Uncovering the Spatiotemporal Evolution and Driving Factors of Flash Flood in the Qinghai–Tibet Plateau

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study systematically investigates the spatiotemporal evolution patterns and driving factors of flash floods in the Qinghai-Tibet Plateau region based on multi-source remote sensing data and historical disaster records. The research topic holds significant scientific importance and practical value. The overall framework of the paper is complete, but some details require further clarification and refinement. It is recommended to revise and improve the following content.

1. The paper mentions using historical flash flood records from 1950 to 2015, but the text (Section 5.1) also notes that research before 1949 is largely absent, and records before 1977 are scarce. Please clearly explain the sources, collection methods, and potential biases of the data from 1950 to 1977, and discuss whether the incompleteness of this early data may affect the robustness of long-term evolution trends (such as the exponential growth conclusion). It is suggested to briefly address this limitation in the Discussion section.
2. The temporal resolution (annual, monthly?) and spatial resolution of the eight driving factors mentioned in the text (such as soil moisture, Human Activity Intensity Index (HAII), Runoff Concentration Index (QCI), etc.) need to be clearly specified. Were these factors precisely matched with flash flood events in terms of spatiotemporal scales? For example, was soil moisture taken from a specific period before a flash flood event or as a monthly average? This directly affects the accuracy of the driving mechanism analysis.
3. For the Random Forest model, please supplement the justification for key parameter settings (e.g., number of decision trees, maximum depth, etc.). This will help readers assess the model’s optimization level and reproducibility.
4. Regarding the classification thresholds for the spatial change rate of flash flood density (Figure 9, 0.28×10⁻⁵, 1.44×10⁻⁴), please explain the basis for their determination to enhance scientific rigor.
5. When describing the migration trajectory of the center of gravity (Section 4.2.2), it is suggested to supplement the specific coordinates of the gravity center point of each month in the text or figure caption to make the trajectory description more accurate.
6. Please further discuss how soil moisture specifically promotes flash flood occurrence. Does the SHAP analysis reveal such effects?
7. The paper conceptualizes human activity as a weighted aggregation of land-use changes. Please specify more concretely which types of human activities (e.g., urban expansion, infrastructure construction, agricultural reclamation, etc.) were identified as key in the study.
8.  It is recommended to compare and discuss the main findings of this study (e.g., exponential growth trend, dominant driving factors) with relevant published research on the Qinghai-Tibet Plateau, highlighting the innovations and consistencies/differences of this study.

9. In the Conclusion section, it is recommended to briefly reiterate the most important research findings in bullet points and clearly provide specific suggestions for flash flood monitoring, early warning, and risk management to make the conclusion more targeted.
10. The overall language of the paper is fluent, but there are minor grammatical errors and spelling mistakes (e.g., "frequency" spelled as "frequence," "predominant" spelled as "predominate," etc.). Please proofread the entire text carefully.
11. Chart Optimization:

(1) It is recommended to label major river names in Figure 1.

(2) In Figure 2, it is suggested to indicate the unit of each parameter. In addition, is the maximum value of slope only 13.42?

(3) The labeling (a)(b) in Figure 10 needs to be checked for consistency with the description in the main text (SHAP as (a), Random Forest as (b)).

12. The reference format needs to be standardized and unified. Some entries lack complete information (e.g., volume, issue, page numbers). Please carefully review and supplement them according to the journal’s requirements.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The paper explores the spatiotemporal patters of flash floods and attempts to attribute these events to climatic and human factors. The topic is relevant for regional disaster risk management, but presents significant methodological issues related to data consistency, which need to be addressed before the paper can be considered for publication.

There's an inconsistency between the disaster inventory and datasets. The author's analyze flash floods from 1950 to 2015, but the Land Cover data used to determine the HAII is only available from 1985 to 2015. How did the Random Forest model account for the period prior to 1985? 

The Abstract and Results sections claim that flash floods show "exponential growth". It is highly likely that this trend reflects administrative capacity rather than a purely physical increase in their frequency.

The paper should not present this trend as uniquely physical and need to report this bias in the Abstract and Conclusions rather than just in the Discussion.

Another concern is that the analysis ends in 2015. The paper is considered for publication in 2025 and the data is a decade old. Given the fast climate changes, the lack of data from 2015 onwards significantly reduces the relevance of the results, therefore please update the dataset or justify the interval.

