the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Mar 2025
Probabilistic Analysis of Future Drought Propagation, Persistence, and Spatial Concurrence in Monsoon-Dominant Asian Region under Climate Change
Abstract. This study examines future drought propagation (the temporal transition from meteorological to agricultural droughts), persistence (inter-seasonal droughts), and spatial concurrence (simultaneous occurrence of monsoonal droughts across regions) under climate change using a multivariate copula approach in Monsoon-dominant Asia. Under the worst-case emission scenario (SSP5-8.5), South Asia (excluding Western and Peninsula India) and Eastern China are projected to experience intensified drought propagation compared to the historical period (1975–2014). In addition to increased propagation in these regions, the propagated agricultural droughts are expected to persist across seasons in the future. In terms of the return period, all-season droughts that historically occurred once in more than 50 years could happen as frequently as every five years by the far-future (2061–2100) at the hydrologically significant Tibetan Plateau. Random Forest models indicate that the temperature is a key driver of future agricultural droughts in nearly half of the study area. The spatial concurrence of monsoonal agricultural droughts between region pairs such as South Asia (SAS), East Asia (EAS), Southeast Asia (SEA), and Tibetan region (TIB) was also assessed. Based on bivariate return periods of spatial concurrence, frequent future spatial drought concurrence is anticipated between populous SAS and EAS compared to the historical timeframe, posing risks to water and food security. Conversely, SEA is projected to experience reduced spatial drought concurrence with other regions, which could encourage greater regional cooperation. Overall, this comprehensive approach which integrates three aspects of drought dynamics, offers valuable insights for climate change mitigation, planning, and adaptation.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(9057 KB)
-
Supplement
(3258 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(9057 KB) - Metadata XML
-
Supplement
(3258 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-522', Anonymous Referee #1, 11 Apr 2025
General Comments:
The authors did a thorough analysis of the future droughts in Monsoon-dominant Asian under the worst-case emission scenario of SSP5-8.5. The analysis of propagation from meteorological to hydrological droughts is new, and the use of bivariate copula function for analyzing drought propagation and spatial concurrence is interesting. I only have a few comments as listed below.
Specific comments:
- Abstract: since the propagation from meteorological drought to agricultural drought is a highlight of this work. It’s helpful to indicate in the abstract that meteorological and agriculture droughts are measured using SPI and SSI, respectively.
- Line 61: I believe there are quite a few studies on drought analysis under climate change. I am not sure whether “only” is the most accurate or appropriate term in this case. Citing these works can help readers better understand the current state of research on this topic.
- Lines 155-160, a common practice in climate impact studies is to use bias correction techniques and correct the biases in GCM output before any further analysis. Do you think this can help reduce the errors in Figure 2c and Fig. S1? How about the difference between observation and GCM output in Fig. S2?
- Line 245: why is R^2 the only performance metrics for soil moisture prediction? With some R^2 values lower than 0.5 in the results, how to justify the accuracy of the RF model or the reliability of its feature importance results?
- Line 250: The predictors that are important in the RF model, are the ones that are important for the estimation of soil moisture. Are they necessarily the same as the ones that may lead to soil moisture deficit? How could the feature importance results be best interpreted?
Technical corrections:
- Line 100: the "+" sign suggests a summation of these variables inside the function, which is not a rigorous expression. Since SM depends on these variables separately, the notation f(T, Pr, H, VC, SR100, W, TS) would be more appropriate.
Citation: https://doi.org/10.5194/egusphere-2025-522-RC1 - AC1: 'Reply on RC1', Xiaosheng Qin, 23 May 2025
-
RC2: 'Comment on egusphere-2025-522', Anonymous Referee #2, 17 Apr 2025
General comments:
The authors investigate the impacts of climate change on drought propagation from meteorological to agricultural droughts in monsoon-dominant Asian regions. Understanding drought propagation mechanisms under climate change is crucial, particularly in assessing temporal transitions in drought propagation, persistence, and spatial concurrence.
While the study addresses an important research gap, several critical issues undermine the validity of the findings:
- L55-64: The manuscript does not adequately justify the need to study drought propagation, persistence, and spatial concurrence together. Additionally, the claim that studies on meteorological-to-agricultural drought propagation are lacking is inaccurate, as several previous studies already exist (Ding et al., 2021; Dai et al., 2022; Fawen et al., 2023; Xu et al., 2023). Lastly, since this study also focuses on a regional scale, the authors should explicitly clarify its novel contributions and regional significance.
- Unclear Monsoon-Based Classification for Drought Persistence: The rationale for categorizing drought persistence into pre-monsoon, monsoon, and post-monsoon periods is not well-explained. It remains unclear whether this classification is tied to SPI-SSI propagation dynamics or solely based on SSI thresholds (e.g., SSI < -0.5). A stronger theoretical or empirical basis for this approach is needed. Justify the focus on monsoonal seasons for drought propagation implications.
