Flood vulnerability mapping and urban sprawl suitability using FR, LR, and SVM models

Area of Science:

• Environmental Science
• Geographic Information Systems (GIS)
• Urban Planning

Background:

• Floods pose significant destructive risks to human life, property, and economic stability.
• Predicting flood-prone areas is challenging due to the dynamic nature of floods.
• Effective land use planning requires accurate flood vulnerability assessments.

Purpose of the Study:

• To create a flood vulnerability map for the Taif catchment area, Saudi Arabia.
• To develop a suitability map for future urban development based on flood risk.
• To compare the performance of bivariate, multivariate, and machine learning models in flood susceptibility mapping.

Main Methods:

• Utilized thirteen flood-contributing parameters for model development.
• Employed bivariate (FR), multivariate (LR), and machine learning (SVM) models.
• Integrated field surveys, historical data, RADAR (Sentinel-1A), and Google Earth imagery (2013-2020).
• Validated models using 70% of flood inventory data for training and 30% for testing.

Main Results:

• The Support Vector Machine (SVM) model demonstrated the highest accuracy at 96.2%.
• Flood susceptibility maps were classified into five zones: very low to very high.
• The final suitability map identified highly suitable areas for development in the east and northeast of the Taif Basin with low flood risk.

Conclusions:

• The SVM model provides an accurate flood susceptibility assessment for the Taif area.
• The developed suitability map aids in informed land use planning for urban development.
• This research offers crucial insights for engineers and authorities to mitigate flood hazards.