Dear authors, both reviewers agree, that the manuscript has been improved considerably. However, both reviewers also agree, that your conclusions of the model performance are too strong, or cannot be understood from your results discussion.

There are also some minor points:

In the tables, what is MB and MNB? I did not find an definition in the paper. Please make sure, that all abbreviations are defined before use. In case of a table, it makes sense, to define them again in the caption to foster readability

Page 8, line 9. Define “RRTMG”.

Page 8, line 6-8. Could you indicate the reasons why you need to use initial and horizontal boundary conditions from another model?

Page 11, line 23-24. I guess you mean “CMAQ_base” and “CMAQ_B”

Figure 3. I suggest changing y label to “Ratio”.

Major:

Page 13, line 5. I don’t think this conclusion can be made based on the figure. The locations of different sources of NH3 are still not clear to me based on the figure.

Page 13, line 9. The r2 of 0.64 is not a well-correlated case to me. I don’t think “CMAQ_base inventory does a good job of capturing the spatial distribution of NH3 emissions near Bakersfield” based on the r2 of 0.64 and mean bias of -2.67 ppbv. I suggest the authors either soften the words or remove this conclusion.

Also consider the attached review for more minor and major points. I also agree with the reviewer, that your statistics is not very clear. As for the correlation it is not clear to me, if the correlations are distinct ffrom 0 (i.e. the hypothesis that the data are not correlated). I consider this especially true in your table 1. A correlation of 0.001 seems to be a non-correlation. But than fitting a line becomes sort of arbitrary. So I agree with reiewer 1, that the statistical reasoning deserves some explanation.

The referee 1 (attached documents) suggests two possible ways for changing the manuscript. If you choose neither, please justify your choice.

Kind regards
Mathias Palm