This manuscript updates the treatment of oxidation products of VOCs in a general circulation model EMAC, by integrating a comprehensive VOC oxidation chemistry model Mainz Organic Mechanism (MOM) and calculating the gas/particle partitioning of the oxidation products with the ORACLE model. The treatment of organic oxidation products here is different from many previous studies in that the identitities of the oxidation products are explicitly tracked, whereas many previous studies lump the oxidation products by volatility with the VBS framework. This allows direct comparison between modeled and measured atmospheric concentrations of species of interest (e.g., CH3OH, HCOOH etc). The model predictions are then evaluated against atmospheric measurements from a wide range of sources (ground stations, aircraft, satellite and assimilated model results) for a number of important atmospheric species, including, most importantly, PM2.5, organic aerosol, and aerosol optical depth (AOD). The model generally caputres PM2.5, AOD and ground organic aerosol, but significantly underestimates organic aerosol for the free troposphere. Nevertheless, this study still represents an important step towards a complete VOC oxidation mechanism in atmospheric models to simulate organic aerosol loading, compositition, and other properties (e.g., optical or hygroscopicity). This reviewer recommends publication. Some comments for the manuscript are:



1. line 16: use "secondary organic aerosols" instead of just "aerosols", because it is already said that the aerosols discussed here are from VOCs.

2. line 37: Can the authors elaborate in the manuscript why explicit chemical identities would be useful in atmospheric simulations here?

3. For the OA treatment (starting from line 103), has the MOM + ORACLE approach been tested against chamber SOA experiments of, for example, alpha-pinene, to see if the model correctly captures SOA mass production in the experiments? This reviewer trusts that the two submodels have been well tested in previous studies but just wonders if this can be done for closure.

4. For section 4.2.2, the authors do not seem to offer a potential explanation for the underestimation. Can this be added to the text?

5. In the outlook section, the authors discuss potential improvements to the model. Since SOA lifecycle in the atmosphere can also be affected by aerosol phase state, heterogeneous reaction with oxidant, etc., which are still highly uncertain, could the authors include these processes in the discussion as well, and talk about how they can be captured in the MOM+ORACLE framework?

This paper provides a description of a new configuration of the global EMAC chemistry-climate model using the detailed Mainz Organic Mechanism which includes 700 compounds and 2000 reactions.  The model configuration is briefly described and the comparison of the simulation to satellite, aircraft and surface observations is presented for several VOCs, OH, AOD and aerosol composition.  The differences between the model and observations are suitably quantified and the discussion of agreement and differences is reasonable.  

However, I feel 2 major aspects are lacking from this work. 

1. No comparisons are made to previous model versions. The authors have made no arguments for which conditions such a complex mechanism is needed. This work ought to show what difference, and particularly improvement, is achieved with this complex chemistry scheme over standard, more reduced schemes. 

2. The VOC measurements used for evaluation of the model are really not adequate and numerous additional observations could have been used to evaluate the VOCs.  The Emmons compilation of aircraft observations only includes data through 2000, and the quality and quantity of in-situ (aircraft) measurements of VOCs have increased dramatically since then.  For example, the ATom missions provide global coverage of the compounds presented here (HCs, CH3CHO, CH3COCH3, etc.).  Numerous other aircraft campaigns led by NASA, NOAA and NSF in the U.S. provide measurements over polluted regions (e.g., INTEX-NA, DISCOVER-AQ, SEAC4RS, SENEX, KORUS-AQ, FIREX-AQ, etc., etc.).  These datasets are all freely available and in standard, similar formats which making them fairly easy to use, though it does require a bit more effort than taking the single profiles provided by the Emmons climatology. 

The IASI retrievals are used extensively, but much more should be included about their accuracy and the results of published validation results. Perhaps they have much greater uncertainty at high latitudes, or remote regions - this should be discussed here.

Thus, I recommend that these 2 aspects be addressed before the paper is accepted for publication. 

The authors presented a comprehensive evaluation of an updated and expanded EMAC model. The implementation of explicit mechanisms and the gas-particle condensation module to the chemistry GCM can be very useful for studying organic vapors and their chemistry in the atmosphere. The model evaluation was made on OH, CO, some key VOCs, AOD and the chemical components of PM2.5. Significant biases present in OVOC simulations, whilst the model performance for aerosol composition is similar to that of previous model configurations. The reasons that lead to the model biases were explored but I feel there are still somewhat insufficient which may be improved with more discussions. Overall I think this is a relevant, important paper and recommend it to be accepted after addressing the following comments.

Specific comments:

(1) Line 81: Are those emissions reasonable compared with other global model inputs or estimates? It is better to add a few sentences to discuss it. It looks like the work did not apply regional emission inventories (e.g., NEI for US, EMEP for Europe, MIX for East Asia or MEIC for China). Would this affect the simulations in anthropogenic-influenced regions? In particular, for aerosol composition, I think the regional inventories may matter significantly.

