the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Understanding summertime peroxyacetyl nitrate (PAN) formation and its relation to aerosol pollution: Insights from high-resolution measurements and modeling
Abstract. Peroxyacetyl nitrate (PAN), a key indicator of photochemical pollution, is generated through a process similar to ozone (O3), involving the photochemical reactions of specific volatile organic compounds (VOCs) in the presence of nitrogen oxides. Notably, PAN has been observed at unexpectedly high concentrations (maximum: 3.04 ppb) during summertime that the daily maximum values of PAN were better correlated to black carbon (BC) (R2=0.85) than ozone (O3) (R2=0.75), suggesting that summertime haze and photochemical pollution were deeply connected. We addressed the puzzle of summertime PAN formation and its association with aerosol pollution under high ozone conditions by analyzing continuous high temporal resolution data utilizing box modeling in conjunction with the master chemical mechanism (MCM). With an IOA value of 0.75, the MCM model proves to be an ideal tool for investigating PAN photochemical formation. The model performed better during the clean period (R2: 0.6782, slope K: 0.9097) than during the haze period (R2: 0.4708, slope K: 0.7477). Through the machine learning method of XGBoost, we found that the top three factors leading to simulation bias were NH3, NO3, and PM2.5. Moreover, the net production rate of PAN becomes negative with PAN constrained, suggesting the existence of an unknown compensatory mechanism. Both RIR and EKMA analyses indicate that PAN formation in this region is VOC-controlled. Controlling emissions of VOCs, particularly alkenes, C5H8, and aromatics, would mitigate PAN pollution. RIR results also show that during the clean period, PAN is more sensitive to changes in various pollutants than during the haze period, underscoring the importance of deep emission reductions. PAN promotes OH and HO2 while inhibiting the formation of O3, RO2, NO, and NO2. This study deepens our comprehension of PAN photochemistry while also offering scientific insights for guiding future PAN pollution control strategies.
- Preprint
(1574 KB) - Metadata XML
-
Supplement
(2836 KB) - BibTeX
- EndNote
Status: open (until 23 Oct 2024)
-
RC1: 'Referee Comments – Understanding summertime peroxyacetyl nitrate (PAN) formation and its relation to aerosol pollution: Insights from high-resolution measurements and modeling (https://doi.org/10.5194/egusphere-2024-2631)', Anonymous Referee #1, 07 Oct 2024
reply
Referee Comments – Understanding summertime peroxyacetyl nitrate (PAN) formation and its relation to aerosol pollution: Insights from high-resolution measurements and modeling (https://doi.org/10.5194/egusphere-2024-2631)
General Overview:
The manuscript (egusphere-2024-2631) presents an attractive aspect of the connection of peroxyacetyl nitrate (PAN) to summertime haze and photochemical air pollution. As claimed by the manuscript, summertime PAN formation is quite an important topic. The topic also falls within the scope of the journal Atmospheric Chemistry and Physics (ACP). The authors present data from field observations with high temporal resolution as well as those from a box model and a machine learning model. The authors not only discuss the key factors and mechanisms of PAN formation based on the presented data but also tested the sensitivity of PAN to several other chemical species in the atmosphere. This manuscript is laid out well and shows the knowledge gap it fills. This manuscript is recommended to be published after addressing the concerns and comments below with minor revisions.
Major Concerns:
- Line 227 – 229: It is stated in the manuscript that “The daily variation of PAN exhibits a clear unimodal pattern, with concentrations starting to rise after sunrise and decreasing after 12:00 caused by thermal decomposition of PAN at high temperatures”. However, the diurnal variations of PAN, O3, and UV during the clean period as shown in Figure 2 (a) seems to exhibit bimodal patterns, which might not be consistent with and might not support the statement in the manuscript.
- Line 276 – 277: It is stated in the manuscript that “reactions without considered in MCM may enhance PAN generation during hazy periods”. This statement seems to be based on the fact that the slope for the hazy period in Figure 3 (c) is less than 1. However, this statement might be challenged by the facts that the R2 value is only 0.4708 and that there are multiple simulated PAN concentrations higher than observed PAN concentrations, which make the statement less convincing. It would probably be safer and more convincing to state that some reactions related to PAN generation might be missing in the MCM during the hazy period.
- Line 385 – 387 and 391 – 393: It is stated in Line 385 – 387 that “decreases in NO led to strong negative RIR” and in Line 391 – 393 that “increased NO level would inhibit the production of PAN”. These 2 statements seem to be contradictory to each other. It would be great if these 2 statements could be explained in more detail to make sure that the statements in this manuscript are consistent with each other.
