Ming Ou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ming Ou has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
Yuhang Yao
13 shared publications- Min Li
College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou, China, 225000
12 shared publications Changman Yin
11 shared publications- Min Li
扬州大学(中国江苏省扬州市
4 shared publications Zilong Feng
JiLin Province Water Resource and Hydropower Consultative Company of P.R CHINA, Changchun, 130012, China
4 shared publications- Yuhang Yao
Yangzhou University
3 shared publications
Latest Publications
- Open Access
- 7 September 2026
- Published by Elsevier in SSRN Electronic Journal
A novel framework for identifying drought propagation pathways: a case study in the Huai River Basin
Drought propagation exhibits pronounced nonlinearity and spatial heterogeneity, yet the multi-pathway physical mechanisms remain inadequately quantified. Taking the Huai River Basin as the study area, we propose an improved convergent cross mapping (ICCM) framework that integrates Empirical Mode Decomposition-Detrended Fluctuation Analysis (EMD-DFA) denoising and Singular spectrum analysis (SSA) deperiodization to identify nonlinear propagation pathways. Coupled with Mann–Kendall trend testing, a copula-based Bayesian threshold model, and ridge regression, the performance of Pearson Correlation Coefficient (PCC), Convergent Cross Mapping (CCM), and ICCM is compared to systematically reveal spatiotemporal differentiation, thresholds, and driving mechanisms of multi-type drought propagation. Results indicate that conventional PCC and CCM yield spatially homogeneous identification dominated by a single propagation pathway: meteorological-hydrological-agricultural (100% and 92.9% of cases, respectively), with a lenient propagation threshold (-0.5) causing over-identification. In comparison, ICCM resolves a diversified propagation pathway pattern co-dominated by meteorological-hydrological-agricultural (50.0%) and meteorological-agricultural (46.4%), with propagation thresholds ranging from -0.68 to -1.90, effectively discriminating between rapid-response zones (2 months) and extreme-lag zones (6 months) across the basin. The causal trends of drought propagation exhibit pronounced spatial heterogeneity, with an overall pattern of intensifying drought coupling in the central and eastern basin and weakening propagation in the northern basin due to strong evapotranspiration suppression. The 2-m dew point temperature is the core dominant factor governing drought propagation, and the synergistic driving mechanisms of local hydrological processes and large-scale climate oscillations are elucidated. ICCM overcomes the limitations of conventional methods, providing scientific support for refined drought assessment and risk mitigation in complex basin.
- 7 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA novel framework for identifying drought propagation pathways: a case study in the Huai River Basin
Drought propagation exhibits pronounced nonlinearity and spatial heterogeneity, yet the multi-pathway physical mechanisms remain inadequately quantified. Taking the Huai River Basin as the study area, we propose an improved convergent cross mapping (ICCM) framework that integrates Empirical Mode Decomposition-Detrended Fluctuation Analysis (EMD-DFA) denoising and Singular spectrum analysis (SSA) deperiodization to identify nonlinear propagation pathways. Coupled with Mann–Kendall trend testing, a copula-based Bayesian threshold model, and ridge regression, the performance of Pearson Correlation Coefficient (PCC), Convergent Cross Mapping (CCM), and ICCM is compared to systematically reveal spatiotemporal differentiation, thresholds, and driving mechanisms of multi-type drought propagation. Results indicate that conventional PCC and CCM yield spatially homogeneous identification dominated by a single propagation pathway: meteorological-hydrological-agricultural (100% and 92.9% of cases, respectively), with a lenient propagation threshold (-0.5) causing over-identification. In comparison, ICCM resolves a diversified propagation pathway pattern co-dominated by meteorological-hydrological-agricultural (50.0%) and meteorological-agricultural (46.4%), with propagation thresholds ranging from -0.68 to -1.90, effectively discriminating between rapid-response zones (2 months) and extreme-lag zones (6 months) across the basin. The causal trends of drought propagation exhibit pronounced spatial heterogeneity, with an overall pattern of intensifying drought coupling in the central and eastern basin and weakening propagation in the northern basin due to strong evapotranspiration suppression. The 2-m dew point temperature is the core dominant factor governing drought propagation, and the synergistic driving mechanisms of local hydrological processes and large-scale climate oscillations are elucidated. ICCM overcomes the limitations of conventional methods, providing scientific support for refined drought assessment and risk mitigation in complex basin.
- Open Access
- 26 June 2026
- Published by Copernicus GmbH in Hydrology and Earth System Sciences
- Vol. 30 (12) , 3925-3944
Daily drought prediction in the Huaihe River Basin using VMD-informer-LSTM
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R2, RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 d, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 d prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model's predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.
- 26 June 2026
- Published by Copernicus GmbH in Hydrology and Earth System Sciences
- Vol. 30 (12) , 3925-3944
Open AccessDaily drought prediction in the Huaihe River Basin using VMD-informer-LSTM
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R2, RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 d, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 d prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model's predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.
- Open Access
- 12 May 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Comment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- 12 May 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Open AccessComment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- Open Access
- 11 May 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Comment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- 11 May 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Open AccessComment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- Open Access
- 6 April 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Comment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- 6 April 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-33
Open AccessComment on egusphere-2026-1033
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- Open Access
- 2 April 2026
- Published by Copernicus GmbH in EGUsphere
- Vol. 2026, 1-33
A hybrid model based on Boruta feature selection and neural network for forecasting hydrological drought
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- 2 April 2026
- Published by Copernicus GmbH in EGUsphere
- Vol. 2026, 1-33
Open AccessA hybrid model based on Boruta feature selection and neural network for forecasting hydrological drought
Accurate hydrological drought prediction is vital for water management. This study proposes a hybrid model combining Boruta feature selection, convolutional neural network (CNN), and bidirectional long short-term memory (BiLSTM) methods, to predict hydrological drought in the Huaihe River Basin of China. The Boruta algorithm selected key predictors from 31 potential drought-influencing factors. By comparing the established model Boruta-CNN-BiLSTM with other models, including Boruta-CNN-LSTM, Boruta-CNN-XGBoost, Boruta-BiLSTM, Boruta-LSTM, and Boruta-XGBoost, the results show Boruta significantly enhances all models. The Boruta-CNN-BiLSTM model has achieved the highest accuracy across 28 basin grid regions, exhibiting the largest performance gains. Furthermore, the prediction performance of the model is mainly influenced by factors such as precipitation, volumetric soil water (0–7 cm), volumetric soil water (7–28 cm) and surface net solar radiation. The model's prediction performance is most affected by precipitation, followed by volumetric soil water (0–7 cm), volumetric soil water (7–28 cm), and surface net solar radiation has the least impact. It provides enhanced support for basin-scale drought risk assessment and water resources management.
- Open Access
- 3 March 2026
- Published by Elsevier in SSRN Electronic Journal
Spatial Compounding of Flash Droughts at different time scales in the Huaihe River Basin, China
Flash droughts, characterized by rapid onset and intensification, have received increasing attention. Previous studies focused on their identification, evolution, and climatic drivers, while the spatial co-occurrence and connectivity of flash droughts (i.e., compound flash droughts) remain poorly understood, limiting regional impact assessment. Based on a case study of the Huaihe River Basin, we develop a framework for identifying compound flash drought events and analyzing their spatial connectivity. Utilizing multi-dimensional flash drought indicators, connectivity and clustering analyses are conducted at annual and seasonal scales, and the dominant driving factors are further examined. The results revealed spatial heterogeneity of compound flash droughts across the basin. High-connectivity areas are concentrated in the northern, north-western, and central regions, with substantially lower connectivity in the southeast. The basin is divided into four types of compound flash droughts based on annual clustering. The high-intensity and high-deficit types are primarily located in the northern and northwestern regions, forming a southeast–northwest increasing gradient. Autumn exhibits the strongest spatial differentiation and the most frequent and widespread compound flash droughts based on seasonal clustering. The dominant driving factors differ among compound flash drought types. This study advances flash drought understanding from a spatial organization and connectivity perspective. It provides new support for basin-scale drought risk assessment and water resources management.
- 3 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSpatial Compounding of Flash Droughts at different time scales in the Huaihe River Basin, China
Flash droughts, characterized by rapid onset and intensification, have received increasing attention. Previous studies focused on their identification, evolution, and climatic drivers, while the spatial co-occurrence and connectivity of flash droughts (i.e., compound flash droughts) remain poorly understood, limiting regional impact assessment. Based on a case study of the Huaihe River Basin, we develop a framework for identifying compound flash drought events and analyzing their spatial connectivity. Utilizing multi-dimensional flash drought indicators, connectivity and clustering analyses are conducted at annual and seasonal scales, and the dominant driving factors are further examined. The results revealed spatial heterogeneity of compound flash droughts across the basin. High-connectivity areas are concentrated in the northern, north-western, and central regions, with substantially lower connectivity in the southeast. The basin is divided into four types of compound flash droughts based on annual clustering. The high-intensity and high-deficit types are primarily located in the northern and northwestern regions, forming a southeast–northwest increasing gradient. Autumn exhibits the strongest spatial differentiation and the most frequent and widespread compound flash droughts based on seasonal clustering. The dominant driving factors differ among compound flash drought types. This study advances flash drought understanding from a spatial organization and connectivity perspective. It provides new support for basin-scale drought risk assessment and water resources management.
