This manuscript presents a thorough evaluation of the performance of the GEOS global atmospheric chemistry model with the GOCART aerosol parameterization. The model output is compared with in situ, airborne measurements of aerosol compositional, microphysical, and optical properties made in the vicinity of the Philippines as part of the CAMP2EX project in 2019.

The output from the model is compared with the observations over project-wide averages, and for some specific flights in biomass burning smoke, and includes both extensive and some intensive parameters. The importance of assimilation of AOD values is examined, as are the effects of a newer convective parameterization scheme. While not comprehensive, this analysis is enough to identify discrepancies in the assumptions regarding aerosol composition, hygroscopicity, and size distribution in the GOCART parameterization. These discrepancies are large enough to be a concern, and indicate that perhaps the GOCART approach, with its one-size-fits-all static parameterization of the properties of different aerosol components, should be replaced with a more physically based and interactive aerosol parameterization.

This manuscript is suitable for publication in ACP following revisions that fall somewhere between major and minor (I've described them as minor). I have one significant comment below that I would definitely like to see addressed prior to publication, followed by a number of minor comments. Despite the number of comments, this is a generally well written paper, and an interesting and enjoyable one to read.

Major Comment:

Much of the analysis relies on comparison of the model output with extinction values at ambient RH. These are referred to as "measurements", but in fact they are derived values that rely implicitly on an underlying assumption regarding aerosol hygroscopicity. As documented in the headers of the data files from the LARGE measurements for this campaign, scattering (which vastly dominates extinction here) was measured at two fixed RH values: ~20% and ~80% RH. The ambient scattering is then calculated by assuming a functional form (the gamma parameterization), which is just a power law fit. In other words, the ambient scattering (hence extinction) is based on fitting a parameter to the two measurement points, then extrapolating to the ambient RH. The error in this approach is probably not large for RH values lying between or close to the two measurement RH values (20% and 80%). However, in the CAMP2EX profiles, the ambient RH exceeds 90% in the upper half of the boundary layer, and this is where the greatest contribution to extinction lies. In this high-RH region, the power-law hygroscopic growth curve pitches very sharply upwards, and small errors in the assumed shape of this growth curve can amplify to very large errors in calculated ambient extinction. So this comparison is not optimal, because the ambient extinction values will have large, and unknown, uncertainties in the high-RH region that dominates AOD.

One of the goals of the comparison of the measured and modeled extinction is to determine if the model hygroscopic growth is consistent with the measurements. To accomplish this, it might make more sense to compare the model to the measurements for the actually measured extinction values at the 20% ("dry") and 80% ("wet") conditions. That way the comparison is not between the modeled values and those derived from the measurements with an assumed shape to the hygroscopic growth curve. Alternatively, you could plot the full f(RH) curve the model would produce (for a given location and time), from 20% RH to ~95% RH, and compare that with the same curve provided by the gamma parameter calculated from the LARGE measurements. This would be informative, and would allow you to compare the extinction at the measured 20% and 80% RH values as well as look at the response at higher RH values. The ratio of these two curves would indicate where there might be relative biases (although whether these lie in the model or in the gamma parameterization assumption might not be so clear).

Minor Comments:

There are a few places where there are typos or where the manuscript could be edited for clarity.

1) Lines 81-85. The FIMS and HSRL2 are introduced by name, but the text doesn't say what they measure. That information appears in lines 100-104. I was trying to figure out what a FIMS measured throughout the text between these lines. Suggest moving the latter section up to lines 81-85.

2) Line 117. Change "hydroscopic" to "hygroscopic".

3) Line 123. Define sigma as the geometric standard deviation.

4) Line 130. Change to, "The underlying meteorology from GEOS is used for horizontal and vertical transport and deposition of all of the aerosol species, as well as wind-driven emissions of dust and sea salt."

5) Line 193. Define "lidar ratio".

6) Lines 211 and 223. "Optical array" has a particular meaning in optical design. Suggest changing to "optical properties instruments" or something similar.

7) Line 213. Change to "fine particles that are efficiently sampled".

