Review of  wes-2024-119 “Tall Wind Profile Validation Using Lidar Observations and Hindcast Data”

Summary

The submitted paper, ``Tall Wind Profile Validation Using Lidar Observations and Hindcast Data'' by Etienne Cheynet et al., analyses the accuracy of three different wind atlases (NEWA, NORA3 and ERA5) by comparing them to long-range wind lidar measurements.

Various error metrics such as bias, RMSE, EMD, $\mathrm{R^2}$ are evaluated for five different sites including offshore (FINO1, FINO3), coastal onshore (Sola, Lista) and complex terrain (Bjerkeim).



The study focuses on wind conditions at heights between 100 m and 500 m, a range not feasible for traditional met masts or short-range lidar profilers, making it directly applicable to tall wind turbines and airborne wind energy (AWE) systems.

Additionally, the authors assess the estimated capacity factors of reference wind turbines (NREL 5 MW, NREL 18 MW, and IEA 15 MW) and AWE systems (3 MW fixed-wing and 100 kW EnerKíte semi-rigid), comparing simulated and measured wind data as a further quality metric.



The authors conclude that all three wind atlases perform well offshore, with NORA3 and ERA5 showing slightly better performance above 200 m. Onshore, NORA3 consistently outperforms ERA5 and NEWA at all heights, emphasizing that the choice of a suitable wind model depends on specific application, location, and height requirements.

The paper addresses important topics, such as the adequate choice of wind data for initial wind resource assessment, validation of wind models at heights relevant for future wind energy systems, and the need to develop DWL profilers to reliably measure these heights over the long term.

General comments

Here are some general comments I would like to see addressed before publishing the paper.

However, some of these comments are a matter of personal preference, and I would appreciate hearing the authors' opinion if they choose not to implement them.

• Please rethink the title ``Tall Wind Profile Validation Using Lidar Observations and Hindcast Data''. This title sounds like you are validating tall wind profiles using lidar and hindcast data. I am not familiar with the term Hindcast, but it is my understanding that only NORA3 is Hindcast data, ERA5 is climate reanalysis data and NEWA derived from WRF and WAsP with boundary conditions from ERA5. Are you using Hindcast and reanalysis interchangeably? It would good if the title would reflect that you are validating and comparing lidar measurements with different wind models up to higher altitudes.
  
     
• Please clarify the writing. Several sentences are difficult to understand or can be understood in various ways. See attached commented document.
• Appendix: Why was the section moved to the appendix? It is my understanding that while this approach delivers good interpolation, it is not used because the improvement doesn't justify the increased afford? I think you could remove it for clarity and only focus on the approach you took, but I am open to hearing your onion.
• You compare bias, $\mathrm{R^2}$, RMSE and EMD in your paper. Could you rank or classify them in terms of there meaningfulness for wind energy assessment, i.e. what metric is most important for the application? This could be different for different applications such as stastical predictions, wind profile fitting etc.
• You determine bias, $\mathrm{R^2}$, RMSE and EMD only in terms of horizontal wind speed. Did you also investigate the directional difference between the models and measurements and are they meaningful or significant?
• Wind data: Please clarify how you compared the model and measurement data. You mention a spatial and temporal interpolation of model data to the lidar location, height and time. How did you interpolate between 30 or 60 min modeled wind data to 10 min increments as the measurements? Is it a linear interpolation? Considering how quickly the wind changes that leads to significant differences. I think it would be better if you averaged the 10 min measurements to 30 min or 60 min instead.
• Measurement campaigns: The duration of measurement campaigns, particularly at Sola and Bjerkreim are very short and seems to have a lot of data missing (Figure: 6). Please comment on what the reason for this is and add a brief statement in Section 2.2. that these measurements are not representative of the typical, annual wind variations at these sites.
• Why are you comparing wind data ``at the range gate nearest to 150 m''? Why this height and not 200 m or 300 m which is closer to operating heights of tall wind turbines and AWEs? Why did you not interpolate to a specific height to compare them better?
• Please introduce the Taylor diagrams a bit more and the what conclusions you can draw from them.
• Do you really need section 5? I think you could merge it with section 6, but I would like to hear your opinion too.
• Please spend a few sentences introducing the different AWE models. Introduce the system design, size, soft-kite, rigid-wing or semi-rigid wing, operating conditions, limitations and model assumptions.

Specific comments

Please see the attached document for additional comments.



Technical corrections

• Please revise the figures to improve readability and clarity (see comments in PDF document)!
• I believe that it is good practice not to have empty sections before a subsection title, e.g. Sections 3 or 5. Either remove the subsection titles or write a very brief summary of the section before the first subsection title.
• Try to formulate mote active voice sentences. Some are mentioned in the attached document.
• You can remove several unused abbreviations and introduce $z$ for height or $\overline{u}$ for average horizontal wind speed
• Please add hyperlinks to the references, e.g. citation A, Tab.1, Fig.2, Eq. 3         
• Capitalize ``fig.'' and ``table'' in the entire the paper