Probabilistic, Multivariable Flood Loss Modeling on the Mesoscale with BT-FLEMO.

Area of Science:

• Hydrology and Water Resources
• Environmental Risk Assessment
• Computational Modeling

Background:

• Flood loss modeling is crucial for risk management but often relies on deterministic methods with high uncertainty.
• Existing depth-damage functions and multivariable models struggle to accurately capture complex flood damage processes.
• Quantifying uncertainty in flood loss estimation is essential for effective decision-making.

Purpose of the Study:

• To develop a probabilistic, multivariable model for improved flood loss estimation in residential buildings.
• To introduce the Bagging decision Tree Flood Loss Estimation Model (BT-FLEMO) for quantifying prediction uncertainty.
• To enhance flood risk analyses and decision support systems.

Main Methods:

• Developed a probabilistic, multivariable Bagging decision Tree Flood Loss Estimation Model (BT-FLEMO).
• Applied and validated BT-FLEMO in 19 municipalities affected by the 2002 River Mulde flood in Saxony, Germany.
• Compared BT-FLEMO against six deterministic loss models and official loss data.

Main Results:

• BT-FLEMO demonstrated superior model accuracy compared to deterministic, univariable, and multivariable models.
• The model provides a probability distribution of estimated losses, quantifying prediction uncertainty.
• Despite high uncertainty, BT-FLEMO's probabilistic output effectively represents the range of estimates from other models.

Conclusions:

• BT-FLEMO offers a significant advancement in flood loss estimation by providing probabilistic outputs and quantifying uncertainty.
• The model enhances the reliability of flood risk assessments and supports better-informed management decisions.
• Probabilistic approaches like BT-FLEMO are vital for addressing the inherent uncertainties in flood loss modeling.