Modeling organic aerosol over Central Europe: uncertainties linked to different chemical mechanisms, parameterizations, and boundary conditions

Bartík, Lukáš; Huszár, Peter; Peiker, Jan; Karlický, Jan; Vlček, Ondřej; Vodička, Petr

doi:https://doi.org/10.5194/egusphere-2025-167

Preprints

https://doi.org/10.5194/egusphere-2025-167

Preprints

25 Apr 2025

| 25 Apr 2025

Modeling organic aerosol over Central Europe: uncertainties linked to different chemical mechanisms, parameterizations, and boundary conditions

Lukáš Bartík, Peter Huszár, Jan Peiker, Jan Karlický, Ondřej Vlček, and Petr Vodička

Abstract. This study explores the uncertainties in modeling organic aerosol (OA) over Central Europe, focusing on the roles of chemical mechanisms, emission parameterizations, and boundary conditions. Organic aerosols, particularly secondary organic aerosols (SOAs), significantly influence climate, health, and visibility, comprising up to 90 % of submicron particulate matter. Using the Comprehensive Air Quality Model with Extensions (CAMx) coupled with the Weather Research and Forecast Model, sensitivity analyses were conducted to assess the impact of intermediate-volatility organic compounds (IVOCs), semi-volatile organic compounds (SVOCs), and chemical boundary conditions on primary and secondary organic aerosol concentrations.

Results showed that including source-specific IVOC and SVOC emissions significantly improved CAMx's performance in reproducing observed OA levels, mainly when using the 1.5-dimensional Volatility Basis Set framework with activated chemical aging. For example, the domain-averaged SOA concentrations increased by up to 1.17 μg m^-3 during summer when both IVOC and SVOC emissions were included. Furthermore, incorporating OA into the boundary conditions enhanced model predictions, with the accuracy of modeled organic carbon concentrations improving by up to 100 % during summer at some monitoring sites. Despite these improvements, challenges remain due to uncertainties in emission estimates, parameterization schemes, and the spatial resolution of the models.

The findings underscore the importance of refined parameterizations for IVOC and SVOC emissions, higher temporal and spatial resolution in chemical boundary conditions, and better representation of chemical aging. Addressing these gaps in future studies will further enhance the understanding and prediction of OA dynamics in regional air quality modeling.

Received: 14 Jan 2025 – Discussion started: 25 Apr 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 1483 KB)

Supplement (740 KB)

Download & links

Lukáš Bartík, Peter Huszár, Jan Peiker, Jan Karlický, Ondřej Vlček, and Petr Vodička

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-167', Anonymous Referee #2, 07 May 2025
The study presents some key challenges in accurately modeling organic aerosol concentrations over Central Europe. By using an advanced regional model along with the different approaches for atmospheric aging of organic matter, it identifies the importance of including explicit SVOC and IVOC parameterizations in emission inventories to achieve higher agreement with observations. Moreover, it points out that the share between primary and secondary organic aerosol considered for the boundary conditions is a critical choice for accuracy. This study explores a region that has not been the main focus of air quality research in Europe and would therefore be a useful reference for the community. Before I can suggest it for final publication, there are some major and minor issues that should be first addressed. You can find them below :

Major Comments :

What I found missing was a comprehensive round-up in Section 4 regarding the rate of improvement among the different model setups that were tested. For example, it is mentioned in Section 3 that the first sensitivity experiment provided increased improvement for the winter period, but the second one did so for the summer period. However different location types were taken into account as well (Rural/Urban), and it was not made clear whether one particular setup is more suitable depending on those conditions. The authors should expand the conclusions in order to clearly present which model configuration is the most important option depending on location type and season.

Minor Comments :

Line 12 : “the accuracy of modeled organic carbon concentrations improving by up to 100 %”. This sentence should be rephrased as the improvement corresponds to the FAC2 metric specifically, and can otherwise be misleading.

Line 47 : The authors should consider presenting the saturation concentrations of LVOCs and SVOCs in a similar format as those for the IVOCs.

Line 55 : “ both the original 1.5-D VBS and its various modifications”. What do these modifications alter in the original 1.5-D VBS framework ?

Line 121 : What was the reasoning behind using 2 different chemical mechanisms in your simulations, since one is more comprehensive than the other ?

Lines 127 – 128 : Is the aerosol size distribution in the CAMx model bimodal too ? Are there any kinetic limitations taken into account for gas diffusion ?

Lines 150 – 152 : Does the REZZO inventory have a different spatiotemporal resolution than CAMS ? What do you mean by 'temporal disaggregation and speciation ' ?

Section 2.3 : I recommend that the authors change the titles of subsections 2.3.1 & 2.3.2 in order to make clear what was changed in each sensitivity analysis.

Table 1 : Why was the VBS framework not combined with the SAPRC07TC chemical mechanism at all for the CVb and CVa experiments ?

Lines 223 & 224 : Some of the factors for SOA redistribution in Tables S3 & S4 are higher in the winter than in the summer. Since typically SOA formation peaks in the warmer periods, is this something you expect ?

Lines 224 & 225 : Similarly, some of the factors for POA redistribution in Table S5 are higher in the summer than in the winter. Since, typically during the colder period POA emissions peak due to combustion generated for heating demands, is this something you expect ?

Section 2.4 : What is the frequency of the model output regarding aerosol concentrations ? Does it produce average (i.e. daily) or instantaneous values ?

Lines 245 & 246 : Were daily values extrapolated for the validation in this case, or was only the model output corresponding to the measurement dates used ?

Line 279 : “the model typically tends to overestimate them more or less”. It would look better if a more accurate measure for the overestimation was stated here.

Figures 2 & 3 : The authors should consider adding a legend with information about what a positive/negative difference corresponds to, similarly to how Figures 6 & 7 are presented.

Lines 291 – 292 : The underestimation of both temperature and relative humidity over the domain, also plays a role in that result. It would be important to point that out.

Line 328 : Does that mean that the concentrations in the CVb & CVa experiments are double those of the CSnI experiment ?

Line 329 : What do these “similarities” refer to ?

Figure 5 : The relative increase in SOA concentrations by the CVb experiment is much more drastic (1 order of magnitude compared to CSnI) in winter than in summer, around the Po Valley. How can this be explained ?

Line 361 : Please change ‘he’ to ‘they’ when referring to a study.

Lines 362 – 365 : If CSnI is taken as the reference experiment in this sensitivity, why is the comparison here with the results of Meroni et al., (2017) made with the CSwI experiment ?

Lines 375 & 376 : In Table 1 it states that this particular experiment did include SVOC emissions. Please clarify.

Section 3.2.2 : For the 2^nd paragraph please provide percentages as a measure for the differences. The authors should also consider changing the order of paragraphs 3 & 4 to match the order of presented results by figures and tables.

Line 416 : Shouldn’t Tables S9 and S10 be referenced here, instead of S4 and S6 ?

Lines 474 – 478 : Is this behavior driven solely by the fractions of PAP0 and PAP1 ? And if so, is it expected ? Are the impacts of PFP0,PFP1 and PFP2 not as important ?

Line 494 : “likely linked to changes in other pollutant(s)”. Could you provide some examples ?

Line 497 : Shouldn’t it be ‘Sp0s100’ instead of ‘Sp100s0’ ? And similarly, shouldn’t it be ‘Vp0s100’ instead of ‘Vp50s000’ ?

Lines 500 & 501 : Why would the evaporation of POA be relevant for the p0s100 scenarios ?

Lines 501 – 505 : Is this behavior driven solely by the fractions of PAS0/PBS0 (VBS module) and SOA4/SOPB (SOAP module) ? Are the impacts of others not as important ?

Table S2 : What is the size distribution of organic matter, black carbon and sulfate ? Did it have to change for the mapping ? Where there any CBCs considered for ammonium, nitrate and mineral cations besides sodium, since they are also treated by ISORROPIA ? Was dust considered in a bulk only state without any particular chemical composition ?

Table S3 : How are the 6 SOA surrogate species allocated ? In Appendix A it is mentioned that SOAP considers 3 species (2+1) from anthropogenic and another 3 (2+1) from biogenic. Can you specify which is used for what ?

Tables S3, S4 and S5 : The authors should consider inserting footnotes to explain what all the species abbreviations refer to.

Table S6 : If I understood correctly, the measurements from the Prague-Schudol station correspond to PM₁₀. This should be clarified in the footnotes.

Table S9 : What do the last 3 columns correspond to ?

