Hindcasts of Sea Surface Wind around the Korean Peninsula Using the WRF Model: Added Value Evaluation and Estimation of Extreme Wind Speeds

Round 1

Reviewer 1 Report

This paper evaluates dynamically downscaled surface wind over the ocean against observations from satellite and buoys. The WRF model was used to hindcast from reanalysis dataset. Considering the importance of having accurate high spatiotemporal resolution wind surface dataset especially over the ocean that influences waves, the paper meets the scope of the journal and expected to be referred for its various applications. I believe the paper is well prepared and deserves acceptance. I only have few minor comments as below.

 

Minor comments:

 

It is not clear to me how gridded reanalysis dataset that provides initial and boundary conditions were used for data assimilation with the 3DVAR that was mentioned in discussion and abstract but not elsewhere.

 

Line 51-53 “many studies have investigated the quantitative added value of the boundary data obtained by dynamic downscaling to the RCM”: More studies could be considered in addition.

 

Author Response

Thank you very much for making very kind and thoughtful comments. We revised our manuscript based on your comments. We hope that revised version of our manuscript and replies are satisfactory. Please let us know if you have further questions. Detailed responses are attached

Author Response File: Author Response.pdf

Reviewer 2 Report

In this paper the authors aim to produce and verify long-term hindcast sea surface wind data with high spatiotemporal resolutions around the Korean Peninsula with the use of WRF model. Additionally, the added value of the hindcast data were evaluated and extreme wind speeds were also estimated. Overall, I find this study quite thorough and useful, but I recommend to revise some major points and I would also like to address a few explanatory questions. More specifically:

1. In section 2.1 the authors refer to a sensitivity test that is not shown. Was it performed by the authors for this study or it was part of a previous study of the authors?
2. The sensitivity test was performed only for PBL? The other parameterizations schemes used in this study have been tested for this area before?
3. In Figure 1, it would be more clear to the reader if the authors showed, either with two lines or with description in the caption, that the black rectangle is (b) map
4. The authors use different time periods for each step. For example, 1979-2017 (WRF hindcast data), 2008-2017 (DASCAT comparison against buoy), 1997-2017 (WRF comparison against buoy). Which one is the research period? Please provide an explanation for this
5. Just to clarify, the dynamical downscaling was performed directly from ERA-Interim 0.25^ox0.25^o (ICBC) or the simulation was initially performed in a horizontal resolution of 25km and then was nested for the downscaling?
6. I cannot totally understand the WRF part. The authors reinitialized the model each day (30h run) with a 6h spin up time and then combined each day to create the research period? Is 6h spin up time long enough for the model to stabilize and fully reach physical equilibrium state each day, in order to create a time period with data similar to a continuous non-partitioned run? Please provide an explanation for this
7. Discussion: Are there more studies in order for the authors to compare their results?

Author Response

We appreciate for making very kind and thoughtful comments. We revised our manuscript based on your comments. We hope that the revised version of our manuscript is satisfactory. Let us know if you have further questions or comments.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear authors,

please have a look on my general and specific comments. They are attached.

Comments for author File: Comments.pdf

Author Response

The authors very appreciate the kind editing and the opportunity to revise the manuscript. We also sincerely appreciate your great efforts and valuable suggestions on the manuscript. We hope that the revised version of our manuscript and replies are satisfactory. Please let us know if you have further questions. The revised manuscript and replies are attached.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The comments I provided in the previous stage were well resolved. I recommend accepting the manuscript.

Author Response

The authors appreciate your kind editing. Thank you very much.

Reviewer 2 Report

The authors have taken into consideration and addressed my suggestions and comments. The manuscript is now more clear and easy to understand.

Two minor word mistakes that I found:
Line 90: Consequentially -> Consequently
Line 100: "A discussion presented..." -> "A discussion is presented..."

And a small suggestion:
In section 2.3 (In-situ data) the authors could write in parenthesis the units of each parameter, or could also provide a table with the parameters and their units

Author Response

Thank you very much for making very kind and thoughtful comments. 

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear authors,

many thanks for addressing my concerns. The quality of the paper was clearly improved. However, some minor comments are left:

Results

I have problems to understand Fig. 2. I think in that in a.) the mean wind speed is shown and in b.) the wind speed variability based on DASCAT. This should be also clearly stated in the figure caption.

I realized now that the observational datasets (satellite and buoy observations) are also quite different. Thus, there is also a relatively large uncertainty regarding the reference data sets used for WRF evaluation. This is a common problem for evaluation of RCM simulations. Thus, I think that it is also important to work on the observation dataset to get a better reference dataset for model evaluation and to account the observational uncertainty. I think this could be also included in the discussion section if this has been not addressed, so far.   

By the way, the quantiles are not identical and the corresponding sentence in the description of the results is contradictory ( … identical quantiles, but bias … , … see also below my comments)

Fig.5, Tab.3 and 5: I would not use “unitless” in the labelling. A variable which has no unit is usually indicated either by [-], [./]. But please have a look on the guidelines for publication in atmosphere. However, it is more important to include units where units are also available e.g. for bias, mean, etc. They are still missing in Tab. 3 and Tab.  4.

The added value of the downscaled wind speeds in clearly visible. But why do you lose so much performance in the offshore areas/open seasons. Here the modified BSS is negative and I think no explanation is given in the paper.  

 

Discussion:

“Regardless of the area analyzed, a positive bias with a value of 415< 2 m·s-1was observed, and the quantile showed an identical distribution compared to the 416KMA buoy data.”

This is not true compared to Fig. 4. The quantiles are not the same, otherwise the red line must be completely on the 1:1 line. There are (strong) deviations especially for the extremes but also for the lower tails of the distribution. Please have look on the distribution based on a logarithmic scale for the wind speeds. The statement is also contradictory. If there is a bias, the quantile cannot be identical.

Author Response

Thank you very much for making kind and thoughtful comments. We appreciate your efforts on our manuscript. Please see the attachment.

Author Response File: Author Response.pdf