- Ambiguity in Drought Concurrence Analysis: The assessment of spatial drought concurrence relies on SSI thresholds but does not explicitly link to SPI-driven propagation. The authors should clarify whether the observed concurrence reflects independent agricultural droughts or is influenced by meteorological drought propagation.
- L93: Random Forest Model Design: The use of soil moisture as the predictand (rather than drought propagation metrics) limits the model’s ability to identify key drivers of propagation. Restructuring the model to treat propagation as the predictand (with climatic variables as predictors) would better address the study’s primary objective.
Minor comments:
L13-L16: “ In terms of the return period, all-season droughts that historically occurred once in more than 50 years could happen as frequently as every five years by the far-future (2061-2100) at the hydrologically significant Tibetan Plateau.” The statement is confusing. Rephrase for clarity.
L17-L19: “The spatial concurrence of monsoonal agricultural droughts between region pairs such as South Asia (SAS), East Asia (EAS), Southeast Asia (SEA), and Tibetan region (TIB) was also assessed.” Replace methodological descriptions (e.g., region pairs assessed) with concrete findings.
L27: Cite specific literature comparing disaster impacts.
L40-41: “While comparing propagation between basins of different climate zones, Zhang et al. (2021) found arid basin to have lower propagation durations compared to humid and sub-humid basins” Rephrase for clarity.
L64-65: “To address the aforementioned gaps, this study proposes a comprehensive copula-based multivariate probabilistic approach, utilizing climate model projection data.” Separate the copula method and climate model applications to avoid logical gaps.
2.1.1: Describe monthly precipitation, SPI/SSI calculations, and soil moisture datasets.
2.1.3: Please provide a detailed description of the hyperparameters in the random forest model (e.g., the number of trees, maximum depth), explaining the parameter tuning process to validate the robustness of the model. Furthermore, it is recommended to include other evaluation metrics (e.g., RMSE) to more accurately demonstrate model performance. Please provide the cross-validation of RF models and shows the R2 of different regions.
Fig.2b: Provide citations for the basis of regional divisions, which is crucial to spatial concurrence analysis.
Fig. 2c: Compare GLDAS and GCM-derived SPI monthly drought characteries (not average precipitation) to validate GCM reliability for drought propagation.
Datasets: Replace GLDAS precipitation/soil moisture with more reliable datasets (e.g., MSWEP for precipitation, GLEAM for soil moisture).
References:
Ding Y, Gong X, Xing Z, et al. Attribution of meteorological, hydrological and agricultural drought propagation in different climatic regions of China. Agricultural Water Management, 2021, 255: 106996.
Xu Z, Wu Z, Shao Q, et al. From meteorological to agricultural drought: Propagation time and probabilistic linkages. Journal of Hydrology: Regional studies, 2023, 46: 101329.
Dai M, Huang S, Huang Q, et al. Propagation characteristics and mechanism from meteorological to agricultural drought in various seasons. Journal of Hydrology, 2022, 610: 127897.
Fawen L, Manjing Z, Yong Z, et al. Influence of irrigation and groundwater on the propagation of meteorological drought to agricultural drought. Agricultural Water Management, 2023, 277: 108099.
Citation: https://doi.org/10.5194/egusphere-2025-522-RC2 - AC2: 'Reply on RC2', Xiaosheng Qin, 23 May 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-522', Anonymous Referee #1, 11 Apr 2025
General Comments:
The authors did a thorough analysis of the future droughts in Monsoon-dominant Asian under the worst-case emission scenario of SSP5-8.5. The analysis of propagation from meteorological to hydrological droughts is new, and the use of bivariate copula function for analyzing drought propagation and spatial concurrence is interesting. I only have a few comments as listed below.
Specific comments:
- Abstract: since the propagation from meteorological drought to agricultural drought is a highlight of this work. It’s helpful to indicate in the abstract that meteorological and agriculture droughts are measured using SPI and SSI, respectively.
- Line 61: I believe there are quite a few studies on drought analysis under climate change. I am not sure whether “only” is the most accurate or appropriate term in this case. Citing these works can help readers better understand the current state of research on this topic.
- Lines 155-160, a common practice in climate impact studies is to use bias correction techniques and correct the biases in GCM output before any further analysis. Do you think this can help reduce the errors in Figure 2c and Fig. S1? How about the difference between observation and GCM output in Fig. S2?
- Line 245: why is R^2 the only performance metrics for soil moisture prediction? With some R^2 values lower than 0.5 in the results, how to justify the accuracy of the RF model or the reliability of its feature importance results?
- Line 250: The predictors that are important in the RF model, are the ones that are important for the estimation of soil moisture. Are they necessarily the same as the ones that may lead to soil moisture deficit? How could the feature importance results be best interpreted?
Technical corrections:
- Line 100: the "+" sign suggests a summation of these variables inside the function, which is not a rigorous expression. Since SM depends on these variables separately, the notation f(T, Pr, H, VC, SR100, W, TS) would be more appropriate.