Figure S3 : I believe that this figure is redundant and can be omitted, as this is more general information that can be easily accessed in statistics books.

References
Meroni, A., Pirovano, G., Gilardoni, S., Lonati, G., Colombi, C., Gianelle, V., Paglione, M., Poluzzi, V., Riva, G., and Toppetti, A.: Investigating the role of chemical and physical processes on organic aerosol modelling with CAMx in the Po Valley during a winter episode, Atmospheric Environment, 171, 126–142, https://doi.org/10.1016/j.atmosenv.2017.10.004 , 2017.
Citation: https://doi.org/10.5194/egusphere-2025-167-RC1
- AC2: 'Reply on RC1', Lukáš Bartík, 01 Aug 2025
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-167/egusphere-2025-167-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2025-167-AC2
RC2:
'Comment on egusphere-2025-167', Anonymous Referee #1, 15 May 2025
The manuscript “Modeling organic aerosol over Central Europe: uncertainties linked to different chemical mechanisms, parameterizations, and boundary conditions” by Bartík et al. combines CAMx model simulations with observations from sites located in Czech Republic. They investigate the sensitivity of organic aerosol concentrations simulated with CAMx on assumptions regarding IVOC and SVOC emissions and model boundary conditions. The topic fits within the scope of ACP. The model evaluation and sensitivity analysis presented in the manuscript are valuable work towards improving model representation of organic aerosols. However, the manuscript contains shortcomings in the description of methods and results and these should be addressed before the manuscript can be recommended for publication.

Major comments
1. The main aspects of the model related to the presented analysis should be included for the reader to be able to understand the work. I find that following aspects of the model should be described better:
L121-126: Why were two chemistry mechanisms used? Could the authors please explain here some basics of what kind of chemistry these two mechanisms include and what are the main similarities/differences between them or otherwise explain the use of two mechanisms. For example, do the two mechanisms include essentially different precursors and/or reaction products?

L127-134: Since aerosols are in focus in this study, please describe the basics of aerosol representation in the model, e.g.: What aerosol dynamics processes are included? Is condensation calculated based on equilibrium partitioning or some other way? It is said that ISORROPIA is used to predict composition and physical phase of inorganic aerosols. Are the organic and inorganic aerosols assumed externally mixed?

L159-160: What are these concentrations of the chemical species based on?

Table S1: The SOAP and VBS schemes seem to consider monoterpenes and sesquiterpenes but these species do not seem to be included in the boundary conditions. If that is the case, why is it so? At least the overall picture of what is and what isn't included in the boundary conditions should be described in the main text.

L180-183: Authors should clarify here that the aging is included in CVb for POA and anthropogenic SOA. In appendix on L574 it is stated that "The gas-phase hydroxyl radical reaction rates for the chemical aging of POA and anthropogenic SOA, except those originating from biomass burning, are assumed to be 4×10−11 and 2×10−11 cm3 molecule−1 s−1, respectively. In contrast, the chemical aging of biogenic SOA and SOA originating from biomass burning (both anthropogenic and biogenic) is disabled." There authors should make it clear that this disabling of aging refers to the reference run, not for all runs with VBS scheme.

L185-190: Why were different emissions used for simulations with different SOA mechanisms? Also, were all POA emissions really replaced with POM_SV emissions? Does that mean that all POA was assumed to be semivolatile? Use of POM_SV in the model is one of the main focus points in the manuscript and Table 2 lists parameterizations used for these, but explanation of how these POM_SV are treated in the model is missing. Does all POM_SV have same volatility in the model?

L560-563: What does "more-volatile" and "less-volatile" mean concretely in terms of volatilities?

2. Description of the observational data used for the model evaluation would need more information:
A map showing the locations of the observational sites would be helpful for a reader. This could be a separate map or the locations could be marked in e.g. in the Fig. 1.

L237-238: According to the Table S6, the length of each of these measurement campaigns was only about one month. Please mention that in the main text.

L250-252: Please mention how long time period was considered from these data.

3. Some clarifications or explanations would be needed in the results section:
L367-368: Could you please explain why you have chosen different emission estimates?

L430: What does the "similar conclusion" refer to here? Does it refer to the conclusion in the previous sentence about wind speed inaccuracy in the model being possibly the explanation for the underestimated OC? Why would that affect CVb and CVa most?