8) Lines 222-234. Here is where you could expand on the model/measurement comparison of extinction as a function of RH and focus on the actual full frh curve or compare the extinctions at 20% and 80% RH.

9) Line 241. I don't understand this sentence. "Analysis increment"?

10) Line 258. Please add a comma between "region" and "or".

11) Line 273. Please change to, ". . .represented that percentage of organic carbon well; however, it struggled with. . . ."

12) Line 281. In the heading for Table 2, I believe this in not "total aerosol mass" because it excludes BC.

13) Line 299. I believe the values for the extinction measurements are 20% and 80% RH, at least according to the file headers.

14) Line 313. Saying the GEOS values for SSA are "below 0.96" doesn't do justice to the magnitude of the discrepancy, which is quite large. Maybe say they range from "0.9 to 0.96". The co-albedos disagree by a factor of 2-4; this is quite large in the context of direct radiative effects.

15) Line 323-324. It's true that the extinction could be juiced up by making the modal diameter larger and the width wider, but the assumed standard deviation of ~2+ is already quite a bit larger than literature values would support. You might want to lead into the next sentence by saying, "We will examine the in situ measurements to see if such changes could be justified. We begin by looking at . . . ."

16) Line 337. The sentence beginning "The primary peak" is confusing. Please define what you mean by the "primary peak? It's being "shifted toward a larger radius" than what? Are these the measured or GEOS size distributions you're talking about? Please clarify.

17) Line 368. I'm not sure what "mass piling up at the top of the PBL" means. I don't think mass (of air or of aerosol) can "pile up", at least not without increasing air density a lot!

18) Lines 391-395. These sentences are unclear. What "concern" about the size distribution could be "rectified"? Please be specific, e.g., "agreement between modeled and measured extinction at high RH could be improved if the modeled size distribution had a larger modal radius and/or a larger standard deviation. However, these adjustments are not supported by the FIMS size distribution measurements. In fact, the observed mode radius is in excellent agreement. . . ."

19) References. Please ensure that all references are compliant with Copernicus formatting guidelines. For example, some of the references (e.g., Burton et al., 2012) have capitalized titles, while most do not. This is a result of reference management software, which always needs to be thoroughly checked manually.

20) References. I'm not sure Schill et al. is citable--it's an unpublished conference presentation.

21) Figures. The figures are generally very nicely done, clearly labeled. However, I'm not sure of the ability of those with color impairment to read the shaded vertical profiles (e.g., Figs. 3-6). Also, in these figures the horizontal lines indicating HSRL2 MLH and GOES PBLH are difficult to discern; the black and blue colors are quite close. Could one line be made dotted? In Fig. 4  you might need a 3rd line type for clarity.

22) Figures 13, 14. It might be nice to fit a lognormal to the FIMS peaks, then you could directly compare the mode diameter and standard deviation with the model. My eye says that the standard deviation for the measurements is less than the very broad modeled values, but I can't be sure without fitting. I much prefer the aspect ratio of Fig. 14 to that of Fig. 13; could Fig. 13 be made with side-by-side plots, rather than vertically stacked ones? This would make Fig. 13 and 14 look more similar.

The authors evaluate the effects of recent modifications in the GEOS model (version 5.22, 5.25 and GOCART2G) on aerosol-related estimations. To that effect, they use remote sensing and in situ measurements from the recent CAMP²EX airborne field campaign (Philippines, Aug-Oct 2019). Their study focuses on the evaluation of modeled Biomass Burning (BB) aerosol speciated mass and total backscatter, scatter, extinction, single scattering albedo and size distribution as well as modeled relative humidity, temperature, and planetary boundary layer height. This paper is well structured, its results important but needs clarification in many places. It will be worthy of publication once the issues below are properly addressed.

Major comments:

. We recommend the authors combine a few or move some figures to the appendix, especially the ones that are barely analyzed in the text (e.g., Fig. 9).

. The recent presence of nitrate and Brown Carbon (BrC) aerosols in GOCART and GOCART2G needs more references and descriptions. We recommend that the authors include a table of microphysical and optical properties for all the species present in the model.

