the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Jun 2026
Global budgets of atmospheric primary and secondary organic aerosols constrained by full-volatility-range organic emissions
Abstract. Organic aerosol (OA) constitutes a major fraction of tropospheric submicron particulate matter, with primary (POA) and secondary (SOA) components exhibiting different physicochemical properties and health impacts. The POA and SOA budgets are however highly uncertain, and the results from different model studies are confusing because of more or less consideration of the volatility distributions of organic precursor emissions and inconsistent attributions of model tracers in model-observation comparisons. Here we develop an OA simulation framework in GEOS-Chem that resolves the full volatility spectrum of organic precursor emissions from anthropogenic sources and open biomass burning. The model reasonably reproduces the observed OC, POA, and SOA concentrations from comprehensive surface, shipborne, and airborne datasets, providing a consistent global validation. The model simulations suggest greater POA (0.5 Tg) and SOA burdens (2.0 Tg) and potentially stronger and more widespread impacts of the OA components on air quality, health, and radiation than previous estimates, led by both of the emission updates and the revised OA scheme. The simulated global SOA production is about 106 Tg in 2018, 46 % of which is contributed by open biomass burning. The results demonstrate distinct regional variations in the dominant source types and population exposure distributions of POA and SOA, highlighting the needs for SOA mitigation, multi-sector control measures, and clean energy replacements for long-term health-oriented air quality improvements globally. The model results are sensitive to the emissions and wet-deposition parameterization, calling for more measurement constraints on local emission factors and the deposition fluxes of OA and its components.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics.
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.- Preprint
(2078 KB) - Metadata XML
-
Supplement
(1464 KB) - BibTeX
- EndNote
Status: open (until 04 Aug 2026)
- RC1: 'Comment on egusphere-2026-3160', Anonymous Referee #1, 14 Jul 2026 reply
-
RC2: 'Comment on egusphere-2026-3160', Anonymous Referee #2, 20 Jul 2026
reply
This manuscript describes the implementation and evaluation of a full-volatility range OA scheme in the GEOS-Chem model. This represents an important effort to expand the representation of OA sources, especially from anthropogenic and biomass burning S/LVOCs. The paper is relatively straight-forward and clear. The main deficiency is a lack of detail in the text that make it difficult to understand how the results compare to the previous scheme in GEOS-Chem and what contributes to these changes. The authors should be able to easily expand the text and figures to provide these details.
I also note that the paper is heavily dependent on a new anthropogenic emission inventory (cited on p5, line 150-151) – until that paper has been peer-reviewed, this current manuscript should not be accepted for final publication.
Major comments:
- Section 2.2: This section is a bit confusing. Given that the authors are comparing their simulation with the default “complex OA” scheme in GEOS-Chem, it seems they should start with a complete description of this scheme (i.e. biogenic SOA formation is not mentioned in the first paragraph which seems to focus on the default scheme).
- Section 2.3: The section is missing details on the measurement techniques used (other than AMS), the particle size sampling, and the uncertainties. On P7, line 199: was the PMF performed consistently across all AMS/ACSM datasets? What is the uncertainty on these assigned factors? It is also a bit puzzling that the authors made such a comprehensive effort to assemble surface observations, but then selected only two aircraft campaigns. Why not more? Why these two campaigns?
- It is strange to present model comparisons with observations prior to describing the new model budget. I would suggest that the authors invert their ordering in Section 3 (i.e. start with what is now Section 3.3).
- The plots do not provide sufficient comparison with the default scheme (for the model version, year and emissions that the authors have simulated). For example, the statements on p10 line 294-295 and p11 line 346 are not currently supported. Observational comparison should show default and new (Figures 3, 4, 5, 6). As should all budget plots (Figure 9 and 10).
- The plots do not sufficiently detail which sectors contribute to the simulated changes. The manuscript should include plots comparing the default POA and SOA from biogenic, fire, and anthropogenic sources, respectively. In particular, while the primary advance of the scheme is the focus on anthropogenic and biomass burning full-volatility sources, the authors seem to have made substantial changes to the biogenic SOA formation. These are not quantified or discussed in the text (or even mentioned in the abstract).
