Rahimi et al. (2026) Integrating geospatial intelligence and machine learning for flood susceptibility mapping

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-02-23
• Authors: Mehdi Rahimi, Bahram Malekmohammadi, Mohammad Karimi Firozjaei, Reza Kerachian, Jamal Jokar Arsanjani, Mou Leong Tan, Joseph L. Awange, Dragan Savić, Qingyun Duan, Amir AghaKouchak
• DOI: 10.1038/s41598-026-41014-3

Research Groups

• Graduate Faculty of Environment, University of Tehran, Tehran, Iran
• College of Management, University of Tehran, Tehran, Iran
• School of Civil Engineering, College of Engineering, University of Tehran, Tehran, Iran
• Geoinformatics Research Group, Department of Planning, Aalborg University Copenhagen, Copenhagen, Denmark
• GeoInformatic Unit, Geography Section, School of Humanities, Universiti Sains Malaysia, Minden, Penang, Malaysia
• Environmental and Atmospheric Sciences Research Group, Scientific Research Center, Al-Ayen University, Nasiriyah, Iraq
• Department of Land Surveying and Geo-Informatics, The Hong Kong Polytechnic University, Hong Kong, China
• Centre for Water Systems, University of Exeter, Exeter, UK
• KWR Water Research Institute, Nieuwegein, The Netherlands
• College of Hydrology and Water Resources, Hohai University, Nanjing, Jiangsu, China
• Department of Civil and Environmental Engineering, University of California, Irvine, CA, USA
• Department of Earth System Science, University of California, Irvine, CA, USA
• Institute for Water, Environment and Health, United Nations University, Richmond Hill, ON, Canada

Short Summary

This study evaluated five machine learning algorithms and an ensemble voting model for flood susceptibility mapping, demonstrating that the ensemble approach significantly improves accuracy and reliability in identifying flood-prone areas.

Objective

• To evaluate and compare the performance of five individual machine learning algorithms (Extreme Gradient Boosting, Decision Tree, Random Forest, Light Gradient Boosting Machine, and Generalized Linear Model) and their ensemble voting combination for flood susceptibility mapping.

Study Configuration

• Spatial Scale: Global (implied by Global Flood Database) for data collection, with application to spatial flood susceptibility analysis.
• Temporal Scale: Flood extent data from 2000 to 2018.

Methodology and Data

• Models used: Extreme Gradient Boosting (XGBoost), Decision Tree (DT), Random Forest (RF), Light Gradient Boosting Machine (LightGBM), Generalized Linear Model (GLM), and an ensemble voting model (integrating RF, XGBoost, LightGBM, DT, and GLM).
• Data sources: Flood extent data from the Global Flood Database (GFD). Ancillary spatial data related to climate, topography, hydrological, and land cover collected from multiple sources.

Main Results

• Individual machine learning models exhibited varying predictive performances: XGBoost (AUC = 0.985), Random Forest (AUC = 0.984), LightGBM (AUC = 0.982), Decision Tree (AUC = 0.972), and Generalized Linear Model (AUC = 0.879).
• The ensemble voting model significantly outperformed all individual algorithms, achieving the highest accuracy (AUC = 0.994).
• The ensemble approach improved mapping accuracy and increased reliability in identifying high-susceptibility areas.
• Advanced machine learning techniques, particularly ensemble frameworks, are highly effective tools for spatial flood susceptibility analysis and risk management.

Contributions

• Comprehensive evaluation and comparison of five diverse machine learning algorithms for flood susceptibility mapping.
• Demonstration of the superior performance of an ensemble voting model, achieving an AUC of 0.994, which significantly improved mapping accuracy and reliability compared to individual models.
• Highlighting the effectiveness of advanced machine learning techniques and ensemble frameworks as robust tools for spatial flood susceptibility analysis and risk management.

Funding

• This research received no external funding.

Citation

@article{Rahimi2026Integrating,
  author = {Rahimi, Mehdi and Malekmohammadi, Bahram and Firozjaei, Mohammad Karimi and Kerachian, Reza and Arsanjani, Jamal Jokar and Tan, Mou Leong and Awange, Joseph L. and Savić, Dragan and Duan, Qingyun and AghaKouchak, Amir},
  title = {Integrating geospatial intelligence and machine learning for flood susceptibility mapping},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-41014-3},
  url = {https://doi.org/10.1038/s41598-026-41014-3}
}

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