. The changes applied to the different models in Table 1 should be further described and the authors should focus on explaining the potential effects of these changes on the modeled aerosol microphysics, spatial distribution, optical properties etc.

. A high-level diagram illustrating the different modules in the model as well as the many changes in Table 1 would be helpful. The diagram could also emphasize what this paper has investigated in more detail (e.g., RH).

. Throughout the paper, we recommend a clear discussion of all the error sources in the model.

Detailed Comments:

. Line 14: “serving as cloud condensation nuclei”. As written, it seems that this is the only way BB impacts radiative forcing. We recommend either re-wording or adding direct and semi direct radiative effects as well.

. Line 19: The authors should be clearer on which satellite/ ground-based sensor(s) is(are) used and in which model version.

. Line 24: Why not say “Aerosol extinction within GEOS is a function of the mass of different aerosol species, the ambient relative humidity, the assumed spectral optical properties and particle size distribution per species”.

 . Line 25: “aggressive” is not usually used in that case. Maybe “high” or “overestimated”.

. Line 27: “a mode radius” – does GEOS assume only one size mode for its particle size distribution of OC? Aerosols are usually (at least) bi-modal so this should be discussed.

. Line 31: See comment for line 14.

. Line 37: “smoke and biomass burning aerosol” reads as if these two things were different.

. Line 49: SSA and its link to aerosol light absorption needs to be briefly described; and we recommend writing “… due to different assumptions for aerosol…”.

. Line 51: We recommend simplifying and writing “An additional source of uncertainty would be the biomass burning aerosol emissions”.

. Line 59: This should be “anthropogenic”. The distinction between “white (anthropogenic) and brown (biomass burning, BB) OC” is not clear. Some BB aerosols can be labelled anthropogenic (e.g., prescribed fires) and it’s not clear what the authors mean by “white” OC. This needs more description.

. Line 66: “two moment cloud microphysics” is mentioned abruptly here, with no obvious link to what was written previously or afterwards. If kept in the text, “two moment cloud microphysics” should also be described.

. Line 73: “future” is written twice.

. Line 80: “over the Philippine Sea”.

. Line 82: A table listing the instruments, measurements, size ranges, resolutions and references is recommended here, like Table 1 in Edwards et al., [2021].

. Line 85: Is the AMS instrument operated by the LARGE team during CAMP²EX? It usually isn’t the case so that would be new. It does not seem to be the case in Stahl et al. [2021]. Please check and clarify.

. Line 92: “50% uncertainty” needs a reference or “[personal communication from …]” and if true, this should be better explained. Again, it was not mentioned in Stahl et al. [2021].

. Line 93: “inconsistencies between measured mass concentrations and optical properties”. This is not clear. It should be explained/ illustrated/referenced.

. Line 117: GOCART (legacy) usually uses [e.g., Chin et al., 2002, 2009, 2014] as references. Also, GOCART was already introduced in line 62 (with the right reference i.e., Chin et al., 2002).

. Line 117: should read “hydrophilic and … hydrophobic” or “hygroscopic and… hydrophobic”.

. line 118: “size bins per model species” is not clear here. It should be added that nitrate was not originally included in GOCART [e.g., Chin et al., 2002, 2009, 2014] but was developed more recently in the GMI model [Bian et al., 2017], which is an off-line chemistry model. It was then later implemented in the GEOS/GOCART model, followed by GOCART2G. As for BrC, the authors should specify that is it present in both GOCART and GOCART2G? If there is a specific reference describing GOCART2G, the authors should mention it. If not (Colarco et al., 2017 might be the only “indirect” reference), they should describe what is meant by BrC chemically and optically.

. Line 121: “The optics look up tables for each aerosol species are the same as described by Colarco et al. (2017)”. We can’t seem to find these look up tables in Colarco et al. (2017). We recommend that this paper adds a table describing these optical properties.

. Line 125: it would be better to write “bias corrected AOD observations from the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard Terra and Aqua are assimilated in all the models of Table 1 except “No GAAS”. Is MODIS the only sensor that is assimilated in all models and is AERONET only assimilated in GEOS 5.22 and not the rest?

