Luo et al. (2026) Disentangling the impacts of climate, catchment, and morphological characteristics on hydrological drought propagation and recovery

Identification

• Journal: CATENA
• Year: 2026
• Date: 2026-01-12
• Authors: Xuan Luo, Xuan Ji, Yi Zou, Siqi Wang, Xinbei Liu, Xiaodong Wu, Jianxing Li, Yungang Li
• DOI: 10.1016/j.catena.2026.109799

Research Groups

• Institute of International Rivers and Eco-Security, Yunnan University, Kunming 650504, China
• Yunnan Key Laboratory of International Rivers and Transboundary Eco-security, Yunnan University, Kunming 650504, China
• Ministry of Education Key Laboratory for Transboundary Eco-security of Southwest China, Yunnan University, Kunming 650504, China

Short Summary

This study developed an attribution approach integrating process-based modeling and machine learning to disentangle multi-factor controls on hydrological drought propagation and recovery in the Lancang-Mekong River Basin. It found that climate is the primary driver for both propagation and recovery times, with recovery time consistently exceeding propagation time.

Objective

• To disentangle the relative contributions and roles of climate, catchment, and morphological characteristics on hydrological drought propagation and recovery dynamics in the Lancang-Mekong River Basin.

Study Configuration

• Spatial Scale: Lancang-Mekong River Basin, including its sub-basins.
• Temporal Scale: Event-based analysis of drought propagation and recovery. (Specific overall study period not detailed in the provided text).

Methodology and Data

• Models used: Variable Infiltration Capacity (VIC) model, Optimal Parameter-based Geodetector (OPGD), Recursive Feature Elimination (RFE), Extreme Gradient Boosting (XGBoost), SHapley Additive exPlanations (SHAP).
• Data sources: Original data from "Impacts of Hydrological Drought Propagation and Recovery" dataset, used to reconstruct natural runoff.

Main Results

• Hydrological drought recovery lag time (RT) (mean: 105.0 days) consistently exceeded meteorological-to-hydrological drought propagation time (PT) (mean: 34.1 days) across all sub-basins.
• Drought duration and severity were found to amplify both PT and RT.
• A hybrid OPGD-RFE-XGBoost model achieved a 50% increase in R² for PT and a 42% increase for RT, utilizing only 18% and 35% of the original features, respectively, compared to a standalone XGBoost model.
• SHAP analysis identified climate as the primary controlling factor for both PT (contribution: 45.5%) and RT (51.8%).
• Catchment and morphological characteristics exhibited subordinate influence on PT (41.3%) and RT (32.7%).

Contributions

• Developed a novel attribution approach integrating process-based hydrological modeling with advanced machine learning techniques (OPGD, RFE, XGBoost, SHAP) to quantify multi-factor controls on hydrological drought propagation and recovery.
• Established an analytical framework that effectively decouples climate-catchment interactions in drought cascades by incorporating spatially structured predictors and explainable machine learning.
• Provided quantitative insights into the relative contributions of climate, catchment, and morphological characteristics to hydrological drought dynamics in a major transboundary river basin.
• Highlighted the significant role of underlying catchment and morphological characteristics in modulating drought propagation and recovery processes.

Funding

• Not specified in the provided text.

Citation

@article{Luo2026Disentangling,
  author = {Luo, Xuan and Ji, Xuan and Zou, Yi and Wang, Siqi and Liu, Xinbei and Wu, Xiaodong and Li, Jianxing and Li, Yungang},
  title = {Disentangling the impacts of climate, catchment, and morphological characteristics on hydrological drought propagation and recovery},
  journal = {CATENA},
  year = {2026},
  doi = {10.1016/j.catena.2026.109799},
  url = {https://doi.org/10.1016/j.catena.2026.109799}
}