An Adaptive Surrogate-Assisted Simulation-Optimization Method for Identifying Release History of Groundwater Contaminant Sources
Round 1
Reviewer 1 Report
This manuscript presents an ideal
adaptive technique and an appropriate surrogate model for the current numerical model for identifying the history of groundwater contaminant sources. The authors represent two cases for identifying the release history of contaminant sources to investigate the model accuracy. The new method represents good results and effectively identifies the history of groundwater contaminant sources with a higher degree of accuracy. Additionally, the article represents an adaptive surrogate technique for reducing most computation. Therefore, I recommend accepting the manuscript with minor revisions.
Hereinafter, the minor recommendation for the authors in the text is only a sample of the issues.
In the abstract section lines 12 and 13, rephrase this paragraph. Use another word than “promising” for example “appropriate solution”.
- Cite this author in the introduction section
Eissa et al., 2018, represent an analytical model to address local and regional solute and contaminant transport in groundwater assuming steady-state flow as an initial condition in an isotropic and homogeneous medium.
Integrative management of saltwater intrusion in poorly-constrained semi-arid coastal aquifer at Ras El-Hekma, Northwestern Coast, Egypt. Groundwater for Sustainable Development 6 (2018) 57–70.
- In Line 104, “csnt” more degree of subscription for the “nt”, the same for line 106 “crnt”.
- I like Figure 2, it is a good and represented figure, however, be careful about the color resolution “Blue, Pink and brown”.
- Citations for Equations 2 and 3.
- In Figure 4, indicate in the figure the distance unit for the x and y axes.
- sed on what you have used the longitudinal dispersity αL is 40m, and the transverse dispersity αT is 5m.
- In Figure 5, use different symbols for each monitoring well better than different colors, or a mix between symbols and colors.
- In Line 291, The true release fluxes of the source are listed in Table 1. It looks like the authors missed Table 1. Table 1 is not inserted in the text.
- Line 332, indicates that you are dealing with a “homogenous” confined aquifer as you set one value for the hydraulic conductivity (18m/day).
- In Figure 7, indicate the unit for the upper scale which ranges from 0 to 45 (mg/l?? or what).
- Figure 9, is not clear. You have good space on the right side, enlarge the figure and use good font sizes for the x and y axes.
Best Regards
Author Response
Response to Reviewer 1 Comments
Thank you for your comments concerning our manuscript entitled " An adaptive surrogate-assisted simulation-optimization meth-od for identifying release history of groundwater contaminant sources" (ID: water-1692841). Those comments are very valuable and helpful for revising and improving the quality of this paper, and they are of important guiding significance to our research as well. The co-authors and I sincerely appreciate your approval of our innovations. Actually, we have revised our paper carefully including correcting some minor errors, enriching the dicussion about the obtained results. Revised portion is marked in the "Revised Manuscript". We sincerely hope the submitted version that meets the standards for publication. The corrections responding to each comment are as follows:
Point 1: In the abstract section lines 12 and 13, rephrase this paragraph. Use another word than “promising” for example “appropriate solution”.
Response 1: Thanks for this comment. We have corrected in Line 14 of the revised manuscript.
Point 2: Cite this author in the introduction section
Eissa et al., 2018, represent an analytical model to address local and regional solute and contaminant transport in groundwater assuming steady-state flow as an initial condition in an isotropic and homogeneous medium.
Integrative management of saltwater intrusion in poorly-constrained semi-arid coastal aquifer at Ras El-Hekma, Northwestern Coast, Egypt. Groundwater for Sustainable Development 6 (2018) 57–70.
Response 2: Thanks for the sharing with this paper. We have cited this paper in Line 37 of the revised manuscript.
Point 3: In Line 104, “csnt” more degree of subscription for the “nt”, the same for line 106 “crnt”.
Response 3: Thanks for this comment. We use the more understandable expression to replace the origin expression. Actually, the “csnt” is the two-dimensional array like “Aij”. The ‘n’ of “csnt” denotes the index of the moniter well while the ‘t’ denotes the index of the stress period.
Point 4: I like Figure 2, it is a good and represented figure, however, be careful about the color resolution “Blue, Pink and brown”.
Response 4: Thanks for the aggrement to this picture. And this comment really improve the readability of our manuscript. We have applied more obvious “Pink and Black” to correct it in Line 149.
Point 5: Citations for Equations 2 and 3.
Response 5: Thanks. We have added the citations in Lines 159 and 166.
Point 6: In Figure 4, indicate in the figure the distance unit for the x and y axes.
Response 6: Thanks. We have corrected it in Line 283.
Point 7: In Figure 5, use different symbols for each monitoring well better than different colors, or a mix between symbols and colors.
Response 7: Thanks for this comment. After several attempts, we are sorry about the fact that using a mix between symbols and colors or different symbols is more unaesthetic than the origin. But we adjust the colors of this figure in Line 300. It seems better than before.
