Accurate Storm Surge Prediction with a Parametric Cyclone and Neural Network Hybrid Model

Round 1

Reviewer 1 Report

Thanks for the author's efforts to work on this manuscript. I have some comments and concerns for the authors' list below:

1. In this paper, you develop a hybrid NN method to predict storm surge. Also, you collect many cases and list them in Table 1. But why do you only discuss the case of Korsa & Haitang? What are the results for other cases?
2. Again, this paper adds and compares the parametric cyclone with the original BKPNN method published in 2020. What's the new or significant improvement? Because I thought the original BKPNN already performed pretty well. Why do you propose the hybrid method? Figure 12 shows not much progress between SLPVRF & E4 during the validation. Do you think these results are very significant?
3. I would suggest the authors present different case results in this paper or add statistical results from all your cases, and it would help convince the reader.
4. Another important question is, why not get a real dynamic model TC result to test? Any difficult? It would be interesting to look at the result.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Thank you for submitting your manuscript to the Water journal. Generally, the topic fits into the scope of the journal and the
structure respects scientific best practise. From my point of view,
the manuscript is of good quality and is written in a good and
understandable style. I have only minor comments.

The introduction is fine. In the section "Materials and Methods",
I strongly recommend to place in the beginning of that section a
flow chart that illustrates the methodology in working steps.
Regarding calibration and validation of the model should be given
more detailed information. Further, I recommend to include another
scheme that illustrates the structure of the applied model. I strongly recommend to separate the section "Results and discussion"
into two separate sections for higher clarity.

In the discussion section, please refer to the reliability and uncertainty of the modeling results. In the conclusions, in addition
to summarising the actions taken and results, please strengthen the
explanation of their significance. It is recommended to use
quantitative reasoning comparing with appropriate benchmarks,
especially those stemming from previous work.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper presents a parametric cyclone and neural network hybrid model for an accurate long lead-time storm surge prediction. The paper reads well and can be accepted after minor revisions. See the below:

1. Line 12: add an before 'accurate'
2. line 12-14: "For case application, the northeastern coastal region (Longdong station) of Taiwan was chosen as the study area and a total of historical 13 typhoon events was collected for model training and validation"-Re-write.
3. line 16: see the color
4. line 20: change the word 'gave'
5. Figure 1: remove the background grid.
6. Need to mention the advantages and limitations of the different models.
7. Figures 2-9: remove the background grid.
8. line 484-507: Need to revise. Only a few sentences will be enough. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I have no more comments. Thanks for, author's reply. And I am looking forward to more results. Also, I hope this method could apply to 2D instead of the 1D series model.

Author Response

Please see the attachment.

Author Response File: Author Response.docx