Are the concentrations in the map figures (e.g. Fig. 4) surface level concentrations or, e.g., averaged through the vertical layers of the model?

L493-495: “The observed impacts in these simulations are likely linked to changes in other pollutant(s) at the boundaries of the model domain, which influence SOA chemistry.” Could the authors please explain what these other pollutants are, how/why they changed and how that would affect the organic aerosol in the model?

L495-496: “The spatial distributions of the mean seasonal impacts on SOA concentrations in Sp50s50 and Sp0s100 (and similarly in Vp50s50 and Vp0s100) exhibit structures akin to those observed for the mean seasonal impacts on POA concentrations in Sp50s50 and Sp100s0 (and likewise in Vp50s50 and Vp50s000) (Figs. 8c and d) during both seasons.” Could the authors please comment if this is an expected result? Or does this point towards the boundary conditions defining too much the concentrations over the simulated area?

L522-529: Why does the improvement with adding the OA at boundaries differ between the stations and seasons? Also, is it reasonable to assume that the OA at boundaries is only or mostly POA, i.e. do the authors expect that adding the OA as POA at boundaries is getting model results closer to the measured values because it is making the model representation of organic aerosols more accurate, or is the agreement better just because there is a large underestimation in the reference simulation and adding the OA at boundaries as POA happens to increase OA concentration most?

4. This study includes sensitivity analysis on estimates of IVOC and SVOC emissions and OA boundary conditions, as well as comparisons using two different SOA schemes and chemistry schemes. Is it possible to conclude which of the analyzed factors/assumptions, or their uncertainties, are most important from the point of view of modelling OA in Central Europe with this model?

Minor comments
Please mention in the abstract that the evaluation of the model simulations is focused on Czech Republic. Currently the reader finds out quite late in the text that the evaluation is not for wider Central Europe but only for one country.
L21: “have an undoubted environmental footprint” Please check the choice of word. In my understanding environmental footprint term is used for the impact of e.g. organization or products on environment, so for the source of aerosols one could talk about environmental footprint, but not for aerosols themselves. I did not find the term “environmental footprint” from the reference given for this statement, therefore it is not clear what the authors mean by this term.
L136-137: I would suggest moving this essential information from the Appendix to the main text.
Some of the figures, e.g. Figure 2, are missing y-axis labels. I recommend adding y-axis labels.
Tables S3, S4 and S5 contain acronyms for surrogate SOA/POA species in the model. Please add explanation of what these species are.
Citation: https://doi.org/10.5194/egusphere-2025-167-RC2
- AC1: 'Reply on RC2', Lukáš Bartík, 01 Aug 2025
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-167/egusphere-2025-167-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2025-167-AC1

Lukáš Bartík, Peter Huszár, Jan Peiker, Jan Karlický, Ondřej Vlček, and Petr Vodička

Supplement

https://doi.org/10.5194/egusphere-2025-167-supplement

Lukáš Bartík, Peter Huszár, Jan Peiker, Jan Karlický, Ondřej Vlček, and Petr Vodička

Viewed

Total article views: 574 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
495	60	19	574	34	17	31

HTML: 495
PDF: 60
XML: 19
Total: 574
Supplement: 34
BibTeX: 17
EndNote: 31

Views and downloads (calculated since 25 Apr 2025)

Month	HTML	PDF	XML	Total
Apr 2025	64	6	3	73
May 2025	80	21	5	106
Jun 2025	50	6	4	60
Jul 2025	35	5	0	40
Aug 2025	103	17	6	126
Sep 2025	163	5	1	169

Cumulative views and downloads (calculated since 25 Apr 2025)

Month	HTML	PDF	XML	Total
Apr 2025	64	6	3	73
May 2025	80	21	5	106
Jun 2025	50	6	4	60
Jul 2025	35	5	0	40
Aug 2025	103	17	6	126
Sep 2025	163	5	1	169

Viewed (geographical distribution)

Total article views: 596 (including HTML, PDF, and XML) Thereof 596 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Sep 2025

Short summary

This study investigates how to better understand and predict organic aerosols, which are tiny particles in the air that can affect our health and climate. By using advanced computer models, we examined the impact of different emissions and environmental conditions on these aerosols in Central Europe. Our findings show that including specific emissions significantly improved the accuracy of our predictions.


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