Characterizing Inter-Seasonal Meteorological Drought Using Random Effect Logistic Regression

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors present a characterization of drought using RELogRM and CFELogRM that is a topic worthy of investigation.

The introduction section should be reorganized, the last paragraph is too long and includes different ideas.

The discussion section should be improved, and the comparison with other techniques should be adequately presented as well as the advantages of the proposed methods. 

The conclusion section should be improved, its contents seem to correspond to the discussion section.

Some of the references must be updated.

Author Response

Comment 1: The authors present a characterization of drought using RELogRM and CFELogRM that is a topic worthy of investigation.

Reply: Thank you for your time, comments, and positive opinion on our study. We did our best to improve our manuscript using your comments. 

Comment 2: The introduction section should be reorganized, the last paragraph is too long and includes different ideas.

Reply: The introduction section was revised.

Comment 3: The discussion section should be improved, and the comparison with other techniques should be adequately presented as well as the advantages of the proposed methods. 

Reply: The discussion section was revised accordingly.

Comment 4: The conclusion section should be improved, its contents seem to correspond to the discussion section.

Reply:  We revised the conclusion section.

Comment 5: Some of the references must be updated.

Reply: The references were updated. Please see the revised references section.

Reviewer 2 Report

Comments and Suggestions for Authors

Manuscript title: Characterizing inter-seasonal meteorological drought using random effect logistic regression

 

       In this manuscript, the spatiotemporal dynamics and inter-seasonal characteristics of meteorological drought across Ankara Province, Turkey, were investigated and compared using the Conditional Fixed Effect Logistic Regression Model (CFELogRM) and Random Effect Logistic Regression Model (RELogRM). The authors use various methods to test the reliability of both models. RELogRM is the optimal model for modeling fall-to-winter season drought dynamics across the study area. Overall, the data quality was reliable, the methodology used was reasonable, and the results obtained were agreeable. However, there are still some shortcomings that need to be resolved. So, I think there should be a revision before acceptance is considered. I will give my specific comments below:

 

Major Comments:

1.     The paper focuses on the specific methodology of the study. However, in the process of describing the equations, there may be some descriptive errors. For example, distinguishing between a and α.

2.     The results of the study are detailed in the paper, but the discussion section is slightly weak. It is recommended to strengthen the comparison with existing studies to improve the reliability of the study.

3.     What are the advantages of Conditional Fixed Effects Logistic Regression Models (CFELogRM) and Random Effects Logistic Regression Models (RELogRM) in helping to analyze the persistence of drought?

Specific Comments:

1.     Introduction

Main issues: In this section, many existing research findings are stated. However, you need to add references for important research findings.

       L40-41: You mentioned four types of drought. Important references should be added here.

       L62: Just quote it normally, deleting ‘e.g.,’. Or to describe it another way.

       L78-80: Such descriptions can be detrimental to the reader's understanding. It is recommended that the names of the people or organizations of the researchers be used instead of the numbers. [22], [13].

       L94: How can the results of the study contribute to the improvement of the early warning system in the capital province of Turkey?

       2. Study area and SPI data

       Figure 1 The scale in the uppermost part of the figure is best set to an integer distribution, with units added for elevation, which also need to be set to an integer distribution.

3. Methods

       The methods section is presented in great detail, but there are minor errors, and some parameters are not well described. The formulae are not formatted uniformly. It is recommended to check them carefully.

       L141: LRχ2, Wχ² . Inconsistency in the format of the number 2. This error appears in several places in the paper, check them carefully.

       L170 upper：Formula 8 What does ‘a’ stand for?

       Formula 23, Formula 24. ‘exp’ Inconsistent format, Italic and Regular.

4. Results

       L254: Indicate whether it's east or west, etc. (E, W, N, S)

       The results section is more informative, but the discussion section is slightly weaker.

5. Discussion

       The discussion section needs to be described in more detail. Deepen the links with other studies.

6. Conclusions

        Revise the conclusions as appropriate, taking into account the changes made to the discussion section.

 References

The references do not reflect the latest research progress. Please add some latest references.

1.    Chen, F., Wang, S., Dong, Q., et al. (2024). Role of Pacific Ocean climate in regulating runoff in the source areas of water transfer projects on the Pacific Rim. npj Climate and Atmospheric Science, 7(1), 153.

2. Esper J, Torbenson M, Büntgen U (2024) 2023 summer warmth unparalleled over the past 2,000 years. Nature 631, 94-97.

3.    Chen, F., Man, W., Wang, S., Esper, J., Meko, D., Büntgen, U., ... & Chen, F. (2023). Southeast Asian ecological dependency on Tibetan Plateau streamflow over the last millennium. Nature Geoscience, 16(12), 1151-1158.

