the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Nov 2022
Real-time measurements of NMVOCs in the central IGB, Lucknow, India: Source characterization and their role in O3 and SOA formation
Abstract. Lucknow is the capital of India’s largest state, Uttar Pradesh, one of South Asia’s most polluted urban cities. Tropospheric photochemistry relies on non-methane volatile organic compounds (NMVOCs), which are ozone and secondary organic aerosol precursors. Using the proton-transfer reaction time of flight mass spectrometer (PTR-ToF-MS) at an urban background site in Lucknow, the chemical characterisation of NMVOCs was performed in real-time from Dec-2020–May 2021. The campaign mean concentrations of the NMVOCs were 125.5 ±37.5 ppbv. The average concentrations of NMVOCs are relatively high during winter. An advanced multi-linear engine (ME-2) model was used to perform the NMVOCs source apportionment using positive matrix factorisation (PMF). It resolves the five main sources contributing to these organic compounds in the atmosphere. They include traffic, two solid fuel combustion factors, secondary volatile organic compounds and volatile chemical products. Biomass burning contributes most to the NMVOCs and SOA formation, while interestingly, traffic sources most influence ozone formation. Significant differences in the composition of the two solid fuel combustion indicate the influence of local emissions and transport of regional pollution to the city. The high temperature during summer leads to more volatilisation of oxygenated VOCs.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1745 KB)
-
Supplement
(958 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1745 KB) - Metadata XML
-
Supplement
(958 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1165', Anonymous Referee #1, 08 Dec 2022
General Comments
This study describes the deployment of a state-of-the-art Proton Transfer Reaction Mass Spectrometer instrument for high time resolution measurements of a large range of volatile organic compounds in an urban location in India. Positive Matrix Factorization was used to apportion the sources of the measured compounds. Relationships with simultaneous measurements of PM2.5 aerosol chemical composition via High Res ToF-AMS as well as Black Carbon (BC), NOx, SO2, Ozone, meteorological parameters and back trajectory analysis were used to support the selected PMF solutions and explore temporal variations.
The ozone formation potential and SOA yield of individual VOCs as well as each identified factor were estimated. This study identified traffic, solid fuel combustion, secondary VOC formation and volatile chemical products associated with industry as the dominant sources of VOCs at the sampling site. VOCs associated with traffic and solid fuel combustion had the highest ozone formation potential and estimated SOA yield.
This work is an addition to other recent studies using HR-PTR-MS in Ahmedabad (e.g. Sahu et al 2015, 2016, 2017) and Delhi (e.g. Wang et al 2020, Tripathi et al 2022, Jain et al 2022) with associated studies using a HR-ToF-AMS (Shukla et al 2021, Lalchandani et al 2021, Tobler et al 2020) undertaken with the researchers from the Indian Institute of Technology Kanpur. While the present study is of relevance to national and regional air quality management and population health studies, the manuscript requires further work to demonstrate novelty and impact for the wider atmospheric chemistry and physics domain. In particular:
- Co-located measurements by Hr-ToF-AMS and PTR-ToF-MS are uncommon and offer a novel opportunity to characterise total atmospheric organic carbon and relationships between gas and aerosol species. A more full presentation of the AMS measurements and an exploration of relationships between gas and aerosol organics would significantly enhance the novelty of this manuscript.
- More in depth comparison with similar previous studies in Indian / other Asian cities would enhance the wider impact of this manuscript – synthesise common factors and their key trace species in VOC and NR-PM2.5 composition emerging from these studies. What are the common factors identified in these studies, what unique factors emerge in individual studies from different regions. The work by Zhang et al 2007 (https://doi.org/10.1029/2007GL029979) may provide a useful example of synthesis.
- While the purpose of these studies is to understand drivers of poor air quality and impact on population health, this is not highlighted in the discussion. The concentrations of individual air toxic VOCs (e.g. benzene, formaldehyde) could be compared to National / WHO air quality objectives. The relative contribution of SOA to total PM2.5 burden could also be presented. Identify days of poorest air quality – what factors were the major contributors?
- Some measurement issues need to be better addressed in the Methodology and Supplement to provide confidence in the interpretation (see below)
- Discussion on the influence of meteorological conditions and photochemistry associated with transport and aging on the airmasses sampled needs to be expanded.
- Manuscript requires a general proofread and some statements require supporting references. Be aware of use of past and present tense.
Specific Comments
Abstract
- Line 19 “the average concentrations of NMVOCs are relatively high during winter”. Use quantitative statements ie NMVOCs were X-X% higher in winter than in summer.
- Comment on which were dominant VOCs / VOC families that comprised NMVOC.
- Comment on % contributions of each factor to NMVOC.
- Line 23 “ Biomass burning contribute most of the NMVOCs and SOA formation, while interestingly traffic sources most influence ozone formation”.
- Use consistent terms ie Biomass burning or SFC
- Is the contribution of biomass burning factors to total OA (and SOA) from the AMS data reported?
- The potential SOA yield from the NMVOC factors was only estimated. Suggest rewording ie Aged and fresh emissions from Solid Fuel combustion (SFC 1 and 2) was the dominant contributor to total NMVOC and compounds related to these factors had a high SOA formation potential.
- Likewise for traffic sources and OFP.
- Line 26 “ The high temperature during summer leads to more volatilisation of oxygenated VOCs.” Again be consistent – does oxygenated VOCs here refer to both SVOC and VCP factors or just VCP? Ie Higher temperatures in summer were associated with more volatilisation of oxygenated Volatile Chemical Products from industry sources.
- The significance/ specialty of the study needs to be highlighted in the abstract – what is its practical significance (ie to the atmospheric measurement community), what is its significance to understanding /management of air quality in this region?
- Introduction
- Suggest presenting only a summary of direct and indirect effect of VOCs on air quality as these will be familiar concepts to most ACP readers. Dedicate more of the introduction to summarising previous studies of NMVOCs and NR-PM2.5 composition in Indian and other Asian cities.
- Move info regarding general characteristics of the sampling region (ie population, industry, land use categories etc) from Section 2.1 into the intro section – ie lines 95 - 106.
- Methodology
- Sampling Site Description
- Move details of inlet and instrument in to next section.
- Were met paramters measured at the site? Provide more detail on meteorology differences between seasons temp, RH, for this location.
- Instrumentation and data analysis
- Suggest rename to PTR-ToF-MS measurements of NMVOCs
- Suggest further reducing general info on PTR-ToF-MS method and focus on presenting specific measurement details.
- More inlet detail – what was the inner diameter of the PFA inlet ? Was the flow down this inlet 60ml/min total or did the PTRMS just sub-sample 60ml/min from a higher sample inlet flow? ie what was the residence time in inlet?
- Provide specific info on calibration and zero measurements here and in supplement ie mean +- stdev of compound specific sensitivities, how were background corrections applied, range of MDLs ?
- Did you use measured sensitivity or calculated sensitivity based on k rates and transmission for all VOCs reported or only those not in cal std? comment on uncertainty associated with these approaches?
- Why were formaldehyde and methanol excluded from compounds reported? These were likely to be very significant?
- PTRMS sensitivity to formaldehyde is low (quasi-thermoneutral) and humidity dependent – applying a standard k-rate approach to this species is erroneous – use empirically derived, humidity dependent sensitivity factors or exclude from your analysis.
- Acetic Acid is notoriously difficult to measure due to its stickiness – how did you account for this?
- How were the data analysed (PTRMS viewer), were the data reported here averaged?
- Supporting measurements
- Suggest separate section (2.3) HR-ToF-AMS measurements of aerosol composition providing more detail on HR-ToF-AMS measurements – aerosol inlet, calibrations, acquisition parameters, PMF followed by 2.4 Supporting measurements – BC, NOx, O3, SO2, meteorology
- Source Apportionment
- The authors provide a good explanation of the PMF approach used and logic for selecting optimum solution.
- Significant text is dedicated in this section and section 3.2.1 to describing the PMF method and optimum solution selection process. Suggest merging these sections into one (sect 2.4) and supplement if text limited. Focus results and discussion on the characteristics and behaviour of the selected factors.
- Ozone formation potential and SOA yield of NMVOCs
- Details on the processes driving O3 formation and their non-linearity were discussed in the Intro and do not need to be repeated here. Only details relevant to the OFP estimation method should be presented. The paper that first described this method should be referenced in the opening sentence.
- Describe MIR. How many of the species reported have an MIR – does excluding those without MIR bias the results?
- Line 236 “ the literature determines the SOA yield” – revise sentence. The SOA yields reported by Bruns et al 2016 were used for this analysis. Where SOA yields were not knoe, compounds with >6 carbons were assumed to have an SOA yield of 0.32 based on ….?
- Note this analysis represents the OFP and SOA formation potential of the air mass composition at the sampling site and not the OFP SOA FP of the various emission sources. Ie airmasses dominated by fresh emissions (eg traffic) will have a different OFP and SOA FP than aged airmasses (eg long range transport of BB plumes)
2.6 CWT back trajectory analysis
1) rename section “Concentration weighted back trajectory analysis”
2) 100m of arrival height is repeated x2 in the text.
3) Acknowledge who this method was first described by.
4) Note the reader may require explanation to reconcile why prevailing winds during study as shown in windroses in Fig S1 were predominantly from the SE – SW yet the CWT plots in fig S3 show higher trajectory density from the North.
- Results and Discussion
3.1 NMVOC concentrations and temporal variation.
1) In general this section requires significant revision to clarify the aim of this section and the concepts presented to improve interpretation. Suggest here presenting
- summary stats for NMVOC, dominant species (Acetald., Acetone, Acetic Acid) and relative contributions of these species and each of the VOC families to NMVOC.
- seasonally and diurnally varying patterns
- High pollution events – provide data on NMVOC, specific air toxics eg benz, Pm2.5, ozone – which NMVOCs species/families were dominant in these episodes, what was the prevailing meterology (ie stagnant conditions in winter, high photochemistry conditions in summer?)
2) Figure 2 – include lines to indicate winter and summer periods. Note low data capture for months to Dec and to April may bias these results. In top panel ‘VOC time series’ Consider instead of plotting ‘Other” plot rel. contribution of VOC families
3) Line 264 – 266 “ the highest concentrations of NMVOCs, NR-PM2.5 during the winter months infer their common sources” – meteorology would also play an important role ie calm conditions in winter and lower PBL? Use quantitative statements ie provide NMVOC and NR-PM2.5 concs in brackets.
4) Line 266 – “In contrast, during the summer months, PM2.5 decreases drastically, but NMVOC concentrations are relatively highest, implying additional sources of NMVOCs”. – sentence requires revision.
