommends citing his own article, "Comprehensive spatiotemporal analysis of long-term mobile monitoring for traffic-related particles in a complex urban environment," which is scheduled to appear in Atmospheric Pollution Research only in May 2026. In my humble opinion, that paper has no essential relevance to the articles he comments on. Moreover, Nima Zafarmomen's academic background (hydrometeorology) is far removed from the research directions relevant to the articles he comments on. Given these circumstances, I question the genuine intent behind Nima Zafarmomen's submitted eponymous community comments, and I would kindly ask the handling editor and the manuscript authors to give full consideration to the reliability of his comments.
  
  Zhiheng Liao
  Institute of Urban Meteorology, China Meteorological Administration, Beijing, China
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2026-2447-CC3
  - CC4: 'Reply on CC3', Nima Zafarmomen, 06 Jul 2026 reply
    
    Thank you for sharing your opinion. However, your comment appears to be based on assumptions about my background and motivations rather than on the scientific content of my community comment. Like you, I participated in the discussion by providing my scientific opinion. Whether any suggested reference is relevant is entirely for the authors and the handling editor to decide. If you believe there are scientific flaws in my comments, I would be happy to discuss those specific points.
    Otherwise, I believe it is more constructive to focus on the science rather than on personal assumptions about other contributors.
    
    Reply
    
    Citation: https://doi.org/10.5194/egusphere-2026-2447-CC4
    
    CC5: 'Reply on CC4', Zhiheng Liao, 06 Jul 2026 reply
    
    I have no right to decide; I am only offering my opinion to the handling editor and the authors. They will make the final judgment on the scientific soundness and reliability of your comments.
    
    Reply
    
    Citation: https://doi.org/10.5194/egusphere-2026-2447-CC5
    
    CC6: 'Reply on CC5', Nima Zafarmomen, 06 Jul 2026 reply
    
    Exactly. Then I believe the discussion should focus on the scientific merits of my comment, not on assumptions about my background or intentions. The handling editor and the authors will decide its relevance.
    
    Reply
    
    Citation: https://doi.org/10.5194/egusphere-2026-2447-CC6
CC2:
'Comment on egusphere-2026-2447', Nima Zafarmomen, 05 Jul 2026 reply
The manuscript presents a study on winter nitrate-containing particles in eastern China using a semi-quantitative SPAMS–NMF framework. The authors conducted autumn–winter observations at the Dianshan Lake Atmospheric Superstation near Shanghai and combined single-particle aerosol mass spectrometry with bulk measurements, meteorology, and non-negative matrix factorization to identify nitrate-containing particle classes and source contributions. The study finds that nitrate-containing particles were mainly in the accumulation mode around 0.5–0.7 μm, with the aged NO₃Lv4 class dominating both particle number and nitrate mass contribution. The manuscript is relevant to winter haze formation, nitrate aerosol chemistry, and source apportionment in the Yangtze River Delta.
The manuscript should more clearly explain the novelty of the proposed SPAMS–NMF framework. The authors state that the method is mass-constrained and semi-quantitative, but it would be helpful to clarify how this approach improves upon previous SPAMS-based source apportionment studies and what the main methodological advancement is.

The regression-based conversion from SPAMS particle information to nitrate mass concentration needs further justification. The reported correlation coefficient of the MLR model is moderate, and the authors should discuss the uncertainty this introduces into the subsequent NMF source apportionment and particle-class mass estimates.

The interpretation of the four nitrate-containing particle classes should be made more concise and transparent. The transition from NO₃Lv1 to NO₃Lv4 is described as increasing nitrate enrichment and aging, but the criteria used to distinguish these classes should be summarized more clearly for readers who are less familiar with SPAMS spectral classification.

The discussion of nocturnal heterogeneous nitrate formation should be expressed more cautiously. Since NO₃, N₂O₅, and ClNO₂ were not directly measured, the authors should clearly distinguish between direct observational evidence and inferred mechanisms based on nighttime enhancement, high relative humidity, and stagnant boundary-layer conditions.

Several figures are information-rich but difficult to read, especially Figures 1, 2, 5, 6, and 7. The authors should improve figure readability by increasing font sizes, simplifying legends where possible, and ensuring consistent notation for NO₃ particle classes and source factors.

The authors are strongly recommended to cite recent work on high-spatiotemporal-resolution monitoring of traffic-related particles in complex urban environments. In particular, Yeganeh, B., Shakerdonyavi, A., Zafarmomen, N., and Taheri, A.: Comprehensive spatiotemporal analysis of long-term mobile monitoring for traffic-related particles in a complex urban environment, Atmospheric Pollution Research, 102870, 2025, is highly relevant. That study provides long-term mobile-monitoring evidence for PM2.5 and black carbon variability and can strengthen the discussion of traffic-related particulate pollution, spatial heterogeneity, combustion-related particles, and urban aerosol source characterization.