4.   Zhao, X., Chen, F., Seim, A., Hu, M., Akkemik, Ü., Kopabayeva, A., ... & Kelgenbayev, N. (2024). Global warming leads to growth increase in Pinus sylvestris in the Kazakh steppe. Forest Ecology and Management, 553, 121635.

 

Comments on the Quality of English Language

English must be revised

Author Response

General Comment: In this manuscript, the spatiotemporal dynamics and inter-seasonal characteristics of meteorological drought across Ankara Province, Turkey, were investigated and compared using the Conditional Fixed Effect Logistic Regression Model (CFELogRM) and Random Effect Logistic Regression Model (RELogRM). The authors use various methods to test the reliability of both models. RELogRM is the optimal model for modeling fall-to-winter season drought dynamics across the study area. Overall, the data quality was reliable, the methodology used was reasonable, and the results obtained were agreeable. However, there are still some shortcomings that need to be resolved. So, I think there should be a revision before acceptance is considered. I will give my specific comments below:

Response: Dear reviewer, Thank you so much for addressing the potential points in our manuscript. We appreciate your valuable comments and understand the importance of discussing scientific points with reviewers. We've carefully revised the manuscript and made all the modifications as you commented. 

Major Comments:

1. The paper focuses on the specific methodology of the study. However, in the process of describing the equations, there may be some descriptive errors. For example, distinguishing between a and α.

Response: We have corrected this from the formula. It was α, which was by mistake written as a.

2. The results of the study are detailed in the paper, but the discussion section is slightly weak. It is recommended to strengthen the comparison with existing studies to improve the reliability of the study.

Response: Thanks for the suggestions. We have revised the discussion section substantially.

3. What are the advantages of Conditional Fixed Effects Logistic Regression Models (CFELogRM) and Random Effects Logistic Regression Models (RELogRM) in helping to analyze the persistence of drought?

Response: The major advantage of these techniques is their ability to capture both spatial and temporal aspects of drought persistence. In literature, several researchers applied logistic and random forest for monitoring drought persistence, however, these techniques don’t capture the temporal aspects of the observed data. The proposed methodology is more suitable for the spatiotemporal analysis, and for the first time, the methods were applied for spatiotemporal monitoring of meteorological drought across Ankara. Upon the comment, we mentioned this issue in the first paragraph of the revised discussion section.

Specific Comments:

1. Introduction

Main issues: In this section, many existing research findings are stated. However, you need to add references for important research findings.

Response: The references were added.

1. a)  L40-41: You mentioned four types of droughts. Important references should be added here.

Response: The text was revised and references were added now.

1. b) L62: Just quote it normally, deleting ‘e.g.,’. Or to describe it another way.

Response: Thanks for the suggestion. We corrected it accordingly.

1. c) L78-80: Such descriptions can be detrimental to the reader's understanding. It is recommended that the names of the people or organizations of the researchers be used instead of the numbers. [22], [13].

Response: We revised the relevant sentences accordingly.

1. d) L94: How can the results of the study contribute to the improvement of the early warning system in the capital province of Turkey?

Response: Upon the comment, we revised the relevant sentence and added the following sentence to the revised discussion section to make our goal clearer.

The outcomes of this analysis are anticipated to provide detailed insights into the SPI variation that affects the likelihood of drought in subsequent seasons. Therefore, feature early warning systems and mitigation strategies may use the findings to improve the forecasting accuracy of the evolved models

2. Study area and SPI data

   Figure 1 the scale in the uppermost part of the figure is best set to an integer distribution, with units added for elevation, which also need to be set to an integer distribution.

Response: The study area map was revised accordingly.

3. Methods

 The methods section is presented in great detail, but there are minor errors, and some parameters are not well described. The formulae are not formatted uniformly. It is recommended to check them carefully.

2. a) L141: LRχ2, Wχ² . Inconsistency in the format of the number 2. This error appears in several places in the paper, check them carefully.

Response: Both LRχ2 and Wχ²  are consistently presented in the revised manuscript.

1. b) L170 upper Formula 8 What does ‘a’ stand for?

Response: it was alpha and now corrected.

24. c) Formula 23, Formula 24. ‘exp’ Inconsistent format, Italic and Regular.

Response: Corrected, thanks for the suggestion.

4. Results
5. a) L254: Indicate whether it's east or west, etc. (E, W, N, S)

Response: Table 1 was corrected using Lat. N and Long. E

b) The results section is more informative, but the discussion section is slightly weaker.

Response: Now discussion section revised accordingly

5. Discussion

       The discussion section needs to be described in more detail. Deepen the links with other studies.