6) Line 275 “ diurnal variations of secondary formation, anthropogenic emission level, weather and PBL heights can be explained by OVOC/ benzene ratios to some extent” revise sentence – the ratio of OVOCs/benzene does not explain these factors.
3.2 PMF results
3.2.1 Optimum solution selection
Merge this with discussion of PMF methodology in 2.4.
3.2.2 Profile and diurnal variation
1) Suggest rename / restructure this section ‘3.2 Characteristics of selected PMF factors’
4) General info on outcomes of PMF, outline info that will be presented to characterise each factor and then present detail under Sub-headings 3.2.X for each factor
2) Fig 5 – add formula/name to key peaks. Figures 6 and 7 are useful as is. Fig 8 – check AMS species labels are the same as presented in text (ie MO-OOA and LO-OOA). Figure 9c while r2 value is 0.52 the plot indicates a poor relationship.
3) Add time series of factors – consider adding to Fig 2 to aid comparison with other variables.
5) Use consistent presentation of characteristics for each factor
- Factor identification ie traffic, SFC1, SFC 2
- Marker species and their average % contribution to the factor.
- relationships to other atmospheric species – move discussion from 3.3 under each relevant sub-heading
- diurnal / seasonal patterns which help identify sources eg diurnal patterns that align with peak traffic; seasonal patterns of SFC.
- CWT plots for each factor – do they align with location of known / likely sources?
- comparison with previous studies – similar markers and % contributions?
3.4 OFP and SOA yield from individual sources
1) Is there a relationship between factors/ species with high SOA and O3 potential and measured concentrations of SOA and O3? Consider a time lag in peaks.
2) This section would be improved by comparison with previous studies and discussion on relevance of this section ie for control strategies to reduce O3 and SOA.
3) The limitations of these approaches should be noted – these are estimates of potential for ozone and SOA formation not actual yields of ozone and SOA.
- Conclusion
1) This section should be used to synthesise what has been learnt from this and the previous studies – what factors are common to Indian/Asian cities and which are different- consider a mapped pie chart type plot for NMVOCs and NR-Pm2.5 composition like that shown in Zhang 2007.
2) Line 546 “ The NMVOCs and NOx derive from the formation of ozone and SOA, but there is limited knowledge of their complex relationship” – sentence needs revising. Reverse relationship is true – The formation of ozone and SOA is driven by the oxidation of NMVOCs. The purpose of this statement is not clear.
3) The overall significance of this work in understanding and better managing air quality in Lucknow and other Indian cities needs to be stated.
Citation: https://doi.org/10.5194/egusphere-2022-1165-RC1 -
AC1: 'Reply on RC1', Sachchida Tripathi, 13 Feb 2023
We thank the referees for their valuable comments which have greatly helped us to improve the manuscript.
Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in
the revised manuscript are written in italic.
-
RC2: 'Comment on egusphere-2022-1165', Anonymous Referee #2, 08 Dec 2022
The manuscript entitled: “Real-time measurements of NMVOCs in the central IGB, Lucknow, India: Source characterization and their role in O3 and SOA formation” by Jain et al. investigates air quality in Lucknow, India to understand how local and regional emissions contribute to ozone and SOA formation. They employed PMF to apportion NMVOCs to sources related to traffic, solid fuel combustion, volatile chemical products, and secondary formation. These factors were further investigated and found to be consistent with other key measurements such as organic PM PMF factors. Traffic and solid fuel combustion contributed most to SOA formation and OFP. Overall, this article fits the scope of the journal and addresses important questions regarding the sources of pollution which drive air quality. I would recommend publication after major revisions with regards to the comments below.
General Comments
Hydrocarbons were measured via PTR-TOF-MS in this study, but alkanes and small alkenes cannot be detected. It is important to be transparent regarding this fact as such species may represent significant fractions of the true total NMVOCs as well as potentially your PMF factors’ relative abundance, OFP, and SOA yield. Acknowledgement and discussion of this limitation is necessary.
Key findings regarding factors’ relative importance of OFP and SOA yield (and thus, the dominant source(s) of pollution) rely on some significant assumptions (unknown MIRs in Table S2, assumed SOA yields in Table S3). This limitation is briefly noted as “There are many NMVOCs species with unknown ozone and SOA yield values. More research on the section is needed.” (lines 529 – 530). It is understandable that the analysis presented here uses the available information, but these limitations must be discussed in greater detail in relation to the findings themselves. Additionally, a sensitivity analysis regarding the unknown values is necessary to substantiate these findings. Do these assumptions make a significant impact or, if not, why?
Comparisons of the most abundant NMVOCs, PMF factors, and dominant contributors to OFR and SOA production with other studies of nearby major cities would help put these measurements and conclusions into a greater context.
The goal of this work seems to be focused on the main contributors to air pollution, but there is relatively little discussion. Do any individual NMVOCs present a health risk based on your measurements and air quality standards? How frequently do ozone and PM concentrations exceed standards, and which NMVOC and AMS factors seem to drive these events?
Specific Comments
Line 18: Notably PTR does not detect alkanes and small alkenes. I suggest specifying “measured” or “quantified” NMVOCs.
Lines 142 – 143: To calibrate the PTR-MS signals, “…a typical value of 2 x 10-9 cm3 s-1 of the proton transfer reaction rate coefficient…” was used. Was this for all NMVOCs you did not directly calibrate, or for all NMVOCs you did not have a literature value for? Clarification would be helpful.
Given that these rate constants vary by a factor of ~2, how does the assumption of an average rate constant affect your calibrations and the rest of your analysis? A discussion on the resulting uncertainties is necessary.
What were the sensitivities and limits of detection for your standards? Do estimated LODs have implications for your other measurements (e.g., are your measurements of tetradecane and others above the LOD)?
Was the instrument’s transmission function investigated / were transmission correction applied?
Since calibrations were done in the beginning, middle, and end of the campaign, were the signals normalized to the reagent ion (as is typical for PTR-MS measurements) to account for relative humidity contributions to the water content in the reactor and general instrument variability?
Were instrument background signals measured and applied? If so, how were they measured?
Line 194: There is some inconsistent use of “ions” and “m/z” alongside “mixing ratios.” If calibrated NMVOCs were used in your PMF analysis, they are no longer ions. This also applies elsewhere in the manuscript.
Lines 270 – 271: “…more partitioning of the gas phase during summers relatively to winters.” is somewhat ambiguous regarding the direction of partitioning during colder vs warmer months.
Line 335: It is unclear to me what “±3” refers to in the context of the scaled residuals.
Lines 335 – 336: The text would suggest 3 – 7-factor solutions, but Figure S4 shows timeseries for 2 – 10-factor solutions. The diurnals in Figure S4 also include an 11-factor solution whereas the 11-factor timeseries is absent. Figure 4 includes 3 – 11-factor solutions. Please address these inconsistencies.
Line 339: Estimating from Figure 4, the percent change in Q/Qexp went from ~12% in the 4-factor solution to ~10% in the 5-factor solution to ~8% in the 6-factor solution which does not seem significantly different. Additionally, the total scaled residuals dropped significantly between the 5 and 6 factor solutions. From these parameters alone, one could argue that the 6-factor solution would be better. However, there are other metrics as described previously in the same paragraph that could be used to rule out the 6-factor solution as nonsensical. Please discuss further why the 5-factor solution was chosen as opposed to a 6-factor solution in relation to these other metrics.
Line 363: A short discussion of Figure S5 (as a whole and for each factor) would be helpful for readers such as myself with limited knowledge of this type of analysis. The slopes suggest an error of 1% or less, but the figure seems to tell us more than that. SFC2 seems to have 2 distinct lobes, one with higher concentrations and lower uncertainties (such that PMF latched onto these measurements to determine this factor) and a second lobe with lower concentrations and a higher slope/uncertainty. This effect is stronger for VCPs, and the SVOC factor has essentially no correlation between spread and concentration. Are these observations meaningful and, if so, what do they mean for the quality of the factors?
Section 3.2.2: More discussion of the CWT analysis is necessary, including references to Figure 1. Do these results align with expectations?
There is abundant discussion of key NMVOCs in each factor and how these factors were identified. How do these factors compare to other studies (distribution of NMVOCs within each factor, most abundant NMVOCs, relative abundance of each factor compared to each other, etc.)
Section 3.4: This section briefly reports the results of the OFP and SOA production results, but requires more discussion to articulate the impact and significance of these findings. They should be compared to other studies for context. Limitations should be restated and interpreted with regards to the results.
Lines 513 – 514: When breaking down the OFP (and later SOA) contributions for each factor, one average value is presented for the full measurement period. Do these distributions (and thus dominant factor) vary significantly between seasons (e.g., winter, late winter, and summer as in Figure S1)?
Figure 5: What is the difference between the bars and dots in the profiles? For the SVOC profile, what do the grey bars represent?
Also, I would suggest creating a new supplementary figure with the factors’ diurnals for each season.
Figure 7a-b: It may be easier to understand this plot with species names as opposed to formulas (or both names and formulas together). Species are already associated with these formulas in Table S1 and “Ethanol” is already provided as a specific compound in this figure. “Unknown” species should be left as formulas.
Figure 10: I believe the third value in each box of (b) and (c) refers to the SOA yield mass concentration and mixing ratios, respectively. Clarification is necessary in the caption. Also, if I am correct, why is the corresponding OFP value not included?
Technical Corrections
Line 101: Please define “MSMEs”.
Line 236: I believe this should be “Equation 5.”
Line 344: Please define “SVOC” (currently defined later in line 368).
Line 345: Please define “random values” / “a”.
Figure 4: there is no clear link between the traces and their corresponding axes. Add info to the caption, e.g., “…total summed scaled residuals (red trace)…” or change right axis color to red (similar to Figure 3).
Figure 10: Assuming the bottom values for (b) and (c) are the SOA yield mass concentration and average mixing ratios, respectively, please include units
Figure S1: Please specify the units of wind speed in the legend.
Figure S5: The color schemes seem off. The SVOC factor is colored according to mean concentration (color gradient from left to right), but the others don’t follow this same convention. Renaming the color bar label (currently just “ppbv” where both x- and y-axes are in ppbv) could help with interpretation.
Figure S6: Please define “T/B ratio.” It does not seem to be mentioned in the main text.
Citation: https://doi.org/10.5194/egusphere-2022-1165-RC2 -
AC2: 'Reply on RC2', Sachchida Tripathi, 13 Feb 2023
We thank the referees for their valuable comments which have greatly helped us to improve the manuscript.
Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in
the revised manuscript are written in italic.
-
AC2: 'Reply on RC2', Sachchida Tripathi, 13 Feb 2023
-
AC3: 'Comment on egusphere-2022-1165', Sachchida Tripathi, 13 Feb 2023
The combined response to the comments by the referee (RC1 and RC2) is attached as a supplement pdf file. Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in the revised manuscript are written in italic.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1165', Anonymous Referee #1, 08 Dec 2022
General Comments
This study describes the deployment of a state-of-the-art Proton Transfer Reaction Mass Spectrometer instrument for high time resolution measurements of a large range of volatile organic compounds in an urban location in India. Positive Matrix Factorization was used to apportion the sources of the measured compounds. Relationships with simultaneous measurements of PM2.5 aerosol chemical composition via High Res ToF-AMS as well as Black Carbon (BC), NOx, SO2, Ozone, meteorological parameters and back trajectory analysis were used to support the selected PMF solutions and explore temporal variations.
The ozone formation potential and SOA yield of individual VOCs as well as each identified factor were estimated. This study identified traffic, solid fuel combustion, secondary VOC formation and volatile chemical products associated with industry as the dominant sources of VOCs at the sampling site. VOCs associated with traffic and solid fuel combustion had the highest ozone formation potential and estimated SOA yield.
This work is an addition to other recent studies using HR-PTR-MS in Ahmedabad (e.g. Sahu et al 2015, 2016, 2017) and Delhi (e.g. Wang et al 2020, Tripathi et al 2022, Jain et al 2022) with associated studies using a HR-ToF-AMS (Shukla et al 2021, Lalchandani et al 2021, Tobler et al 2020) undertaken with the researchers from the Indian Institute of Technology Kanpur. While the present study is of relevance to national and regional air quality management and population health studies, the manuscript requires further work to demonstrate novelty and impact for the wider atmospheric chemistry and physics domain. In particular:
- Co-located measurements by Hr-ToF-AMS and PTR-ToF-MS are uncommon and offer a novel opportunity to characterise total atmospheric organic carbon and relationships between gas and aerosol species. A more full presentation of the AMS measurements and an exploration of relationships between gas and aerosol organics would significantly enhance the novelty of this manuscript.
- More in depth comparison with similar previous studies in Indian / other Asian cities would enhance the wider impact of this manuscript – synthesise common factors and their key trace species in VOC and NR-PM2.5 composition emerging from these studies. What are the common factors identified in these studies, what unique factors emerge in individual studies from different regions. The work by Zhang et al 2007 (https://doi.org/10.1029/2007GL029979) may provide a useful example of synthesis.
- While the purpose of these studies is to understand drivers of poor air quality and impact on population health, this is not highlighted in the discussion. The concentrations of individual air toxic VOCs (e.g. benzene, formaldehyde) could be compared to National / WHO air quality objectives. The relative contribution of SOA to total PM2.5 burden could also be presented. Identify days of poorest air quality – what factors were the major contributors?
- Some measurement issues need to be better addressed in the Methodology and Supplement to provide confidence in the interpretation (see below)
- Discussion on the influence of meteorological conditions and photochemistry associated with transport and aging on the airmasses sampled needs to be expanded.
- Manuscript requires a general proofread and some statements require supporting references. Be aware of use of past and present tense.
Specific Comments
Abstract
- Line 19 “the average concentrations of NMVOCs are relatively high during winter”. Use quantitative statements ie NMVOCs were X-X% higher in winter than in summer.
- Comment on which were dominant VOCs / VOC families that comprised NMVOC.
- Comment on % contributions of each factor to NMVOC.
- Line 23 “ Biomass burning contribute most of the NMVOCs and SOA formation, while interestingly traffic sources most influence ozone formation”.
- Use consistent terms ie Biomass burning or SFC
- Is the contribution of biomass burning factors to total OA (and SOA) from the AMS data reported?
- The potential SOA yield from the NMVOC factors was only estimated. Suggest rewording ie Aged and fresh emissions from Solid Fuel combustion (SFC 1 and 2) was the dominant contributor to total NMVOC and compounds related to these factors had a high SOA formation potential.
- Likewise for traffic sources and OFP.
- Line 26 “ The high temperature during summer leads to more volatilisation of oxygenated VOCs.” Again be consistent – does oxygenated VOCs here refer to both SVOC and VCP factors or just VCP? Ie Higher temperatures in summer were associated with more volatilisation of oxygenated Volatile Chemical Products from industry sources.
- The significance/ specialty of the study needs to be highlighted in the abstract – what is its practical significance (ie to the atmospheric measurement community), what is its significance to understanding /management of air quality in this region?
- Introduction
- Suggest presenting only a summary of direct and indirect effect of VOCs on air quality as these will be familiar concepts to most ACP readers. Dedicate more of the introduction to summarising previous studies of NMVOCs and NR-PM2.5 composition in Indian and other Asian cities.
- Move info regarding general characteristics of the sampling region (ie population, industry, land use categories etc) from Section 2.1 into the intro section – ie lines 95 - 106.
- Methodology
- Sampling Site Description
- Move details of inlet and instrument in to next section.
- Were met paramters measured at the site? Provide more detail on meteorology differences between seasons temp, RH, for this location.