- Line 457 – 458: The expression of the sentence “This further indicates that, despite the high temperatures, there is still a significant concentration of PAN, suggesting the existence of an unknown compensatory mechanism” seems unclear when it is put right after the sentence “Additionally, the net production rate of PAN becomes negative with PAN constrained”. It might take the reader quite some time and efforts to see if these 2 sentences are logically connected. It might be worth trying to rephrase the sentences to make the expressions more clearly.
Minor Concerns:
- Line 31 – 32: Under the context of the abstract, the meanings of acronyms RIR and EKMA are not clear. The full form “Relative Incremental Reactivity” of RIR is not shown until Section 2.2 (Line 129) while the full form “Empirical Kinetic Modeling Approach” of EKMA is not shown until Section 3.3 (Line 397). It would be great if the full form can appear first in the manuscript before a corresponding acronym is used.
- Line 33 and 462: The terminology “deep emission reduction” appears in the manuscript without clear definition. If the definition is similar to that in some previous studies, it would be great to cite the relevant studies with clear definition. Otherwise, it might be helpful to define it in the manuscript.
- Line 43: It might be better to cite the source of reactions R1 – R3 when they first appear in the description.
- Line 58 – 59: While discussing previous studies on wintertime photochemical air pollution in the manuscript, it would be helpful to cite the source of the statement “it is found that aerosol promotes PAN generation”.
- Line 59 – 60: While discussing previous studies on wintertime photochemical air pollution in the manuscript, please“Surprisingly high concentrations of OH radical, particularly under hazy conditions, have been observed and are largely attributed to HONO photolysis”.
- Line 139, 283, and 288: The meaning of the acronym OBM is not clear. The full name of the acronym seems to be missing in the manuscript.
- Line 144: It might be better if the source of the “SHapley Additive exPlanations (SHAP)” approach can be cited.
- Line 144 – 146: It might be better if some of the studies that have successfully applied the SHAP approach can be cited.
- Line 188 – 189: The statement “the precursor concentration of PAN is significantly lower than in the northern region” is not quite clear. Is it meant to be It would be helpful if the statement can be clarified.
- Line 190 – 191figure to support the statement “The correlation between the daily maximum values of PAN and BC is the strongest (R=0.85), followed by O3 (R=0.75)”?
- Line 227: It would be helpful if it could be pointed out that the average diurnal patterns of PAN and related variables for clean and hazy conditions are shown in Figure 2 before the contents in Figure 2 are discussed in detail without clearly stating where the contents are shown.
- Line 274 – 276: Both the R2 and K values are discussed in the manuscript, but only the R2 values are defined (Line 152) and shown (Figure 3 (c)). The K values seem to be not defined in the manuscript. It might take the reader some time and efforts to notice that the K values potentially mean the slopes in Figure 3 (c). I would suggest the authors to clearly define
- Line 279 – 280 (Figure 3): The legend of the figure shows “obs” and “sim” without their definitions. if the legend of the figure could be defined in and be consistent with the caption of the figure. For example, the caption could be modified as “Comparison of observed (obs) PAN and simulated (sim) PAN”.
- Line 282 – 283 and 304 – 305: It might be better if the definition of bias described as “difference between the model simulation values and the observed values” can be expressed mathematically as what values minus what values for the reader to be clear about the mathematic definition.
- Line 283 and 285: Since there are 2 models, a box model and a machine learning model, being used in this study
- Line 285 – 286: It is stated that “NH3 is the most significant parameter affecting bias, contributing 19.68 %”. However, it seems that the number 19.68 % is not shown in Figure 4 (a). It would be helpful to clarify whether the contribution of 19.68 % is on average or obtained in some other ways. It would also be great to show such a value in the corresponding figure as described in the manuscript.
- Line 294 – 295: It is stated that “NO3- is the second most significant parameter influencing the bias between the two, contributing 11.33 %”. However, it seems that the number 11.33 % is not shown in Figure 4 (a). Could the authorsclarify whether the contribution of 11.33 % is on average or obtained in some other ways? it be possible to show such a value in the corresponding figure as described in the manuscript?
- Line 297: It is stated that “PM2.5 is the third most significant parameter, contributing 9.4 %”. However, it seems that the number 9.4 % is not shown in Figure 4 (a). It would be helpful to clarify whether the contribution of 9.4 % is on average or obtained in some other ways. It would also be great to show such a value in the corresponding figure as described in the manuscript.
