Technical note: Refining &delta;<sup>15</sup>N isotopic fingerprints of local NOx for accurate source identification of nitrate in PM<sub>2.5</sub>

Xiao, Hao; Li, Qinkai; Ding, Shiyuan; Dai, Wenjing; Cui, Gaoyang; Li, Xiaodong

doi:https://doi.org/10.5194/egusphere-2024-1621

Preprints

https://doi.org/10.5194/egusphere-2024-1621

Preprints

27 Jun 2024

| 27 Jun 2024

Technical note: Refining δ¹⁵N isotopic fingerprints of local NOx for accurate source identification of nitrate in PM_2.5

Hao Xiao, Qinkai Li, Shiyuan Ding, Wenjing Dai, Gaoyang Cui, and Xiaodong Li

Abstract. Stable nitrogen isotopic composition (δ¹⁵N) has proven to be a valuable tool for identifying sources of nitrates (NO₃^–) in PM_2.5. However, the absence of a systematic study on the δ¹⁵N values of domestic NOx sources hinders accurate identification of NO₃^– sources in China. Here, we systematically determined and refined δ¹⁵N values for six categories of NOx sources in the local Tianjin area using an active sampling method. Moreover, the δ¹⁵N values of NO₃^– in PM_2.5 were measured during pre-heating, mid-heating and late-heating periods, which are the most heavily polluted in Tianjin. Results shown that the representative nature and region-specific characteristics of isotopic fingerprints for six categories of NOx sources in Tianjin. The Bayesian isotope mixing (MixSIAR) model demonstrated that coal combustion, biomass burning, and vehicle exhaust collectively contributed more than 60 %, dominating the sources of NO₃^– during sampling periods in Tianjin. However, failure to consider the isotopic signatures of local NOx sources could result in an underestimation of the contribution from coal combustion. Additionally, the absence of industrial sources, an uncharacterized source in previous studies, may directly result in the contribution fraction of other sources being overestimated by the model more than 15 %. Notably, as the number of sources input to the model increased, the contribution of various NOx sources was becoming more stable, and the inter-influence between various sources significantly reduced. This study demonstrated that the refined isotopic fingerprint in a region-specific context could more effectively distinguish source of NO₃^–, thereby providing valuable insights for controlling NO₃^– pollution.

Received: 30 May 2024 – Discussion started: 27 Jun 2024

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 preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 1429 KB)

Supplement (1643 KB)

Download & links

Hao Xiao, Qinkai Li, Shiyuan Ding, Wenjing Dai, Gaoyang Cui, and Xiaodong Li

Status: closed

RC1:
'Comment on egusphere-2024-1621', Anonymous Referee #1, 24 Jul 2024
Summary:

The authors present a manuscript detailing various stable nitrogen isotope ratio measurements of major nitrogen oxide (NOx) sources and use these values for particulate nitrate source apportionment during a typically heavily polluted period in Tianjin, China. This manuscript provides an impressive wealth of new δ15N(NOx) emission source measurements. Specifically, they have quantified δ15N(NOx) values for previously unmeasured NOx sources, such as industrial emissions, and have made measurements representative of emissions in China, contrasting with most literature values which are predominantly for the US, potentially influencing the results due to different combustion and emission control technologies. They have also employed appropriate measurement techniques (Fibiger et al., 2014), crucial since some previous δ15N(NOx) measurements utilized a wide range of NOx collection techniques that may not all be suitable for δ15N(NOx) source characterization. I strongly support their initiative to enhance the δ15N values of NOx sources.

However, I have major reservations about this work, including its presentation and application for particulate nitrate source apportionment. There appears to be significant confusion in their literature review and citations, which require substantial editing. I have tried to point out some areas that need to be refined, though I suspect that there might be others as well. While the authors are commended for attempting to correct for isotope fractionation, I believe there are issues with their approach. They should have reported the δ18O(NO3-) data and calculated fractionations necessary for accurate source apportionment. This approach to isotope fractionation correction is described in the supplement, but the actual δ18O(NO3-) data is not included, shown, or discussed in the present work. Additionally, it is unclear whether the mixing model results are valid or contribute any meaningful insight into NOx emission source apportionment. The model is significantly under-constrained (4-5 parameters for one variable), making the output nonsensical. In conclusion, I believe this paper has the potential to be publishable after significant revisions and would recommend resubmission.
Major Comments:

Lines 64-67: Double-check these references. Felix and Elliott, 2014 used a passive sampler and collected NO2 only (not total NOx). Walters et al., 2018 report ambient δ15N measurements of NO2 and are not direct source emission measurements.

Lines 73-76: You might also consider citing, (Fibiger and Hastings, 2016) here.

Lines 95-96: Please define “dual isotopic compositions”. Do you mean the 15N and 14N (as dual isotopes of nitrogen), or are you referring to dual isotope deltas of δ 15N and δ18O? If you refer to δ18O, more discussion/explanation is needed here. The δ18O values are known to derive from the oxidants involved in NOx chemistry and do not reflect NOx sources, which is currently implied by the wording of this sentence.

Lines 132-134: Please indicate the location of the auxiliary pollutant concentrations and meteorology parameters in Fig. S1.

Line 174: The KnMnO4/NaOH NOx absorbing solution, has been reported to have a large NO3- blank (Fibiger et al., 2014). It was mentioned earlier in the text that blanks were evaluated. What were the blank levels, and how were they considered in the reported concentration and isotope data?

Lines 177-180: You should cite the original methods paper that describes the denitrifying bacteria method for nitrate isotope analysis (Casciotti et al., 2002; Sigman et al., 2001).

Lines 183-185: Please cite the papers that report on USGS34, USGS34, and IAEA-N3.

Lines 202: How is the mixing model output utilized? Most of the mixing model development papers indicate that the average and standard deviation may not be the best metrics for these mixing models, particularly when utilizing one variable (δ15N) to mix between several parameters (4 or 5 sources). Please justify the use of this type of under-constrained model for source apportionment application. For example, if you apply your mixing model to the tunnel measurements (as a test case), would your mixing model results show that vehicles were the dominant source? Or would it suggest many other sources?

Text S2: This is really important information, and it is somewhat of a shame that it has been moved to the supplement since it is quite important for the source apportionment calculations. One thing to note is that for the nitrogen isotope fractionation for NO2 + OH reaction, the authors only consider the isotope exchange between NO/NO2. However, ambient measurements and models have indicated that this isotope effect is generally quite small and often counterbalanced by NO + O3 reactions (Bekker et al., 2023; Li et al., 2020; Walters et al., 2018), under conditions of high NO2 to NOx molar ratios. The authors need to justify or revisit the use of this fractionation. Further, what about the isotope effect associated with NO2 + OH. Recent modeling work has suggested that this reaction could be important in setting the δ15N of nitrate (Fang et al., 2021). Once the ε value has been determined (which, again, I have some reservations about), how is it applied to back out the δ15N(NOx) data? The authors should have shown the calculated ε values and the backed-out δ15N(NOx) results in the main text (note: it is not included in the supplement either). Further, the authors utilize δ18O data to estimate the nitrate formation pathways. The δ18O data should be included in the manuscript, as well as the estimated nitrate formation pathways since all of this data is critical in the author’s approach for NOx source apportionment using δ15N.