(2) Section 3.2.5: I think the authors only used the EANET data for 2010. Please specify.

(3) Line 238-239: What kind of biogenic CO precursors?

(4) Line 280-281: How high? I am surprised that temperature difference can be significant between the model parameters and the observations. If the overestimation is due to excess of biogenic VOCs, it is better to compare the surface concentrations of isoprene etc. Some measurements are available in those tropical areas and the data are published (e.g., from AMAZE-08).

(5) Consistent model underestimation present in mid- and high latitudes of Northern Hemisphere for CH3OH, CH3COCH3, HCOOH, CH3COOH, and PAN. The authors made some explanations. For example, Biogenic or biomass burning VOCs emissions might be too low during summertime in those regions. In-cloud chemistry may happen. In what regions, such in-cloud chemistry can be a significant source (e.g., HCOOH)? This model bias is too consistent, which made me wonder how this is related to underrepresented emission inventories of VOCs in EDGAR (e.g., contributed by outflow of polluted areas). If biomass burning is not well represented, what does the model-satellite comparison look like for different seasons? Or if it is biogenic, would the simulation become better in winter? I think the current discussion about the potential bias is a bit too general.

(6) Line 389-390: What are those concentrations? The magnitudes seem not matching with annual mean surface concentrations of OH.

(7) Section 4.3: What about the model performance of surface OH in different regions? This would affect VOC chemistry and SOA formation significantly. Surface OH concentrations can be quite high in polluted areas (e.g., in China). A recent paper pointed the model underestimation of OH because of the lack of HONO source [Miao et al., 2021]. Would that be a problem in the updated model presented here?

(8) Section 4.5: Is the model performance lower in winter than in summer? Residential coal burning in China seems to be a big source of PM2.5 associated with large uncertainty in the inventory. Would that affect the comparison in Asia?

(9) Line 470-473: What do the model values in higher vertical layers compare with the observations?

(10) Line 487-488: Nitrate overestimation seems a general problem in chemical transport models. A main reason could be the inappropriate deposition parameters [Miao et al., 2020; Zhai et al., 2021]. But more pronounced bias in winter than in summer is inconsistent with the other studies.

(11) Section 4.6.5: It would be good to make some connections for model biases on species. For example, how would the significant biases for the oxygenated organics discussed in the paper affect the OA simulations? So does OH as I mentioned in #7.

(12) Line 553- 562: The observations in Asia used here for aerosol composition are quite limited. Studies have suggested the emission inventories for semivolatile and intermediate volatility organic compounds are insufficient and lead to the majority of the model biases in simulating OA [Miao et al., 2021; Zhao et al., 2016] in regions that are influenced significantly by anthropogenic emissions. In Europe, residential wood burning also contributes a lot to the primary OA in winter. Would that be a possible bias? For the rural areas, fragmentation of OA is not included in the model, which may contribute to the overestimation over North America. OA includes both primary OA and SOA. It might be more clear if the two cases are separated discussed.

Technical remarks:

Line 15: Update “Volatile Organic Compounds” to be “Volatile organic compounds”.

Line 29: Please provide the species name for “HNCO”.

Line 217: C3H6 instead of C4H6?

Definition of abbreviations such as MM, OM, MSTD and OSTD etc. are repeated for multiple times in the paper (e.g., page 25-26).

Please be consistent when using the term of “column densities”. In many places, “column” is used to stand for column densities or column concentrations (Line 238, 274, and so on).

Line 254: Remove “tends to”.

Line 266-268: Revise what are in the parentheses.

Line 389: “hydroxyl” means “OH”? Also in Line 396, to be consistent, OH should be used instead of “hydroxyl radical”.

Line 466-467: This sentence needs to be revised. Are model values low or observations low?

Figure 16: Colors (light vs dark) are not clear.

Line 543: Do you mean “Table 5” instead of “Fig. 5”?

References:

Miao, R., Q. Chen, M. Shrivastava, Y. Chen, L. Zhang, J. Hu, Y. Zheng, and K. Liao (2021), Process-based and observation-constrained SOA simulations in China: the role of semivolatile and intermediate-volatility organic compounds and OH levels, Atmos. Chem. Phys., 21(21), 16183-16201, doi:10.5194/acp-21-16183-2021.

Miao, R., et al. (2020), Model bias in simulating major chemical components of PM_2.5 in China, Atmos. Chem. Phys., 20(20), 12265-12284, doi:10.5194/acp-20-12265-2020.

Zhai, S. X., et al. (2021), Control of particulate nitrate air pollution in China, Nature Geoscience, 14(6), 389-+, doi:10.1038/s41561-021-00726-z.

Zhao, B., S. Wang, N. M. Donahue, S. H. Jathar, X. Huang, W. Wu, J. Hao, and A. L. Robinson (2016), Quantifying the effect of organic aerosol aging and intermediate-volatility emissions on regional-scale aerosol pollution in China, Sci Rep, 6, doi:10.1038/srep28815.