Uncertainty Analysis of Numerical Simulation of Seawater Intrusion Using Deep Learning-Based Surrogate Model

Round 1

Reviewer 1 Report (Previous Reviewer 2)

the authors attempt to analyze the uncertainty related to numerical simulation of the seawater intrusion. the authors should consider major issues before accepting the current version for publication.

1-Introduction: the authors should improve the literature review to clarify the originality of this work by citing the relevant works as suggested in the attachment

2-Materials and methods. the author should describe the study area and datasets separately.

3-please merge some figures as suggested in the attachment

4-please add a discussion of the results 

5-please revise the conclusions.

6-please see the attachment for many other comments

 

 

 

Comments for author File: Comments.pdf

Author Response

Reviewer#1：Thank you for your comments. We have revised the whole manuscript according to your specific suggestions (attachment).

1. Introduction: the authors should improve the literature review to clarify the originality of this work by citing the relevant works as suggested in the attachment

Thank you for your suggestions. We have added references and descriptions of previous studies and revised this part.

2. Materials and methods. the author should describe the study area and datasets separately.

Thank you. We have revised this part according to your suggestions

3. please merge some figures as suggested in the attachment

It has been revised according to the requirements in the attachment.

4. please add a discussion of the results

We accept your suggestion, please check it in the revised manuscript.

5. please revise the conclusions.

We synthesized the opinions of several reviewers and revised the conclusion.

6.please see the attachment for many other comments

For other details, We have modified them one by one according to the notes in the attachment.

• Lines 40:The factors are locations of cracks and hydrodynamic parameters.
• Regarding the integration of pictures, I think your suggestion is very reasonable. I will unify and merge the editing links to be published later (according to the specific size requirements of publishing editors).

Author Response File: Author Response.docx

Reviewer 2 Report (New Reviewer)

The manuscript makes an in-depth study on the problem of seawater intrusion under the condition of climate change. Considering the impact of uncertain factors on seawater intrusion, it is innovative and practical. It is recommended to receive it after minor revision. Suggestions for modification are as follows:

1 The author should clarify the source of uncertainty factors, especially the impact of climate change on seawater intrusion, and there should be an independent paragraph in the article.

2 The accuracy of the simulation model is the guarantee of the availability of the surrogate model. It is necessary to supplement the discussion of verifying the reliability of the simulation model.

The simulation model should be used as a separate part, instead of being placed in the part of “Establishing surrogate models of the simulation model”.

3 In Figure 6, “boundary of known flow” is the boundary of groundwater flow, while “zero flux boundary” is the boundary of ground water solute transport. Please express the boundary conditions as solute transport boundary conditions.

4 In Table 1, range of the pumping capacity of well group 1 is 273.00, and that of well group 3 is 300. 273 and 300 is not a range, are the ranges 0-273 and 0-300?

5 There are too many pictures in the manuscript, so it is suggested to put Figure 8-12 in one figure, and it is suggested to make the same modifications to Figure 13-17 and Figure 18-19.

6 The part of conclusion should be reorganized more detailed and scientific.

7 Many analysis results have been obtained through the research. The following contents should be added: How to reduce the salinity effect in your study area? What method will you adopt in your next research?

Author Response

Reviewer#2:Thank you for your comments. We have revised the whole manuscript according to your specific suggestions (attachment).

1 The author should clarify the source of uncertainty factors, especially the impact of climate change on seawater intrusion, and there should be an independent paragraph in the article.

Thank you for your advice. We have added this description on lines 53-54.

2 The accuracy of the simulation model is the guarantee of the availability of the surrogate model. It is necessary to supplement the discussion of verifying the reliability of the simulation model.

In the manuscript, the content of reliability analysis of alternative models is specially set, and the position is located in lines 294-323.

3. The simulation model should be used as a separate part, instead of being placed in the part of “Establishing surrogate models of the simulation model”.

Thank you for your suggestions. We have set up a separate chapter "establishment of surrogate model" to better highlight the corresponding contents.

4 In Figure 6, “boundary of known flow” is the boundary of groundwater flow, while “zero flux boundary” is the boundary of ground water solute transport. Please express the boundary conditions as solute transport boundary conditions.

We have changed it to impermeable boundary.

5 In Table 1, range of the pumping capacity of well group 1 is 273.00, and that of well group 3 is 300. 273 and 300 is not a range, are the ranges 0-273 and 0-300?

The pumping capacity refers to the average value of many years, not a range.

6 There are too many pictures in the manuscript, so it is suggested to put Figure 8-12 in one figure, and it is suggested to make the same modifications to Figure 13-17 and Figure 18-19.

Thank you. Other reviewers have also raised similar questions, and I will make unified revisions according to the size requirements of subsequent publications editors.

7 The part of conclusion should be reorganized more detailed and scientific.

We have made necessary additions and adjustments to the conclusions. Thank you for your advice.

8 Many analysis results have been obtained through the research. The following contents should be added: How to reduce the salinity effect in your study area? What method will you adopt in your next research?