- Instrumentation and data analysis
- Suggest rename to PTR-ToF-MS measurements of NMVOCs
- Suggest further reducing general info on PTR-ToF-MS method and focus on presenting specific measurement details.
- More inlet detail – what was the inner diameter of the PFA inlet ? Was the flow down this inlet 60ml/min total or did the PTRMS just sub-sample 60ml/min from a higher sample inlet flow? ie what was the residence time in inlet?
- Provide specific info on calibration and zero measurements here and in supplement ie mean +- stdev of compound specific sensitivities, how were background corrections applied, range of MDLs ?
- Did you use measured sensitivity or calculated sensitivity based on k rates and transmission for all VOCs reported or only those not in cal std? comment on uncertainty associated with these approaches?
- Why were formaldehyde and methanol excluded from compounds reported? These were likely to be very significant?
- PTRMS sensitivity to formaldehyde is low (quasi-thermoneutral) and humidity dependent – applying a standard k-rate approach to this species is erroneous – use empirically derived, humidity dependent sensitivity factors or exclude from your analysis.
- Acetic Acid is notoriously difficult to measure due to its stickiness – how did you account for this?
- How were the data analysed (PTRMS viewer), were the data reported here averaged?
- Supporting measurements
- Suggest separate section (2.3) HR-ToF-AMS measurements of aerosol composition providing more detail on HR-ToF-AMS measurements – aerosol inlet, calibrations, acquisition parameters, PMF followed by 2.4 Supporting measurements – BC, NOx, O3, SO2, meteorology
- Source Apportionment
- The authors provide a good explanation of the PMF approach used and logic for selecting optimum solution.
- Significant text is dedicated in this section and section 3.2.1 to describing the PMF method and optimum solution selection process. Suggest merging these sections into one (sect 2.4) and supplement if text limited. Focus results and discussion on the characteristics and behaviour of the selected factors.
- Ozone formation potential and SOA yield of NMVOCs
- Details on the processes driving O3 formation and their non-linearity were discussed in the Intro and do not need to be repeated here. Only details relevant to the OFP estimation method should be presented. The paper that first described this method should be referenced in the opening sentence.
- Describe MIR. How many of the species reported have an MIR – does excluding those without MIR bias the results?
- Line 236 “ the literature determines the SOA yield” – revise sentence. The SOA yields reported by Bruns et al 2016 were used for this analysis. Where SOA yields were not knoe, compounds with >6 carbons were assumed to have an SOA yield of 0.32 based on ….?
- Note this analysis represents the OFP and SOA formation potential of the air mass composition at the sampling site and not the OFP SOA FP of the various emission sources. Ie airmasses dominated by fresh emissions (eg traffic) will have a different OFP and SOA FP than aged airmasses (eg long range transport of BB plumes)
2.6 CWT back trajectory analysis
1) rename section “Concentration weighted back trajectory analysis”
2) 100m of arrival height is repeated x2 in the text.
3) Acknowledge who this method was first described by.
4) Note the reader may require explanation to reconcile why prevailing winds during study as shown in windroses in Fig S1 were predominantly from the SE – SW yet the CWT plots in fig S3 show higher trajectory density from the North.
- Results and Discussion
3.1 NMVOC concentrations and temporal variation.
1) In general this section requires significant revision to clarify the aim of this section and the concepts presented to improve interpretation. Suggest here presenting
- summary stats for NMVOC, dominant species (Acetald., Acetone, Acetic Acid) and relative contributions of these species and each of the VOC families to NMVOC.
- seasonally and diurnally varying patterns
- High pollution events – provide data on NMVOC, specific air toxics eg benz, Pm2.5, ozone – which NMVOCs species/families were dominant in these episodes, what was the prevailing meterology (ie stagnant conditions in winter, high photochemistry conditions in summer?)
2) Figure 2 – include lines to indicate winter and summer periods. Note low data capture for months to Dec and to April may bias these results. In top panel ‘VOC time series’ Consider instead of plotting ‘Other” plot rel. contribution of VOC families
3) Line 264 – 266 “ the highest concentrations of NMVOCs, NR-PM2.5 during the winter months infer their common sources” – meteorology would also play an important role ie calm conditions in winter and lower PBL? Use quantitative statements ie provide NMVOC and NR-PM2.5 concs in brackets.
4) Line 266 – “In contrast, during the summer months, PM2.5 decreases drastically, but NMVOC concentrations are relatively highest, implying additional sources of NMVOCs”. – sentence requires revision.
6) Line 275 “ diurnal variations of secondary formation, anthropogenic emission level, weather and PBL heights can be explained by OVOC/ benzene ratios to some extent” revise sentence – the ratio of OVOCs/benzene does not explain these factors.
3.2 PMF results
3.2.1 Optimum solution selection
Merge this with discussion of PMF methodology in 2.4.
3.2.2 Profile and diurnal variation
1) Suggest rename / restructure this section ‘3.2 Characteristics of selected PMF factors’
4) General info on outcomes of PMF, outline info that will be presented to characterise each factor and then present detail under Sub-headings 3.2.X for each factor
2) Fig 5 – add formula/name to key peaks. Figures 6 and 7 are useful as is. Fig 8 – check AMS species labels are the same as presented in text (ie MO-OOA and LO-OOA). Figure 9c while r2 value is 0.52 the plot indicates a poor relationship.
3) Add time series of factors – consider adding to Fig 2 to aid comparison with other variables.
