Dear Yumeng Chen,

I am very pleased to inform you that the revised version of your manuscript has been accepted for publication after a few minor corrections. I would like to ask you to address points 1) to 3) listed below. Point 4) is optional and not mandatory for publication.

Best regards,
Andrea Stenke

1) Line 129: the definition of the horizontal divergence is incorrect. Note that divergence is not a vector as defined here, and the second term of the divergence operator needs to have an added cos(phi) term in the numerator: (v cos(phi)). Since the divergence is actually never used in the following equations, I suggest just leaving it out.

2) Line120 onwards: It seems that the current notation with the density rho implicitly assumes (for the purpose of the analysis) that the tracer mixing ratio is 1 (as defined in line 109) and that the flow is non-divergent, as it was done in Lin and Rood (MWR, 1996) for the FFSL scheme. Clarify whether the non-divergent flow condition is assumed here.

3) Line 192: the subscript 'a' in the symbol \phi_a is undefined. Add a definition.

4) Figs. (22)-(24): The authors switched to a different color scheme which is appreciated. However, it is obvious that the colors saturate and the range of values in the simulation is not captured. In addition, discrete colors are a better choice (as in the earlier version of the manuscript) instead of the new gliding color scheme. The most adequate color scheme here would use non-equidistant spacings that are tailored to the simulation data. Overall, the current plots convey the main message that the AMR grid captures certain regions, therefore changing the colors again is not a must-do correction. It would be a nice-to-have enhancement/improvement of these figures.