General Comments

In this manuscript, the authors present a systematic comparison of the Tropospheric Monitoring Instrument (TROPOMI) version 1.2 and Ozone Monitoring Instrument (OMI) QA4ECV tropospheric NO2 column through global chemical data assimilation (DA) integration. The comparison of the impact of TROPOMI and OMI tropospheric NO2 on global chemical data assimilation is comprehensive. The topic of the manuscript fits the scope of  AMT. The manuscript is mostly well written. However, some details of observation data and discussions are needed. The paper can be published after some minor revisions.

The study is based on only two months (the period April–May 2018). To my knowledge, TROPOMI has strong negative bias in wintertime. If the study is conducted for the winter period or other months, will the conclusion be different? The discussion is missing in the paper.

Specific comments:

L71: typo: rfaction to fraction.

L88-93 section 2.2.1, can you please provide more details about the NO2 observations in the Atom aircraft-campaign? Such as: what is the time window of the no2 observations on each day? The frequency of the no2 observations, per minute? Per hour?

L169: spatial representativeness error, should it be  √σ_m²+ σ_r²  ?

L 265-261: How did you compare the vertical profiles between aircraft measurements to the model simulation? Can you give more details? Did you average the profiles over the area?

L417-450: The study time period of the data assimilation is April-May not the whole year. Please mention this in the conclusion.  If you get the same conclusions or not when you include winter period. It could be nice that you can add some discussion here.

L426: It is not accurate to conclude the global change of NOx emissions per year. Please rephrase the sentence or mention the time period.

The manuscript by Sekiya et al. compared the global chemical data assimilation results when using NO2 retrievals from TROPOMI and OMI. The TROPOMI posterior NO2 shows better agreement with NO2 observations and smaller magnitude than the OMI one. The manuscript is generally well-written. The topic fit the scope of AMT. The result is important in interpreting existing NOx data assimilations. I suggest publication after addressing the following comments.

L7, if TROPOMI NO2 is biased generally low, would the comparison with independent data improved for the wrong reason?

Figure 1. Please provide the resolution these data are gridded to in the figure description. How much do precision error and the number of observations in the super-observation grid each contribute to the smaller super-observation errors in TROPOMI data?

Line 221-222, it would be clearer to first explain what the range of chi-square is, and what do values larger and smaller than 1 generally mean.

Figure 2, I am a bit surprised that a large portion of the TROPOMI DA improvement is over the ocean, where there is no emissions. Please explain what possibly causes this.

L250-251, I am confused about the “regardless of the TROPOMI low bias” part. Is this only true because you calculate RMSE against the TROPOMI observations?

Figure 4, Please provide more information on what is being optimized in the DA. Are both NO2 concentrations and emissions optimized at the same time? Are emissions all in the surface layer? If not, how are they distributed vertically, and how does DA adjust emissions differently at different layers?

L287, a similar comment as a previous one, if there are systematic low biases in TROPOMI data, why do its DA results have better agreement with independent data?

L330, could you also add a figure showing the changes in NO2 concentrations from the two DA?

L339-351, based on the low biases in TROPOMI NO2 retrievals and the comparisons here, what is the implication for existing DA and inversion results using this version of TROPOMI NO2?

L446-447, would you expect the low biases in TROPOMI NOx emissions reduce using this new product, and by how much?