Application of RNN-LSTM in Predicting Drought Patterns in Pakistan: A Pathway to Sustainable Water Resource Management
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Editor,
The reviewer reviewed the manuscript (MS) titled “Application of RNN-LSTM in Predicting Drought Patterns in Pakistan: A Pathway to Sustainable Water Resource Management” submitted to Water journal in detail to meet the scientific requirements.
In this study the authors employed RNN-LSTM model to enhance the prediction capability for drought analysis in Pakistan. Authors highlighted this model as the first attempt to forecast drought in Pakistan. They also proposed this study as a guide for decision makers.
This is a country specific study and employs a method to forecast drought in Pakistan. Actually, a new method is not introduced or proposed in this study. It may be submitted as a case study since the authors focus on Pakistan. This study may have a potential for possible publication in the journal after a revision.
Subscripts or superscripts should be reviewed and corrected where necessary.
Keywords should be updated considering the study outcomes and methods.
The manuscript is well-organized. In section 1, the aim of the study is highlighted. However, local studies are employed in this study. It may be avoided and different study areas may be incorporated.
Limitations of the study, advantages or disadvantages of the adapted methodology may be discussed in the Introduction section.
Location map needs to be included in the study area section.
Figure captions must be included after figures and are referred in the text. The first figure is not considered.
Figures should be improved. Excell graphs should not be pasted in picture format.
Metric system units must be preferred.
How can one interpret the model results? For example, in figure 5/8, there may be assessment criteria to predict goodness of fit. Authors should employ some statistical parameters to evaluate the predictions.
Are the result sensitive to training parameters? What do the authors think about uncertainty of the predictions? Model architecture and training sets are vital for accuracy of the estimations.
How do the authors determine overfitting problem?
Conclusion section is required to make a conclusion on the study. This section is not a summary of the manuscript.
Finding from the study should be included in the Abstract.
Comments on the Quality of English LanguageGenerally fine, but it should be edited by professionals.
Author Response
Point-by-point Reviewer Responses are attached in the word file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsDear Authors,
my detailed comments are in the attached file
Comments for author File: Comments.pdf
Author Response
Point-by-point Reviewer Responses are attached in the word file.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsReview of Manuscript ID -water-2991816
This paper investigates the impact of climate change on drought frequency, intensity, and geographical distribution globally, with a focus on Pakistan. Employing Recurrent Neural Network-Long Short-Term Memory (RNN-LSTM) modeling, it aims to enhance drought forecasting using secondary data on rainfall, temperature, and water supplies. The study emphasizes the necessity of early warning systems for vulnerable areas, highlighting increasing temperatures, fluctuating rainfall, and population growth straining water resources. The findings underscore the importance of sustainable water management amidst climate challenges and demographic pressures, advocating for long-term water security and development in Pakistan. The paper is well structured, however there are some major concerns given below, which are need to be addressed by the authors to improve the quality of the paper.
1. While the abstract mentions the aim to improve water security and foster sustainable development in Pakistan, it does not delve into specific implications or recommendations arising from the study findings. Including a brief discussion of how the research outcomes can inform policy-making or resource management strategies would add depth to the abstract.
2. Line no. 36 – Unit for rising sea-level?
3. The flow of the introduction could be improved to strengthen the argumentation. Some sentences are disjointed and could be rephrased for clarity and coherence. Additionally, the transition between different topics, such as climate change, population growth, and water scarcity, could be smoother to maintain reader engagement.
4. The introduction concludes by mentioning the objective of the paper, which is to forecast water resources in Pakistan under climate change using the LSTM model. While this provides a general overview, it would be beneficial to clearly outline specific research objectives and hypotheses to guide the study.
5. Some paragraphs contain lengthy sentences and complex phrases, which may hinder readability and comprehension. Simplify the language where possible to ensure clarity and conciseness, particularly when explaining technical concepts or research findings.
6. While the literature review covers a broad range of topics, including population growth, climate change impacts, and water management policies, there is a lack of critical analysis and synthesis of existing literature. Provide more in-depth analysis and discussion of key research findings, identifying gaps, contradictions, or areas of consensus in the literature.