. Line 131: ‘… as well as deposition and wind-driven emissions of dust and sea salt”. Why not wind-driven emissions of BB or urban pollution?

. Line 137: This sentence is not clear and should be re-written.

. Line 155: What is meant by “The diurnal evolution of … the lower troposphere”?

. Line 157: “Relative humidity was selected for this evaluation since it is used in the optics lookup tables for aerosols”. We recommend specifying “in the model” here. This is where more information on the model and its different modules would be helpful in the introduction. Why not plot modeled PBLH and measured MLH on Figure 1.

. Line 165: The angstrom exponent and its link to aerosol size should be explained

. Line 169: “or localized urban emissions not in the CEDS emissions dataset” This is not clear. The first part of the sentence is about AERONET measurements (and sources of errors in the Level 1.5 data) and the second part seems to be about the model.

. Line 176: How are the MLH vs PBLH computed? Wouldn’t we not expect MLH and PBLH to be the same? This paper is focused on evaluating modeled BB aerosol composition, microphysics, and optical properties. The authors should explain why they are also evaluating the simulated PBLH (e.g., impact on modeled aerosol vertical distribution).

. Line 177: Some figures show the three model versions of Table 1 and some do not. We recommend consistency. We recommend adding “(not shown here)” after “indistinguishable”

. Line 179: “This trend … ”.

. Line 182: We recommend describing the link between the spectral dependance of the aerosol backscatter and the size of particles.

. Line 185: quantify “slight improvement”

. Line 189: “there is a larger impact of the change in relative humidity between GEOS 5.22 and GEOS 5.25 than the aerosol updates implemented in GOCART2G” This sentence is not clear and illustrated.

. Line 191: “compares well”

. Line 192: “is located too high due to the height of the boundary layer in the model.” The PBL is defined as a strong gradient in the aerosol scattering profile. This feels like circular reasoning.

. Line 193: Why not show lidar ratios as well?

. Line 197: “based on the region of interest” makes it sound as if HSRL aerosol type is based on the location, but it is not.

. Line 199: “the GEOS aerosol speciation for each HSRL2 derived aerosol type” – as the authors compare GEOS aerosol speciation for different aerosol types, why not analyze the results and evaluate whether the composition agrees well with the types in a quantitative way? Are the GEOS species “correctly translating” the HSRL aerosol types? This would be similar to the work of Kacenelenbogen et al. [2022]

. Line 200: “drastic decrease in the sample size above 2 km” should be quantified. And “There is also a focus placed on the GOCART2G” can be replaced by “We focus on…”.

. Line 202: Figure 5e – The authors should explain the differences between “fresh smoke” and “smoke” from HSRL

. Line 203: “smallest sample size of the aerosol types” should be quantified.

. Line 204: “This could indicate deficiencies in the model’s optical properties for smoke, the transport, meaning the smoke plume is not in the correct location without the data assimilation, or uncertainties in the emissions” could be changed to “This could indicate deficiencies in the model’s smoke optical properties and transport (i.e., the smoke plume is not in the correct location without the data assimilation), or uncertainties in the BB emissions.”

. Line 207: “HSRL2 can have difficulty distinguishing between the two” This sentence needs more information and a reference.

. Line 211: “LARGE optical array is in situ and can provide a direct comparison between extinction and aerosol composition” this might be misleading if the AMS instrument is not operated by the LARGE group. Also, the extinction is for the total aerosol and composition is per species. The authors should clarify. The authors should also describe how these in situ measurements are selected i.e., airborne vertical profiles, constant altitude legs etc.

. Line 213: “representative of fine particles that are efficiently sampled by the inlet”; “were subsampled such that only particles with an aerodynamic diameter less than 5 μm were included…”

. Line 234: “total aerosol mass concentration overestimated in GEOS”. We recommend the authors show and evaluate the modeled total aerosol mass concentration profile.