Lines 210-214: What about fuel types? There is evidence that gasoline vs diesel-powered vehicles have different δ15N(NOx) values (Fibiger et al., 2014; Miller et al., 2017; Walters et al., 2015a, b). Could this have impacted the results if the vehicle-fleet distributions varied between these locations?

Lines 214-215: Generally, gasoline has very low nitrogen content, such that “fuel NOx” should be low. Is there evidence that the fuel content utilized in the study location had significant amounts of nitrogen?

Lines 222-225: This could only be true if the oil/gas had significant amounts of nitrogen. If the N content is low and NOx is primarily derived from thermal NOx, then the differences between regions would potentially reflect vehicle fleet differences, combustion differences, and/or combustion technology differences, which would have a direct impact on δ15N(NOx). It is also important to note here that according to Fig. 1, there is a wide range in emitted NOx concentration for vehicles (similar to Walters et al., 2015b). In this case, it might be better to utilize a mass-weighted δ15N(NOx) rather than an unweighted option since the NOx emission will be weighted towards the heavier NOx emitters.

Lines 228-229: Again, how is “significant” defined? Also, in the mentioned Figure S4b, it appears to have an error. Felix et al., 2012 report a δ15N(NOx) from coal combustion up to 21.0 ‰ but this graph shows it only goes up to ~12‰. Please double-check this figure.

Lines 247-249: In Walters et al., 2015a, both a residential furnace and a natural gas power plant were measured.

Lines 323-326: Nitrate is noted to be positively correlated with PM2.5. Please provide the correlation statistics.

Lines 335-337: Was there a noticeable change in nitrate production mechanisms elucidated from δ18O(NO3-) data?

Lines 366-369: The authors argue that the δ15N(NO3-) differences between these three periods could be related to NOx source differences. However, what about isotope fractionation? They previously mentioned that the NOx oxidation efficiency changed as elucidated by NO3- and NO2 concentration trends (in which NO3- concentrations didn’t increase, but NO2 did during the heating period). Including δ18O(NO3-) data, as well as the fractionation corrected δ15N(NOx), would be important here to normalize the influence of potential chemical changes that might also impact δ15N(NO3-).

Lines 415-419: Adding more parameters to an under-constrained mixing model will make the results more under-constrained. All the mixing model results have large standard deviations/uncertainties, and I’m not sure if it is warranted to discuss the differences between various model simulations, as all of the source estimates appear to overlap. For example, soil emissions are estimated to be as important as some combustion-related NOx emissions. This tends to invalidate the mixing model results, in my opinion, as these emissions should be rather low for a highly urbanized location during a cooler period.

Lines 585-589: Again, I think weighted averages might be better to report here than unweighted.

Table 1. Walters et al., 2015b did not report coal combustion δ15N(NOx). For soil emissions, you may also consider adding (Yu and Elliott, 2017).

Fig 3. It would be helpful if the stacked bar plots followed the legend order for easier visual interpretation. Furthermore, it appears that this is an error in the figure caption, as Scenarios 1 and 3 are defined twice, while Scenarios 2 and 4 are undefined.
Technical Corrections:

Technical Comment 1: Through the manuscript, please change NOx to NOx (with the “x” formatted as a subscript and in italics).

Technical Comment 2: Please write all quantity symbols (including δ, n, etc) in italics.

Technical Comment 3: Lines 59-61, this is an incomplete sentence.

Technical Comment 4: Lines 138-141: You can delete the word “Initially” here.

Technical Comment 5: Lines 156-157: Please change “cutted” to “cut”

Technical Comment 6: Lines 166-170: Neutralized is used twice in this sentence. I would suggest deleting the “to neutralize…” part of the sentence.

Technical Comment 7: Lines 170-171: You can delete “salt” from this sentence

Technical Comment 8: Lines 179: “Pseudomonas aureofaciens” should be italicized.

Technical Comment 9: Text S2: Nitrate-forming reactions aren’t typically expected to be “aqueous” reactions, such as in cloud reactions. Instead, I think the authors mean heterogeneous, and I would recommend switching these terms.

Technical Comment 10: Line 222: “Significantly” is used here and in other places in the manuscript. What significant test was utilized and what are the p-values to indicate significance?

Technical Comment 11: Lines 363-366: Please provide p-values to indicate whether these differences are significant.
References:
Bekker, C., Walters, W. W., Murray, L. T., and Hastings, M. G.: Nitrate chemistry in the northeast US – Part 1: Nitrogen isotope seasonality tracks nitrate formation chemistry, Atmospheric Chemistry and Physics, 23, 4185–4201, https://doi.org/10.5194/acp-23-4185-2023, 2023.

Casciotti, K. L., Sigman, D. M., Hastings, M. G., Böhlke, J. K., and Hilkert, A.: Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method, Analytical Chemistry, 74, 4905–4912, 2002.

Fang, H., Walters, W. W., Mase, D., and Michalski, G.: i_\textrmNRACM: incorporating 15N into the Regional Atmospheric Chemistry Mechanism (RACM) for assessing the role photochemistry plays in controlling the isotopic composition of NO_\textrmx, NO_\textrmy, and atmospheric nitrate, Geoscientific Model Development, 14, 5001–5022, https://doi.org/10.5194/gmd-14-5001-2021, 2021.

Felix, J. D. and Elliott, E. M.: Isotopic composition of passively collected nitrogen dioxide emissions: Vehicle, soil and livestock source signatures, Atmospheric Environment, 92, 359–366, 2014.

Felix, J. D., Elliott, E. M., and Shaw, S. L.: Nitrogen Isotopic Composition of Coal-Fired Power Plant NOx: Influence of Emission Controls and Implications for Global Emission Inventories, Environmental Science & Technology, 46, 3528–3535, 2012.

Fibiger, D. L. and Hastings, M. G.: First Measurements of the Nitrogen Isotopic Composition of NOx from Biomass Burning, Environmental Science & Technology, 50, 11569–11574, https://doi.org/10.1021/acs.est.6b03510, 2016.

Fibiger, D. L., Hastings, M. G., Lew, A. F., and Peltier, R. E.: Collection of NO and NO2 for Isotopic Analysis of NOx Emissions, Analytical Chemistry, 86, 12115–12121, https://doi.org/10.1021/ac502968e, 2014.

Li, J., Zhang, X., Orlando, J., Tyndall, G., and Michalski, G.: Quantifying the nitrogen isotope effects during photochemical equilibrium between NO and NO₂: implications for δ15N in tropospheric reactive nitrogen, Atmospheric Chemistry and Physics, 20, 9805–9819, https://doi.org/10.5194/acp-20-9805-2020, 2020.

Miller, D. J., Wojtal, P. K., Clark, S. C., and Hastings, M. G.: Vehicle NOx emission plume isotopic signatures: Spatial variability across the eastern United States, Journal of Geophysical Research: Atmospheres, 122, 2016JD025877, https://doi.org/10.1002/2016JD025877, 2017.