5) Use consistent presentation of characteristics for each factor
- Factor identification ie traffic, SFC1, SFC 2
- Marker species and their average % contribution to the factor.
- relationships to other atmospheric species – move discussion from 3.3 under each relevant sub-heading
- diurnal / seasonal patterns which help identify sources eg diurnal patterns that align with peak traffic; seasonal patterns of SFC.
- CWT plots for each factor – do they align with location of known / likely sources?
- comparison with previous studies – similar markers and % contributions?
3.4 OFP and SOA yield from individual sources
1) Is there a relationship between factors/ species with high SOA and O3 potential and measured concentrations of SOA and O3? Consider a time lag in peaks.
2) This section would be improved by comparison with previous studies and discussion on relevance of this section ie for control strategies to reduce O3 and SOA.
3) The limitations of these approaches should be noted – these are estimates of potential for ozone and SOA formation not actual yields of ozone and SOA.
- Conclusion
1) This section should be used to synthesise what has been learnt from this and the previous studies – what factors are common to Indian/Asian cities and which are different- consider a mapped pie chart type plot for NMVOCs and NR-Pm2.5 composition like that shown in Zhang 2007.
2) Line 546 “ The NMVOCs and NOx derive from the formation of ozone and SOA, but there is limited knowledge of their complex relationship” – sentence needs revising. Reverse relationship is true – The formation of ozone and SOA is driven by the oxidation of NMVOCs. The purpose of this statement is not clear.
3) The overall significance of this work in understanding and better managing air quality in Lucknow and other Indian cities needs to be stated.
Citation: https://doi.org/10.5194/egusphere-2022-1165-RC1 -
AC1: 'Reply on RC1', Sachchida Tripathi, 13 Feb 2023
We thank the referees for their valuable comments which have greatly helped us to improve the manuscript.
Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in
the revised manuscript are written in italic.
-
RC2: 'Comment on egusphere-2022-1165', Anonymous Referee #2, 08 Dec 2022
The manuscript entitled: “Real-time measurements of NMVOCs in the central IGB, Lucknow, India: Source characterization and their role in O3 and SOA formation” by Jain et al. investigates air quality in Lucknow, India to understand how local and regional emissions contribute to ozone and SOA formation. They employed PMF to apportion NMVOCs to sources related to traffic, solid fuel combustion, volatile chemical products, and secondary formation. These factors were further investigated and found to be consistent with other key measurements such as organic PM PMF factors. Traffic and solid fuel combustion contributed most to SOA formation and OFP. Overall, this article fits the scope of the journal and addresses important questions regarding the sources of pollution which drive air quality. I would recommend publication after major revisions with regards to the comments below.
General Comments
Hydrocarbons were measured via PTR-TOF-MS in this study, but alkanes and small alkenes cannot be detected. It is important to be transparent regarding this fact as such species may represent significant fractions of the true total NMVOCs as well as potentially your PMF factors’ relative abundance, OFP, and SOA yield. Acknowledgement and discussion of this limitation is necessary.
Key findings regarding factors’ relative importance of OFP and SOA yield (and thus, the dominant source(s) of pollution) rely on some significant assumptions (unknown MIRs in Table S2, assumed SOA yields in Table S3). This limitation is briefly noted as “There are many NMVOCs species with unknown ozone and SOA yield values. More research on the section is needed.” (lines 529 – 530). It is understandable that the analysis presented here uses the available information, but these limitations must be discussed in greater detail in relation to the findings themselves. Additionally, a sensitivity analysis regarding the unknown values is necessary to substantiate these findings. Do these assumptions make a significant impact or, if not, why?
Comparisons of the most abundant NMVOCs, PMF factors, and dominant contributors to OFR and SOA production with other studies of nearby major cities would help put these measurements and conclusions into a greater context.
The goal of this work seems to be focused on the main contributors to air pollution, but there is relatively little discussion. Do any individual NMVOCs present a health risk based on your measurements and air quality standards? How frequently do ozone and PM concentrations exceed standards, and which NMVOC and AMS factors seem to drive these events?
Specific Comments
Line 18: Notably PTR does not detect alkanes and small alkenes. I suggest specifying “measured” or “quantified” NMVOCs.
Lines 142 – 143: To calibrate the PTR-MS signals, “…a typical value of 2 x 10-9 cm3 s-1 of the proton transfer reaction rate coefficient…” was used. Was this for all NMVOCs you did not directly calibrate, or for all NMVOCs you did not have a literature value for? Clarification would be helpful.
Given that these rate constants vary by a factor of ~2, how does the assumption of an average rate constant affect your calibrations and the rest of your analysis? A discussion on the resulting uncertainties is necessary.
What were the sensitivities and limits of detection for your standards? Do estimated LODs have implications for your other measurements (e.g., are your measurements of tetradecane and others above the LOD)?
Was the instrument’s transmission function investigated / were transmission correction applied?
Since calibrations were done in the beginning, middle, and end of the campaign, were the signals normalized to the reagent ion (as is typical for PTR-MS measurements) to account for relative humidity contributions to the water content in the reactor and general instrument variability?
Were instrument background signals measured and applied? If so, how were they measured?
Line 194: There is some inconsistent use of “ions” and “m/z” alongside “mixing ratios.” If calibrated NMVOCs were used in your PMF analysis, they are no longer ions. This also applies elsewhere in the manuscript.