- How much of the fire source is POA vs SOA? P6 suggests that the emission of IVOC and S/LVOC from fires are large (exceeding the anthropogenic emissions). Can the authors comment on how this much SOA from fires is consistent with previous field studies? (e.g. Cubison et al., 2011)
- The fire source would be highly variable year-to-year (and thus statement on p7, line 207 is false). Given that the study appears to find a large fire source, it’s not clear that one year of simulation is sufficient to characterize a contribution from fires. The authors could either simulate additional years (5-10 years?) to provide some variability estimate, or clearly state this uncertainty up-front on their single-year values (in abstract and throughout).
Additional comments:
- P2, line 56: define C* upon first usage
- P4, line 127: provide reference for all loss correction factors
- P5, line 146-147: is fragmentation not considered? If not, might the scheme overestimate the oxidation products of S/LVOCs?
- P6, lines 173, 190: These figures (Figure S3 and S4) are new results of this scheme and should be shown in the main text, not the SI.
- Figure 1: missing a description of circles vs triangles. Also unclear if oxidation arrows are pointing to specific circles or groups (e.g. the kOH, SVOCs arrow is placed next to the gas phase C*=1 bin). Where do emissions of BVOCs fit on this diagram?
- Figure 2: only shows surface. Perhaps the authors could also show a cross-section like in Figure 7d-f in the SI?
- P8, line 232: The figures are shown in absolute concentrations, so these percentages do not match the information provided. I suggest if the authors want to cite percentages that they provide percent difference plots in the SI.
- Figure 5e-h: missing stats on panels
- Figure 8: missing units on the cumulative probability distribution plots
Citation: https://doi.org/10.5194/egusphere-2026-3160-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 57 | 52 | 12 | 121 | 31 | 7 | 10 |
- HTML: 57
- PDF: 52
- XML: 12
- Total: 121
- Supplement: 31
- BibTeX: 7
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper investigates the impact of full-volatility-range organic emissions on the global budgets of organic aerosols using the GEOS-Chem atmospheric chemistry transport model. The authors highlight the importance of incorporating a wide volatility spectrum of organic precursor emissions from anthropogenic sources and open biomass burning on global SOA production. This approach leads to greater POA and SOA burdens compared to the previous model configuration. Overall, the proposed developments appear to improve the model's agreement with a range of observational datasets used for the present study, compared to the previous organic aerosol representation.
General comments:
The manuscript provides a detailed description of the emissions used for this study. I acknowledge the authors' effort in presenting all of the different sensitivity simulations; however, the aim of this strategy is not clearly explained until the end of the manuscript, which makes it challenging for the reader. The discussion of the results is well-organized, though some repetitions exist at the beginning of certain sections. Aside from these minor concerns, the discussion of the modeling methods could have been more explicit in some cases to better justify the conclusions of this study, e.g., regarding whether the assumptions used are driven by observational/laboratory constraints, computational/modeling limitations, or both. Therefore, I recommend a revision to address such issues before the submission is accepted.
Major comments:
The authors present their new scheme as a "full-volatility-range" representation. However, it appears to be more of a hybrid volatility-based scheme that combines traditional VBS theory and empirical SOA yield parameterizations. The latter is also supported by the authors' statement in the Conclusions that "the SOA scheme remains a volatility-lumped and yield-based approach." Stating this explicitly in the Methods, rather than in the Conclusion of the manuscript, would improve the transparency of the methodology section and help the reader better understand these developments.
The authors also state in the Methods that "Under low-NOx conditions, the IVOC-SOA is treated as nonvolatile mass with a constant SOA yield of 73%." As far as I understand, this suggests that a constant yield is applied for every IVOC category in this case, which is a strong assumption. IVOC oxidation is parameterized completely differently between high- and low-NOx conditions; i.e., under high-NOx conditions, oxidation products are distributed among separate volatility bins, while under low-NOx conditions, SOA-IVOCs are parameterized as nonvolatile with a fixed yield. It would be very useful for the reader if the authors discussed more explicitly why they use different parameterizations under high- and low-NOx conditions and what the added value of this assumption is compared to other modeling studies. I would also propose further discussing the implications for SOA partitioning/evaporation under different NOx conditions in the model (and what thresholds are used).
