The authors have made a set of substantial additional analyses to improve the article, which are well represented within the scope of the manuscript. They addressed most of the reviewer comments adequately and proposed changes accordingly in their revised article. Also included explicit statements to support the significance of the study, which were missing in the earlier version. The title of the manuscript is changed to best match the content of the manuscript.

Requested revision: clarify the following issue.
Fig.4 in acp-2020-540-AC2-supplement.pdf:
Panels (a and b): The model's performance (temperature and wind speed) is too GOOD here (even R becomes 1 in (a))!
Is it because of data assimilation/nudging as explained in Sect. 2.1 (preprint version)?
["We also used the surface analysis nudging and observation nudging options to assimilate the National Centers for Environmental Prediction (NCEP) operational global upper-air (ds351.0) and surface (ds461.0) observation weather station data ...."]

If authors use the data from these sites for data assimilation, it is not proper to present Fig. 4a-c(+Fig.S2a-c) as part of the model evaluation.

You mention the high correlation of the SAC data with that of the nearby stations which are included in the data assimilation. I think you need to be careful saying they are 'independent' observations, as they are not independent from a strict statistical perspective.

You could try this from line 23 on page 6:
"In addition to the CO2 measurements, the hourly air temperature, wind speed and wind direction are measured at a height of 100 meters above ground level at the SAC station. The meteorological data from SAC are not included in the data assimilation process of the NCEP operational global weather observation subsets used in the WRF nudging program (section 2.1). We therefore use them in the evaluation of the model performance in simulating the meteorological fields."

Thanks