I  reviewed the manuscript titled "Investigating Lab-scaled Offshore Wind Aerodynamic Testing" submitted to Wind Energy Science. First and foremost, I would like to commend that this is a valuable work and particularly the innovative approach you employed for anomaly detection in lab-scale wind turbine testing. The topic is both relevant and insightful, and the methodologies presented have great potential to enhance safety and reliability in experimental campaigns.

During my review, I identified some areas where improvements could further strengthen the manuscript. To ensure clarity and precision, I have noted a few points related to language accuracy, sentence structure, and the overall presentation of some technical aspects. Below are the main suggestions and comments:

1. Language and Grammar:

   - Par 25: Some sentences contain grammatical errors and awkward phrasing, making it challenging to understand the intended meaning. For instance, the use of the term "acute care" in a technical context is not entirely appropriate. I suggest replacing it with more suitable phrases such as "careful handling" or "close attention," depending on the tone you want.

   -There are a few instances where the text contains repetitive expressions, such as "as a consequence..." and "consequently," within the same paragraph. Reducing repetition will help enhance readability.

- Par 25. There are sentences that need to be fixed, for example, the sentence starting with "Furthermore" should read "Furthermore, due to tight mass considerations, particularly for floating models, system redundancy in the case of equipment malfunctions is generally not implemented, as noted by Parker (2022)."

- Par 150 is really difficult to understand (Especially the sentence starting with "The combinations..").  Both in terms of English and some parameters, such as DE|E and DE&E conditions, are not clearly explained.

- Par 160: FOR is written in capital.  The sentence "As is...." Should be rewritten.

- Par 165: No verb in the first sentence.

- Par 190: Sentence "This work ...." should be rewritten.

2. Clarity and Technical Accuracy:

   - Par 35: In some sections, the explanation of velocity measurement as a vibration-based monitoring technique is not entirely clear. A more detailed rationale or supporting reference would improve comprehension.

- Par 95: What does rotating a matrix mean?  Do you mean transposing it?  Needs to be clarified.

- Par 130:  Information given here is mentioned earlier and is repeated here.  This reduces the rigor, conciseness, and precision of the entire text. 

   - Tables 1 and 2 have a central importance to the paper. But they are not adequately explained or referenced in the text.  They should be explained and discussed thoroughly.  Providing more context, especially when discussing essential results or comparisons, would enhance the reader's understanding.

- Uncertainty estimates for the P, R, and FI in Tables 1 and 2 would definitely help the technical rigor of the paper.  For example, those parameters are given with 3 decimal resolution.  Can this be justified?

- Par 170: Increased frequency would help to make a faster detection, not reduced.  

3. Structural and Visual Presentation:

   -The arrangement of the three graphs in Figure 9 can fit in a single line, which would help in better page management and improve readability.

4. Additional Considerations:

   - It would be valuable to discuss the physical or mathematical reasons behind why the 1PC model performed better in anomaly detection.

I believe that addressing these points would significantly enhance the manuscript's quality and impact. 

The paper is well written and contains potentially useful information for wind energy researchers. However, the level of details provided in the paper is hinders repeatability and reduces the impact of the paper.

Please add the following:

1) Details and example of data-preprocessing with time-series signals, idle sensors and exact data cleaning method used forat least one representative case. What channels are used and why were they picked.

2) Traing - testing methodology: Type and size of RNN model used,  time / resources needed for training.

3) How many different faulty and non faulty cases are considered in simulation and how do 1PC and MPC compare in terms of false positive and negative detection. If this has not been studied yet, please add comments about this and/or add this as future work.