General comments:

This paper conducts downscale modeling for the Europe-Denmark-Copenhagen domain by coupling HIRLAM, CAMx and M2UE models. Based on two urban monitoring stations, the authors illustrate that the HIRLAM-CAMx system captured well the diurnal cycle of NO_x and O₃. During September 4 to 5, 2011, the street level NO_x pollution (mainly from road traffic) was 2-fold higher on weekend and more than 5 times higher during working day with high pollution episode.

Overall, the downscaling modelling study is quite interesting. However, with the lack of data and figures of both observation and simulation results, the accuracy of this downscaling modeling system needs to be further verified. In addition, several parts of the manuscript should be improved before it can be considered for publication. Specific comments are listed below.

Specific  comments:

1. Line 128. In the M2UE simulation, is the 500 m horizontal extent sufficient to represent the atmospheric environment of the street? I wonder if the airflow around the monitoring site is still substantially influenced by the boundary conditions. The extent of the airflow around the street that is largely affected by boundaries needs to be further investigated.
2. Line 122. What chemical components and their reactions are included in the M2UE?
3. Are the emission data used in the M2UE simulation equal to the emissions in the corresponding grids in the CMAx simulation? If not, how the difference between the emission data affects the simulation results?
4. Line 150. Are the inlet or outlet properties of each boundary fixed in the 48 h simulations? How many grids in HIRLAM-S03 and CMAx-04 are chosen to drive the M2UE model? Are the inlet/outlet properties derived from these coarse grids consistent with each other?
5. Line 175. Why do the authors select 2011 as the study period?
6. Figure 4. Why only the NO_x concentration is validated in the M2UE simulation? More chemical species are suggested to be evaluated against obervations.
7. For the M2UE modeling, simulation output is analyzed at only one site. Since the results of a single grid may not be enough to support the conclusion, it is recommended to add more analysis on the horizontal and vertical planes.
8. Line 222 and Figure 4. Time-series data/figures of traffic volumes, emissions and boundary conditions are suggested to be provided. Do these data also show two peaks at 31 and 34 h? If so, the interesting difference between the observed concentrations and emissions may need to be discussed.
9. Line 246. Typo “/Pearson”.
10. Line 247. Typo “domians/”.
11. Typos: “μg” instead of “ug”.

General comments: This paper proposes a downscaling modelling system for prediction of weather and atmospheric composition from the regional scale down to street scale, by running a computational fluid dynamics model. This is an interesting topic, and the study is relevant, making the purpose of the research an important work within the scope of the Atmospheric Chemistry and Physics journal. The paper is well written and structured, however more detailed the results could be presented, strengthen the discussion. Overall, the discussion of results could be improved providing information on the main findings of the work, how they compare with papers already published. In my opinion, some improvements and clarifications need to be made before publication. Specific comments: 1. Is there a “double-counting” of the urban processes in the CFD simulations, considering that those simulations are initialized with meteorological fields from HIRLAM model with activated BEP module? 2. Are those emissions provided by TNO and used in CAMx simulations the most updated emissions for all the SNAPs? 3. Which kind of k-epsilon turbulence schemes are available in M2EU model? In addition, which schemes are available for meshing? 4. Authors mentioned that “the horizontal and vertical extent of the computational domain is 500 x 500 x 100 m.” There is a lack of information about the dimensions of the domain (e.g. height of the tallest building in the domain). Did the authors follow the COST Action 732 guidelines 5. Authors should provide a figure plotting the mesh to be more specific about this information “the horizontal and vertical numerical grids with local refinement of grid-cells ranging from 1 m at the Jagtvej 130 Street to 5 m near the domain’s boundaries was used.” How about the distance between the boundaries and the build-up area? Again, are you considering the COST Action guidelines or similar? 6. “The Jagtvej Street emission inventory was derived from hourly traffic counts (Berkowicz et al., 2004) and emissions for different types of road-vehicles (Chao et al., 2018).” Therefore, I am assuming you are only considering the road traffic emissions for the M2EU simulations. Am I right?