- Line 307 – 308 and 320 – 321: Since there are data from field observations and model simulations being used in this study, it would be helpful if it could be clearly stated whether the average production and destruction rates of PAN during clean and haze periods are observed or simulated by what model. Although the caption of Figure 5 mentions that they are simulated, it would be great if it could be clearly stated in the paragraph of description as well.
- Line 312 – 313: I would appreciate itwhether the net production rate of PAN is simulated net production rate or observed net production rate.
- Line 313 – 314: Is the diurnal variation of PAN based on observation or simulations? I would suggest the authors toit here.
- Line 327 – 329: It is stated that “We conducted a correlation analysis of the net production rate of PAN with temperature, PAN concentration, VOCs, and NO2”. However, it seems that only the correlation between the simulated net production rate of PAN and observed PAN concentration is shown in Figure S7. It would be great if the other correlations mentioned in the manuscript can be provided to support the statement.
- Line 330 – 332: It is stated that the sensitivity experiments are shown in Figure 5, but it seems that the sensitivity experiments are actually shown in Figure 6.
- Line 331, 333, 335, 337, and 347 (Figure 6): Could the authorsstate in the text and caption of the figure whether the net production rate of PAN is simulated net production rate of PAN or not?
- Line 349 – 350: It is stated that “budget analysis of PA’s production and consumption pathways is frequently used”. However, only 1 study is cited to support this statement, which might not be convincing. It might be better if more studies are cited to support the statement that the method is frequently used. Otherwise, it might be safer and more convincing to state that the method has been used with only 1 citation.
- Line 351: It would be helpful if it is clearly stated whether the diurnal patterns are simulated diurnal patterns or not.
- Line 356 – 357: It is stated that “the conversion of PAN into PA radical through thermal decomposition had high correlations with temperature during both haze and clean days”. It would be appreciated if a figure of correlations can be provided to support the statement.
- Line 364 – 366: It would be clearer to the reader if it could be stated what “these four pathways” mentioned in the statement are. The pathways are shown in Figure 7, but it would be clearer if they could be described in the paragraph as well.
- Line 377: It is stated that “The primary contributor to the PAN destruction rate was the reaction between PA and NO2”, but it seems that this sentence is meant to describe the PA destruction rate instead of the PAN destruction rate.
- Line 381 (Figure 7): Are
- Line 416 and 421: It is stated that “ΔHO2 and ΔOH are positive for most periods” and that “ΔRO2, ΔNO2, and ΔNO are negative for most periods”. It might be better if there are specific numbers to quantitatively support the statements since there are also multiple periods with negative values of ΔHO2 and ΔOH and multiple periods with positive values of ΔRO2, ΔNO2, and ΔNO shown in Figure 9.
- Line 417, 419, 423, and 424: The statistic term “significant difference” appears multiple times in the manuscript. It would be more convincing if the significance levels to determine whether there would be statistically significant differences or not are clearly stated in this study.
- Line 420: The meaning of the acronym AOC is not clear. The full name of the acronym seems to be missing in the manuscript.
- Line 429 – 430: the definition of “The difference of HO2, OH, RO2, NO2, and NO between base scenario with PAN mechanism and scenario without PAN mechanism” could be expressed mathematically as what values minus what values for the reader to be clear about the mathematic definition.
- Line 432 – 434: The term “inhibition rate” appears in the manuscript without clear definition. It would be helpful to define it in the manuscript.
- Line 440: It would be clearer if it could be stated that the precursors mentioned here are precursors of what specific chemical species.
- Line 440: It would be clearer if it could be stated what “their” means in the sentence with the term “their secondary formation”.
- Line 451: It would be clearer if the units for the number 0.009 could be stated.
- Line 456: It would be helpful if it is clearly stated whether the net production rate of PAN is simulated net production rate of PAN or not.
- Figure S6term “their product” in the caption should be clearly described along with the mathematic expression O3´JO1D to be consistent with the axis labels of the figure.
- Figure S7: It seems that the axis labels of abscissa and ordinate are missing.
- Figure S8: It might be clearer for the reader to see the variations of time series if the axis limits could be adjusted closer to the minimum and maximum of each time series.
Technical Comments:
- Line 76: “Ximen” seems to be a typographical error of “Xiamen”.
