Prediction of Sea Surface Temperature by Combining Interdimensional and Self-Attention with Neural Networks

Round 1

Reviewer 1 Report

The authors developed LSTM models with attention-like mechanism to predict the sea surface temperature (SST) by the marine environmental factors including wind speed and sea level pressure. The proposed method can be applied to complement and extend the existing measured and calculated dataset. This is an interesting paper but a lot of the concepts can be explained more and better. In the following some comments

1. The manuscript needs extensive revision for language. There are many typos and grammar issues

-- Word spelling mistakes in the title ‘combing’, ‘Datsets’, ‘selction’ and ‘optimizaiton’ in the Figure 1.

-- Please check carefully if some abbreviation needs to be defined, like ERA5, ECMWF in the abstract, AVHRR, AMSR etc. in line 55 and 56

--Some abbreviations have been used again in the text, so you can remove them

2. In line 150, what does “interval” mean in this context ? And larger compared with what? Consider clarifying these points.

3. In line 319, it is not clear what “them” indicates here. Please clarify.

4. Use a different heading for subsection 4.2 instead of repeating “Results and Discussion”.

5. Please check the page range or DOI for references 2, 8, 19, 24, and 30.

6. As we know, marine environmental forecasting including the sea surface temperature is mainly based on the numerical model at present. What are the main disadvantages of data-driven models? Have you considered combining the data-driven models with numerical models?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The study " Prediction of Sea Surface Temperature by Combing Inter-Dimensional and Self-Attention with Neural Network" proposed a multi-variable Long Short-Term Memory (LSTM) model to predict a 1-month monthly mean SST by utilizing the historical SST, wind speed, and air pressure at sea level as input.

Overall the topic is practical and meaningful, But in the present form, the presentation of this work is inferior, and that poor presentation makes this work dull. The abstract of this work should be completely reconstructed. The language of this work already contains several flaws; later, long sentences without any citation/reference make this work dull and seem like a report. This work required extensive English editing and a gentle touch to consider this work.

1.      Although in the introduction section, the authors have presented valid arguments and reasons behind, But long sentences without any reference to those arguments seem these all the author's assumptions. Section 1. Introduction should be revised scientifically. 

2.      Even though the authors obtained good results, their presentation and discussion of the results are so dull and poor. So first, authors should discuss their results gently. Then it's strongly recommended to separate the discussion section and discuss/validate these findings with previous studies by generating valid arguments in line with their research and presenting gaps that the authors covered in their study.

3.      Am I missing something, or authors didn't report the rules of the proposed model?

4.      Ahead of the conclusion section, it's recommended to report the "Limitations" of this work.

5.      I am a bit confused. Figures 3, 4, and 5 presented in the datasets section are the author's results. And why does it need to be presented in the datasets where even authors haven't discussed/opened the methodology yet?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

-The paper should be interesting ;;;

-it is a good idea to add a block diagram of the proposed research (step by step);;;;;;

-it is a good idea to add more photos of measurements;;

-What is the result of the analysis?;;

-figures should have high quality;;; 

-references to figures in the text should be added;;;

-SI units should be added to figures for example Time [...]

-Figure 6 please add axes to figures what is what;;

-please add photos of the application of the proposed research, 2-3 photos (if any) ;;; 

-what will society have from the paper?;;

-Please compare the proposed method with other approaches/other methods;;

-references should be from the web of science 2020-2022 (50% of all references, 30 references at least);;;

-Conclusion: point out what have you done;;;;

-please add some sentences about future work;;;

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The author's responses are satisfactory. About the author's confusion about my previous Comment # 3:

 

"As the authors proposed a multi-variable long short-term memory (LSTM) model, my previous comment # 3 was about the limitations of that proposed method. But I already have found that answers in Line 116-117 ~ of the revised version."

Author Response

Please see the attachment.

Reviewer 3 Report

more photos of the application of the proposed method should be added;;

The paper should be interesting;;;

Author Response

Please see the attachment.

Author Response File: Author Response.pdf