Comments on the Quality of English Language

Minor improvements

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Review Report-R-1

Journal	Remote Sensing (ISSN 2072-4292)
Manuscript ID	remotesensing-4101010
Type	Article
Title	Uncovering the spatiotemporal evolution and driving factors of flash flood in the Earth's Third Pole
Final decision	Major revision

 

Best regards,

01-25-2026

 

Dear authors,

Thank you for your work, it seems interesting to me. I have thoroughly examined the text, methodology, and findings of your publication "Uncovering the spatiotemporal evolution and driving factors of flash flood in the Earth's Third Pole,". The main criticisms and particular recommendations arranged to enhance the papers and scientific sound are listed below.

The evolution and causes of flash floods in the high-altitude, data-poor Qinghai-Tibet Plateau (TP) are a crucial topic covered in this work. To provide a comprehensive evaluation, it uses an impressive combination of machine learning (Random Forest and SHAP), historical records, and multi-source remote sensing data. To support the results, a few points need to be clarified or improved.

Major Comments

• Explanation of Temporal Gaps: The research makes use of flash flood data from 1950 to 2015. Nevertheless, the land cover data utilized spans the years 1985 to 2015. You should specifically address how the results, particularly the human activity intensity (HAI) analysis, can be impacted by the absence of land cover data during the first 35 years of the research period (1950–1985).
• Resolution Discrepancy: While other datasets (such as runoff or soil moisture) are probably at considerably coarser scales, the elevation data is at a resolution of 30 meters. To prevent spatial bias in the machine learning models, provide a section describing how these various spatial resolutions were harmonized (e.g., resampling techniques).
• Random Forest Performance: The model utilized in this work lacks performance indicators (e.g., RMSE, MAE, or cross-validation scores), despite the publication claiming that Random Forest is robust and produces high accuracy.
• SHAP Implementation: One advantage of interpretability is the application of SHAP. The study should make it clearer if the "dominant drivers" soil moisture and anthropogenic intensity are merely plateau-wide averages or if they varied considerably among the TP's several subregions.
• The findings show that flash floods have "increased exponentially" in comparison to reporting bias. Authors should examine if this is a physical rise in risk or if it is somewhat attributable to better "historical disaster records" and reporting mechanisms in more recent decades (particularly, 2008–2015).
• The paper accurately points out that improved preparedness does not cause fatalities to rise proportionately with flood disasters. A succinct examination of governmental adjustments or infrastructural advancements in China in the late 2000s that aided in this resilience could bolster this.
• Human Activity Intensity (HAI): The discovery that human activity is a "predominant driver" is noteworthy. Authors should be more explicit about the human activities that have the biggest impact on these high-altitude headwaters, such as road development, overgrazing, or urbanization.
• Climate Change Context: Although soil moisture and runoff are emphasized, the role of glacier melting and permafrost degradation two distinctive aspects of the TP described in the introduction could be more clearly connected to the "soil moisture" driver in the discussion.

Extra comments

• Make sure that the text provides a clear definition of the "Eight flash flood drivers" in Figure 2, particularly indices such as the Runoff Concentration Index (QCI) and Precipitation Concentration Index (PCI).
• Mathematical Notation: The presented sample appears to have a typo in Equation 1 (Gravity Model), where the formula for overline Y uses Xi instead of Yi. I suggest: Proofread every equation thoroughly.
• "Earth's Third Pole" is mentioned in the abstract, yet "Qinghai-Tibet Plateau (QTP)" is used in the study region section. To prevent misunderstanding for readers from other countries, use consistent primary terms throughout the manuscript, even if they are synonymous.

 

Finally, the paper offers a useful, repeatable guide for studying flood dynamics. This study will be elevated to the high standards of remote sensing by addressing the inconsistent data resolution, offering model validation measures, and expanding the conversation on reporting bias.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

 

1. The paper mentions using historical flash flood records from 1950 to 2015, but the text (Section 5.1) also notes that research before 1949 is largely absent, and records before 1977 are scarce. Please clearly explain the sources, collection methods, and potential biases of the data from 1950 to 1977, and discuss whether the incompleteness of this early data may affect the robustness of long-term evolution trends (such as the exponential growth conclusion). It is suggested to briefly address this limitation in the Discussion section.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for addressing my concerns. Paper looks good!

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors,

Thank you for your modifications.

Author Response

Please see the attachment.

Author Response File: Author Response.docx