. Line 237: “an additional buildup of black carbon”

. Line 241: “This results in positive values for the analysis increment for black carbon mass” This sentence is not clear and should be re-written.

. Line 245: “Since brown carbon originates as a portion of what was organic carbon prior to GOCART2G, it is being included as organic carbon in the figure.” This sentence is also not clear and should be re-written.

. Line 247: “In general, there is not enough aerosol for these two species in the model.” Could be replaced by “In general, these two aerosol species are underestimated in the model.”

. Line 255: Figure 7d

. Line 265: Figure 9 seems to have minimal value in this paper and could be replaced by 1-2 sentences in the text. The relative humidity plot is not discussed, and the lowest altitude is not quantified on Fig. 9.

. Line 268: “… optics look up tables are unchanged” -- This should be emphasized in Table 1 and in the description of Table 1.

. Line 269: “the aerosol mass concentration and relative humidity have the potential to differ in each of the model simulations” -- This should also clearly be stated in Table 1 and its description.

. Line 270: “the relationship between the two and the optical properties remain the same.” This sentence is not clear.

. Line 274: “the ratio of sea salt” – the authors should consider replacing “ratio” by “fraction”

. Line 277: “given the preference for coarse mode sea salt in GEOS (Bian et al., 2019)”. This sentence is not clear.

. Line 279: We recommend adding “in the model” after “The deficiency in sulphate and nitrate”

. Table 2: Instead of “LARGE Observations”, the authors should write the name of the instrument e.g., “Aerodyne HR-ToF-AMS”; and the four digit in “0.0677:1” do not seem necessary.

. Line 286: “… are displayed in Figure 10”

. Line 287: “… which is always positive and representative of dry conditions”

. Line 289: “It is evident that GEOS needs a large bias in the mass concentration of organic carbon to accurately represent dry extinction.” This sentence is not clear. The authors should rephrase. Also, they should consider quantifying the bias by providing an envelope around the 1:1 line and a percentage of points within this envelope.

. Line 310: The authors should provide the ranges of SSA values in Pistone et al. [2019]

. Line 312: “Nearly all observations have…”

. Line 324: “section 3.2”

. Line 326: The authors should describe the “chemical influence flag” and which gases it uses.

. Line 330: The authors should explain why the size distribution is bi-modal for FIMS and unimodal per species in GEOS

. Line 357: “… parameterizations as well as…”

. Line 384: “there is evidence of this in the FIMS observations from CAMP2Ex” – this sentence needs more information.

References:

Bian, Huisheng, et al. "Investigation of global particulate nitrate from the AeroCom phase III experiment." Atmospheric Chemistry and Physics 17.21 (2017): 12911-12940. https://doi.org/10.5194/acp-17-12911-2017

Chin, Mian, et al. "Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite and Sun photometer measurements." Journal of the atmospheric sciences 59.3 (2002): 461-483.

Chin, Mian, et al. "Light absorption by pollution, dust, and biomass burning aerosols: a global model study and evaluation with AERONET measurements." Annales Geophysicae. Vol. 27. No. 9. Copernicus GmbH, 2009

Chin, Mian, et al. "Multi-decadal aerosol variations from 1980 to 2009: a perspective from observations and a global model." Atmospheric Chemistry and Physics 14.7 (2014): 3657-3690

Edwards, Eva-Lou, et al. "Assessment of NAAPS-RA performance in Maritime Southeast Asia during CAMP 2 Ex." Atmospheric Chemistry and Physics Discussions (2021): 1-46. https://acp.copernicus.org/preprints/acp-2021-870/acp-2021-870.pdf

Kacenelenbogen, Meloë SF, et al. "Identifying chemical aerosol signatures using optical suborbital observations: how much can optical properties tell us about aerosol composition?." Atmospheric Chemistry and Physics 22.6 (2022): 3713-3742.

Stahl, Connor, et al. "Total organic carbon and the contribution from speciated organics in cloud water: airborne data analysis from the CAMP 2 Ex field campaign." Atmospheric Chemistry and Physics 21.18 (2021): 14109-14129, https://acp.copernicus.org/articles/21/14109/2021/