Xafoulis et al. (2026) Evaluation of a Computational Simulation Approach Combining GIS, 2D Hydraulic Software, and Deep Learning Technique for River Flood Extent Mapping

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

Identification

• Journal: Hydrology
• Year: 2026
• Date: 2026-01-09
• Authors: Nikolaos Xafoulis, Evangelia Farsirotou, Spyridon Kotsopoulos, Aris Psilovikos
• DOI: 10.3390/hydrology13010026

Research Groups

Not explicitly stated in the provided text. The study was implemented in the Enipeas River basin, located within the Thessalia River Basin District, Greece.

Short Summary

This study develops a computational simulation approach for rapid flood extent prediction by integrating Geographic Information Systems, 2D hydraulic modeling, and deep learning (U-Net CNN). Applied to the Enipeas River basin, the methodology demonstrates close spatial and quantitative agreement (differences below 8%) with traditional hydraulic simulations and official flood maps, proving its efficiency for flood risk mapping.

Objective

• To develop and validate a computational simulation approach for rapid flood extent prediction by integrating Geographic Information Systems, 2D hydraulic modeling, and deep learning techniques.

Study Configuration

• Spatial Scale: Enipeas River basin, located within the Thessalia River Basin District, Greece.
• Temporal Scale: Synthetic flood scenarios for a 1000-year return period event.

Methodology and Data

• Models used:
  • Hydrological analysis: HEC-HMS (version 4.12)
  • Hydraulic simulations: HEC-RAS 2D
  • Deep learning: U-Net (Convolutional Neural Network - CNN) architecture
• Data sources:
  • Open access geospatial data: Digital Elevation Model (DEM), slope, flow direction, stream centerline, land use.
  • Simulated flood extents from HEC-RAS 2D (used for training and validation).
  • Official Flood Risk Management Plan maps (used for validation).

Main Results

• The proposed methodology, integrating GIS, 2D hydraulic modeling, and deep learning, successfully predicts flood extents.
• Model outputs show close spatial and quantitative agreement with both 2D hydraulic simulations and official Flood Risk Management Plan maps.
• Flood extent area differences between the deep learning model outputs and reference data are consistently below 8%.
• The methodology is identified as a potential and efficient tool for rapid flood risk mapping.

Contributions

• Development of an integrated computational simulation approach combining GIS, 2D hydraulic modeling, and a U-Net deep learning architecture for flood extent prediction.
• Demonstration of the high accuracy and efficiency of the deep learning approach for rapid flood risk mapping, validated against both detailed hydraulic simulations and official flood risk maps.
• Provides a novel tool for quick assessment of flood risk, potentially aiding disaster management and planning.

Funding

Not explicitly stated in the provided text.

Citation

@article{Xafoulis2026Evaluation,
  author = {Xafoulis, Nikolaos and Farsirotou, Evangelia and Kotsopoulos, Spyridon and Psilovikos, Aris},
  title = {Evaluation of a Computational Simulation Approach Combining GIS, 2D Hydraulic Software, and Deep Learning Technique for River Flood Extent Mapping},
  journal = {Hydrology},
  year = {2026},
  doi = {10.3390/hydrology13010026},
  url = {https://doi.org/10.3390/hydrology13010026}
}