Sigman, D. M., Casciotti, K. L., Andreani, M., Barford, C., Galanter, M., and Böhlke, J. K.: A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater, Analytical chemistry, 73, 4145–4153, 2001.

Walters, W. W., Tharp, B. D., Fang, H., Kozak, B. J., and Michalski, G.: Nitrogen isotope composition of thermally produced NOx from various fossil-fuel combustion sources, Environmental science & technology, 2015a.

Walters, W. W., Goodwin, S. R., and Michalski, G.: Nitrogen Stable Isotope Composition (δ15N) of Vehicle-Emitted NOx, Environmental Science & Technology, 49, 2278–2285, https://doi.org/10.1021/es505580v, 2015b.

Walters, W. W., Fang, H., and Michalski, G.: Summertime diurnal variations in the isotopic composition of atmospheric nitrogen dioxide at a small midwestern United States city, Atmospheric Environment, 179, 1–11, https://doi.org/10.1016/j.atmosenv.2018.01.047, 2018.

Yu, Z. and Elliott, E. M.: Novel Method for Nitrogen Isotopic Analysis of Soil-Emitted Nitric Oxide, Environmental Science & Technology, 51, 6268–6278, https://doi.org/10.1021/acs.est.7b00592, 2017.
Citation: https://doi.org/10.5194/egusphere-2024-1621-RC1
- AC1: 'Reply on RC1', Hao Xiao, 08 Sep 2024
  
  Response: We would like to express our sincerest gratitude for your invaluable feedback, which has been instrumental in enhancing the quality of our manuscript. We have endeavored to optimize the manuscript and have incorporated the requisite changes, highlighted in red in the revised paper, which will not impact the content or framework of the paper. We hope that these amendments will meet with your approval. Firstly, we have revised both the literature review and inappropriate citations, including but not limited to errors noted by the reviewers. Second, we have added a discussion of the data on δ¹⁸O-NO₃^- to the revised manuscript. In addition, we refer to previous studies and add a description of the isotopic fractionation calculation process, which have been widely used in similar studies. Finally, we also refine the discussions in validity of the quantitative results of the model as well and add additional evidence. It is acknowledged that the present study is not without shortcomings, including the fractionation calculations of isotopes and the constraints of the model. However, every effort has been made to improve these aspects in the revised manuscript. Furthermore, more in-depth studies will be conducted in the future to address these shortcomings. We also summaries these deficiencies in section 3.5 Limitations and outlook of the manuscript. We would like to reiterate our appreciation for your comments and suggestions. Below, we will provide detailed one-on-one revisions and responses to deficiencies in the original manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1621-AC1
RC2:
'Comment on egusphere-2024-1621', Anonymous Referee #2, 08 Aug 2024

This work reports the measurement of the nitrogen isotopes (d15N) from major domestic NOx sources such as vehicle exhaust, industrial emissions, and natural gas combustion in Tianjin, China. The authors directly collected d15N samples from many NOx sources using active sampling methods and compared their values with previous studies. Different d15N(NOx) values from the previous studies were obtained in this area due to the local characteristics of NOx sources. They also measured inorganic ions of PM2.5 over three different heating periods to understand the contribution of NOx to nitrate formation in Tianjin. Coal combustion is reported as the main source of NO3- during heating periods in this area. In addition, the contributions from different NOx sources to NO3- in PM2.5 were quantified using a stable isotope mixing model depending on many different scenarios. In general, the study primarily focuses on refining the d15N(NOx) values from various NOx sources in Tianjin, China, suggesting that these values reflect the local characteristics. While I appreciate the effort to collect and determine d15N(NOx) values from various NOx sources and quantify NO3- formation in China, I believe several issues should be addressed and further discussion should be added when considering this study for publication in the ACP.
First, the manuscript type of technical notes in the ACP is meant to be peer-reviewed publications that report new developments, significant advances, or novel aspects of experimental and theoretical methods and techniques -- I feel like this study doesn’t meet the standard of ACP publication as a technical note. Rather, it looks closer to a measurement report since they mainly reported the measured d15N(NOx) values depending on the NOx sources and inorganic ions of PM2.5. Additionally, their major discussion focuses on the isotope mixing model to estimate the contribution of NOx sources, but their outputs are not validated. The isotope mixing model is commonly used in the isotope field to identify and quantify the contributions of multiple sources using potential source end-members. However, this statistical approach has limitations in accurately quantifying source portions. In this study, the measurement values used as inputs show high standard deviations and wide ranges. Given the isotope mixing model is sensitive to input parameters, large variabilities can lead to significant uncertainties in the output. Further, there are known large fractionations of nitrogen isotope between NOx to NO3- conversion (Li et al., 2020; Bekker et al., 2023), but these isotope effects are not considered in this study. It is also unclear which values are used for the mixing model. The model output should be validated to ensure its reliability through NOx emission inventory data or other types of observations from the Tianjin area. Lastly, the authors mentioned that a systematic study of d15N(NOx) values from domestic NOx sources is crucial for accurately identifying nitrate sources (Lines 15-16). The distinct values obtained in this study reflect the local characteristics of NOx emissions, indicating that these d15N(NOx) values might have limitations when applied to other regions. Therefore, it would be helpful if the author provided recommendations on how to best use these values and clarified their scientific relevance to other regions.

Major Comments:
1. This manuscript aims to explore NOx emission sources in the Tianjin area using isotope analysis. The author mentioned in the introduction that this area predominantly has high NOx emissions from coal combustion for heating during the heating period (Lines 93-94). However, there is no mention of other periods, such as preheating and late heating, or other potential emission sources, even though the highest NO3-and PM2.5 concentrations were observed during the preheating period. What would be the potential sources of NO3- and PM2.5 during these three different periods? It would be helpful if this argument could be supported or compared with NOx emission inventory data or another type of observation from the Tianjin area.
2. In this study, PM2.5 samples were collected for the day and night time from Oct to Apr, but the diurnal or seasonal pattern of NOx and NO3- are not considered at all in this study. Meteorological factors, especially temperature, have a significant effect on nitrogen isotope fractionation and nitrate formation. Further, the major sources of NOx emission could be variable depending on meteorological conditions. How do the meteorological factors affect the NOx and NO3- in this study?
3. Lines 381-382: As previously mentioned, given that the Coal Replacement Project has led to more natural gas usage in the recent year (Lines 356-357), I suspect natural gas combustion might be an important NOx source in your area too. Also, in a previous study, it was reported that NOx emissions driven by natural gas (i.e., liquid-fuel) combustion are one of the important NOx sources in urban areas, accounting for a larger portion than soil and biomass burning (e.g., Bekker et al., 2023). Please clarify why four sources except for natural gas specifically are chosen for the mixing model run (Line 381-382).
4. Lines 386-389: The author argued that if the mixing model is run for the entire sampling period, the results for scenarios 1 and 2 are insignificantly different. However, the results for scenario 1 are slightly lower than those for scenario 2 during the pre-heating periods. This part is confusing and there is no reported statistical analysis. Considering the larger uncertainties, especially for scenario 2, it is likely that the difference is not significant, but this should be more thoroughly addressed. In addition, what does it mean if the scenarios do not produce different results? How can you state that scenario 2 leads to fewer inaccuracies (Lines 398-399) through the mixing model results?
5. Lines 420-430: Again, it was mentioned that natural gas combustion could not be negligible in this area and the measured d15N(NOx) values from natural gas combustion were compared with the previous study. It would be helpful if you added more explanation about why your area shows more negative values compared to the previous study. Also, please add the referred values or table 1 reference in the main text.
6. The Figure 3 is missing in the manuscript. Also, please clarify the x-axis of the graphs.
7. Lines 444-454: The mixing model was run for different scenarios, with scenarios 3 and 4 using different d15N(NOx) values of natural gas combustion (i.e., scenario 3 is the previous values from other study and 4 is the measured value in this study). After that, the author compared scenario 3 and 4 results from the mixing model (Line 431-448) and made a conclusion that natural gas combustion is important for NO3- formation in this area. This part lacks an explanation in drawing the conclusion. Please clarify why natural gas is important from the two scenario results even though they account for a smaller portion than the other sources.
8. Lines 462-465: how is the d15N(NOx) value from the iron and steel industry distinct from other sources? And why? In Figure 1, d15N(NOx) from the industrial emission shows the largest range, encompassing all the values. Please clarify how you can differentiate this value from others.
9. Line 198: it was mentioned that the nitrogen isotope fractionation coefficient during NOx to NO3- conversion is calculated in Text S2, but it is unclear how these calculations and final values are applied to the d15N values for the mixing model input.