- Line 199 – 201: The sentence “PM2.5 concentrations during the haze period were significantly higher than during the clean period, being 2.49 times that of the clean period” might be corrected as “PM2.5 concentrations during the haze period were significantly higher than those during the clean period, being 2.49 times those of the clean period”.
- Line 204 – 205: The sentence “During the haze period, ozone concentrations were also significantly higher than during the clean period, being 2.04 times that of the clean period” might be corrected as “During the haze period, ozone concentrations were also significantly higher than those during the clean period, being 2.04 times those of the clean period”.
- Line 435: The “Fig. S11” seems to be meant as “Fig. S8” since there are only 8 figures in the supporting information.
- Figure S5: The axis label of the abscissa “maximum daily ozone concentration (PAN)” seems to be a typographical error of “maximum daily ozone concentration (O3)”.
Citation: https://doi.org/10.5194/egusphere-2024-2631-RC1 -
RC2: 'Corrections for Referee Comments – Understanding summertime peroxyacetyl nitrate (PAN) formation and its relation to aerosol pollution: Insights from high-resolution measurements and modeling (https://doi.org/10.5194/egusphere-2024-2631)', Anonymous Referee #1, 07 Oct 2024
reply
Referee Comments on the following lines have been updated and corrected due to some formatting issues while they were posted earlier.
• Line 59 – 60: While discussing previous studies on wintertime photochemical air pollution in the manuscript, please cite the source for the statement “Surprisingly high concentrations of OH radical, particularly under hazy conditions, have been observed and are largely attributed to HONO photolysis”.
• Line 188 – 189: The statement “the precursor concentration of PAN is significantly lower than in the northern region” is not quite clear. Is it meant to be “the precursor concentration of PAN is significantly lower than that in the northern region” or “the precursor concentration of PAN is significantly lower in the northern region”? It would be helpful if the statement can be clarified.
• Line 190 – 191: Could the authors provide a figure to support the statement “The correlation between the daily maximum values of PAN and BC is the strongest (R=0.85), followed by O3 (R=0.75)”?
• Line 274 – 276: Both the R2 and K values are discussed in the manuscript, but only the R2 values are defined (Line 152) and shown (Figure 3 (c)). The K values seem to be not defined in the manuscript. It might take the reader some time and efforts to notice that the K values potentially mean the slopes in Figure 3 (c). I would suggest the authors to clearly define the K values in the manuscript.
• Line 279 – 280 (Figure 3): The legend of the figure shows “obs” and “sim” without their definitions. It would be great if the legend of the figure could be defined in and be consistent with the caption of the figure. For example, the caption could be modified as “Comparison of observed (obs) PAN and simulated (sim) PAN”.
• Line 283 and 285: Since there are 2 models, a box model and a machine learning model, being used in this study, it would be appreciated if the “target” and the “features” mentioned here are of which model can be clearly stated.
• Line 294 – 295: It is stated that “NO3- is the second most significant parameter influencing the bias between the two, contributing 11.33 %”. However, it seems that the number 11.33 % is not shown in Figure 4 (a). Could the authors clarify whether the contribution of 11.33 % is on average or obtained in some other ways? Would it be possible to show such a value in the corresponding figure as described in the manuscript?
• Line 312 – 313: I would appreciate it if it could be clearly stated whether the net production rate of PAN is simulated net production rate or observed net production rate.
• Line 313 – 314: Is the diurnal variation of PAN based on observation or simulations? I would suggest the authors to clearly state it here.
• Line 331, 333, 335, 337, and 347 (Figure 6): Could the authors clearly state in the text and caption of the figure whether the net production rate of PAN is simulated net production rate of PAN or not?
• Line 381 (Figure 7): Are the PA radical production and destruction rates simulated or not?
• Line 429 – 430: It would be helpful if the definition of “The difference of HO2, OH, RO2, NO2, and NO between base scenario with PAN mechanism and scenario without PAN mechanism” could be expressed mathematically as what values minus what values for the reader to be clear about the mathematic definition.
• Figure S6: The term “their product” in the caption should be clearly described along with the mathematic expression O3×JO1D to be consistent with the axis labels of the figure.Citation: https://doi.org/10.5194/egusphere-2024-2631-RC2
-
RC3: 'Comment on egusphere-2024-2631', Anonymous Referee #2, 09 Oct 2024
reply
Detailed comments can be found in the supplement.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
142 | 34 | 13 | 189 | 13 | 3 | 1 |
- HTML: 142
- PDF: 34
- XML: 13
- Total: 189
- Supplement: 13
- BibTeX: 3
- EndNote: 1
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1