the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Aug 2025
Evaluating Simulations of Organic Aerosol Volatility and Degree of Oxygenation in Eastern China
Abstract. Volatility and oxygen-to-carbon (O/C) molar ratios are critical properties of organic aerosols (OA), influencing their viscosity, hygroscopicity, and light absorption thereby resulting in impacts on air quality and climate. While atmospheric models often track these properties to simulate OA evolution, their performance remains insufficiently evaluated. This study assessed OA volatility and O/C simulations by comparing CMAQ model outputs using official AERO7i and community-contributed two-dimensional volatility basis set (2D-VBS) schemes, against two field measurements in eastern China. Apart from baseline modelling, two additional simulations using AERO7i incrementally incorporated low-volatility/semi-volatile/intermediate-volatility organic compound (L/S/IVOC) emissions and enhanced anthropogenic secondary organic aerosol (SOA) yields. An optimized 2D-VBS simulation further constrained O/C ratios of primary organic aerosol (POA) emissions using observations. The results showed that OA mass concentrations were underestimated by 24 % in 2D-VBS and 27–34 % with updated AERO7i, likely due to underrepresented vehicular POA emissions and nighttime SOA formation. All simulations captured the substantial contribution of low-volatility products (C* <0.1 μg m⁻3) but failed to reproduce the detailed volatility distributions within this range. Simulated O/C ratios were biased low in aged air masses (notably with 2D-VBS) and slightly overestimated in areas with more local emissions using updated AERO7i. Misrepresentations of OA volatility primarily led to biases in viscosity predictions, while the hygroscopicity parameter played a more important role. These findings highlight the need to better constrain OA volatility and O/C in models to improve projections of OA air quality and climate impacts.
- Preprint
(3829 KB) - Metadata XML
-
Supplement
(7459 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-2879', Anonymous Referee #1, 21 Aug 2025
-
RC2: 'Comment on egusphere-2025-2879', Anonymous Referee #3, 03 Nov 2025
This study presents a comprehensive evaluation of the CMAQ model's performance in simulating OA mass, volatility, and O/C ratios at two observation sites in eastern China. The study compares model outputs using different OA schemes against field measurements, providing valuable insights into the model's strengths and limitations. Overall, the methodology employed in this study is solid and relevant, contributing to the understanding of OA properties and their impacts on air quality and climate. Here are my specific comments.
Major comments
- A detailed comparison of the similarities and differences between the two observation sites (DY and GZ) would strengthen the study. For example, the authors state that DY is influenced by aged air masses while GZ is more impacted by local emissions. However, there is no evidence presented in the study to support this statement. How did the authors come to this conclusion? Additionally, comparing the absolute magnitudes and relative contributions of POA, S/IVOCs, and VOCs from different emission sources at these two sites would help explain the observed and simulated differences between the two sites.
- Did the authors include L/S/IVOC emissions from the FINN in their simulations that include L/S/IVOC emissions? Relevant information is not available in Table S3.
- Model performance evaluation. Why was SO2 not evaluated? The correlation of O3, PM2.5, and NO2 is particularly low at the GZ site, and NO2 and PM2.5 are substantially underestimated. What are the potential causes for this? Was it due to overestimated wind speed at GZ or bias related to the wind direction?
- The authors provided several potential explanations for the underestimated SOA concentrations, such as insufficient formation of organic nitrates and underprediction of aqueous-phase formation pathways. However, POA is also underestimated in all cases. How much would the underestimation of POA contribute to the SOA underestimation, given the gas-particle partitioning relationship? Besides emission uncertainties, could biases in meteorological simulations cause underestimated OA, SOA, and POA?
- Table S10. How were the metrics calculated in Table S10? Were they based on hourly pairs of simulation results and observations or were they based on the averaged diurnal data presented in Figure 1?
- Lines 273-275: The authors mentioned that BBOA and COA were excluded from the observations when comparing to the simulation results. Did the authors also exclude the simulated POA from biomass burning and cooking? If so, how did the authors exclude simulated POA from these two sources? By performing brute force simulations or using another method? If not, then the POA is even more underestimated.
- After comparing mass concentration, O/C ratio, Tg, and viscosity, can the authors make specific conclusions or recommendations on parameters or configurations that could improve model performance?
Minor comments
- There is a reference error in the title of Figure S4. It states "Chen et al. (2024a); (Chen et al., 2024b) and Feng et al. (2023)." However, the legend mentions Chen et al. and Zheng et al.
- Can the authors explain why the L/S/IVOC emissions agree well with Chen et al. (2024) but not Zheng et al. (2023)? Is it due to different methods used?
- Line 146-147: Please elaborate on “However, the source contributions and volatility distributions of L/S/IVOCs were slightly different.”
- Figure S13: Add distributions based on observations.
- Line 91: CMAQ should be spelled out in the first place mentioned.
Citation: https://doi.org/10.5194/egusphere-2025-2879-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 952 | 64 | 25 | 1,041 | 47 | 27 | 28 |
- HTML: 952
- PDF: 64
- XML: 25
- Total: 1,041
- Supplement: 47
- BibTeX: 27
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Overview
This study evaluates the performance of CMAQ model with SAPRC07-aero7 mechanism in simulating organic aerosol (OA) mass concentrations, volatility distributions, and O/C ratios with two field measurements in eastern China. The authors conducted several sensitivity simulations, including adding emissions of IVOC and updating SOA formation mechanisms. The simulations were well designed but still difficult to capture the properties of measured OA, like the mass concentrations, volatility distributions, and O/C ratios. These limitations affected predictions of OA physicochemical properties such as glass transition temperature (Tg), viscosity, and hygroscopicity. The findings highlight the need for better constraints to improve model accuracy in simulating air quality and OA properties.
The manuscript is well-structured, and its conclusions are insightful, offering valuable guidance for future research. It is recommended for publication with some revisions.
Major comments:
The authors carefully explored potential reasons for the underestimation of SOA at two sites in China. However, underpredicted POA emissions (including POC and PNCOM) could significantly affect SOA partitioning and contribute to the observed biases. To evaluate this, I recommend an additional sensitivity simulation in which POA emissions are increased to match observational levels—particularly at the GZ site—to examine whether SOA predictions improve as a result. Furthermore, comparing the diurnal variations of observed and estimated emissions may help identify missing sources and better constrain emission uncertainties. While this would be a sensitivity test, inaccuracies in emission inventories are a well-known issue affecting model performance. In reality, the underprediction of SOA is likely due to a combination of underestimated POA emissions and missing or incomplete SOA formation pathways.
Minor comments: