In this paper, formaldehyde (HCHO) and nitrogen dioxide (NO2) vertical profiles are retrieved from MAX-DOAS observations at three sites in Asia from January 2017 through December 2018. The three sites are Phimai in Thailand, Pantnagar in India, and Chiba in Japan. They correspond to rural, semi-urban, and urban conditions, respectively. The NO2 and HCHO concentrations in the 0-4km altitude range show consistent seasonal variations throughout the investigated period, which are interpreted in terms of dry and wet seasons and, in the case of Phimai and Pantnagar, biomass burning episodes. The HCHO to NO2 concentration ratios together with MAX-DOAS ozone retrieval results are also used to infer the ozone sensitivity to NOx and VOCs at the three sites. It is found that reasonable estimates of transition regions between the NOX-limited and VOC-limited ozone production regimes can be derived when the NO2-HCHO chemical feedback is accounted for.

In the second part of the study, the MAX-DOAS observations of NO2 and HCHO are used to assess the CHASER global CTM at the three sites. CHASER shows reasonably good performances in reproducing the abundances of both trace gases in Phimai and Pantnagar but not in Chiba. Comparison results are interpreted in terms of model resolution, emission inventories, and contributions of the different emission sources.

This study fits with the scope of ACP. However, there are a lot of aspects of the work that should be further clarified and/or discussed prior to final publication. Those aspects are detailed below. Moreover, as already raised during the quick review, the overall presentation quality is questionable, largely due to the poor English language used throughout the manuscript but also to repeated errors in the axes and title labellings of several figures. This should be improved in the revised version of the paper.

Important specific comments:

*Lines 225-235: The VCD retrieval is based on several assumptions that are poorly discussed and justified. For instance, did you check that the dependence of the Abox profiles on the trace gas concentration profiles is indeed minimal? Did you test other a priori VCD values? How valid is assuming an Angstrom exponent value of 1.00?

*Lines 234-236: You should describe how these averaging kernels are calculated. Looking at Figure 3, HCHO and NO2 VCD averaging kernels seem to be close to unity but it is not the case for f1 averaging kernels, and especially f2 and f3 averaging kernels which are close zero. Does it mean that you can basically retrieve only the VCDs from your measurements and that for f1, f2, and f3 the retrieval essentially reproduces the a priori? Also, are similar averaging kernels obtained for the two other stations? These points should be further discussed in the revised manuscript.

*Line 284: Anthropogenic emissions used in the CHASER model were based on the HTAP_v2.2 2008 inventory. Why didn’t you use more recent inventories like the REAS v3 one (see https://acp.copernicus.org/articles/20/12761/2020/acp-20-12761-2020.pdf)? How can it affect the results and conclusions of your study, especially for Pantnagar and Chiba?

*Figure 9: How do you explain such a large effect when model profiles are smoothed with the MAX-DOAS averaging kernels, especially in the altitude range (0-2km) where the MAX-DOAS retrievals have a maximum of sensitivity.

*Figure 10 and related discussion: What would be also interesting to show are model profiles at both 2.8°x2.8° and 1.4°x1.4° resolution smoothed by the MAX-DOAS AVK. I think only this comparison allows to discuss quantitatively the effect of the model resolution on the CHASER/MAX-DOAS agreement. Since the 2.8°x2.8° and 1.4°x1.4° model profiles have a significantly different shape, we can expect a different impact when those profiles are smoothed with the AVKs.

*Section 3.2.3: Given the coarse horizontal resolution of the CHASER model (2.8°x2.8°), how valid is the assessment of NO2 and HCHO from this model for the Pantnagar station which is located in a region (Himalayan foothills) with highly varying topography? I would suggest to remove Pantnagar from the model evaluation since the topography is not properly taken into account in your analysis.

Minor comments:

*Line 125: You should indicate here which types of industries are located in the Pantnagar region.

*Lines 151-152: The use of the 70°EL instead of the 90°EL for the reference spectra should be better justified. How the use of 70°EL (instead of 90°EL) can minimize variations in the measured signals. Also what do you mean by ‘variations in the measured signals’?

*Line 196: You should give here the AEC value at 100km you used, as well as the scaling height of your exponentially decreasing a priori profile.

*Lines 201-203: The parameterization of Irie et al. (2008a) does not provide information on the vertical resolution and measurement sensitivity. Then it is said that ‘The retrievals and simulations conducted by other groups for similar geometries (i.e., Frieß et al., 2006) are used to overcome such limitations’. I don’t understand this latter sentence. Do you mean that you used previous studies based on the optimal estimation method to estimate the vertical resolution and sensitivity of your own parameterized retrieval? Could you please clarify?

*Lines 206-208: You should describe in a table the settings (pressure and temperature profiles, wavelength, surface albedo, etc) you used for the calculation of your box air mass factors LUT.

*Lines 244-245:  For the estimation of the systematic errors, uncertainties of 30% and 50% on the retrieved AOD are assumed. Where these uncertainty values come from?

*Lines 246-247: Did you try to estimate the presence of an EL bias e.g. by performing horizon scans on a regular basis?

*Lines 252-254: The criteria used for the cloud screening should be justified. How do you determine them?

*Lines 289-290: Where these emission values come from? References or justification are needed here.

*Lines 397-398: In Figure 5, only the O3 concentrations for SZA < 50° are used to minimize stratospheric effects. Does it mean that only HCHO and NO2 data corresponding to SZA lower than 50° have been selected for these plots? If not, this means that HCHO and NO2 retrieval results does not timely coincide with the O3 concentrations. This point should be clarified.

*Line 398: It is stated that the JM2 O3 product showed good agreement with ozonesonde measurements. Has such verification been done at the three stations involved in the present study? Also, the Irie et al. (2021) reference is missing in the list.

*Figure 7(a): Even if they both correspond to high O3 concentration conditions, I am surprised to see that the Rfn vertical profiles at Phimai and Pantnagar have both the same shape. Could you comment on this point? Also, why the Rfn vertical profiles from the CHASER model are not included in Figures 7(a) and (b)?

*Section 3.2.1: I think it would be useful to show the seasonally-averaged MAX-DOAS AVK corresponding to the climate classifications of each site in the Supplement. This would support the discussion here.

*Figure 8: given the very large error bars on the MAX-DOAS vertical profiles, I think it is important to say that the CHASER with AK – MAX-DOAS differences are not statistically significant.

*Section 3.2.3: Why no CHASER versus MAX-DOAS profile comparisons are shown for NO2 and HCHO for Pantnagar? This is not consistent to what is presented at the Phimai and Chiba stations.

*Line 744: Is it 1.1° or 1.4°?

Technical corrections:

*Line 24: ‘variation’ -> ‘variations’

*Line 29: ‘good performances reproducing’ -> ‘good performances in reproducing’

*Line 48; ‘the lifetime’ -> ‘the lifetime of HCHO’

*Line 78: ‘satellite retrieval’ -> ‘satellite data retrievals’

*Lines 97-98: ‘in three atmospheric environments’ -> ‘in three different atmospheric environments’.

*Figure 1, page 6: I would use ‘concentration’ instead of ‘concentrations’ in the legend of the color bar.

*Line 144: ‘campaign’ -> ‘campaigns’

*Line 147: ‘consist’ -> ‘consists’

*Line 164: ‘following equation.’ -> ‘following equation:’

*Lines 174-175: ‘cross section data’ -> ‘cross section data sources’

*Line 181: ‘using the optimal estimation method (Irie et al., 2008a; Rogers, 2000)’ -> ‘using the approach developed by Irie at al. (2008a) which is based on the optimal estimation method (Rogers, 2000).’

*Line 182: ‘In this approach, the measurement vector y….are defined as’

*Line 188: ‘window’ -> ‘windows’

*Line 192: ‘compromise’ -> ‘includes’

*Figures 5 and 6: It is not clear to me why the y-axis scales of the three plots are not the same in both figures. Please comment. Also, to my opinion, only the transition lines should change between figures 5 and 6, so one unique figure including the three transition lines should be fine.

*Line 458: ‘clarify’ -> ‘support’

*Page 554: ‘imitate’ -> ‘reproduce’

*Figure 9: ‘HCHO’ should be changed to ‘NO2’ in the x-axis label of all plots.

*Figure 10(b): I guess the blue and green curves should be inverted (green curve should be in blue and the blue curve in green).

*Figure 11: the same x-axis scale should be used in the four plots.

*Line 822: ‘Biogenic’ -> ‘biogenic’

*Legend of Figure 14(b): ‘no anthrpogenic’ -> ‘no anthropogenic’

The manuscript by Hoque et al. shows MAX-DOAS measurements of NO₂ and HCHO at three sites in Asia, namely Phimai (Thailand), Pantnagar (India) and China (Japan). The MAX-DOAS measurements are compared with the global chemistry model CHASER simulated concentration in the near-surface layer as well as the profiles. An attempt was made to use the ratio of Formaldehyde and NO₂ concentrations to derive ozone production sensitivity.