Thank you for your comments. We added the corresponding contents in the conclusion part.

Author Response File: Author Response.docx

Reviewer 3 Report (New Reviewer)

The study applied Monte Carlo method to analyze influence of randomness on the uncertainty in the results of numerical predictions of seawater intrusion. The overall content of the manuscript is relatively complete, discussing an interesting subject. However, there are still several issues to be addressed before considering publication.

Major comments:

1. There are still some language issues in the manuscript, I have listed some in the specific comments.

2. Structural problem. The manuscript in current form makes it difficult for readers to read. I suggest that the main text be divided into Introduction, Methods, Results, Discussion and Conclusions.

3. The introduction: From line 68-224. This part is more like a manual. I think the author's intention is to include this part in the introduction, which should introduce the latest research progress in related areas.

4. The manuscript lacks discussion which may include interpretation of your study results, discussion of limitations of the methods, and comparison with other research results.

5.Overall, very few references are cited. Some references are earlier, and it is suggested to add more recent ones so that the total number of references should be at least 40.

Specific comments:

1. There are many first-person descriptions in the full text. I suggest that the author revise it.

2. Line 20: Delete “in it”.

3. Line 46: “No study to date has examined predictive simulations of seawater intrusion by …”. I don't recommend using such a strong tone.

4. Line 68-224: I suggest compressing this part and adding peer research progress.

5. Which line do your results and discussions start from? Line 225? I still insist that the author revise the manuscript structure.

6. Line 233: Hydrogeological conditions are critical to groundwater models and the authors need to add the hydrogeological conditions and conceptual models of the study area.

7. Line 409: In front of the conclusion, authors need to add a real discussion.

8. The manuscript involves a lot of content, such as Deep learning application, surrogate modeling, uncertainty analysis, numerical modeling, seawater intrusion, and raising of sea water level, However, the corresponding conclusions are relatively few. It is suggested to increase the support of the conclusion part, and at least explain which work is auxiliary to other work.

9. As for the source of uncertainty, the relationship between uncertainty and randomness should be distinguished, and it is suggested to add explanations.

10. How can the sea water level be random？It is suggested to specify.

11. Some figures are of poor quality , like Figure 18 and 19.

12. Better to show some numerical results of your work in the conclusion section.

Author Response

Reviewer#3

1. There are still some language issues in the manuscript, I have listed some in the specific comments.

Thank you. We have revised them one by one according to specific opinions.

2. Structural problem. The manuscript in current form makes it difficult for readers to read. I suggest that the main text be divided into Introduction, Methods, Results, Discussion and Conclusions.

Thank you for your comments. We have optimized the chapters based on the suggestions of the reviewers of the Planning Commission.

3. The introduction: From line 68-224. This part is more like a manual. I think the author's intention is to include this part in the introduction, which should introduce the latest research progress in related areas.

I accept your suggestion. In this part, a comprehensive revision was made according to the third specific comments, including the addition of references and the review of research progress.

4. The manuscript lacks discussion which may include interpretation of your study results, discussion of limitations of the methods, and comparison with other research results.

As for the discussion of the results, we put this part in the result analysis part. The limitations of the method are discussed in the part of precision comparison of alternative models. In general, your comments are very practical. Thank you. 

5.Overall, very few references are cited. Some references are earlier, and it is suggested to add more recent ones so that the total number of references should be at least 40.

Thank you and accept your suggestions. In the process of revising the full text, we added the references of recent references.

Specific comments:

1. There are many first-person descriptions in the full text. I suggest that the author revise it.

OK, we have tried our best to correct and replace.

2. Line 20: Delete “in it”..

Ok.

3. Line 46: “No study to date has examined predictive simulations of seawater intrusion by …”. I don't recommend using such a strong tone.

There is few research report on predictive simulations of seawater intrusion by considering the rise in sea level in the context of climate change, where this is an important issue.

4. Line 68-224: I suggest compressing this part and adding peer research progress.

I accept your suggestion. In this part, a comprehensive revision was made according to the third opinion, including the addition of references and the review of research progress.

1. Which line do your results and discussions start from? Line 225? I still insist that the author revise the manuscript structure.

We have integrated the opinions of several reviewers. This manuscript is the analysis and discussion of the conclusion from line 325 to the conclusion. We hope that our adjustment of structure and content can meet your requirements.

2. Line 233: Hydrogeological conditions are critical to groundwater models and the authors need to add the hydrogeological conditions and conceptual models of the study area.

Thank you for your valuable comments. We added this part in lines 240-248.

7 Line 409: In front of the conclusion, authors need to add a real discussion.

Thank you. We added 438-442 lines of summary, and summarized and combed the contents of the previous part and put them in lines 325-338 for the overall discussion of the results.

8. The manuscript involves a lot of content, such as Deep learning application, surrogate modeling, uncertainty analysis, numerical modeling, seawater intrusion, and raising of sea water level, However, the corresponding conclusions are relatively few. It is suggested to increase the support of the conclusion part, and at least explain which work is auxiliary to other work.

