A Novel Approach Based on Machine Learning and Public Engagement to Predict Water-Scarcity Risk in Urban Areas

Round 1

Reviewer 1 Report

 The study aimed to develop a novel approach in the urban water field to overcome the limitations and challenges of traditional approaches and take advantage of the good characteristics of ML, MCDA, and GIS, that were mentioned in previous studies.  the trained model was evaluated by the five ML models (NB, RF, KNN, SVM, and MLP) using a testing dataset via a confusion matrix. In general, the article is acceptable however some comments need the authors' attention:

- The introduction section lacks the essential competence in presenting literature, research novelty, and the research aims.

- Graphics of the article are sufficient and complete. However, an enhanced resolution is needed. For instance, the histogram is not in good resolution.

- The conducted methodology is good, however, more explanation of the adopted feature selection is needed. More information is required for this section.

- The results of the prediction statistical analysis reported in the section “Results and Discussion” need further explanation and proper justification.

- The manuscript is required a new section presenting the practical base of the current research findings.

- The article is required a solid recommendation for the water scarcity risk in urban areas.

-      The authors have to check the spelling and grammar structures throughout the paper. There should be some minor improvements to English grammar.

Author Response

Please, The Author's reply in the PDF file attached.

Author Response File: Author Response.pdf

Reviewer 2 Report

In the increasingly widespread urban living environment, water has an important impact on life maintenance, environmental health, and even cultural innovation. This paper was carried out in Nasiriyah City, with detailed method interpretation and result analysis, which shows special practical significance. However, I also proposed the following suggestions:

 

1. Please use the full name when the abbreviation（such as WSR）first appeared in the abstract.

 

2. PP is of great significance to accurately identify and predict the risk of WRS. It is suggested to further explain the organizational process and evaluation content of PP in the Methodology.

 

3. The presentation of Figures 6 to 10 has great potential for improvement. There are meaningless elements (plan patches, remote sensing backgrounds, etc.) and the text annotations are too small, which makes reading difficult

 

4. In my opinion, Tables 3 to 7 can be combined into a standard table.

 

5. Suggested that the styles and annotations of all the charts should be similarly standardized

 

6. The sources of water security risks include water quality, water supply, human needs, and social-economic indicators, etc. In my opinion, the risk of Water Shortage Risk cannot be equated with the Water Security Risk. It is suggested that the differences should be reflected in the article.

 

7. I am curious about how to calculate the Data accuracy (1m) of Site serving work in Table 1.

Author Response

Please, Author's reply in PDF File attached 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The authors adopted a novel machine learning approach to predict water scarcity risk in urban areas. The experiment process was designed to be complete and rigorous and this paper was well-written. However, more explanation and demonstration are required. My comments are as follows:

 

1). The multiple model indicators proposed by the author can be collected, but water scarcity risk is a comprehensive risk that cannot be directly measured. How did the author collect model sample data of water scarcity risk, and how did the sample data be processed and input into the machine learning model?

2) Several models compared by the author are commonly used machine learning models, and the parameters of the models need to be explained in the text.

3) The quality of the picture, especially the resolution needs to be improved.

4) The spatial applicability of the model needs to be analyzed in the discussion, especially whether the model with higher accuracy obtained in this study also achieves better accuracy in other water scarcity risk estimation studies.

Author Response

Please, The Author's reply in the PDF file attached.

Author Response File: Author Response.pdf