Liu et al. (2025) AI-Driven GIS Modeling of Future Flood Risk and Susceptibility for Typhoon Krathon under Climate Change

Identification

• Journal: Computer Modeling in Engineering & Sciences
• Year: 2025
• Date: 2025-01-01
• Authors: Chih‐Yu Liu, Cheng‐Yu Ku
• DOI: 10.32604/cmes.2025.070663

Research Groups

• Department of Harbor and River Engineering, National Taiwan Ocean University, Taiwan
• Center of Excellence for Ocean Engineering, National Taiwan Ocean University, Taiwan
• Keelung City Fire Department, Keelung City Government, Taiwan

Short Summary

This study develops a Random Forest (RF)-based GIS model to assess flood susceptibility in Keelung City using data from Typhoon Krathon (2024) and projects future risks under IPCC AR5 RCP8.5 climate scenarios.

Objective

• To evaluate current flood susceptibility in Keelung City and simulate future flood risk increases resulting from projected climate-driven changes in precipitation intensity.

Study Configuration

• Spatial Scale: Keelung City, Taiwan.
• Temporal Scale: 2024 (Typhoon Krathon event) and future projections based on IPCC AR5 RCP8.5 scenarios.

Methodology and Data

• Models used: Random Forest (RF) algorithm.
• Data sources:
  • Geospatial/Environmental variables: Average elevation, slope, Topographic Wetness Index (TWI), Normalized Difference Vegetation Index (NDVI), flow accumulation, and drainage density.
  • Hydrometeorological data: Frequency of cumulative rainfall threshold exceedance (collected during Typhoon Krathon).
  • Climate projections: IPCC AR5 RCP8.5 scenarios (+2°C to +4°C warming).

Main Results

• Model Performance: The RF model achieved a high accuracy of 97.45%.
• Feature Importance: NDVI was identified as the most critical predictor of flood susceptibility, followed by flow accumulation, TWI, and rainfall frequency.
• Climate Impact: Under the RCP8.5 scenario, the 50-year return period rainfall in Keelung City is projected to increase by 42.40% to 64.95% under a warming of +2°C to +4°C.
• Future Risk: Simulation results indicate that two specific districts in Keelung City will experience the most significant increase in flood risk.

Contributions

• Integrates real-time typhoon event data with AI-driven GIS modeling to provide a high-accuracy susceptibility map.
• Bridges the gap between current flood observations and future climate projections (RCP8.5) to identify specific urban areas requiring targeted climate adaptation.

Funding

• Not specified in the provided text.

Citation

@article{Liu2025AIDriven,
  author = {Liu, Chih‐Yu and Ku, Cheng‐Yu and Tsai, Ming-Han and You, Jia-Yi},
  title = {AI-Driven GIS Modeling of Future Flood Risk and Susceptibility for Typhoon Krathon under Climate Change},
  journal = {Computer Modeling in Engineering & Sciences},
  year = {2025},
  doi = {10.32604/cmes.2025.070663},
  url = {https://doi.org/10.32604/cmes.2025.070663}
}