The paper presents a thorough evaluation of mobile car-mounted turbulence measurements near the surface. The mobile measurements are compared with corresponding stationary tower data, which shows that the mobile system can provide satisfactory mean and turbulence data following a proper procedure for flow distortion correction. Furthermore, it is shown that using wavelet analysis for calculating higher order statistics of the mobile measurements can be more appropriate than the traditional eddy-covariance technique. The paper is well written and I recommend publication after minor review.

Specific comments:

1. It is not clear how many measurement passes are made for each track.

2. ln 243: The authors should clarify how exactly the mobile data can have "a time series with a temporal length 11 times that of" the 1-km variance. Since the track length is 1 km, where does the additional data (the temporal equivalent of 10 km) come from?

3. lns 605-607 and 617: The interpretation of the confidence interval should be clarified. Why is one standard deviation related to the 95% confidence interval? Why is the confidence related to "not significantly different than 0" at ln 607 and "consistent with the tripod" at ln 617?

Technical comments:

- ln 365: the sentence looks unfinished?

- ln 394: "Figure 6(a), (b) and (c) show..." - Such statements should be clear from figure captions and are not needed in the main text. Similar holds for other figures (e.g. Fig. 8).

- ln 478: delete one occurrence of "of the".

- Fig. 12b: x-axis should say "... sonic...".

General comments.

In this work the authors compare turbulence measurements made on a car instrumented with a sonic anemometer with the same measurements taken on a fixed tripod at the side of the road. The choice of the site, with lateral obstructions but with light traffic, is appropriate to the purposes of the comparison. The growing need of spatially extended data over inhomogeneous terrain makes mobile measurements an important topic in turbulence measurements. The most intriguing part is a wavelet based approach to reduce eddy-covariance measurements when they substantially differ from the ones from the fixed instruments and to remove the effect of intersecting vehicles on the measurements. The paper is clear and well written, sometimes a bit heavy to read for someone not familiar with all the correction methods described throughout. I recommend publication after the authors addressed my minor comments.

Major comments.

• Section 2.5. As said above, this may be the most interesting part of the paper, since it offers a solid correction method for car measurements. However, while the wavelet analytical formulation is very clear, how the wavelet is applied is far less clear. I struggled a bit in understanding what is the averaging time scale on which the measurements are compared. At line 403 it seems that the maximum track length (in seconds) is around 40-60s while 5 to 8 minutes averaging was used before for wind directions and speeds. Was a different averaging time used for variances and covariances to compare with wavelet analysis or was only T set to 40-60 s as wavelet max-scale to reduce low-frequency contribution?  
• The comparison between the turbulent heat flux is very interesting and well discussed. Would not be the case to compare the temperature variances seen by the tripod and the car?

Minor comments.

• The paper would benefit a table with the number of measurements records analyzed each day.
• In Section 2.4 what is the averaging time of the data presented?