7. The literature review briefly mentions the gap in research concerning the application of advanced predictive models like Recurrent Neural Networks (RNN) with Long Short-Term Memory (LSTM) capabilities. However, this gap could be further emphasized to underscore the novelty and significance of the proposed research approach.
8. While the literature review cites some relevant sources, there is a lack of integration of recent studies or publications. Include recent research findings and developments in the field of water resources management and climate change adaptation to ensure the review reflects the current state of knowledge.
9. While the description outlines Pakistan's geographical coordinates and general climate classification, it lacks detailed information about specific regions or provinces within Pakistan. Providing more granular details about regional variations in climate, topography, and water resources would enhance the understanding of the study area.
10. The description mentions the classification of Pakistan into ten agroecological zones by the Pakistan Agricultural Research Council (PARC) but does not elaborate on the significance of these zones or their implications for water resources management. Including information about the unique features and challenges of each zone would provide valuable context for the study.
11. Incorporating maps or visual aids depicting the geographical distribution of precipitation, agroecological zones, and water resources infrastructure would aid readers in visualizing the study area and understanding spatial patterns and relationships.
12. The description references data from the Pakistan Agricultural Research Council (PARC) but does not specify the timeframe or sources of the data. Including references to recent studies or datasets would ensure the accuracy and currency of the information presented.
13. The methodology lacks detailed explanations of key steps, such as data preprocessing techniques, hyperparameter selection rationale, and model evaluation criteria. Providing more detailed descriptions of these aspects would enhance the reproducibility and transparency of the study.
14. The description of the normalization process is vague and lacks clarity on the specific normalization techniques used. Additionally, the rationale behind selecting three as the number of steps in each sequence (n_steps) is not adequately justified. More information on the normalization method and its implications for model training would improve the understanding of the data preprocessing stage.
15. The description of the LSTM model architecture is brief and lacks justification for the choice of specific parameters, such as the number of neurons in the LSTM layer and the activation function used. Providing rationale for these choices and discussing potential alternatives would enhance the robustness of the model architecture.
16. The methodology briefly mentions the training process over 200 epochs but lacks information on how the model performance was monitored during training and whether any regularization techniques were employed to prevent overfitting. Including details on the optimization process and strategies for preventing overfitting would strengthen the methodological rigor.
17. While the methodology outlines a proposed framework for forecasting, it lacks specificity on how the LSTM model will be utilized for forecasting and how the model outputs will be interpreted in the context of water resource management. Providing a more detailed explanation of the forecasting framework and its application to real-world decision-making processes would enhance the practical relevance of the study.
18. While the conclusion outlines the general aims of the study and the utilization of the RNN-LSTM model, it could benefit from providing more specific details about the study's key findings and insights. Including specific results or projections generated by the model would enhance the clarity and impact of the conclusion.
19. The conclusion briefly mentions the model's application in strategic planning for water resources management but does not delve into the potential policy implications or recommendations based on the study findings. Providing concrete recommendations for policymakers or stakeholders would enhance the practical relevance of the conclusion.
Comments on the Quality of English LanguageMinor improvements required
Author Response
Point-by-point Reviewer Responses are attached in the word file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsLocation map should be revised. Study area should be illustrated in Pakistan map.
Figure number and caption are not given in section 2.6.
Punctuations should be checked. the manuscript should be checked grammatically.
Author Response
Please find the attached responses to reviewer 1.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsDear Authors, after introducing the corrections, the text has become much better. Unfortunately, the lines in your answers do not match and this is difficult to check. Please, be careful about this when reviewing your next manuscripts.
You still haven't underlined the diagram on page 6/7. I propose:
2.6 Proposed Framework for Forecasting (as a diagram):
Figure 2 - start with a capital letter.
My biggest criticism concerns the Conclusions chapter. This is still a summary. Does your research really prove nothing? See the goal of your research.
Author Response
Please find the attached responses to reviewer 2.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsAccept
Author Response
We are thankful to worthy reviewer for providing valuable comments and suggestions, which have improved the quality of the manuscript.
