Jia et al. (2026) Dynamic Monthly Population-Exposure-Based Flash-Flood Risk Mapping Using an Explainable XGBoost Framework

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-09-21
• Authors: Jingyi Jia, Qing Li, Naizheng Shen, Haoran Yang, Zhiwei Shi, Jinqi Wang, Haonan Deng, Dong Yingbo, Xiaoxuan Xia, Meihong Ma
• DOI: 10.3390/rs18183253

Research Groups

• Institute of Mountain Hazards and Environment, Chinese Academy of Sciences (IMHE-CAS)
• Hunan Provincial Bureau of Meteorology

Short Summary

This study developed a monthly flash-flood susceptibility model for Hunan Province, China, integrating precipitation, topographic, and environmental factors with population exposure. The framework provides insights into seasonal variations in population-exposure-based flash-flood risk.

Objective

• Investigate the relationship between monthly flash-flood susceptibility and population exposure in Hunan Province, China

Study Configuration

• Spatial Scale: Provincial scale (Hunan Province, China)
• Temporal Scale: Monthly scale for 1990–2016 and retrospective application for January to December 2024

Methodology and Data

• Models used: XGBoost-based model with SHAP analysis
• Data sources:
  • Flash-flood inventory (1990–2016)
  • Multisource topographic and environmental factors
  • Monthly precipitation data
  • Population exposure data

Main Results

• The XGBoost- and SHAP-based monthly flash-flood susceptibility model showed good performance (AUC = 0.82) in independent temporal testing for 2013–2016.
• Monthly maximum 1-day precipitation (M1P) was identified as the most important factor (23.0%), followed by elevation (21.1%) and topographic wetness index (14.7%).
• Population-exposure-based flash-flood risk (FFR) increased markedly in June–July, peaking in July with high- and very-high-risk areas covering 26.86% of the province.

Contributions

• This study provides a framework for understanding seasonal variations in population-exposure-based flash-flood risk.
• The results support improved understanding of spatial risk patterns and can be used to inform flood risk management strategies.

Funding

• This research was supported by the National Key Research and Development Program of China (Grant No. 2018YFC1508401) and the Chinese Academy of Sciences (CAS) Interdisciplinary Science Project (Grant No. XDA19040201).

Citation

@article{Jia2026Dynamic,
  author = {Jia, Jingyi and Li, Qing and Shen, Naizheng and Yang, Haoran and Shi, Zhiwei and Wang, Jinqi and Deng, Haonan and Yingbo, Dong and Xia, Xiaoxuan and Ma, Meihong},
  title = {Dynamic Monthly Population-Exposure-Based Flash-Flood Risk Mapping Using an Explainable XGBoost Framework},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18183253},
  url = {https://doi.org/10.3390/rs18183253}
}