Response: we revised and enhanced the discussion section, and we provided a more detailed analysis that strengthens the connection between our findings and existing literature. We contextualized our results within the broader framework of drought persistence studies and compared them with similar research to highlight both corroborating and divergent findings. This included a discussion of the methodologies and statistical models used in related studies, emphasizing how our approach contributes new insights to the field. Additionally, we explored the practical implications of our findings for water resource management and policy-making, particularly in regions with similar climatic conditions. By identifying gaps in the current literature and suggested avenues for future research, we clearly articulated the novel contributions of our study and its potential impact on drought monitoring and mitigation strategies. Additionally, we highlighted specific areas for future research, building on the gaps identified in the discussion, to guide ongoing and future investigations in drought monitoring and related fields. Please see the revised discussion section.

6. Conclusions

        Revise the conclusions as appropriate, taking into account the changes made to the discussion section.

      Response: We revised the conclusions and clearly articulated the novel contributions our study makes to the scientific literature.

 References

The references do not reflect the latest research progress. Please add some latest references.

1. Chen, F., Wang, S., Dong, Q., et al. (2024). Role of Pacific Ocean climate in regulating runoff in the source areas of water transfer projects on the Pacific Rim. npj Climate and Atmospheric Science, 7(1), 153.
2. Esper J, Torbenson M, Büntgen U (2024) 2023 summer warmth unparalleled over the past 2,000 years. Nature 631, 94-97.
3. Chen, F., Man, W., Wang, S., Esper, J., Meko, D., Büntgen, U., ... & Chen, F. (2023). Southeast Asian ecological dependency on Tibetan Plateau streamflow over the last millennium. Nature Geoscience, 16(12), 1151-1158.
4. Zhao, X., Chen, F., Seim, A., Hu, M., Akkemik, Ü., Kopabayeva, A., ... & Kelgenbayev, N. (2024). Global warming leads to growth increase in Pinus sylvestris in the Kazakh steppe. Forest Ecology and Management, 553, 121635.

Response: Upon the comment, we studied these and some newly published papers and used the most relevant ones to update our references list.  Among the suggested papers, Esper et al. (2024) and Zhao et al. (2024) were used in the present study.

Reviewer 3 Report

Comments and Suggestions for Authors

Decision: Accept with Minor Revision

I have reviewed the manuscript titled "Characterizing inter-seasonal meteorological drought using random effect logistic regression". The manuscript investigates drought persistence and variability in Ankara Province, Turkey, employing advanced Binary Outcome Panel Data (BOPD) models, specifically Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM). By analyzing spatiotemporal drought patterns and applying rigorous statistical tests, the study demonstrates that RELogRM is effective in modeling drought dynamics across all four seasons. Key findings focus on how variations in precipitation (SPI-3) influence the likelihood of drought in subsequent seasons. The insights from this study enhance early warning systems and water resource management, offering valuable contributions to understanding and mitigating the impacts of drought, and I can only give an overview of the main issues:

Scope of the Study:

Strength:

The manuscript effectively integrates theoretical foundations with empirical results, providing a comprehensive approach that offers readers a well-rounded understanding of the topic. The study utilized data covering a substantial time period and an appropriate geographic scope for the nature of the research.

Improvement Point:

The scope of the study could be expanded to include additional geographic regions beyond Ankara, providing a comparative analysis across different areas and addressing drought challenges on a broader scale.

Empirical Aspect:

Strength:

The study utilized advanced Binary Outcome Panel Data (BOPD) models, such as Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM), which enhance the accuracy of analyzing drought dynamics across different seasons. Spatiotemporal drought patterns were analyzed, and rigorous statistical tests were applied, providing detailed insights into how precipitation variations affect the likelihood of drought in subsequent seasons.

Improvement Point:

I recommend to add more modern analytical techniques, such as Bayesian analysis or machine learning methods, it could be employed to enhance model accuracy and provide a deeper understanding of drought dynamics and persistence across seasons, thus improving the extracted results.

Results and Discussion:

Strength:

The study contributes to improving early warning systems and water resource management by providing insights into how changes in precipitation (SPI-3) affect drought, thereby enhancing strategies for mitigating climatic drought occurrences.

Improvement Point:

The study could be enhanced by incorporating additional climatic variables, such as temperature and humidity, into the models for a more comprehensive and accurate analysis. Expanding the scope to include areas beyond Ankara would provide comparative analysis across different regions, addressing drought challenges on a broader scale.

References:

 Improvement Point:

The study could be further enhanced by incorporating additional reviews of relevant previous research.

Conclusion:

In conclusion, the study is strong and distinguished due to its use of advanced Binary Outcome Panel Data (BOPD) models, such as Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM), which provide a precise analysis of drought dynamics across seasons. However, the study could be further enhanced by incorporating additional reviews of relevant previous research, expanding the scope to include other geographic areas, and adding variables such as temperature and humidity. Nevertheless, these additions do not diminish the study's excellence.

Author Response

Decision: Accept with Minor Revision

I have reviewed the manuscript titled "Characterizing inter-seasonal meteorological drought using random effect logistic regression". The manuscript investigates drought persistence and variability in Ankara Province, Turkey, employing advanced Binary Outcome Panel Data (BOPD) models, specifically Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM). By analyzing spatiotemporal drought patterns and applying rigorous statistical tests, the study demonstrates that RELogRM is effective in modeling drought dynamics across all four seasons. Key findings focus on how variations in precipitation (SPI-3) influence the likelihood of drought in subsequent seasons. The insights from this study enhance early warning systems and water resource management, offering valuable contributions to understanding and mitigating the impacts of drought, and I can only give an overview of the main issues:

Response: Thank you so much for your time, comprehensive review, and comments. We are pleased that you found our study effective in modeling drought dynamics and appreciate your positive response to our study. Upon your comments, we improved our manuscript.

Scope of the Study:

Strength: The manuscript effectively integrates theoretical foundations with empirical results, providing a comprehensive approach that offers readers a well-rounded understanding of the topic. The study utilized data covering a substantial time period and an appropriate geographic scope for the nature of the research.

Response: Thank you so much for highlighting the strengths of our manuscript. 

Improvement Point: The scope of the study could be expanded to include additional geographic regions beyond Ankara, providing a comparative analysis across different areas and addressing drought challenges on a broader scale.

Response: This is an important point that can be done for all the provinces, separately or integrated. However, adding another case study area requires substantial time and revision in the entire manuscript which was out of the scope of this research. Upon your comment, comparative analysis across different areas and addressing drought challenges on a broader scale were suggested in the revised conclusion section as a potential topic for future studies.

Empirical Aspect:

Strength: The study utilized advanced Binary Outcome Panel Data (BOPD) models, such as Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM), which enhance the accuracy of analyzing drought dynamics across different seasons. Spatiotemporal drought patterns were analyzed, and rigorous statistical tests were applied, providing detailed insights into how precipitation variations affect the likelihood of drought in subsequent seasons.

Response: Thank you so much for highlighting the strengths of our methodology. 

Improvement Point: I recommend to add more modern analytical techniques, such as Bayesian analysis or machine learning methods, it could be employed to enhance model accuracy and provide a deeper understanding of drought dynamics and persistence across seasons, thus improving the extracted results.

Response: Implementation of advanced Bayesian analysis or machine learning methods, can be considered as a potential topic for future studies.

Results and Discussion:

Strength: The study contributes to improving early warning systems and water resource management by providing insights into how changes in precipitation (SPI-3) affect drought, thereby enhancing strategies for mitigating climatic drought occurrences.

Response: Thank you so much for highlighting the strengths of our results and discussion. 

Improvement Point: The study could be enhanced by incorporating additional climatic variables, such as temperature and humidity, into the models for a more comprehensive and accurate analysis. Expanding the scope to include areas beyond Ankara would provide comparative analysis across different regions, addressing drought challenges on a broader scale.

Response: Thank you for your insightful suggestion. We agree that incorporating additional climatic variables, such as temperature and humidity, would enhance the models and provide a more comprehensive analysis. In our study, we focused on the Standardized Precipitation Index (SPI) as it is a well-established tool for analyzing drought based on precipitation data. However, we recognize that including additional variables could improve the accuracy and depth of the analysis. As mentioned in the discussion section, we plan to develop and utilize new indices and additional variables in future research to address these aspects. Furthermore, expanding the scope to include regions beyond Ankara would indeed allow for comparative analysis across different areas, offering valuable insights into drought challenges on a broader scale. We appreciate your feedback and will consider these aspects in our ongoing and future research.

References:

 Improvement Point: The study could be further enhanced by incorporating additional reviews of relevant previous research.

Response: We added previous research for comparison purposes.

Conclusion: In conclusion, the study is strong and distinguished due to its use of advanced Binary Outcome Panel Data (BOPD) models, such as Random Effect Logistic Regression (RELogRM) and Conditional Fixed Effect Logistic Regression (CFELogRM), which provide a precise analysis of drought dynamics across seasons. However, the study could be further enhanced by incorporating additional reviews of relevant previous research, expanding the scope to include other geographic areas, and adding variables such as temperature and humidity. Nevertheless, these additions do not diminish the study's excellence.

Response: Thank you for your positive assessment of our study. We acknowledge that incorporating reviews of relevant previous research, expanding the geographic scope, and adding variables like temperature and humidity would further enhance the study. As noted, we plan to integrate these elements into future research to provide a more comprehensive analysis and broader applicability. Your feedback is valuable and reinforces our commitment to advancing the study’s scope and depth while maintaining its excellence.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I am satisfied with the revisions. I think it can be published.

Comments on the Quality of English Language

English still needs further revision

Author Response

Comment: I am satisfied with the revisions. I think it can be published.

Response: Thank you so much for your time.