Citation: https://doi.org/10.5194/egusphere-2025-522-RC1 - AC1: 'Reply on RC1', Xiaosheng Qin, 23 May 2025
-
RC2: 'Comment on egusphere-2025-522', Anonymous Referee #2, 17 Apr 2025
General comments:
The authors investigate the impacts of climate change on drought propagation from meteorological to agricultural droughts in monsoon-dominant Asian regions. Understanding drought propagation mechanisms under climate change is crucial, particularly in assessing temporal transitions in drought propagation, persistence, and spatial concurrence.
While the study addresses an important research gap, several critical issues undermine the validity of the findings:
- L55-64: The manuscript does not adequately justify the need to study drought propagation, persistence, and spatial concurrence together. Additionally, the claim that studies on meteorological-to-agricultural drought propagation are lacking is inaccurate, as several previous studies already exist (Ding et al., 2021; Dai et al., 2022; Fawen et al., 2023; Xu et al., 2023). Lastly, since this study also focuses on a regional scale, the authors should explicitly clarify its novel contributions and regional significance.
- Unclear Monsoon-Based Classification for Drought Persistence: The rationale for categorizing drought persistence into pre-monsoon, monsoon, and post-monsoon periods is not well-explained. It remains unclear whether this classification is tied to SPI-SSI propagation dynamics or solely based on SSI thresholds (e.g., SSI < -0.5). A stronger theoretical or empirical basis for this approach is needed. Justify the focus on monsoonal seasons for drought propagation implications.
- Ambiguity in Drought Concurrence Analysis: The assessment of spatial drought concurrence relies on SSI thresholds but does not explicitly link to SPI-driven propagation. The authors should clarify whether the observed concurrence reflects independent agricultural droughts or is influenced by meteorological drought propagation.
- L93: Random Forest Model Design: The use of soil moisture as the predictand (rather than drought propagation metrics) limits the model’s ability to identify key drivers of propagation. Restructuring the model to treat propagation as the predictand (with climatic variables as predictors) would better address the study’s primary objective.
Minor comments:
L13-L16: “ In terms of the return period, all-season droughts that historically occurred once in more than 50 years could happen as frequently as every five years by the far-future (2061-2100) at the hydrologically significant Tibetan Plateau.” The statement is confusing. Rephrase for clarity.
L17-L19: “The spatial concurrence of monsoonal agricultural droughts between region pairs such as South Asia (SAS), East Asia (EAS), Southeast Asia (SEA), and Tibetan region (TIB) was also assessed.” Replace methodological descriptions (e.g., region pairs assessed) with concrete findings.
L27: Cite specific literature comparing disaster impacts.
L40-41: “While comparing propagation between basins of different climate zones, Zhang et al. (2021) found arid basin to have lower propagation durations compared to humid and sub-humid basins” Rephrase for clarity.
L64-65: “To address the aforementioned gaps, this study proposes a comprehensive copula-based multivariate probabilistic approach, utilizing climate model projection data.” Separate the copula method and climate model applications to avoid logical gaps.
2.1.1: Describe monthly precipitation, SPI/SSI calculations, and soil moisture datasets.
2.1.3: Please provide a detailed description of the hyperparameters in the random forest model (e.g., the number of trees, maximum depth), explaining the parameter tuning process to validate the robustness of the model. Furthermore, it is recommended to include other evaluation metrics (e.g., RMSE) to more accurately demonstrate model performance. Please provide the cross-validation of RF models and shows the R2 of different regions.
Fig.2b: Provide citations for the basis of regional divisions, which is crucial to spatial concurrence analysis.
Fig. 2c: Compare GLDAS and GCM-derived SPI monthly drought characteries (not average precipitation) to validate GCM reliability for drought propagation.
Datasets: Replace GLDAS precipitation/soil moisture with more reliable datasets (e.g., MSWEP for precipitation, GLEAM for soil moisture).
References:
Ding Y, Gong X, Xing Z, et al. Attribution of meteorological, hydrological and agricultural drought propagation in different climatic regions of China. Agricultural Water Management, 2021, 255: 106996.
Xu Z, Wu Z, Shao Q, et al. From meteorological to agricultural drought: Propagation time and probabilistic linkages. Journal of Hydrology: Regional studies, 2023, 46: 101329.
Dai M, Huang S, Huang Q, et al. Propagation characteristics and mechanism from meteorological to agricultural drought in various seasons. Journal of Hydrology, 2022, 610: 127897.
Fawen L, Manjing Z, Yong Z, et al. Influence of irrigation and groundwater on the propagation of meteorological drought to agricultural drought. Agricultural Water Management, 2023, 277: 108099.
Citation: https://doi.org/10.5194/egusphere-2025-522-RC2 - AC2: 'Reply on RC2', Xiaosheng Qin, 23 May 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 328 | 61 | 14 | 403 | 21 | 10 | 13 |
- HTML: 328
- PDF: 61
- XML: 14
- Total: 403
- Supplement: 21
- BibTeX: 10
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dineshkumar Muthuvel
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(9057 KB) - Metadata XML
-
Supplement
(3258 KB) - BibTeX
- EndNote
- Final revised paper