While I have mentioned some critical concerns about the significance of this study with respect to the Journal in my short review before the discussion phase, I provide my elaborate review here. Most likely, the short review prior to the discussion phase is not available in the interactive discussion; I append that here and expect it to be addressed.

Broadly the paper covers two separate aspects, namely MAX-DOAS measurements and comparisons with the global model. On the one hand, there are some shortcomings in both aspects of this study; I also find it difficult to motivate the readers, why such a comparison should be made in 2022, and what do we expect to learn from it. For a comprehensive evaluation of the global model, a global dataset (e.g. NDACC) should be used, which are also recently employed to evaluate TROPOMI data products (e.g. (De Smedt et al., 2021; Lerot et al., 2021; Verhoelst et al., 2021) ). If the study is focused on south-east Asia, why a regional model with a better spatial resolution is not used?

Several previous studies have used high resolution (few km), global models, for comparison with MAX-DOAS measurements and emphasised the need to even go for higher spatial resolution (sub km). This study, on the other hand, presents the model results at 2.8° resolution in the base case and 1.4° in the improved resolution case, which in my opinion, is too coarse for comparison with MAX-DOAS measurements.

Concerning the drawbacks related to MAX-DOAS retrievals, I find the vertical grid resolution (1km) too coarse, which limits the usability and interpretation of such data for air pollution-related studies. There are some technical issues related to the measurements as well, but those should be discussed in a detailed review if the editor deems the manuscript suitable for discussion in ACPD.

Detailed review:

Introduction:

1. The authors motivate the readers about the current study in a way that MAX-DOAS measurements of near-surface concentrations and profiles are used to evaluate a global model CHASER (lines78-98). A study with such motivation is more suited for GMD (model evaluation papers). At least in the introduction, I could not find motivation for understating the atmospheric chemistry of the region of interest.
2. Line 62 – I think it is more accurate to replace “radiation” with “radiance”.
3. Lines 68-78: In my opinion, MAX-DOAS is, a powerful independent technique for monitoring atmospheric constituents, and I would mention it first before stating that it is complementary to in situ and satellite measurements. Observation, dataset and methods1.     In my opinion, the climate classification for Pantnagar should be done in a different way. The current classification does not consider summer as a separate season and is rather partly combined in spring and summer monsoon. The months Apr-June are extreme summer months in the Indo-Gangetic plain, with daytime temperatures above 40°C and an average temperature above 30 °C. 

Observation, dataset and methods:

4. Figure1: As the study focuses on the evaluation of the model over the south and east Asian region, I would recommend restricting the map boundaries to only relevant regions. The colour codes show the surface volume mixing ratios (VMR) and not concentrations. The colour bar legend should be corrected accordingly.
5. lines 139-141: Campaign is used two times in the same sentence.
6. MAX-DOAS system: What is the spectral range of the spectrometer used in these measurements. I am keen to know why the higher wavelength window of 460-490nm was chosen for NO₂ retrieval. The instrument used for this study participated in the CINDI and CINDI-2 campaign, and there the fit interval used for NO₂ retrieval was 425-490nm or 411-445nm.
7. lines 148-149 Why would you want to minimise the variations in measured signals for various off-axis measurements. According to the DOAS principle, reference measurements should be taken at a 90° elevation angle to account for stratospheric contribution in the dSCDs. If the 90° measurements could not be taken due to any physical restrictions, this should be stated accordingly.
8. How would the additional off-axis measurements at elevation angle > 10°reduce the systematic errors in the fitting results. In my opinion, measurements at some elevation angles (e.g. 15° and 30°), provide important information regarding the trace gas and aerosol profiles during inversion and should not be skipped if possible. Moreover, later in this study, the authors analyse profiles at high altitudes (> 2Km), and measurements at high elevation angles are necessary for the accuracy of such retrieval. Even the surface layer used in this study has a thickness of 1km, and measurements at high elevation angles are crucial for this layer as well.
9. Lines 170-171 and Figure 2: How does the DOAS fit for O4 look like in the two wavelengths window used in this study. An intercomparison of O4 dSCDs retrieval from the two fit windows should also be shown (at least in the appendix).
10. Line 187. It was difficult for me to visualise what the profile shapes look like for different values of F. It would be nice to have example plots showing the profile shapes for some values of F (similar to that shown in Fig 1 of Beirle et al., 2019 for h and s)
11. It will be more accurate to save that VMRs are “calculated” using the partial VCDs rather than “converted”. Though in this study, the height of the box is fixed, in general, it would be better to also mention that this conversion also considers the height of the box.
12. Why the heights of the boxes are chosen to be so wide at 1km. Several studies (e.g. Kumar et al 2020) indicate a strong gradient in NO2 profiles in the lowest 1km. As the MAX-DOAS measurements in this study are used to evaluation of near surface VMRs of trace gases from the global model, higher vertical resolution in the profile retrieval should be more relevant.
13. Lines 222-223: In lines 170-171, the authors state that significant O4 absorption in 460-490nm was used to retrieve the O4 ΔSCD. Then why an aerosol retrieval in the same wavelength window is not performed? Rather an Angstrom exponent was used to retrieve the AOD at 470nm.
14. Lines 223-223: What is the basis of the assumption of Angstrom exponent = 1. How does the choice of Angstrom exponent affect the retrieval?
15. Line 261: Please cite the latest version of CHASER and mention the model version number.
16. Lines 262-264: What is the name of the chemical mechanism used for CHASER simulation?
17. Please provide specific details of biomass burning emissions. Which product of ECMEF (might be GFAS?)
18. Lines 278-285: Please provide an estimate of NOx and VOCs emissions from different sectors in the regions of interest. This is important to understand and confirm the important emissions sectors speculated in the subsequent sessions.
19. Line 286: It would be nice to already mention here, what is the purpose of multiple CHASER simulations?

Results and discussion:

18. Lines 301 and 306 (and also at several sections of the manuscript): Figure 4 shows volume mixing ratios (not concentrations).
19. Figure 4: Please use the same y-axis scale for all the subplots. Also, in lines 302 and 348, it is important to mention that the standard deviations (or error bars) show the variability (not to be confused with measurement uncertainty).
20. Lines 326-328: How do these mixing ratios and the seasonality compare with the other studies reported in the Indo-Gangetic plain or other sites in India) (e.g. Biswas and Mahajan 2021, AAQR, Kumar et al 2020, ACP).
21. Section 3.1.3.1: In my understanding, the HCHO and NO2 indicator ratios (R_FN) indicator proposed by Martin et al., 2004 and Duncan et al., 2010 are based on the tropospheric vertical column densities (VCDs) and NOT concentrations. As the authors work with the MAX-DOAS system in this study, why they have chosen to calculate the ratio based on concentration and not the VCDs?
22. Lines 390-391 and Figure 5: What is the person correlation coefficient of the scatter plots shown here. I wonder, how robust are the calculations drawn based on slopes of the scatter plot if the correlation is poor.
23. Figure 5: Please show a similar plot colour-coded according to solar radiation (radiance at a selected wavelength). This would enable the authors to evaluate the contribution of chemistry in ozone production independent of available solar radiation.
24. How do the R_FN values compare to previous studies (based on model, satellite and MAX-DOAS observations) in India (or Indo-Gangetic plain)?
25. Line 472: I was wondering if the boundary layer height directly from the model simulations or reanalysis data products (e.g. ERA5) can be used and more suitable.
26. Section 3.1.3.2: It is difficult for me to understand the need to calculate the factor “F” (column to surface conversion factor, equation 9) in the context of this study. Authors use and discuss “F” to get column integrated values (i.e. concentration). However, the MAX-DOAS retrieval also provided the vertical column densities, which is a much simpler approach.
27. Lines 501ß503 It might be true that there is no relevant literature in the south and south-east Asia presenting “F” values. But there is sufficient literature discussing both the surface concentrations and the vertical column densities, from which “F” can be derived.
28. Lines 511-512: Averaging kernels are highly sensitive to atmospheric conditions, and hence these should be applied to individual profiles, and the averaging should be performed rather than using an averaged averaging kernel for a season.
29. Lines 515-516: How and between which parameters are the R values calculated? Are the R-Values calculated using individual measurements, daily average or seasonal mean?
30. Figure 8: can the authors explain why the application of averaging kernels significantly decreases the column in Phimai, but results in an increase in Chiba?
31. Lines 533, 550, 590, 652, 653, 655: The MAX-DOAS profile retrievals are performed at a vertical resolution of 1Km, and hence it is not appropriate to quantitively evaluate the model profiles at intermediate layers (e.g. 0.5 km or 200m).
32. Line 546: Please provide appropriate reference justifying the model overestimation of biogenic emissions.
33. Lines 554-558: If biogenic emissions are overestimated in the model (as mentioned before), I would expect a higher increase in simulated HCHO than observed between January and August.
34. I am surprised to learn that emissions due to wintertime heating is not included in the anthropogenic emission inventory. From the EDGAR website (https://edgar.jrc.ec.europa.eu/dataset_htap_v2#p1), it seems that the sectors “htap_6 Residential” and “htap_3 Energy” include the wintertime heating emissions.
35. Sector 3.2.3 could be merged with 3.2.1 and 3.2.2.
36. Lines 678-680: It is not clear for me, why observations above 1.8km are compared with the model. Both MAX-DOAS profile retrievals and model simulation are performed above the ground level.
37. Lines 685-686: This brings me to the previous comment. Why in the first place, measurements are restricted to elevation angles less than 10°.
38. Lines 687-691: What is included in the whole IGP. Please show it on a map. What are the limitations of comparing the measurements at a point (representative of a few Km) to the entire IGP?.
39. Figure 11: Please use the same y-axis range for subplots of HCHO and NO2.
40. Line 711-712: If the biogenic emissions are overestimated, how come the simulated isoprene concentrations are reasonable?
41. Figure 12 and line 735: In my opinion, it will be better to show the time series at the three stations rather than the zonal mean if the inferences are made with respect to observation at the three sites.
42. Lines 738:740: 10% is the average, and based on this, one can not infer that the comparison result will improve by at least 10%.
43. Line 748: How did the authors estimate that the impact of model resolution is 20%.
44. Line 755: What stops the authors from using an updated emission inventory if those are already available.
45. Line 782: How are the biogenic emissions optimised?
46. Lines 789-791: Please provide an estimate of NOx emissions from different sectors based on the emission inventory used for the simulations.

References

Beirle, S., et al. (2019). "The Mainz profile algorithm (MAPA)." Atmos. Meas. Tech. 12(3): 1785-1806.

De Smedt, I., Pinardi, G., Vigouroux, C., Compernolle, S., Bais, A., Benavent, N., Boersma, F., Chan, K. L., Donner, S., Eichmann, K. U., Hedelt, P., Hendrick, F., Irie, H., Kumar, V., Lambert, J. C., Langerock, B., Lerot, C., Liu, C., Loyola, D., Piters, A., Richter, A., Rivera Cárdenas, C., Romahn, F., Ryan, R. G., Sinha, V., Theys, N., Vlietinck, J., Wagner, T., Wang, T., Yu, H., and Van Roozendael, M.: Comparative assessment of TROPOMI and OMI formaldehyde observations and validation against MAX-DOAS network column measurements, Atmos. Chem. Phys., 21, 12561-12593, 10.5194/acp-21-12561-2021, 2021.

Lerot, C., Hendrick, F., Van Roozendael, M., Alvarado, L. M. A., Richter, A., De Smedt, I., Theys, N., Vlietinck, J., Yu, H., Van Gent, J., Stavrakou, T., Müller, J. F., Valks, P., Loyola, D., Irie, H., Kumar, V., Wagner, T., Schreier, S. F., Sinha, V., Wang, T., Wang, P., and Retscher, C.: Glyoxal tropospheric column retrievals from TROPOMI, multi-satellite intercomparison and ground-based validation, Atmos. Meas. Tech. Discuss., 2021, 1-48, 10.5194/amt-2021-158, 2021.

Verhoelst, T., Compernolle, S., Pinardi, G., Lambert, J. C., Eskes, H. J., Eichmann, K. U., Fjæraa, A. M., Granville, J., Niemeijer, S., Cede, A., Tiefengraber, M., Hendrick, F., Pazmiño, A., Bais, A., Bazureau, A., Boersma, K. F., Bognar, K., Dehn, A., Donner, S., Elokhov, A., Gebetsberger, M., Goutail, F., Grutter de la Mora, M., Gruzdev, A., Gratsea, M., Hansen, G. H., Irie, H., Jepsen, N., Kanaya, Y., Karagkiozidis, D., Kivi, R., Kreher, K., Levelt, P. F., Liu, C., Müller, M., Navarro Comas, M., Piters, A. J. M., Pommereau, J. P., Portafaix, T., Prados-Roman, C., Puentedura, O., Querel, R., Remmers, J., Richter, A., Rimmer, J., Rivera Cárdenas, C., Saavedra de Miguel, L., Sinyakov, V. P., Stremme, W., Strong, K., Van Roozendael, M., Veefkind, J. P., Wagner, T., Wittrock, F., Yela González, M., and Zehner, C.: Ground-based validation of the Copernicus Sentinel-5P TROPOMI NO2 measurements with the NDACC ZSL-DOAS, MAX-DOAS and Pandonia global networks, Atmos. Meas. Tech., 14, 481-510, 10.5194/amt-14-481-2021, 2021.