Refining Wildland-urban interface delineation: A building-level exposure assessment framework

Area of Science:

• Environmental Science
• Geospatial Analysis
• Wildfire Risk Management

Background:

• Conventional wildland-urban interface (WUI) mapping uses fixed thresholds, potentially misrepresenting localized wildfire risks.
• Existing methods may not fully capture building-specific fuel exposure in diverse landscapes.

Purpose of the Study:

• To develop and validate a novel method for WUI delineation using building-level wildfire exposure.
• To enhance spatial accuracy in identifying areas vulnerable to wildfire.

Main Methods:

• Employed two random forest models to assess building wildfire exposure status and intensity.
• Utilized Shapley Additive Explanations (SHAP) and partial dependence plots (PDPs) to identify key predictive factors and thresholds.
• Incorporated building density, Vegetation Fire Spread Index (VFS), and vegetation density as primary predictors.

Main Results:

• The proposed method identified a larger WUI area (5.25%) compared to the Federal Register-based framework (2.42%).
• A higher percentage of historically exposed buildings (24.48%) were included in the new WUI delineation versus the traditional method (18.65%).
• Building density and VFS were key for exposure status; vegetation density influenced relative intensity.

Conclusions:

• The building-level exposure framework offers a more precise spatial basis for wildfire risk assessment.
• This approach facilitates localized screening and differentiated mitigation strategies for wildfire threats.
• Linking building exposure to surrounding fuels improves wildfire preparedness in WUI areas.