Minor comments:
Line 2: Please make the subscript x for NOx in the entire manuscript and supplementary.

Lines 42-44: Please check the sentence.

Lines 58-62: Please make clear these lines since however are repeated in every sentence.

Lines 123-124: I would suggest adding a detailed site description in this part, including information on the population of Tianjin and its size, to provide a better understanding of the area.

Lines 132-135: It would be helpful to mark the monitoring stations on the map.

Line 430: Can you check if the figure number is correct? Can’t find (d) in Figure 4.

Lines 431-446: It is quite hard to compare which fraction values represent scenarios 3 and 4 in parentheses. Please clarify these values.

Line 524: What does mean “these two sources”? Please clarify this.

Lines 902-905: the description of IE is missing in the caption.

Figure 6: (a) and (b) appear to be repetitive of Fig 5. Please revise to avoid redundancy.

Refereces
Bekker, C., Walters, W. W., Murray, L. T., and Hastings, M. G.: Nitrate chemistry in the northeast US–Part 1: Nitrogen isotope seasonality tracks nitrate formation chemistry, Atmospheric Chemistry and Physics, 23, 4185–4201, 2023.
Li, J., Zhang, X., Orlando, J., Tyndall, G., and Michalski, G.: Quantifying the nitrogen isotope effects during photochemical equilibrium between NO and NO2: implications for δ15N in tropospheric reactive nitrogen, Atmospheric Chemistry and Physics, 20, 9805–9819, 2020.

Citation: https://doi.org/10.5194/egusphere-2024-1621-RC2
- AC2: 'Reply on RC2', Hao Xiao, 08 Sep 2024
  
  Response: We thank the reviewers for investing the time to thoroughly evaluate our initial manuscript and for the constructive comments.
  First, the reviewers suggested that our study resembles a measurement report. Initially, our manuscript was indeed submitted to "Measurement Report." However, since the focus is on an enhanced framework for identifying aerosol nitrate sources through stable nitrogen isotopic analysis of local sources, the editor recommended that the paper be considered a "Technical Note." We are open to publishing the manuscript as a "Measurement Report" if feasible.
  Second, reviewers were concerned about the results of our model quantification. One of the benefits to conducting mixture models in a Bayesian framework is that information from other data sources can be included via informative prior distributions (Parnell et al., 2013;Moore and Semmens, 2008). Once an informative prior for the proportional contribution of sources is established, MixSIAR can accept the prior as an input during the model specification process. Although some of them propose to the results may be inaccurate when utilizing one variable (δ¹⁵N) to mix between several parameters (4 or 5 sources). Generally, prior beliefs about proportional source contributions (f_q) are defined using the Dirichlet distribution, with an interval of [0, 1], and the sum of the contributions of the sources of all input models defaults to 1. For example, if only 3 sources are input to the model, the model will evaluate the source contributions based on the sum of their contributions being 1. However, if the sum of the contributions from these three sources is much less than 1 in a mixed environment, then data from more sources will need to be entered to more accurately estimate the corresponding contribution from each source. In other words, the absence of the number of source input models may lead to increased uncertainty in the quantitative results of the model. More important, a growing number of recent studies have suggested the need to increase the number of sources in the model to eliminate the interactions between the various sources (Lin et al., 2024;Lin et al., 2021;Zhang et al., 2024). In the SIAR model, the Monte Carlo approach was used to quantify the emission sources of nitrate aerosols (Zong et al., 2017), it is also widely used in similar studies (Wu et al., 2024;Cheng et al., 2022;Chen et al., 2022;Song et al., 2021). To enhance the reliability of the study results, the model generated 10,000 potential scenarios for each evaluated potential source (Song et al., 2019;Fan et al., 2020). Finally, the posterior distributions and stability for the proportional contribution of each Scenarios were compared (Figure 4 and Table S2). In result, as the number of sources input to the model increased, the contribution of various NOx sources was becoming more stable, and the inter-influence between various sources significantly reduced. This implied that is no significant interinfluence in terms of estimated source apportionments when the more emission sources were considered in SIAR model. Overall, we believe that the results quantified by the model in this study are acceptable.
  Third, the effect of the fractionations of nitrogen isotope between NO_x to NO₃^- conversion has been added in the revised version. In this study, the δ¹⁸O-NO₃⁻ values helped constrain the fractionation factor from NO_x to NO₃⁻ (Xiao et al., 2020), but only two primary pathways, hydroxyl radical oxidation and nitrogen pentoxide hydrolysis, were taken into account. Previous research supports the view that these pathways account for up to 95% of NO₃⁻ production (Lin et al., 2021;Xiao et al., 2020), implying that alternative pathways might exert a relatively minor impact on εN calculations. Nonetheless, future measurements of Δ¹⁷O- NO₃⁻ are essential to elucidate the isotopic fractionation coefficients comprehensively during the formation of NO₃⁻.
  Finally, we acknowledge the limitations of our study and offer recommendations for future research. We analyzed only six samples from typical NO_x sources in Tianjin, without examining all NO_x sources in China. Variations in δ¹⁵N background values, combustion processes, and NO_x emission standards affect the δ¹⁵N signal of NO_x emissions from different sources. Given the uniformity of industrial standards across Chinese cities, these values could also help determine the source of NO₃⁻ in other locations. Our study emphasizes the need for a comprehensive determination of δ¹⁵N values for typical NO_x sources. Refining NO_x source types and improving δ¹⁵N values for other NO_x sources would enhance NO₃⁻ source apportionment in future research. Besides, the absence of constraints in the model may introduce some uncertainty into the results of this study. Consequently, further refinement may be necessary in the future to address this issue. Relevant content can be found in section 3.5 of the revised manuscript.
  We have carefully modified and proofread the manuscript. Below, we will provide detailed one-on-one revisions and responses to deficiencies in the original manuscript. All changes are marked by the red font in the revised manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1621-AC2

