Mapping the Wildland-Urban Interface in the United States Using Building Point Locations

Abstract:

The wildland-urban interface (WUI) refers to areas where human development intermingles or exists in close proximity to wildland vegetation. Due to this intermingling, the WUI is of particular concern for risks posed by the natural environment to urban populations (such as wildfires) or vice versa (such as habitat fragmentation and biodiversity loss). Previous work has mapped WUI extent in the conterminous United States using spatially continuous land cover and census data. In this study, we generated a new national WUI map based on locations of human-made structures, based on a Microsoft® dataset representing nearly 125 million individual building footprints. We used centroids of building footprints to classify WUI across the US based on point densities of buildings and pixel densities of wildland vegetation within varying moving-window sizes. Wildland vegetation was determined based on the National Land Cover Dataset. The structure-based method resulted in greater WUI extent compared to previous maps based on census data (increase of 40-60%, depending on moving-window size). However, overall patterns were similar between the two methods, with the greatest WUI extent mapped in the eastern US. The largest differences were in areas with low population densities and moderate agricultural cover, particularly in the Great Plains and Midwest, while the smallest differences were for the Northeast/New England. WUI classifications based on individual building locations offer advantages over census-based approaches in terms of greater spatial resolution, particularly in areas of low population density where census blocks may aggregate over large areas.