Xu et al. (2025) Objectivization of an expert assessment framework for drought monitoring

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-11-17
• Authors: Haiting Xu, Yonghua Zhu, Jianhui Wei, Haishen Lv, Ying Pan, Yingying Xu, Yuan Yao, Di Liu, Harald Kunstmann
• DOI: 10.1016/j.jhydrol.2025.134592

Research Groups

• National Key Laboratory of Water Disaster Prevention, College of Hydrology and Water Resources, Hohai University, China.
• College of Geography and Remote Sensing, Hohai University, China.
• Institute of Meteorology and Climate Research (IMKIFU), Karlsruhe Institute of Technology (KIT), Germany.
• Engineer school, Qinghai Institute of Technology, China.
• Institute of Geography and Centre for Climate Resilience, University of Augsburg, Germany.

Short Summary

The study develops a Comprehensive Drought Monitoring Model (CDMM) that objectivizes the expert-based U.S. Drought Monitor (USDM) framework using the Random Forest algorithm. The model successfully reproduces USDM drought categories and demonstrates high transferability by effectively capturing regional drought dynamics across China.

Objective

• To objectivize the subjective expert assessment framework used in drought monitoring (specifically the USDM) to create a consistent, automated, and transferable model for multi-scale drought risk recognition.

Study Configuration

• Spatial Scale: Continental United States (for model training and validation) and China (for regional application and transferability assessment).
• Temporal Scale: Long-term regional scales, focusing on drought frequency and duration across different climatic regions.

Methodology and Data

• Models used: Comprehensive Drought Monitoring Model (CDMM) based on the Random Forest (RF) machine learning algorithm.
• Data sources: U.S. Drought Monitor (USDM) drought category data, and multiple integrated drought monitoring indices covering meteorological, agricultural, and hydrological perspectives.

Main Results

• The CDMM accurately reproduces the spatial distributions and drought categories established by the USDM expert framework.
• In China, the model identified a higher frequency of short-duration drought events in the eastern monsoon region.
• The northwest and southwest regions of China were found to experience more prolonged drought events compared to the east.
• The model proved effective at capturing drought dynamics across diverse climatic regions and multiple timescales.

Contributions

• Successfully "objectivizes" subjective expert experience, allowing the robust USDM assessment framework to be applied to other regions without requiring constant expert intervention.
• Provides a transferable machine-learning framework that integrates multifaceted drought types (meteorological, agricultural, and hydrological) for improved monitoring.
• Enhances the ability to track drought evolution and risks at regional scales, particularly in areas lacking established expert-led monitoring systems.

Funding

• Not provided in the source text.

Citation

@article{Xu2025Objectivization,
  author = {Xu, Haiting and Zhu, Yonghua and Wei, Jianhui and Lv, Haishen and Pan, Ying and Xu, Yingying and Yao, Yuan and Liu, Di and Kunstmann, Harald},
  title = {Objectivization of an expert assessment framework for drought monitoring},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134592},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134592}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)