Long et al. (2025) Reconstruction of drought propagation pathways: A global analysis of multitype propagation chains and nonlinear mechanisms

Identification

• Journal: Global and Planetary Change
• Year: 2025
• Date: 2025-10-31
• Authors: Junchen Long, Changchun Xu, Hongyu Wang, Zhiyi Li, F. R. Xu
• DOI: 10.1016/j.gloplacha.2025.105144

Research Groups

• Xinjiang Key Laboratory of Oasis Ecology, Xinjiang, China
• College of Geography and Remote Sensing Science, Xinjiang University, Xinjiang, China

Short Summary

This study reconstructs global drought propagation pathways by combining copula functions, a Bayesian framework, and multi-scale drought indices, revealing that abnormal evapotranspiration often initiates drought and quantifying the nonlinear roles of natural and anthropogenic drivers using interpretable machine learning.

Objective

• To elucidate the actual pathways of drought propagation by analyzing the timing of drought signals in precipitation, evapotranspiration, soil moisture, and vegetation, and to quantify the dynamic roles of natural and anthropogenic drivers in these processes.

Study Configuration

• Spatial Scale: Global
• Temporal Scale: Focuses on the timing and dynamic changes of drought signals and propagation events.

Methodology and Data

• Models used: Copula functions, Bayesian framework, multi-scale drought indices, XGBoost (eXtreme Gradient Boosting) with SHAP (SHapley Additive exPlanations) values.
• Data sources: Drought signals derived from precipitation, evapotranspiration, soil moisture, and vegetation data.

Main Results

• In 69 % of regions, abnormal evapotranspiration initiates drought propagation, rather than reduced precipitation.
• In 67 % of areas, soil moisture changes precede vegetation response.
• Three forms of drought propagation were identified: sequential, reverse, and leapfrogging.
• Four distinct drought propagation chains were constructed globally.
• Evapotranspiration–soil moisture drought processes primarily dominate the spatial differences in global drought chains.
• Most influencing factors (natural and anthropogenic drivers) exhibit nonlinear, threshold, and diminishing effects on drought propagation.

Contributions

• Provides a novel method combining copula functions, a Bayesian framework, and multi-scale drought indices to reconstruct actual drought propagation pathways.
• Identifies and characterizes multitype drought propagation chains (sequential, reverse, leapfrogging) globally.
• Quantifies the dynamic and nonlinear roles of natural and anthropogenic drivers in drought propagation using interpretable machine learning (XGBoost with SHAP).
• Offers a scientific basis for dynamic, region-specific drought forecasting and water resource management.

Funding

Not specified in the provided text.

Citation

@article{Long2025Reconstruction,
  author = {Long, Junchen and Xu, Changchun and Wang, Hongyu and Li, Zhiyi and Xu, F. R.},
  title = {Reconstruction of drought propagation pathways: A global analysis of multitype propagation chains and nonlinear mechanisms},
  journal = {Global and Planetary Change},
  year = {2025},
  doi = {10.1016/j.gloplacha.2025.105144},
  url = {https://doi.org/10.1016/j.gloplacha.2025.105144}
}

Generated by BiblioAssistant using gemini-2.5-flash (Google API)