Wildland-Urban Interface (WUI) Mapping for Mainland Portugal Using Microsoft's Global Building Footprints Data

The expansion of the Wildland-Urban Interface (WUI) highlights the critical need for precise mapping to improve wildfire risk management. A key challenge, however, is the scarcity of high-resolution, nationwide building footprint data. To bridge this gap, we developed a semi-automated, multi-criteria filtering framework designed to enhance the quality of open-source global building datasets—specifically Microsoft’s Global Building Footprints (MSB)—for mainland Portugal. Our methodology combines regional adaptability with spatial analysis techniques, including area-based thresholds and proximity rules, using Portugal’s official Building Geographic Location Database (BGE) as a reference. To optimize residential representation, the framework iteratively removes non-residential outliers (e.g., industrial facilities, solar farms, transmission infrastructure) through dynamically adjusted thresholds applied across administrative levels (municipalities and NUTS-2 regions). As a result, the filtering process reduced the original dataset from approximately 5.6 million to 3.0 million building footprints. This dataset provides WUI maps for Mainland Portugal, generated using Microsoft’s Global Building Footprints. The geodatabase include WUI maps, original building footprints, and filtered versions for analysis. Our WUI maps are composed of 11 classes: Classification of WUI types: 1 - Intermix 2 - Interface Classification of building density in non-WUI areas: 3 - Very Low 4 - Low 5 - Medium 6 - High Classification of Land Cover: 200 - Agriculture 300 - Forest 400 - Shrubland 500 - Without Vegetation 600 - Water