Status: closed

RC1:
'Comment on egusphere-2024-1621', Anonymous Referee #1, 24 Jul 2024
Summary:

The authors present a manuscript detailing various stable nitrogen isotope ratio measurements of major nitrogen oxide (NOx) sources and use these values for particulate nitrate source apportionment during a typically heavily polluted period in Tianjin, China. This manuscript provides an impressive wealth of new δ15N(NOx) emission source measurements. Specifically, they have quantified δ15N(NOx) values for previously unmeasured NOx sources, such as industrial emissions, and have made measurements representative of emissions in China, contrasting with most literature values which are predominantly for the US, potentially influencing the results due to different combustion and emission control technologies. They have also employed appropriate measurement techniques (Fibiger et al., 2014), crucial since some previous δ15N(NOx) measurements utilized a wide range of NOx collection techniques that may not all be suitable for δ15N(NOx) source characterization. I strongly support their initiative to enhance the δ15N values of NOx sources.

However, I have major reservations about this work, including its presentation and application for particulate nitrate source apportionment. There appears to be significant confusion in their literature review and citations, which require substantial editing. I have tried to point out some areas that need to be refined, though I suspect that there might be others as well. While the authors are commended for attempting to correct for isotope fractionation, I believe there are issues with their approach. They should have reported the δ18O(NO3-) data and calculated fractionations necessary for accurate source apportionment. This approach to isotope fractionation correction is described in the supplement, but the actual δ18O(NO3-) data is not included, shown, or discussed in the present work. Additionally, it is unclear whether the mixing model results are valid or contribute any meaningful insight into NOx emission source apportionment. The model is significantly under-constrained (4-5 parameters for one variable), making the output nonsensical. In conclusion, I believe this paper has the potential to be publishable after significant revisions and would recommend resubmission.
Major Comments:

Lines 64-67: Double-check these references. Felix and Elliott, 2014 used a passive sampler and collected NO2 only (not total NOx). Walters et al., 2018 report ambient δ15N measurements of NO2 and are not direct source emission measurements.

Lines 73-76: You might also consider citing, (Fibiger and Hastings, 2016) here.

Lines 95-96: Please define “dual isotopic compositions”. Do you mean the 15N and 14N (as dual isotopes of nitrogen), or are you referring to dual isotope deltas of δ 15N and δ18O? If you refer to δ18O, more discussion/explanation is needed here. The δ18O values are known to derive from the oxidants involved in NOx chemistry and do not reflect NOx sources, which is currently implied by the wording of this sentence.

Lines 132-134: Please indicate the location of the auxiliary pollutant concentrations and meteorology parameters in Fig. S1.

Line 174: The KnMnO4/NaOH NOx absorbing solution, has been reported to have a large NO3- blank (Fibiger et al., 2014). It was mentioned earlier in the text that blanks were evaluated. What were the blank levels, and how were they considered in the reported concentration and isotope data?

Lines 177-180: You should cite the original methods paper that describes the denitrifying bacteria method for nitrate isotope analysis (Casciotti et al., 2002; Sigman et al., 2001).

Lines 183-185: Please cite the papers that report on USGS34, USGS34, and IAEA-N3.

Lines 202: How is the mixing model output utilized? Most of the mixing model development papers indicate that the average and standard deviation may not be the best metrics for these mixing models, particularly when utilizing one variable (δ15N) to mix between several parameters (4 or 5 sources). Please justify the use of this type of under-constrained model for source apportionment application. For example, if you apply your mixing model to the tunnel measurements (as a test case), would your mixing model results show that vehicles were the dominant source? Or would it suggest many other sources?

Text S2: This is really important information, and it is somewhat of a shame that it has been moved to the supplement since it is quite important for the source apportionment calculations. One thing to note is that for the nitrogen isotope fractionation for NO2 + OH reaction, the authors only consider the isotope exchange between NO/NO2. However, ambient measurements and models have indicated that this isotope effect is generally quite small and often counterbalanced by NO + O3 reactions (Bekker et al., 2023; Li et al., 2020; Walters et al., 2018), under conditions of high NO2 to NOx molar ratios. The authors need to justify or revisit the use of this fractionation. Further, what about the isotope effect associated with NO2 + OH. Recent modeling work has suggested that this reaction could be important in setting the δ15N of nitrate (Fang et al., 2021). Once the ε value has been determined (which, again, I have some reservations about), how is it applied to back out the δ15N(NOx) data? The authors should have shown the calculated ε values and the backed-out δ15N(NOx) results in the main text (note: it is not included in the supplement either). Further, the authors utilize δ18O data to estimate the nitrate formation pathways. The δ18O data should be included in the manuscript, as well as the estimated nitrate formation pathways since all of this data is critical in the author’s approach for NOx source apportionment using δ15N.

Lines 210-214: What about fuel types? There is evidence that gasoline vs diesel-powered vehicles have different δ15N(NOx) values (Fibiger et al., 2014; Miller et al., 2017; Walters et al., 2015a, b). Could this have impacted the results if the vehicle-fleet distributions varied between these locations?

Lines 214-215: Generally, gasoline has very low nitrogen content, such that “fuel NOx” should be low. Is there evidence that the fuel content utilized in the study location had significant amounts of nitrogen?

Lines 222-225: This could only be true if the oil/gas had significant amounts of nitrogen. If the N content is low and NOx is primarily derived from thermal NOx, then the differences between regions would potentially reflect vehicle fleet differences, combustion differences, and/or combustion technology differences, which would have a direct impact on δ15N(NOx). It is also important to note here that according to Fig. 1, there is a wide range in emitted NOx concentration for vehicles (similar to Walters et al., 2015b). In this case, it might be better to utilize a mass-weighted δ15N(NOx) rather than an unweighted option since the NOx emission will be weighted towards the heavier NOx emitters.

Lines 228-229: Again, how is “significant” defined? Also, in the mentioned Figure S4b, it appears to have an error. Felix et al., 2012 report a δ15N(NOx) from coal combustion up to 21.0 ‰ but this graph shows it only goes up to ~12‰. Please double-check this figure.

Lines 247-249: In Walters et al., 2015a, both a residential furnace and a natural gas power plant were measured.

Lines 323-326: Nitrate is noted to be positively correlated with PM2.5. Please provide the correlation statistics.

Lines 335-337: Was there a noticeable change in nitrate production mechanisms elucidated from δ18O(NO3-) data?

Lines 366-369: The authors argue that the δ15N(NO3-) differences between these three periods could be related to NOx source differences. However, what about isotope fractionation? They previously mentioned that the NOx oxidation efficiency changed as elucidated by NO3- and NO2 concentration trends (in which NO3- concentrations didn’t increase, but NO2 did during the heating period). Including δ18O(NO3-) data, as well as the fractionation corrected δ15N(NOx), would be important here to normalize the influence of potential chemical changes that might also impact δ15N(NO3-).