Minor comments:
Abstract: The transition to "population exposure" and "air quality improvements" seems quite abrupt, as this is not the focus of this study. I suggest moving this sentence to the conclusion/discussion section, rather than the abstract, which should focus on the actual developments the paper presents.
Lines 61-62: It would be helpful to briefly refer to why an overestimation appears for the nonvolatile POA assumption.
Section 1: An outlook paragraph presenting what each of the following sections discusses will be helpful to the reader.
Section 2.2: A table presenting the global annual emissions of S/LVOCs and IVOCs from anthropogenic sources and biomass burning sources used for each scheme (i.e., the simple and complex OA schemes, etc.) would be useful.
Lines 129-133: Here, the authors refer to heterogeneous reactions for IEPOX and glyoxal/methylglyoxal, but they also include a sentence about neglecting isoprene SOA formation under low-NOx conditions. Please explain briefly why low-NOx isoprene SOA formation is neglected and whether this impacts heterogeneous SOA production. If this is not the case, it would be better to move this sentence earlier to improve the flow to avoid confusion.
Line 137: Please provide information about the enthalpy of vaporization used for all components.
Lines 148-149: Please clarify whether POA and SOA are allowed to re-evaporate, particularly under low-NOx conditions.
Lines 204-212: The authors appear to attempt to justify why they compare observations from 2011–2022 with a single model simulation as the reference year. Although this approach includes several assumptions, it is common practice in such modeling studies, focusing mostly on the "climatological" evaluation of the model rather than interannual variability. However, it is unclear from the discussion whether only one year of observation (that of the model simulation) is used, or if the model is run with "average-year" emissions for comparison with observations from several nearby years. Note, however, that since the authors state that biomass burning contributes 46% to total SOA formation, this can vary significantly between years due to climate variability (e.g., El Niño). The earlier statement, therefore, appears more valid for anthropogenic emissions than for biomass burning emissions. Moreover, in Sect. 4, lines 403-404, the authors mention: "Significant changes in SOA production are expected in the future because of the reduction of anthropogenic emissions driven by tightened regional air quality standards (e.g., in China) and the intensified wildfires." This appears somewhat inconsistent with the earlier claim that using a single year would counteract potential biases. Overall, I recommend explicitly stating your methodology for the model evaluation and simply discussing the limitations of using a single reference year.
Lines 355-356: Since this is not a climate study, it is better to move this sentence to the discussion section.
Lines 356-357: The authors imply that roughly 31% of the total SOA burden in the model comes from anthropogenic SOA. How does this finding compare with other studies? Please discuss.
Lines 360-361: Such a sentence should also be placed in the introduction to justify the later choice of an additional sensitivity simulation focusing on the wet removal efficiency. Otherwise, this comes very abruptly in the text.
Section 2.4: Does the model separately track the different species calculated by VBS? How many (additional) species does the model use for the different sensitivity simulations? How much does the computational cost increase depending on the simulation setup?
Section 4: Please briefly discuss which of the assumptions used in the model have the greatest impact on the reported increases in POA and SOA burdens.
Technical Comments:
Line 33: validation → evaluation
Line 45: "Immediately?" Under what ambient atmospheric conditions? Please remove or rewrite by explaining in more detail.
Line 400: Better split into a new paragraph.
Table S1: What do OPOAs stand for? Also add some explanation for other species’ names.
Manuscript: The names of the main simulations, i.e., "Default" and "Base," are somewhat confusing since they both imply “reference” simulations. You can consider adopting more descriptive names/abbreviations or explicitly note their definitions throughout the manuscript (esp. figures/tables) to avoid confusion.