Thank you for your comments. We added an introduction to the relationship between methods in the abstract, and also described the corresponding relationship between methods in the first paragraph of the uncertainty analysis.

9. As for the source of uncertainty, the relationship between uncertainty and randomness should be distinguished, and it is suggested to add explanations.

Uncertainty is the description of simulation results, which is caused by the randomness of inputs in the model.In line 51 of the manuscript, the title and subsequent paragraphs explain the source of uncertainty. In the simulation, the uncertainty of sea level rise is generalized as a random variable.

10. How can the sea water level be random？It is suggested to specify.

The future sea level height in this manuscript is predicted by other studies. Under the background of global warming, the future sea level rise height cannot be given a specific value, so it is considered as a random variable within the existing prediction range.

11. Some figures are of poor quality , like Figure 18 and 19.

We have replaced the corresponding picture

12. Better to show some numerical results of your work in the conclusion section.

Thank you for your advice. We have appropriately added the corresponding content in the conclusion. The specific data are more detailed in the analysis section.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report (Previous Reviewer 2)

The revised manuscript has been improved. However, the systematic review is still weak, notably the review about the application of ML models in predicting Chloride concentration in coastal aquifers. I suggest the authors enrich the systematic review to show the originality of their work. 

Author Response

Thank you for your suggestions. I have added relevant descriptions in the introduction and corresponding references are added. I hope my understanding of your suggestions is correct.

Reviewer 3 Report (New Reviewer)

I would still like to see the author's revisions to the manuscript structure, in the order of "Introduction, Methods, Results and Discussion, Conclusion". I mentioned this in my previous comment, but apparently the authors didn't revise it. To be honest, the current structure is a bit confusing.

Author Response

I accept your suggestions and have made necessary adjustments to the structure of the article. I hope it can meet your standards.

Author Response File: Author Response.pdf

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

Comments for author File: Comments.pdf

Reviewer 2 Report

Seawter intrusion is a serious  issue in managing aquifier systems. The authors proposed attempt to analyze the uncertainties associated to sea level rise as main uncertainty sources. they used Conventional Neural Netwotk to predict Cloride concentration as indice. the idea is good. However, there are several issues should be considred before publication.

1. Introduction: please check the begining of the introduction it is out of the scope. i think it is related to journal template. The systematic review about the predictiong of chloride concentration in coastal aquifier is absent. please enrich the references by citing relevant references. for instances, https://doi.org/10.1016/j.gsd.2018.11.014 and https://doi.org/10.1016/j.eti.2021.101641

2. M and M: the authors begin with the uncertainty of numerical models. What are these models? This became apparent unexpectedly. 1 reconsider the organization of this section. 2 Describe the models which are used to get the results presented in Figure 7. 3 describe the data used and their sources. What is this unit (104m³/a). let space between numerical value and unit throughout the manuscript. a flowchart can be added to illustrate the proposed methodology and to merge the figures 1 to 5 in one figure.

why you used Chloride concentration as limit to decide that there is Seawater intrusion?

3. results. Figures (from 8 to 12) and (From 13 to 17) can be merged in two figures to optimise reading. the resolution of figures 18 and 19 should be improved.

*. Discussions: There is no discussion of the results 

4.Conclusions: "In the uncertainty analysis by simulation model, the processor of the computer is i7- 444 4790k (4.0ghz), the memory is 16GB, and the system is win7 (x64). Running gms10.0 for 445 solving the simulation model takes about 370s" this is not a conclusion please deplace it to the M and M section.

Reviewer 3 Report

Review the manuscript entitled “Uncertainty analysis of numerical simulation on seawater intrusion based on Deep Learning surrogate model.”

 

This manuscript adopted SEAWAT program to investigate the seawater intrusion due to the variations of sea level caused by climate scenario. The writing, structure, and content are very poor and the novelty and innovation are poor, too. The authors need to pay more attention to the manuscript content before it is submitted for review. Some comments and suggestions are listed below for reference.

 

Comments

1.      The abstract does not include any results.

2.      The first several sentences in the introduction are the content of the sample of introduction. The following content of the introduction does not match the title and text and it does not provide the motivation of this study. This section needs to be totally rewritten.

3.      The materials and methods section includes some results. They should be separated clearly. The content also does not have a good structure and not easy to follow.

4.      The figures are not suitable. For example, figures 1 and 2 are not necessary. Figures 6 and 7 do not include sufficient information. Figures 8-17 can be combined. Figures 18 and 19 can have higher resolution.

5.      The results section is very confusing. It is not clear where do the results come from? The figures are simulated by SEAWAT or deep learning surrogate model? How does the surrogate model work is also not clear.

6.      The conclusions section also very confusing. Why provide the compute comparison? They are not mentioned in anywhere of the manuscript beyond the conclusion. The conclusions also do not recall to the problem of the title.