Lines 415-419: Adding more parameters to an under-constrained mixing model will make the results more under-constrained. All the mixing model results have large standard deviations/uncertainties, and I’m not sure if it is warranted to discuss the differences between various model simulations, as all of the source estimates appear to overlap. For example, soil emissions are estimated to be as important as some combustion-related NOx emissions. This tends to invalidate the mixing model results, in my opinion, as these emissions should be rather low for a highly urbanized location during a cooler period.

Lines 585-589: Again, I think weighted averages might be better to report here than unweighted.

Table 1. Walters et al., 2015b did not report coal combustion δ15N(NOx). For soil emissions, you may also consider adding (Yu and Elliott, 2017).

Fig 3. It would be helpful if the stacked bar plots followed the legend order for easier visual interpretation. Furthermore, it appears that this is an error in the figure caption, as Scenarios 1 and 3 are defined twice, while Scenarios 2 and 4 are undefined.
Technical Corrections:

Technical Comment 1: Through the manuscript, please change NOx to NOx (with the “x” formatted as a subscript and in italics).

Technical Comment 2: Please write all quantity symbols (including δ, n, etc) in italics.

Technical Comment 3: Lines 59-61, this is an incomplete sentence.

Technical Comment 4: Lines 138-141: You can delete the word “Initially” here.

Technical Comment 5: Lines 156-157: Please change “cutted” to “cut”

Technical Comment 6: Lines 166-170: Neutralized is used twice in this sentence. I would suggest deleting the “to neutralize…” part of the sentence.

Technical Comment 7: Lines 170-171: You can delete “salt” from this sentence

Technical Comment 8: Lines 179: “Pseudomonas aureofaciens” should be italicized.

Technical Comment 9: Text S2: Nitrate-forming reactions aren’t typically expected to be “aqueous” reactions, such as in cloud reactions. Instead, I think the authors mean heterogeneous, and I would recommend switching these terms.

Technical Comment 10: Line 222: “Significantly” is used here and in other places in the manuscript. What significant test was utilized and what are the p-values to indicate significance?

Technical Comment 11: Lines 363-366: Please provide p-values to indicate whether these differences are significant.
References:
Bekker, C., Walters, W. W., Murray, L. T., and Hastings, M. G.: Nitrate chemistry in the northeast US – Part 1: Nitrogen isotope seasonality tracks nitrate formation chemistry, Atmospheric Chemistry and Physics, 23, 4185–4201, https://doi.org/10.5194/acp-23-4185-2023, 2023.

Casciotti, K. L., Sigman, D. M., Hastings, M. G., Böhlke, J. K., and Hilkert, A.: Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method, Analytical Chemistry, 74, 4905–4912, 2002.

Fang, H., Walters, W. W., Mase, D., and Michalski, G.: i_\textrmNRACM: incorporating 15N into the Regional Atmospheric Chemistry Mechanism (RACM) for assessing the role photochemistry plays in controlling the isotopic composition of NO_\textrmx, NO_\textrmy, and atmospheric nitrate, Geoscientific Model Development, 14, 5001–5022, https://doi.org/10.5194/gmd-14-5001-2021, 2021.

Felix, J. D. and Elliott, E. M.: Isotopic composition of passively collected nitrogen dioxide emissions: Vehicle, soil and livestock source signatures, Atmospheric Environment, 92, 359–366, 2014.

Felix, J. D., Elliott, E. M., and Shaw, S. L.: Nitrogen Isotopic Composition of Coal-Fired Power Plant NOx: Influence of Emission Controls and Implications for Global Emission Inventories, Environmental Science & Technology, 46, 3528–3535, 2012.

Fibiger, D. L. and Hastings, M. G.: First Measurements of the Nitrogen Isotopic Composition of NOx from Biomass Burning, Environmental Science & Technology, 50, 11569–11574, https://doi.org/10.1021/acs.est.6b03510, 2016.

Fibiger, D. L., Hastings, M. G., Lew, A. F., and Peltier, R. E.: Collection of NO and NO2 for Isotopic Analysis of NOx Emissions, Analytical Chemistry, 86, 12115–12121, https://doi.org/10.1021/ac502968e, 2014.

Li, J., Zhang, X., Orlando, J., Tyndall, G., and Michalski, G.: Quantifying the nitrogen isotope effects during photochemical equilibrium between NO and NO₂: implications for δ15N in tropospheric reactive nitrogen, Atmospheric Chemistry and Physics, 20, 9805–9819, https://doi.org/10.5194/acp-20-9805-2020, 2020.

Miller, D. J., Wojtal, P. K., Clark, S. C., and Hastings, M. G.: Vehicle NOx emission plume isotopic signatures: Spatial variability across the eastern United States, Journal of Geophysical Research: Atmospheres, 122, 2016JD025877, https://doi.org/10.1002/2016JD025877, 2017.

Sigman, D. M., Casciotti, K. L., Andreani, M., Barford, C., Galanter, M., and Böhlke, J. K.: A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater, Analytical chemistry, 73, 4145–4153, 2001.

Walters, W. W., Tharp, B. D., Fang, H., Kozak, B. J., and Michalski, G.: Nitrogen isotope composition of thermally produced NOx from various fossil-fuel combustion sources, Environmental science & technology, 2015a.

Walters, W. W., Goodwin, S. R., and Michalski, G.: Nitrogen Stable Isotope Composition (δ15N) of Vehicle-Emitted NOx, Environmental Science & Technology, 49, 2278–2285, https://doi.org/10.1021/es505580v, 2015b.

Walters, W. W., Fang, H., and Michalski, G.: Summertime diurnal variations in the isotopic composition of atmospheric nitrogen dioxide at a small midwestern United States city, Atmospheric Environment, 179, 1–11, https://doi.org/10.1016/j.atmosenv.2018.01.047, 2018.

Yu, Z. and Elliott, E. M.: Novel Method for Nitrogen Isotopic Analysis of Soil-Emitted Nitric Oxide, Environmental Science & Technology, 51, 6268–6278, https://doi.org/10.1021/acs.est.7b00592, 2017.
Citation: https://doi.org/10.5194/egusphere-2024-1621-RC1
- AC1: 'Reply on RC1', Hao Xiao, 08 Sep 2024
  
  Response: We would like to express our sincerest gratitude for your invaluable feedback, which has been instrumental in enhancing the quality of our manuscript. We have endeavored to optimize the manuscript and have incorporated the requisite changes, highlighted in red in the revised paper, which will not impact the content or framework of the paper. We hope that these amendments will meet with your approval. Firstly, we have revised both the literature review and inappropriate citations, including but not limited to errors noted by the reviewers. Second, we have added a discussion of the data on δ¹⁸O-NO₃^- to the revised manuscript. In addition, we refer to previous studies and add a description of the isotopic fractionation calculation process, which have been widely used in similar studies. Finally, we also refine the discussions in validity of the quantitative results of the model as well and add additional evidence. It is acknowledged that the present study is not without shortcomings, including the fractionation calculations of isotopes and the constraints of the model. However, every effort has been made to improve these aspects in the revised manuscript. Furthermore, more in-depth studies will be conducted in the future to address these shortcomings. We also summaries these deficiencies in section 3.5 Limitations and outlook of the manuscript. We would like to reiterate our appreciation for your comments and suggestions. Below, we will provide detailed one-on-one revisions and responses to deficiencies in the original manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1621-AC1
RC2:
'Comment on egusphere-2024-1621', Anonymous Referee #2, 08 Aug 2024

