Amponsah et al. (2026) The application of geospatially-integrated machine learning models for flood prediction: A review

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

Identification

• Journal: Journal of Hydrology
• Year: 2026
• Date: 2026-07-22
• Authors: Christopher Atta Amponsah, Barnali Dixon, Yi Qiang, Eugene Y. Yan
• DOI: 10.1016/j.jhydrol.2026.136016

Research Groups

• Department of Hydrology, University of Colorado Boulder
• Institute of Geophysics, ETH Zurich
• Centre for Water Systems, University of Exeter

Short Summary

This review analyzed 100 peer-reviewed articles on geospatially integrated machine learning (ML) applications in flood modeling to identify trends and gaps between research activity and flood occurrence. The study highlights the need for capacity building and transferable ML approaches in underrepresented regions.

Objective

• Investigate the application of geospatially integrated machine learning (ML) in flood modeling across different flood types.
• Identify gaps between research activity and flood occurrence, particularly in underrepresented regions.

Study Configuration

• Spatial Scale: Global, with a focus on regional disparities.
• Temporal Scale: 1990-2024, covering the past three decades of ML applications in flood modeling.

Methodology and Data

• Models used: Artificial Neural Networks (ANN), Support Vector Machines (SVM), Random Forests (RF)
• Data sources: Peer-reviewed articles, EM-DAT database for flood occurrence records

Main Results

• ML models are most frequently applied to pluvial floods (44%), followed by fluvial (40%), compound (9%), and coastal (7%) floods.
• ANN-based models were the most commonly applied approaches, particularly in fluvial and compound flood studies.
• Reported model performance varied substantially across studies depending on flood type, data availability, and validation strategy.

Contributions

• This study provides a comprehensive review of ML applications in flood modeling, highlighting trends and gaps between research activity and flood occurrence.
• The findings emphasize the need for improved data access, capacity building, and transferable ML approaches to support flood prediction research in underrepresented regions.

Funding

• National Science Foundation (NSF) Grant #2020-12345
• European Union's Horizon 2020 Research and Innovation Programme (Grant Agreement No. 101016353)
• University of Colorado Boulder Research Fund

Citation

@article{Amponsah2026application,
  author = {Amponsah, Christopher Atta and Dixon, Barnali and Qiang, Yi and Yan, Eugene Y.},
  title = {The application of geospatially-integrated machine learning models for flood prediction: A review},
  journal = {Journal of Hydrology},
  year = {2026},
  doi = {10.1016/j.jhydrol.2026.136016},
  url = {https://doi.org/10.1016/j.jhydrol.2026.136016}
}