Idowu et al. (2026) National-Scale Flood Susceptibility Mapping of Nigeria Using Statistical and Machine Learning Models with Satellite-Driven Validation for Data-Sparse Environments

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

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-09-22
• Authors: Dorcas Idowu, Jessica Boakye, Wendy Zhou
• DOI: 10.3390/rs18193264

Research Groups

• Department of Geosciences, University of Ibadan
• Nigerian Meteorological Agency (NIMET)
• National Space Research Development Agency (NASRDA)

Short Summary

This study evaluates national-scale flood susceptibility in Nigeria using four machine learning models and satellite-informed data, demonstrating the value of combining statistical and machine learning approaches for accurate flood assessment.

Objective

• Investigate the applicability of frequency ratio, logistic regression, Random Forest, and gradient boosting (XGBoost) models for nationwide flood susceptibility assessment in Nigeria

Study Configuration

• Spatial Scale: National scale, covering the entire territory of Nigeria
• Temporal Scale: 100-year floodplain simulation using historical discharge data from Dartmouth Flood Observatory

Methodology and Data

• Models used: Frequency ratio (FR), logistic regression (LR), Random Forest (RF), gradient boosting (XGBoost)
• Data sources: HEC-RAS 100-year floodplain simulation, Dartmouth Flood Observatory satellite-derived discharge, Sentinel-1 SAR flood extent data

Main Results

• XGBoost achieved the highest AUC (0.956) and overall accuracy (0.892) on the held-out test subset
• HAND ranked highest overall across importance analyses
• Consistently high flood-class detection across all four susceptibility models (93.81–94.50%) against independent Sentinel-1 SAR flood extent data

Contributions

• This study provides a nationwide assessment of flood susceptibility in Nigeria using satellite-informed hydrodynamic data and machine learning approaches, addressing the scarcity of spatially explicit flood information in the region.
• The results demonstrate the value of combining statistical and machine learning methods for accurate flood assessment.

Funding

• This research was funded by the Nigerian Federal Ministry of Environment (Project Code: NME/2020/001)
• Supported by the World Bank's Nigeria Erosion, Watershed Management Project (NEWMAP)

Citation

@article{Idowu2026NationalScale,
  author = {Idowu, Dorcas and Boakye, Jessica and Zhou, Wendy},
  title = {National-Scale Flood Susceptibility Mapping of Nigeria Using Statistical and Machine Learning Models with Satellite-Driven Validation for Data-Sparse Environments},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18193264},
  url = {https://doi.org/10.3390/rs18193264}
}