This work reports the measurement of the nitrogen isotopes (d15N) from major domestic NOx sources such as vehicle exhaust, industrial emissions, and natural gas combustion in Tianjin, China. The authors directly collected d15N samples from many NOx sources using active sampling methods and compared their values with previous studies. Different d15N(NOx) values from the previous studies were obtained in this area due to the local characteristics of NOx sources. They also measured inorganic ions of PM2.5 over three different heating periods to understand the contribution of NOx to nitrate formation in Tianjin. Coal combustion is reported as the main source of NO3- during heating periods in this area. In addition, the contributions from different NOx sources to NO3- in PM2.5 were quantified using a stable isotope mixing model depending on many different scenarios. In general, the study primarily focuses on refining the d15N(NOx) values from various NOx sources in Tianjin, China, suggesting that these values reflect the local characteristics. While I appreciate the effort to collect and determine d15N(NOx) values from various NOx sources and quantify NO3- formation in China, I believe several issues should be addressed and further discussion should be added when considering this study for publication in the ACP.
First, the manuscript type of technical notes in the ACP is meant to be peer-reviewed publications that report new developments, significant advances, or novel aspects of experimental and theoretical methods and techniques -- I feel like this study doesn’t meet the standard of ACP publication as a technical note. Rather, it looks closer to a measurement report since they mainly reported the measured d15N(NOx) values depending on the NOx sources and inorganic ions of PM2.5. Additionally, their major discussion focuses on the isotope mixing model to estimate the contribution of NOx sources, but their outputs are not validated. The isotope mixing model is commonly used in the isotope field to identify and quantify the contributions of multiple sources using potential source end-members. However, this statistical approach has limitations in accurately quantifying source portions. In this study, the measurement values used as inputs show high standard deviations and wide ranges. Given the isotope mixing model is sensitive to input parameters, large variabilities can lead to significant uncertainties in the output. Further, there are known large fractionations of nitrogen isotope between NOx to NO3- conversion (Li et al., 2020; Bekker et al., 2023), but these isotope effects are not considered in this study. It is also unclear which values are used for the mixing model. The model output should be validated to ensure its reliability through NOx emission inventory data or other types of observations from the Tianjin area. Lastly, the authors mentioned that a systematic study of d15N(NOx) values from domestic NOx sources is crucial for accurately identifying nitrate sources (Lines 15-16). The distinct values obtained in this study reflect the local characteristics of NOx emissions, indicating that these d15N(NOx) values might have limitations when applied to other regions. Therefore, it would be helpful if the author provided recommendations on how to best use these values and clarified their scientific relevance to other regions.

Major Comments:
1. This manuscript aims to explore NOx emission sources in the Tianjin area using isotope analysis. The author mentioned in the introduction that this area predominantly has high NOx emissions from coal combustion for heating during the heating period (Lines 93-94). However, there is no mention of other periods, such as preheating and late heating, or other potential emission sources, even though the highest NO3-and PM2.5 concentrations were observed during the preheating period. What would be the potential sources of NO3- and PM2.5 during these three different periods? It would be helpful if this argument could be supported or compared with NOx emission inventory data or another type of observation from the Tianjin area.
2. In this study, PM2.5 samples were collected for the day and night time from Oct to Apr, but the diurnal or seasonal pattern of NOx and NO3- are not considered at all in this study. Meteorological factors, especially temperature, have a significant effect on nitrogen isotope fractionation and nitrate formation. Further, the major sources of NOx emission could be variable depending on meteorological conditions. How do the meteorological factors affect the NOx and NO3- in this study?
3. Lines 381-382: As previously mentioned, given that the Coal Replacement Project has led to more natural gas usage in the recent year (Lines 356-357), I suspect natural gas combustion might be an important NOx source in your area too. Also, in a previous study, it was reported that NOx emissions driven by natural gas (i.e., liquid-fuel) combustion are one of the important NOx sources in urban areas, accounting for a larger portion than soil and biomass burning (e.g., Bekker et al., 2023). Please clarify why four sources except for natural gas specifically are chosen for the mixing model run (Line 381-382).
4. Lines 386-389: The author argued that if the mixing model is run for the entire sampling period, the results for scenarios 1 and 2 are insignificantly different. However, the results for scenario 1 are slightly lower than those for scenario 2 during the pre-heating periods. This part is confusing and there is no reported statistical analysis. Considering the larger uncertainties, especially for scenario 2, it is likely that the difference is not significant, but this should be more thoroughly addressed. In addition, what does it mean if the scenarios do not produce different results? How can you state that scenario 2 leads to fewer inaccuracies (Lines 398-399) through the mixing model results?
5. Lines 420-430: Again, it was mentioned that natural gas combustion could not be negligible in this area and the measured d15N(NOx) values from natural gas combustion were compared with the previous study. It would be helpful if you added more explanation about why your area shows more negative values compared to the previous study. Also, please add the referred values or table 1 reference in the main text.
6. The Figure 3 is missing in the manuscript. Also, please clarify the x-axis of the graphs.
7. Lines 444-454: The mixing model was run for different scenarios, with scenarios 3 and 4 using different d15N(NOx) values of natural gas combustion (i.e., scenario 3 is the previous values from other study and 4 is the measured value in this study). After that, the author compared scenario 3 and 4 results from the mixing model (Line 431-448) and made a conclusion that natural gas combustion is important for NO3- formation in this area. This part lacks an explanation in drawing the conclusion. Please clarify why natural gas is important from the two scenario results even though they account for a smaller portion than the other sources.
8. Lines 462-465: how is the d15N(NOx) value from the iron and steel industry distinct from other sources? And why? In Figure 1, d15N(NOx) from the industrial emission shows the largest range, encompassing all the values. Please clarify how you can differentiate this value from others.
9. Line 198: it was mentioned that the nitrogen isotope fractionation coefficient during NOx to NO3- conversion is calculated in Text S2, but it is unclear how these calculations and final values are applied to the d15N values for the mixing model input.

Minor comments:
Line 2: Please make the subscript x for NOx in the entire manuscript and supplementary.

Lines 42-44: Please check the sentence.

Lines 58-62: Please make clear these lines since however are repeated in every sentence.

Lines 123-124: I would suggest adding a detailed site description in this part, including information on the population of Tianjin and its size, to provide a better understanding of the area.

Lines 132-135: It would be helpful to mark the monitoring stations on the map.

Line 430: Can you check if the figure number is correct? Can’t find (d) in Figure 4.

Lines 431-446: It is quite hard to compare which fraction values represent scenarios 3 and 4 in parentheses. Please clarify these values.

Line 524: What does mean “these two sources”? Please clarify this.

Lines 902-905: the description of IE is missing in the caption.

Figure 6: (a) and (b) appear to be repetitive of Fig 5. Please revise to avoid redundancy.

