Spatiotemporal Flood Hazard Map Prediction Using Machine Learning for a Flood Early Warning Case Study: Chiang Mai Province, Thailand
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Editor Nikola Lučić,
It is my pleasure to review the manuscript: sustainability-2948918, entitled “Development of Flood Hazard Map Using Spatiotemporal Data Prediction”. This study proposed a framework which integrated LSTM technique for temporal prediction, IDM for spatial interpolation, and CNN for map generation, to generate a flood hazard map using spatiotemporal data. It is an interesting study with importance and significance to flood risk management.
However, major concerns were address to the structure and discussion of this study. Hence, I would recommend a major revision before this study to be further considered in the Atmosphere.
Please find my specific comments below:
1. It is suggested to revise the title to highlight the methods or findings of this study
2. The current structure of the manuscript is confusing.
It is strongly suggested to adjust the structure with following sections: 1. Introduction, 2. methods, 3 Results, 4. Discussions, 5 Conclusion.
3. Please make sure all the non-original equations can be supported by references
4. It is not recommended to place figures in the Conclusion section. Please move Fig. 6-11 to Results section
5. Please enlarge the figure font to improve the readability.
6. The results are informative and interesting. However, there is a lack of in-depth discussion, including the reason why the predicted flood did not perfectly match the real situation, and how this method can be applied into other occasions, as well as the significance of this study.
7. Please add doi links for all the references
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe paper conducts a spatiotemporal framework for generating flood hazard maps utilising LSTM, IDW, and CNN techniques. I was especially impressed with the clear and thorough way that paper explained the findings related to each section and suggested explanations. The data is sufficient to prove the conclusion and mechanism. This article can be included in publication after minor revisions.
1. May I request additional information regarding the selection criteria that were applied to the LSTM, IDW, and CNN models in your framework?
2. To what extent might the proposed framework operate in areas where spatial and temporal data are scarcer or less precise?
3. Have you encountered any particular obstacles or constraints while integrating the three methodologies (LSTM, IDW, CNN) that you believe future research could investigate and resolve?
4. Could the authors provide more information on how Evaluation? Additional relevant literature is possible. Authors can consult this reference when revising. “Process Safety and Environmental Protection, 2024, 183, 496-504.”
5. The content is extensive, but some reference literature is outdated (within 5 years). Citations for key publications should be complete. This section should also explain your work's novelty and applicability. I recommend authors consult "MDPI" and other related journals for the latest literature. Please limit sustainability citations to 30%. Authors can consult this reference when revising. Safety 2023, 9(4), 84
6. The article under review is a comparative study. The "Abstract" summarizes the article. The "Introduction" chapter provides enough background to summarize the issue. Clear and understandable descriptions of experiments. Due to its detail, Results and Discussions is the focus. The authors clearly know their research field. With the above revisions, this article meets scientific research paper requirements and should be published.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThis manuscript (sustainability-2948918) presents a framework for generating spatiotemporal flood hazard maps in Chiang Mai Province, Thailand, using LSTM for temporal prediction, IDW for spatial interpolation, and a CNN for map creation, demonstrating high precision in forecasting flood-related factors and in map validation metrics.
Sections of the manuscript, as well as its overall readiness for publication, are inadequate. The sections are poorly written, lacking essential components from the introduction to the conclusion (which are nonexistent), lacking coherence, and the statistical analyses are not robust. There is a deficiency in the description of presented elements and data, as well as a critical discussion of the results. What data currently exist on the topic?
Moreover, the figures, legends, and tables are not suitable in terms of formatting, structure, and description. The manuscript exhibits a certain degree of similarity to other works in many sentences, and thus corrections and proper citations are necessary.
Additional keywords should be included, avoiding repetition of those already mentioned in the title. The introduction should be articulated clearly to outline the hypothesis and objectives of the study.
The materials and methods section requires a theoretical framework; names of manufacturers for the equipment and software should be included as per the guidelines and instructions for authors.
All figure captions must be rewritten to accurately reflect the figures and the elements they depict. Currently, the descriptions are inadequate and uninformative. Abbreviations and sample numbers also need to be detailed.
The presented maps lack scale, elevation details, and legends indicating the elements depicted. Figure 5 should be a table, not a figure, and lacks necessary details on what it represents. It does not contain the necessary statistics to confirm the data, similar to Table 1, which is inadequate. Much of the information presented is unnecessary and irrelevant, without appropriate explanation or statistics. Figure 11 shows predictions and observations for only two classes, without drawing conclusions on the manuscript's topic.
Comments on the Quality of English LanguageThe manuscript requires extensive grammatical and spelling corrections.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors have addressed all the concerns, I have no more issues. I would recommend this manuscript to be accepted in the current version.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe manuscript (sustainability-2948918) has significantly improved; however, it still requires modifications to the captions. The authors need to adequately explain each figure and table, providing detailed descriptions of all elements presented, including the sample size and spatial variability (in the case of maps). The sections titled '2. Background' and '3. Proposed Framework' should be merged at the end of the introduction along with the stated objectives. In section '2. Material and Methods' must be written and formulated properly. Note that simply moving sections around or changing their titles is insufficient. The authors must rewrite them to meet the technical language standards required for a scientific article.
Comments on the Quality of English LanguageEnglish requires moderate modifications for grammar and spelling.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