The source interpretations from NMF would benefit from additional uncertainty analysis. Since NMF solutions are not unique, the authors should report how factor number selection was justified and whether sensitivity tests were performed using different factor numbers or initialization settings.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2447-CC2
RC1:
'Comment on egusphere-2026-2447', Anonymous Referee #2, 17 Jul 2026 reply
This study presents a semi-quantitative framework integrating SPAMS observations, chemical measurements, and NMF source apportionment to investigate nitrate containing particles in Shanghai. The approach helps bridge the gap between particle number based single particle measurements and mass-based source apportionment, which represents a useful methodological advancement. The manuscript is generally well organized, and the dataset collected at the Dianshan Lake supersite is valuable. The results provide new insights into the mixing state, size distribution, and source contributions of nitrate containing particles during winter haze episodes. However, several methodological descriptions and interpretations could be clarified. Additional explanations regarding the semi-quantitative calibration framework, the classification of nitrate particle types, and the interpretation of NMF factors would improve the transparency and reproducibility of the study. Some sections of the Results and Discussion could also be streamlined to better distinguish observations from interpretations.
Section 2.4 presents the multiple linear regression equation used to estimate nitrate mass concentrations from SPAMS observations. It would be helpful if the authors briefly explained how the predictor variables were selected and whether multicollinearity among meteorological parameters and PM₅related variables was evaluated. Such clarification would increase confidence in the robustness of the regression model.

The manuscript frequently discusses NO3Lv1–NO3Lv4 particle classes,however, the physical meaning of these four levels is not immediately clear to readers. The authors are encouraged to provide a summary in the main text describing the major spectral characteristics and distinguishing features of each class, rather than relying primarily on the Supplement.

The authors conclude that NO3Lv4 represents a more aged and secondary-processed nitrate particle class. While this interpretation is reasonable, a short discussion linking its dominant nitrate signals, size distribution characteristics, and associated particle types (e.g., K_N, NaK) would further strengthen the conclusion.

Several abbreviations and particle-type names are not always presented consistently throughout the manuscript (e.g., NO3_K_N, K_CN, K_NS, seasalt/SeaSalt, NO3Lv4/NO₃Lv4). Please carefully check the nomenclature and formatting to improve readability and avoid confusion.

The manuscript highlights the importance of EC-containing nitrate particles throughout the observation period. It would be beneficial to briefly discuss whether the observed EC–nitrate association mainly reflects condensation of secondary nitrate onto pre-existing combustion particles or common source influences

Some mechanistic interpretations regarding NO3/N2O5 chemistry and ammonium nitrate formation are emphasized in the Conclusions section. The authors may consider explicitly linking these conclusions to the corresponding observations presented in the Results section to improve the logical flow of the manuscript.

The manuscript would benefit from careful English editing. Several minor grammatical and typographical issues were noted. A thorough language check is recommended before publication. For examples include:

Section 4, Line 370: “simultanous” should be “simultaneous”.
Section 4: “advance understanding nitrate aerosol” could be revised to “advance the understanding of nitrate aerosols”.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2447-RC1
RC2: 'Comment on egusphere-2026-2447', Anonymous Referee #1, 23 Jul 2026 reply

In this study, the authors measured nitrate-containing particles during autumn-winter in eastern China using single-particle aerosol mass spectrometry, i.e., the SPAMS. A multiple linear regression (MLR)-based semi-quantitative method was developed to estimate particulate nitrate mass by combining SPAMS data with MARGA nitrate measurements, PM₂.₅ mass concentrations, and meteorological parameters. Developing such a semi-quantitative approach is technically challenging and valuable for the SPAMS community. The methodology, assumptions, calibration procedures, and uncertainties are well described in the Supplement. However, the manuscript should clearly distinguish particulate nitrate mass from particle mass throughout the text.
The particle classification forms the foundation of the subsequent scientific interpretation. However, the current classification scheme is overly complex, with too many particle classes and subclasses, making the results difficult to follow. I recommend simplifying the classification framework and reducing the number of particle categories where possible to improve clarity and consistency.
Some of the main interpretations are also not sufficiently supported by the presented evidence. In particular, the conclusions regarding nitrate internally mixed with combustion and marine particles are not fully convincing based on the current SPAMS data. In addition, considering the location of the measurement site, lake spray should be discussed as a potential source of Na-containing particles.
Although NMF was used for source apportionment, its methodology and added value are not clearly demonstrated. The SPAMS particle classifications and NMF source factors are presented largely in parallel rather than being effectively integrated, resulting in an unnecessarily complex framework with multiple particle "classes," "types," and source "factors." I recommend simplifying the terminology, presenting NMF as a complementary analysis, and moving detailed NMF results to the Supplement while retaining only the key findings in the main manuscript.
Overall, this manuscript addresses a topic well within the scope of the journal, and the methodology and results have the potential to make a valuable contribution. However, substantial revisions to the particle classification, source apportionment, and scientific interpretation are required before the manuscript is suitable for publication. I therefore recommend Major Revision.
Please refer to the uploaded PDF files for the detailed comments.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-2447-RC2
RC3:
'Comment on egusphere-2026-2447', Anonymous Referee #3, 23 Jul 2026 reply
This manuscript presents a semi-quantitative framework that integrates single-particle aerosol mass spectrometry (SPAMS), bulk aerosol measurements, and non-negative matrix factorization (NMF) to investigate the composition, mixing state, size distribution, and source contributions of nitrate-containing particles during the autumn–winter period at a regional site in the Yangtze River Delta. The study addresses an important scientific issue, namely the increasing dominance of particulate nitrate in PM_2.5across eastern China and the difficulty of quantitatively interpreting SPAMS observations. The proposed SPAMS–NMF framework represents an interesting attempt to bridge particle-number-resolved measurements and mass-based source apportionment.
Overall, the manuscript is well organized, the observational dataset is comprehensive, and the scientific questions addressed are timely and relevant. Compared with conventional SPAMS studies, the integration of the semi-quantitative mass-constrained approach with NMF analysis enhances the interpretation of nitrate-containing particles. The following concerns are intended to improve the clarity and robustness of the presentation.
Throughout the manuscript, both particle number and estimated mass concentrations are discussed. However, the implications of the differences between these two metrics are not always sufficiently explained. A brief discussion clarifying why some particle classes contribute disproportionately to number versus mass would improve the interpretation of Figures 2 and 3.

