This is a very well written paper. It clearly shows that the numerical predictions work well against experimental data. The authors did a very thorough study on the validation and parameter study. I recommend the publicaiton of this paper.

This research article aims at providing wind turbine noise measurement data that can be used for the validation of sound propagation models. It is indeed of great importance to be able to predict the impact of wind turbine noise at relative long ranges, and there not many detailed datasets available to perform systematic model validation. I think the 10 validation cases that have been selected and analyzed by the authors will be of interest for the community. However, as explained in my specific comments, I don't think that atmospheric turbulence can be characterized with the sensors available in this experimental campaign.

The part of the article devoted to the model description and the comparison with measurements is less convincing, and suffers from several shortcomings. In particular, the choice of a source model with 3 point sources need to be discussed, and may be the cause of the discrepancy between the modeled and measured 1/3 octave band spectra. The ground effect is not well characterized, and the modeling of atmospheric turbulence is not convincing. Finally, the use of the total SPL to judge the accuracy of the model is misleading because very small differences between model predictions and measurements can be obtained while the corresponding spectra differ significantly.

I thus recommend major corrections to be done according to the detailed comments described in the attached document.

The overall impression about this paper is positive. The paper represents an important step in the search for suitable data and procedures to compare measurements and model results. The main problem, however, is the selection of suitable measurement data together with all the conditions that are decisive for the generation and propagation of the sound.

The short term variability of meteorological parameters (e.g. wind speed, wind direction, that are part of low frequency turbulent motion) and heterogeneous ground properties is difficult to handle. This may lead to high variance in measured sound data and is difficult to compare to ‘regular’ model data. This variability is not discussed in detail. E.G. how it was determined the absence of a low level jet by mast measurements to 100m height? How the averaging of meteo and accoustic data affect the comparison of data?  Which time range would be recommended (1min,, 5 min, 10 min) ?

Here, the surprisingly good agreement between the model and the measurement gives rise to doubts about the selection of the cases examined. Normally, comparisons of measurement results with very similar conditions often show higher differences. Nevertheless, the results of this paper should be regarded as a first step how to proceed, about input parameters and comparison between measurements and model results.

It is still unclear how a wind turbine should be represented in a sound propagation model. The method chosen here with three alternative sources is one of many possibilities, but there is no evidence that this is the best possible.

The wide range of literature is presented, what may help any reader to advance how to proceed with measurements and comparison to model efforts.