A Method for Estimating Ship Surface Wind Parameters by Combining Anemometer and X-Band Marine Radar Data

Round 1

Reviewer 1 Report (Previous Reviewer 1)

All my concerns have been addressed. The paper is suitable for publication.

Author Response

Thank you for recommending publication, there is no revision required.

Reviewer 2 Report (Previous Reviewer 2)

Thanks to the authors for detailed responses to comments and serious work on the article. All questions that arose during the first round of reviewing have been removed. The article has been significantly improved and can be published in the journal.

Author Response

Thank you for recommending publication, there is no revision required.

Reviewer 3 Report (Previous Reviewer 3)

This paper proposes a method to improve wind power measurement accuracy in the presence of structural interference and designs an approach that combines the wind speed sensor with X-band marine radar (RCRF) to further obtain wind power parameters. Although the author carefully answered the questions raised, there are still some issues that need to be addressed：

1、The topic selection has a certain meaning. However, remote sensing applications and innovation are still insufficient. The innovative nature of marine radar applications should be more fully elucidated.

2、As far as the data processing part is concerned, replacing BP with RF may not be innovative enough.

Author Response

Thank you for recommending publication, it has been revised in the manuscript.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report (Previous Reviewer 3)

Thank you for the careful response and modification. I think the authors have addressed the concerns.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This paper proposed a method of obtaining wind parameters, the authors claimed the method could reduce the impacts of structures on the wind retrievals. The work was properly designed, and the conclusions are informative. However, the novelty of this paper must be comprehensively discussed. Why such a combined estimation is necessary, what is the most significant difference between this method and the previous? The authors should clearly demonstrate the scientific motivation. What's more, the English writing should be improved.

Specific comments:

1. The wind parameters on ship surface (Line 27) and the parameters on the sea surface (Line 34) are different. They were not clearly defined in the introduction.

2. What is the relation between deviation analysis and bias correction (Line 35)? The authors just listed some papers for each methods. It would be useful if the authors could analyze the similarity and the difference between the two aspects and connect their characteristics to the method provided in this paper.

3." the marine radar measurement range is close to the airflow"(Line 66). The meaning of these words are not clear.

4. Usually, retrieving wind parameters from radar images need to be verified by anemometer data. In this paper, you modified the anemometer data by the radar retrievals (Line 80-81). How to explain the " paradox"?

5. It would be interesting if you could compare the performance of RCBP to the method only based on anemometers (Multi-anemometer optimal layout and weighted fusion method for estimation of ship surface steady-state wind parameters. Zhang et al. 2022, OE)

6. In Case A and E (Fig. 11-13), the performance of RCRF seems bad. What is the potential reason?

Author Response

RESPONSE PLEASE SEE ATTACHMENT

Author Response File: Author Response.pdf

Reviewer 2 Report

The article is devoted to improving the accuracy of methods for determining wind speed on a ship. The authors use numerical modeling and machine learning for this purpose. The article is written in too much detail, which makes it difficult to understand the key points of the work. The main notes are as follows:

1. In my opinion, the need for this work is not at all obvious. Wind speed and direction have been measured on ships for a very long time, as well as on helicopter and aircraft carriers, and I have not heard of the need to increase accuracy. The surface wind field is a turbulent 3-dimensional process in which velocity pulsations can reach tens of percent of the average value, and also applies to direction. It is necessary to reformulate the main goal of the work.

2. Remote sensing is the main topic of the RS journal. The article provides one radar image from which, apparently, the direction of waves, and not the wind, is extracted. In this journal I would like to see more remote sensing data, or at least their proper use.

3. In my opinion, in its current form, the article can be sent to another journal, for example JMSE, but cannot be published in this journal.

Author Response

RESPONSE PLEASE SEE ATTACHMENT

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper proposes a method to improve wind power measurement accuracy in the presence of structural interference and designs an approach that combines the wind speed sensor with X-band marine radar (RCRF) to further obtain wind power parameters. However, there are several issues need to be addressed:

1.       This paper uses the fusion of marine radar and anemometer to estimate ship surface wind field parameters. The topic selection has a certain meaning. However, according to the content of the manuscript, numerical simulation and machine learning methods are mainly used, and the marine radar parameters are only used as constraints. The application and innovation of remote sensing are insufficient.

2.       Table1 "Monitoring point" may need to be adjusted appropriately so that "Monitoring" is in one row.

3.       Page 4, Lines 140~141. How are the ship model used and the assumed environmental parameters used in CFD digital simulation determined? Is it consistent with the parameters of the ship and environment where the marine radar is located?

4.       Page 8 Line 230: “The RF wind direction and RF wind speed are taken as output”. However, the source or calculation method of RF wind direction and speed is not mentioned in the text.

5.       Page 9 Line 278. The error value of the training set and the error value of the test set should be added to determine whether overfitting occurs.

6.       The reason for choosing the random forest method is not explained in Section 3.1.2. Consider explaining the advantages over other machine learning methods and other boosting tree methods such as XGBboost.

7.       Figure 9. The meaning of different colors in the legend is not explained; the horizontal and vertical coordinate axes are marked with distance, but in fact, the radial direction in the figure represents distance. This label can easily lead to ambiguity.

8.       Figure 10. The readability of figure 10 needs to be further improved. What do the numbers (65m, 100m, 400m) in the picture represent?

9.       The variables used in Formula 14 are inconsistent with the variables expressed in the text. The meaning of the operators used in Equation 14 is not explained. Does ║·║ represent the norm? Is the second equal sign of the formula true?

10.   Formulas 16 and 17. As commonly used statistical indicators, MAE and MRE may be easily misunderstood by readers in subsequent analysis. Maybe consider increasing the θ subscript for MAE and the v subscript for MRE.

11.    Page 13 Lines 363~364. As a comparative method, RCBP should be introduced more carefully, especially the difference from the method in this paper.

Author Response

RESPONSE PLEASE SEE ATTACHMENT

Author Response File: Author Response.pdf