The manuscript appropriately emphasizes that the proposed method provides semi-quantitative rather than absolute quantitative estimates. Nevertheless, it would be helpful to further discuss the potential applicability of this framework beyond the present study. For example, can the regression model be transferred to other observation sites or seasons, or would site-specific calibration always be required? A brief discussion of the general applicability and limitations would increase the broader value of the methodology.

The uncertainty section has been considerably improved. Nevertheless, since the regression model forms the basis of the subsequent analyses, the authors may briefly comment on how the remaining unexplained variance (R = 0.68) could influence the estimated contributions of different nitrate particle classes. Even a qualitative discussion would help readers better understand the robustness of the derived source apportionment.

The workflow of ART-2a clustering followed by manual consolidation is generally clear. However, it would be beneficial to briefly explain the criteria used during the manual merging process, e.g., spectral similarity, temporal evolution, characteristic marker ions, which would improve the reproducibility of the classification procedure.

The observational support for the inferred nighttime chemical formation mechanism is relatively insufficient. In Section 3.4 and in the Abstract, the authors repeatedly mention that nitrate accumulation during severe pollution episodes, such as PE2 and PE4, was mainly driven by nighttime heterogeneous chemistry, for example N2O5 hydrolysis. However, as the authors acknowledge in the "Limitations" section, this study did not directly measure N2O5, NO3 radicals, or ClNO2. Nevertheless, relatively strong statements are made in the Discussion and Conclusions regarding nighttime N2O5 chemistry, chloride–nitrate transformation, and the formation mechanism of non-ammonium nitrate. The authors are advised to revise these statements using more cautious wording, such as "likely," "suggesting," or "consistent with." To make this key inference more convincing, the authors are also encouraged to conduct semi-quantitative supplementary analyses using the available online observations. For example, the product of NO2 and O3 could be used as an approximate proxy for the nighttime N2O5 production rate, P(N2O5), and its temporal evolution could be analyzed together with relative humidity (RH). Inferring the N2O5 pathway solely from nighttime concentration increases and high-humidity conditions leaves the evidence chain relatively weak.

The trajectory analysis occasionally attributes specific source regions solely based on HYSPLIT trajectories. Since trajectories indicate transport pathways rather than emission sources, the wording should be moderated.

The manuscript still contains a number of grammatical and typos, e.g., "wasfollowed", "particlesstarted", "isare", inconsistent spacing, missing articles, and subject–verb agreement. Careful language editing is recommended.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2447-RC3

Wenfei Zhu, Qinghong Wang, Qingsong Wang, Jialin Shi, Yu Sun, Juntao Huo, Yi Sun, Jia Chen, Yue Zhao, Mei Li, Jun Chen, Shengrong Lou, and Hui Chen

Supplement

https://doi.org/10.5194/egusphere-2026-2447-supplement

Wenfei Zhu, Qinghong Wang, Qingsong Wang, Jialin Shi, Yu Sun, Juntao Huo, Yi Sun, Jia Chen, Yue Zhao, Mei Li, Jun Chen, Shengrong Lou, and Hui Chen

Viewed

Total article views: 83 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
55	15	13	83	17	8	8

HTML: 55
PDF: 15
XML: 13
Total: 83
Supplement: 17
BibTeX: 8
EndNote: 8

Views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	55	15	13	83

Cumulative views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	55	15	13	83

Viewed (geographical distribution)

Total article views: 57 (including HTML, PDF, and XML) Thereof 57 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 23 Jul 2026

Short summary

This study examines nitrate-containing particles during winter haze in eastern China. A framework combining single-particle mass spectrometry and statistical analysis was developed to better quantify nitrate sources. Results show that nitrate is mainly found in aged particles and strongly mixed with combustion and marine emissions. Pollution episodes are driven by stagnant conditions, high humidity, nighttime reactions, and regional transport.


Total:	0
HTML:	0
PDF:	0
XML:	0