Point 8: In Line 291, The true release fluxes of the source are listed in Table 1. It looks like the authors missed Table 1. Table 1 is not inserted in the text.
Response 8: We are sorry for this inappropriate writing. We have corrected it in Line 299.
Point 9: Line 332, indicates that you are dealing with a “homogenous” confined aquifer as you set one value for the hydraulic conductivity (18m/day).
Response 9: Thanks. We have revised it in Line 358, according to your comments.
Point 10: In Figure 7, indicate the unit for the upper scale which ranges from 0 to 45 (mg/l?? or what).
Response 10:We are sorry for it. Actually, it is ‘mg/l’. We have revised it in Line 382.
Point 11: In Figure 7, indicate the unit for the upper scale which ranges from 0 to 45 (mg/l?? or what).
Response 11: Thanks. We have used new picture in Line 432, according to your comments in Line 358. It looks more clearness.
At the end of this letter, thank you again for these full-constructive comments. Your comments truly improve readability of our work. We appreciate your time and efforts devoted to this letter and our manuscript. Finally, we sincerely hope the submitted version that meets the standards for publication.
Best Regards
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript entitled “An adaptive surrogate-assisted simulation-optimization method for identifying release history of groundwater contaminant sources” presents a research aimed to identify the history of groundwater contaminant sources with regards to obtain a higher degree of accuracy and shorter computation time than in traditional methods.
The authors stated, based on the literature review, that each surrogate model has its limitations under some particular situations. The intention of this research was to develop an adaptative method to identify the inverse contamination source issues and in results it was found that the proposed method has more advantages (higher accuracy, computation time, flexibility and feasibility) than the traditional one.
General comments:
The study has a methodological character. The discussion seems to be weak. Therefore, I suggest authors to confront their consideration with other scientific works, especially those the most actual and up to date. Currently, the authors referred to only 11 scientific works that were published within the last 5 years. This is only 27% regarding the entire reference list presented in this paper.
The authors also concluded that the performance of the proposed method is disturbed by the increasing dimensions of the problem. So the question is to what extent of the problem the proposed method could be valid and representative for the problem analysis?
Minor comments:
- Line 159: Is the unit of potentiometric head expressed as T-1 ? Or should it be [L]?
- Line 283: How did the authors set the longitudinal dispersity αL as 40 m, and the transverse dispersity αT as 5 m? What method of the assessment did they use?
- Line 337: The authors mentioned that the porosity of the aquifer is 0.25. Did the authors use the effective porosity value?
- Did the authors consider the value of the specific yield Sy at any stage of the analysis?
Author Response
Response to Reviewer 2 Comments
Thank you for your comments concerning our manuscript entitled " An adaptive surrogate-assisted simulation-optimization meth-od for identifying release history of groundwater contaminant sources" (ID: water-1692841). Those comments are very valuable and helpful for revising and improving the quality of this paper, and they are of important guiding significance to our research as well. Actually, we have revised our paper carefully including correcting some minor errors, adding the lastest citations, and enriching the dicussion portion of our manuscript. Revised portion is marked “red” in the "Revised Manuscript". We sincerely hope the submitted version that meets the standards for publication. The corrections responding to each comment are as follows:
Point 1: The study has a methodological character. The discussion seems to be weak. Therefore, I suggest authors to confront their consideration with other scientific works, especially those the most actual and up to date. Currently, the authors referred to only 11 scientific works that were published within the last 5 years. This is only 27% regarding the entire reference list presented in this paper.
Response 1: Thanks for this comment. It significantly guides us to improve our work. We are deeply aware of the shortcomings of the discussion portion. Actually, we have done some additional works during the submission. And we have refered many high-qulaity paper published on JH, AWR and so on.
For the reviewer’s piont to “The discussion seems to be weak.”, we have tried our best to enrich the dicussion portion of the lateset manuscript. The actions can be concluded:
- To fully investigate the performance of the proposed method, we newly add two compared methods which use the predetermined surrogate. The introduction of them is in Line 256 of the latest manuscript. In the latest manuscript, there are four compared methods to study the advantages and disadvantages of the proposed method.
- We have added the detailed analysis for the representative solutions obtained by all methods. For Case 1, the revision is in Lines 326-347 of the latest manuscript. For Case 2, the revision is in Lines 401-408 and Lines 433-442 of the latest manuscript.
- We aslo add more visualized picture to support our discussion, including Figure 6 in Line 345, Figure 10 in Line 445.
- To better show the basic information of our proposed cases, we conclude the information with the form of Table. For example, Table 2 in Line 285 and Table 7 in Line 367.
For the reviewer’s piont to “This is only 27% regarding the entire reference list presented in this paper.”, we have updated some latest citations to our paper. Now, the percentage is 38%. Thanks to this comment. We have learned more writing skills from these scientific works and revised our manuscript better.
Point 2: The authors also concluded that the performance of the proposed method is disturbed by the increasing dimensions of the problem. So the question is to what extent of the problem the proposed method could be valid and representative for the problem analysis?