Refereces
Bekker, C., Walters, W. W., Murray, L. T., and Hastings, M. G.: Nitrate chemistry in the northeast US–Part 1: Nitrogen isotope seasonality tracks nitrate formation chemistry, Atmospheric Chemistry and Physics, 23, 4185–4201, 2023.
Li, J., Zhang, X., Orlando, J., Tyndall, G., and Michalski, G.: Quantifying the nitrogen isotope effects during photochemical equilibrium between NO and NO2: implications for δ15N in tropospheric reactive nitrogen, Atmospheric Chemistry and Physics, 20, 9805–9819, 2020.

Citation: https://doi.org/10.5194/egusphere-2024-1621-RC2
- AC2: 'Reply on RC2', Hao Xiao, 08 Sep 2024
  
  Response: We thank the reviewers for investing the time to thoroughly evaluate our initial manuscript and for the constructive comments.
  First, the reviewers suggested that our study resembles a measurement report. Initially, our manuscript was indeed submitted to "Measurement Report." However, since the focus is on an enhanced framework for identifying aerosol nitrate sources through stable nitrogen isotopic analysis of local sources, the editor recommended that the paper be considered a "Technical Note." We are open to publishing the manuscript as a "Measurement Report" if feasible.
  Second, reviewers were concerned about the results of our model quantification. One of the benefits to conducting mixture models in a Bayesian framework is that information from other data sources can be included via informative prior distributions (Parnell et al., 2013;Moore and Semmens, 2008). Once an informative prior for the proportional contribution of sources is established, MixSIAR can accept the prior as an input during the model specification process. Although some of them propose to the results may be inaccurate when utilizing one variable (δ¹⁵N) to mix between several parameters (4 or 5 sources). Generally, prior beliefs about proportional source contributions (f_q) are defined using the Dirichlet distribution, with an interval of [0, 1], and the sum of the contributions of the sources of all input models defaults to 1. For example, if only 3 sources are input to the model, the model will evaluate the source contributions based on the sum of their contributions being 1. However, if the sum of the contributions from these three sources is much less than 1 in a mixed environment, then data from more sources will need to be entered to more accurately estimate the corresponding contribution from each source. In other words, the absence of the number of source input models may lead to increased uncertainty in the quantitative results of the model. More important, a growing number of recent studies have suggested the need to increase the number of sources in the model to eliminate the interactions between the various sources (Lin et al., 2024;Lin et al., 2021;Zhang et al., 2024). In the SIAR model, the Monte Carlo approach was used to quantify the emission sources of nitrate aerosols (Zong et al., 2017), it is also widely used in similar studies (Wu et al., 2024;Cheng et al., 2022;Chen et al., 2022;Song et al., 2021). To enhance the reliability of the study results, the model generated 10,000 potential scenarios for each evaluated potential source (Song et al., 2019;Fan et al., 2020). Finally, the posterior distributions and stability for the proportional contribution of each Scenarios were compared (Figure 4 and Table S2). In result, as the number of sources input to the model increased, the contribution of various NOx sources was becoming more stable, and the inter-influence between various sources significantly reduced. This implied that is no significant interinfluence in terms of estimated source apportionments when the more emission sources were considered in SIAR model. Overall, we believe that the results quantified by the model in this study are acceptable.
  Third, the effect of the fractionations of nitrogen isotope between NO_x to NO₃^- conversion has been added in the revised version. In this study, the δ¹⁸O-NO₃⁻ values helped constrain the fractionation factor from NO_x to NO₃⁻ (Xiao et al., 2020), but only two primary pathways, hydroxyl radical oxidation and nitrogen pentoxide hydrolysis, were taken into account. Previous research supports the view that these pathways account for up to 95% of NO₃⁻ production (Lin et al., 2021;Xiao et al., 2020), implying that alternative pathways might exert a relatively minor impact on εN calculations. Nonetheless, future measurements of Δ¹⁷O- NO₃⁻ are essential to elucidate the isotopic fractionation coefficients comprehensively during the formation of NO₃⁻.
  Finally, we acknowledge the limitations of our study and offer recommendations for future research. We analyzed only six samples from typical NO_x sources in Tianjin, without examining all NO_x sources in China. Variations in δ¹⁵N background values, combustion processes, and NO_x emission standards affect the δ¹⁵N signal of NO_x emissions from different sources. Given the uniformity of industrial standards across Chinese cities, these values could also help determine the source of NO₃⁻ in other locations. Our study emphasizes the need for a comprehensive determination of δ¹⁵N values for typical NO_x sources. Refining NO_x source types and improving δ¹⁵N values for other NO_x sources would enhance NO₃⁻ source apportionment in future research. Besides, the absence of constraints in the model may introduce some uncertainty into the results of this study. Consequently, further refinement may be necessary in the future to address this issue. Relevant content can be found in section 3.5 of the revised manuscript.
  We have carefully modified and proofread the manuscript. Below, we will provide detailed one-on-one revisions and responses to deficiencies in the original manuscript. All changes are marked by the red font in the revised manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1621-AC2

Hao Xiao, Qinkai Li, Shiyuan Ding, Wenjing Dai, Gaoyang Cui, and Xiaodong Li

Supplement

https://doi.org/10.5194/egusphere-2024-1621-supplement

Hao Xiao, Qinkai Li, Shiyuan Ding, Wenjing Dai, Gaoyang Cui, and Xiaodong Li

Viewed

Total article views: 451 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
317	91	43	451	59	11	17

HTML: 317
PDF: 91
XML: 43
Total: 451
Supplement: 59
BibTeX: 11
EndNote: 17

Views and downloads (calculated since 27 Jun 2024)

Month	HTML	PDF	XML	Total
Jun 2024	66	28	7	101
Jul 2024	109	17	7	133
Aug 2024	57	19	5	81
Sep 2024	48	20	21	89
Oct 2024	20	3	1	24
Nov 2024	10	1	0	11
Dec 2024	7	3	2	12

Cumulative views and downloads (calculated since 27 Jun 2024)

Month	HTML	PDF	XML	Total
Jun 2024	66	28	7	101
Jul 2024	109	17	7	133
Aug 2024	57	19	5	81
Sep 2024	48	20	21	89
Oct 2024	20	3	1	24
Nov 2024	10	1	0	11
Dec 2024	7	3	2	12

Viewed (geographical distribution)

Total article views: 424 (including HTML, PDF, and XML) Thereof 424 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 13 Dec 2024

Short summary

This study established a refined isotopic fingerprint of NOx sources in local Tianjin, which included previously uncharacterized sources in China. Results shown that the representative nature and region-specific characteristics of isotopic fingerprints for six categories of NOx sources in Tianjin. A reasonable source-resolved structure of NO₃^– could obtained by MixSIAR model using the δ¹⁵N values of NOx source established in this study, suggest the important of the refined isotopic fingerprint.


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

Technical note: Refining δ15N isotopic fingerprints of local NOx for accurate source identification of nitrate in PM2.5

Supplement

Viewed

Viewed (geographical distribution)

Technical note: Refining δ¹⁵N isotopic fingerprints of local NOx for accurate source identification of nitrate in PM_2.5