- Open Access
- 2 February 2026
- Published by Wiley
Spatial Compounding of Flash Droughts at different time scales in the Huaihe River Basin, China
Flash droughts, characterized by rapid onset and intensification, have received increasing attention. Previous studies focused on their identification, evolution, and climatic drivers, while the spatial co-occurrence and connectivity of flash droughts (i.e., compound flash droughts) remain poorly understood, limiting regional impact assessment. Based on a case study of the Huaihe River Basin, we develop a framework for identifying compound flash drought events and analyzing their spatial connectivity. Utilizing multi-dimensional flash drought indicators, connectivity and clustering analyses are conducted at annual and seasonal scales, and the dominant driving factors are further examined. The results revealed spatial heterogeneity of compound flash droughts across the basin. High-connectivity areas are concentrated in the northern, north-western, and central regions, with substantially lower connectivity in the southeast. The basin is divided into four types of compound flash droughts based on annual clustering. The high-intensity and high-deficit types are primarily located in the northern and northwestern regions, forming a southeast–northwest increasing gradient. Autumn exhibits the strongest spatial differentiation and the most frequent and widespread compound flash droughts based on seasonal clustering. The dominant driving factors differ among compound flash drought types. This study advances flash drought understanding from a spatial organization and connectivity perspective. It provides new support for basin-scale drought risk assessment and water resources management.
- 2 February 2026
- Published by Wiley
Open AccessSpatial Compounding of Flash Droughts at different time scales in the Huaihe River Basin, China
Flash droughts, characterized by rapid onset and intensification, have received increasing attention. Previous studies focused on their identification, evolution, and climatic drivers, while the spatial co-occurrence and connectivity of flash droughts (i.e., compound flash droughts) remain poorly understood, limiting regional impact assessment. Based on a case study of the Huaihe River Basin, we develop a framework for identifying compound flash drought events and analyzing their spatial connectivity. Utilizing multi-dimensional flash drought indicators, connectivity and clustering analyses are conducted at annual and seasonal scales, and the dominant driving factors are further examined. The results revealed spatial heterogeneity of compound flash droughts across the basin. High-connectivity areas are concentrated in the northern, north-western, and central regions, with substantially lower connectivity in the southeast. The basin is divided into four types of compound flash droughts based on annual clustering. The high-intensity and high-deficit types are primarily located in the northern and northwestern regions, forming a southeast–northwest increasing gradient. Autumn exhibits the strongest spatial differentiation and the most frequent and widespread compound flash droughts based on seasonal clustering. The dominant driving factors differ among compound flash drought types. This study advances flash drought understanding from a spatial organization and connectivity perspective. It provides new support for basin-scale drought risk assessment and water resources management.
- Open Access
- 28 January 2026
- Published by Copernicus GmbH
- Vol. 2025, 1-31
Comment on egusphere-2025-5645
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R², RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 days, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 day prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model’s predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.
- 28 January 2026
- Published by Copernicus GmbH
- Vol. 2025, 1-31
Open AccessComment on egusphere-2025-5645
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R², RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 days, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 day prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model’s predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.
- Open Access
- 20 January 2026
- Published by Copernicus GmbH
- Vol. 2025, 1-31
Comment on egusphere-2025-5645
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R², RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 days, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 day prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model’s predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.
- 20 January 2026
- Published by Copernicus GmbH
- Vol. 2025, 1-31
Open AccessComment on egusphere-2025-5645
Accurate drought prediction is a key challenge in water resource management and agricultural planning. This study proposes a novel drought prediction framework that integrates Variational Mode Decomposition (VMD), Informer, and Long Short-Term Memory (LSTM) networks to enhance hydrological drought forecasting in the Huaihe River Basin, China. The VMD-Informer-LSTM model decomposes complex non-stationary drought sequences into multi-scale components, effectively extracting long-term trends and short-term fluctuations. Results show that the model outperforms LSTM, Transformer-LSTM, and Informer-LSTM, improving R², RMSE, MAE, and MAPE by 28.4 %, 46.2 %, 46.5 %, and 50.8 %, respectively, over the baseline LSTM. When the prediction period is 30 days, the VMD-Informer-LSTM achieves the highest prediction accuracy. During the 120–180 day prediction period, the prediction accuracy of all models declines, with drought intensity generally underestimated. Misclassifications are mainly concentrated in the transition zones between humid and semi-humid regions, with higher error frequency in semi-humid areas. Prediction accuracy is highest in the upstream and downstream regions, followed by the Yishuisi River Basin, while the midstream region performs poorly due to human interference. Shapley Additive Explanations (SHAP) further reveal that precipitation and temperature are the dominant meteorological drivers, jointly accounting for nearly half of the model’s predictive power. These results confirm that the VMD-Informer-LSTM provides the most accurate predictions among the tested models, offering valuable support for drought risk assessment and water resource management in the Huaihe River Basin and other similar regions.