Dear Authors,

According to the suggestions from the 2 reviewers (who I thank for having greatly contributed to improve the manuscript), it is a pleasure for me to accept your paper for publication in TC. You have improved a lot your paper with this revised version.

However, before final publication, reviewers suggested yet some minor changes and requested to improve your explanation of the "lower" runoff at the north of ice sheet at high resolution. I agree with them that your explanation about clouds is not convincing.
According to your Fig S4, the warm biases of CESM at the north of ice sheet at 700hPa (it should be fine to check this also at 850hPa) is reduced when CESM is run at higher resolution. This is therefore just here an error compensation that explains why CESM_lowReso works better than CESM_highReso and not necessary linked to clouds. Your free atmosphere seems to be warmer in this area when CESM is run at lower resolution. To "prove" this, adding a figure showing Fig S4e - Fig S4d and S4f - Fig S4d could help. Comparing the JJA sensible heat flux should prove this?

Before final acceptation, I'm then waiting from you a revised version (that will be reviewed by myself only), dealing with last remarks/suggestions from the reviewers.

Thanks and best regards,
Xavier F.

Dear Author,

Thanks for this revised version and for your clear and honest explanation of the performance "decrease" of CESM when it is run at higher resolution.

For me, the downscalling technique has originally been developed (and probably tuned) to compensate in part biases when CESM is run at low resolution. As you explain now well, when CESM is run at high resolution, the corrections brought by this downscalling technique are not enough to compensate the CEMS biases. This downscalling technique is particular dependent of the (fixed) vertical lapse rates used to extrapolate TT, LWD, ... to the sub-grid topography. Using other (in fact larger) values for these lapse rates when CEMS is run at higher resolution could fix in part the "decrease" of performance. If you agree with me, feel free to mention this in your conclusion.

Anyway, your paper is ready for me in the present state and could be sent to the Copernicus office for the typesetting.

Best regards,
Xavier F.