General comments

This article details a recently developed offshore wind data set, the NOW-23 data set, aimed at wind climatology studies for power generation, covering most offshore regions around the contiguous USA and the Great Lakes. The manuscript outlines the data set, the processes used to derive the data from numerical modelling, and details the choices (and validation of these choices) in model set up and execution. The authors conclude with a summary of uncertainties in the data set and include a brief description of an accompanying data set, NOW-WAKES.

The manuscript is well written and concise, however, perhaps in their attempt at brevity the authors have missed the opportunity to reinforce their reasoning for producing this work (see comment #1 below). I do not believe any extra sections are required, but some statistics regarding the improved performance of NOW-23 when compared to the previous version (WIND) would help cement the justification for the manuscript. This is partly touched upon in Appendix C, but a table would help clarify these differences.

Regarding the data set, I was able to access the online repository and download a sample set of data (~1.5 Gb), extract it (in the form of a CSV file), and view/analyze the data.

I recommend publication after the authors consider making some minor changes in response to the comments below.

Specific comments - Manuscript

1. Section 1, lines 35 - 40: Not all readers may be aware of the shortcomings of WIND, therefore stating explicitly how long WIND was reforested for compared to the new 20-year period for NOW-23 would be advantageous. Either this could be added to the text, or perhaps the authors could add a table (or modify Table 1.) for the reader to easily to compare these parameters between WIND and NOW-23.

A new table could detail the differences between WIND and NOW-23 in:

• Temporal Resolution
• Forecast period
• Horizontal resolution
• Heights
• etc

2. I found the approach of validating each area independently using the WRF ensembles as thorough and exhaustive, therefore improving the end user’s confidence in the data set. Additionally, section 11 describes well the uncertainty in the modelled wind speed. However, this also raises a question: how ‘bad’ would the ensemble members have to be compared to observations for you not produce a product for a certain region? Has this been considered? I feel it not addressed in the manuscript.

3. Is Section 12 necessary if two previously published articles already detail this information? If word-count is at a premium then this could be a section to consolidate, using perhaps just the first paragraph to outline the product and draw the reader’s attention to the other two articles.

4. I applaud the use the Taylor diagrams. The clear and concise explaination of what would be a perfect score is greatly appreciated by those of us who do not use these diagrams regularly.

5. Section 13, line 575: is it possible the term ‘long-term data’ should be changed to ‘Sample data’ to match the nomenclature in the online data repository?

6. I’m not sure the Appendix figures were meant to be interspersed throughout the References. I am sure the typesetter will fix this, but it is worth double checking to avoid it happening again.

Specific comments – Data set

7. On the NOW-23 repository page, I would not say the statement ‘Examples of using the HSDS Service to Access NOW-23 data. Contains resources that will help users view the data found in this submission.’ is accurate. The link that opened for me was for the WIND Toolkit page in the rex documentation. It may be true that these resources will also allow users to explore the NOW-23 dataset, however, since this is a substantial update of the resource, I would expect the link to send me to a dedicated page detailing how to use it specifically for the NOW-23 dataset. Even if the page is simply copied and elements of the text are changed to make it relevant to the NOW-23 dataset, it may not be necessary to construct an entirely new page. It would be pertinent to update other pages in the rex documentation to reflect its use with NOW-23 (e.g., the home page).

8. What has been done about outliers in the dataset? The first sample data I opened (‘Virginia_lidar’) had a boundary layer height maximum of 6553.5 m, which appeared, when compared to the rest of the sample data, as definitely an outlier (but I could be wrong).

Technical corrections

9. Table 14: Should this entry for WRF4 match the WRF4 from either Tables 12 or 13? If so, LSM should be NOAH-MP. At present, it appears to match the entry in Table 12 for WRF2.

10. Figure 18: Please add the period covered in the caption (I.e., is it for the entire 20 yr. period?).

11. Figure B4: add ‘WRF” to match the other figure captions of the same nature.

12. Figure C1: same comment as #10. Is this for the entire period for each data set, or a shorter period where there is concurrent data.

Reference corrections

13. Shamarock et al. (2019): is there a DOI for this entry available?

14. Wilczark et al. (2015): I believe all authors must be listed (although I assume the copy editor will confirm this).

15. Hahmann et al. (2020): same comment as above

16. Krishnamurthy et al. (2023): is there an updated entry for this reference?

Review of “The 2023 National Offshore Wind data set (NOW-23)” by Nicola Bodini and colleagues. Submitted to ESSD, Manuscript number essd-2023-490

Scope

The scope of the manuscript is well-suited for this journal.

Originality

The question of originality is not critical for a data journal. However, this is very important, and it would encourage other groups to publish the description of their datasets in this manner. 

Scientific rigor

In general, the scientific rigor is adequate. However, I have a few questions and clarification points below.

1. The wind time series uses a 5-minute temporal resolution. Have you checked that the spectra contain energy at this time scale? If not, this should be mentioned somewhere. 
2. L104-109: The ERA5 reanalysis uses the OSTIA SST product. This should be mentioned, and unsurprisingly, forcing the WRF model with the same SST is often advantageous. 
3. How can you explain the observed wind profile in Figure 13? By the way, what is the source of SST for the Great Lakes?  Are these points treated as lakes or seas? 
4. L380. It is not clear what “overestimates atmospheric stability” means. More stable?  Values of temperature gradients?  Please clarify. 

Writing 

The writing is clear and well-structured. 

Length

The manuscript is a bit long.  Also, the structure is tedious, looking more like a report than a scientific paper. I suggest removing the sub-sections for each zone, which often contain only one or two sentences. The eight regions could be combined according to each ocean basin. 

Figures and tables

I suggest redrawing most of the wind profiles and Taylor diagrams.  It is not possible to distinguish the various simulations.  I suggest using a logarithmic y scale in the profiles, which is standard in wind energy. For the Taylor diagrams, it is possible to zoom in the relevant region so that the differences between the runs are highlighted. See, for example, https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2017JD027504.

Are tables 2 and 12 the same? Could the tables be compressed to shorten and facilitate reading? I suggest one table with all the possible ensemble members and a final column stating which domain they are used in.

I also suggest that all figure captions contain the height and time period used in the validation (e.g., Figs 3, 4, 8, 9, 11). The period needs to be added in the caption of Figure 18.

Title 

The title is short and informative.

Abstract

The abstract is concise but contains most of the relevant information. The years and the model used in the simulations should also be included to facilitate future searches. 

References

The references are relevant and comprehensive.

Recommendation

I recommend publishing the article once the suggestions regarding the structure and the figures’ redrawing are considered. 

Minor corrections:

1. Please substitute “WRF” with “WRF model” or “the WRF model” when appropriate. 
2. L79: consider -> considered. Maybe the rest of the sentence should also be in the past?