Lines 270 – 271: “…more partitioning of the gas phase during summers relatively to winters.” is somewhat ambiguous regarding the direction of partitioning during colder vs warmer months.
Line 335: It is unclear to me what “±3” refers to in the context of the scaled residuals.
Lines 335 – 336: The text would suggest 3 – 7-factor solutions, but Figure S4 shows timeseries for 2 – 10-factor solutions. The diurnals in Figure S4 also include an 11-factor solution whereas the 11-factor timeseries is absent. Figure 4 includes 3 – 11-factor solutions. Please address these inconsistencies.
Line 339: Estimating from Figure 4, the percent change in Q/Qexp went from ~12% in the 4-factor solution to ~10% in the 5-factor solution to ~8% in the 6-factor solution which does not seem significantly different. Additionally, the total scaled residuals dropped significantly between the 5 and 6 factor solutions. From these parameters alone, one could argue that the 6-factor solution would be better. However, there are other metrics as described previously in the same paragraph that could be used to rule out the 6-factor solution as nonsensical. Please discuss further why the 5-factor solution was chosen as opposed to a 6-factor solution in relation to these other metrics.
Line 363: A short discussion of Figure S5 (as a whole and for each factor) would be helpful for readers such as myself with limited knowledge of this type of analysis. The slopes suggest an error of 1% or less, but the figure seems to tell us more than that. SFC2 seems to have 2 distinct lobes, one with higher concentrations and lower uncertainties (such that PMF latched onto these measurements to determine this factor) and a second lobe with lower concentrations and a higher slope/uncertainty. This effect is stronger for VCPs, and the SVOC factor has essentially no correlation between spread and concentration. Are these observations meaningful and, if so, what do they mean for the quality of the factors?
Section 3.2.2: More discussion of the CWT analysis is necessary, including references to Figure 1. Do these results align with expectations?
There is abundant discussion of key NMVOCs in each factor and how these factors were identified. How do these factors compare to other studies (distribution of NMVOCs within each factor, most abundant NMVOCs, relative abundance of each factor compared to each other, etc.)
Section 3.4: This section briefly reports the results of the OFP and SOA production results, but requires more discussion to articulate the impact and significance of these findings. They should be compared to other studies for context. Limitations should be restated and interpreted with regards to the results.
Lines 513 – 514: When breaking down the OFP (and later SOA) contributions for each factor, one average value is presented for the full measurement period. Do these distributions (and thus dominant factor) vary significantly between seasons (e.g., winter, late winter, and summer as in Figure S1)?
Figure 5: What is the difference between the bars and dots in the profiles? For the SVOC profile, what do the grey bars represent?
Also, I would suggest creating a new supplementary figure with the factors’ diurnals for each season.
Figure 7a-b: It may be easier to understand this plot with species names as opposed to formulas (or both names and formulas together). Species are already associated with these formulas in Table S1 and “Ethanol” is already provided as a specific compound in this figure. “Unknown” species should be left as formulas.
Figure 10: I believe the third value in each box of (b) and (c) refers to the SOA yield mass concentration and mixing ratios, respectively. Clarification is necessary in the caption. Also, if I am correct, why is the corresponding OFP value not included?
Technical Corrections
Line 101: Please define “MSMEs”.
Line 236: I believe this should be “Equation 5.”
Line 344: Please define “SVOC” (currently defined later in line 368).
Line 345: Please define “random values” / “a”.
Figure 4: there is no clear link between the traces and their corresponding axes. Add info to the caption, e.g., “…total summed scaled residuals (red trace)…” or change right axis color to red (similar to Figure 3).
Figure 10: Assuming the bottom values for (b) and (c) are the SOA yield mass concentration and average mixing ratios, respectively, please include units
Figure S1: Please specify the units of wind speed in the legend.
Figure S5: The color schemes seem off. The SVOC factor is colored according to mean concentration (color gradient from left to right), but the others don’t follow this same convention. Renaming the color bar label (currently just “ppbv” where both x- and y-axes are in ppbv) could help with interpretation.
Figure S6: Please define “T/B ratio.” It does not seem to be mentioned in the main text.
Citation: https://doi.org/10.5194/egusphere-2022-1165-RC2 -
AC2: 'Reply on RC2', Sachchida Tripathi, 13 Feb 2023
We thank the referees for their valuable comments which have greatly helped us to improve the manuscript.
Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in
the revised manuscript are written in italic.
-
AC2: 'Reply on RC2', Sachchida Tripathi, 13 Feb 2023
-
AC3: 'Comment on egusphere-2022-1165', Sachchida Tripathi, 13 Feb 2023
The combined response to the comments by the referee (RC1 and RC2) is attached as a supplement pdf file. Please find below our point-by-point responses (in blue) after the referee comments (in black). The changes in the revised manuscript are written in italic.
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 438 | 156 | 15 | 609 | 55 | 3 | 7 |
- HTML: 438
- PDF: 156
- XML: 15
- Total: 609
- Supplement: 55
- BibTeX: 3
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Vaishali Jain
Sachchida N. Tripathi
Nidhi Tripathi
Mansi Gupta
Vishnu Murari
Sreenivas Gaddamidi
Ashutosh K. Shukla
Andre S. H. Prevot
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1745 KB) - Metadata XML
-
Supplement
(958 KB) - BibTeX
- EndNote
- Final revised paper