Response 2: Thanks to this comment. We think that this comment is very good question. In two proposed cases in our manuscript, the proposed method find the exact solutions in Case 1 while only obtaining the approximated solutions in Case 2. Therefore, the increasing dimensions of the problem truly affects the performance of the method. But in Case 2, from Figure 10 in Line 445, although the solutions obtained by AST-SOM are not exact, the simulated values fits to the observed values. The approximated solutions do not influence practical application. Actually, in our experiments, the proposed methods can effectively solve the problems with dimensions up to 30. We think that high-dimensional problems will involve uncertainty analysis for unkonwn parameters. So we consider that your comment is a very sagacious question. It may contain hopeful future. Nowadays, we have not solved this problem that high dimensions causes “equifinality for different parameters” phenomenon. But we have tried our best to read related latest papers. And we have listed it to be slolved in our furture work (in Line 472).
Figure 10. The errors between observed and simulated values for solutions obtaining the best value.
Point 3: Line 159: Is the unit of potentiometric head expressed as T-1 ? Or should it be [L]?
Response 3: We are sorry for this mistake. We have corrected in Line 161 of the lateset paper.
Point 4: How did the authors set the longitudinal dispersity αL as 40 m, and the transverse dispersity αT as 5 m? What method of the assessment did they use?
Response 4: Thanks. We only set the longitudinal dispersity and the transverse dispersity, referring by the range of study area and some papers.
Point 5: The authors mentioned that the porosity of the aquifer is 0.25. Did the authors use the effective porosity value?
Response 5: Yes, we use the effective porosity value. We are sorry for our poor presentation about this. We have clearly presented these basic information in the form of Tables. For example, Table 2 in Line 285 and Table 7 in Line 367.
Point 6: Did the authors consider the value of the specific yield Sy at any stage of the analysis?
Response 6: No. The proposed cases are two confined aquifers. So, we consider the Storage coefficient Ss. We are sorry for our poor presentation about this. We have clearly presented these in the Table 2 (Line 285) and Table 7 (Line 367).
At the end of this letter, thank you again for these full-constructive comments. Your comments truly improve our work, including correcting some minor errors, adding the lastest citations, and enriching the dicussion portion of our manuscript, and so on. We appreciate your time and efforts devoted to this letter and our manuscript. We sincerely hope that the submitted manuscript can meet your standards.
Best Regards
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors have improved the manuscript significantly and almost all of my concerns have been fully adressed. Nevertheless, it is not clear for me what does it mean that "We only set the longitudinal dispersity and the transverse dispersity, referring by the range of study area and some papers". What papers did the authors mention? Are they cited in the main text?
Best regards
Author Response
Response to Reviewer 2 Comments
Dear Reviewer 2:
Thank you for your comments concerning our manuscript entitled " An adaptive surrogate-assisted simulation-optimization meth-od for identifying release history of groundwater contaminant sources" (ID: water-1692841). We sincerely appreciate for your approval to our previous revision. Anyway, the improvement of the manuscript is inseparable from your full-constructive comments.
For the only comment:
Point 1:
It is not clear for me what does it mean that "We only set the longitudinal dispersity and the transverse dispersity, referring by the range of study area and some papers". What papers did the authors mention? Are they cited in the main text?
Response:
We are sorry for the unclear presentation for the method to identify the longitudinal or transverse dispersity. Actually, in our work, we use the hypothetical cases. As we know, the dispersity is significant difference depending on the aquifer media. So, we refer some representative papers from JH, AWR to design our cases including valuing the longitudinal dispersity and the transverse dispersity. The referred papers are listed as below [1–4]. And we have stated them in Lines 250 to 251 of the latest manuscript.
- Zhao, Y.; Qu, R.; Xing, Z.; Lu, W. Identifying Groundwater Contaminant Sources Based on a KELM Surrogate Model Together with Four Heuristic Optimization Algorithms. Adv. Water Resour. 2020, 138, 103540, doi:10.1016/j.advwatres.2020.103540.
- Zhao, Y.; Lu, W.; An, Y. Surrogate Model-Based Simulation-Optimization Approach for Groundwater Source Identification Problems. Environ. Forensics 2015, 16, 296–303, doi:10.1080/15275922.2015.1059908.
- Ayvaz, M.T. A Linked Simulation–Optimization Model for Solving the Unknown Groundwater Pollution Source Identification Problems. Journal of Contaminant Hydrology 2010, 117, 46–59, doi:10.1016/j.jconhyd.2010.06.004.
- Xing, Z.; Qu, R.; Zhao, Y.; Fu, Q.; Ji, Y.; Lu, W. Identifying the Release History of a Groundwater Contaminant Source Based on an Ensemble Surrogate Model. J. Hydrol. 2019, 572, 501–516, doi:10/gnp6xm.
At the end of this letter, thanks to your time and efforts devoted to our manuscript, again. We really learn a lot from this pleasant experience with your selfless assistance.
Best regards
Author Response File: Author Response.docx
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.
