This article presents an interesting comparison of statistical quantities between a field experiment, LES+ALM+BHawc, and BEM+BHawC. The paper is generally well written and provides many details about the setup.  I enjoyed seeing results from a field campaign as part of the comparison. Some parts of the paper need clarification and a better discussion. I have the following recommendations to improve the manuscript before publication.

1. The authors are comparing two methods (BEM vs LES), but they have quite different inflows between the methods. The authors need to include plots of wind speed (inflow) as function of height and turbulence intensity as function of height for the different cases. The inflow will be the first source of difference between the methods and the authors should discuss this in more depth.
2. The authors select 75 aerodynamic nodes along the blade and value of epsilon for their simulations. The values selected (eps/D) are not expected to provide an optimal representation of the blade loading from an ALM. I recommend the authors to expand this discussion and to highlight some of the limitations from the ALM at the given resolutions.

Specific comments:

“A substantial advantage of this code is the possibility to emulate the actual controller of industrial wind turbines.”

Comment: Have you been able to successfully use an industry controller? Industry typically does not share their controllers. It is possible to use their controllers in other codes as well, but researchers do not have access to these controllers. I would recommend that the authors remove/modify this statement as it might be misleading.

“The latter is usually chosen as twice the maximum cell-size encountered in the rotor region.”

Comment: Can you please expand on the description of epsilon chosen for this work? There is a large dependency of blade loading on epsilon. Please specify the actual value of epsilon (as physical length scale) and why the choice was made. Also, epsilon/dx=2 is typically not enough resolution to properly resolve the flow field around the ALM and the loading is affected by this.

“Yet, the code can only handle one turbine at a time as it considers only one incident flow field. The simulation of an operating wind farm remains feasible by considering different

incident flows for each turbine, mimicking at best the local flow properties.”

Comment: I recommend the authors to remove this statement or elaborate. This statement applies to any code, each turbine has different inflow, so a new instance should always be used per turbine.

“We expect this issue to be at least partially removed by implementing an appropriate dynamic stall model in YALES2. Indeed, simulations carried out with the standalone version

of BHawC showed a link between the divergence of the structural solver and the dynamic stall phenomenon model activation. It is emphasized that the induction is close to being maximum in the mentioned wind speed range, leading to a strong coupling between the structure and the ambient flow. Therefore, the relaxation of aerodynamic forces induced by the dynamic stall

phenomenon is likely to be essential in these cases.”

Comment: This seems to be an interesting hypothesis. However, it is not the focus of the paper. I recommend the authors to avoid these statements unless they relate to the work done in the manuscript.

Sections 4.1 and 4.2

Comment: These sections provide interesting checks to ensure that the code is working as expected. However, they do not add much value to the manuscript and should be moved to an appendix.

“In all simulations performed with the coupled code, the mean turbulence intensity in the rotor-swept area was assessed: obtained values are gathered in Table 2.”

Comment: There is a significant mismatch between the codes. I recommend the authors to include profiles of wind speed and turbulence intensity from the different methods.

“Considering all the aforementioned sources of discrepancies, the results given by the code YALES2-BHawC for the structural part are therefore deemed encouraging and promising. As for the prediction of the main operating parameters and performance of the turbine, the coupled code already provides very accurate results.”

Comment: I recommend the authors to reevaluate these statements. All these discrepancies suggest that there is a substantial amount of error cancellation between models going on.

The paper presents a field data-based validation of an aero-servo-elastic solver coupled with LES simulations. The topic is very interesting and relevant for the readers of Wind Energy Science. The article is well-written. The authors provided detailed explanations regarding how to couple the LES tool YALES2 with the aeroelastic tool BHawC.

I only have some minor comments as follows:

1. page 2. Please define ‘CFD’
2. page 2. ‘… HawC2…” should be HAWC2
3. “To the authors’ knowledge, (Gremmo et al., 2022) is the only reference where a code with these capabilities has been reported.” Can the authors elaborate on what capabilities they have? As far as I know, there are also examples of coupling between EllipSys and HAWC2, which can be easily found in the literature.
4. page 8. “First, it cannot work without the structural solver also having at least the blade element theory implemented within. In other words, a standalone structural solver is not suited for this strategy.” Can the authors elaborate more? Why wouldn’t the standalone structural solver without the blade element work?
5. Figure 8. I don’t really understand this figure. It seems the mean error is so small? Is it true?