The Wildland Urban Interface (WUI), i.e., the area where houses are in or near wildland vegetation, is widespread in the US and in other countries, and growing rapidly, but there is no global WUI hotspots map. Because the WUI is by definition the zone between urban areas and forests or shrublands, it has fallen between the cracks. That is unfortunate, because the WUI is where wildfire problems for society are worst, and where many other environmental problems are concentrated.
The goal of our project is to identify and characterize global hotspots of the Wildland Urban Interface. We will build on our strong record of accomplishment mapping the WUI in the US for federal agencies.
First, we will map potential WUI hotspots from 2011 to 2021 globally using coarse- and medium-resolution satellite data. We define a WUI hotspot as a 10-20,000 km2 landscape where WUI is either widespread, or growing rapidly, or both. Our 1st objective is to analyze multi-year VIIRS nighttime lights to identify 20 potential WUI hotspots globally. Our 2nd objective is to analyze multi-year Landsat and Sentinel-2 data to map change in impervious surface cover and wildland vegetation in these 20 hotspots.
Second, we will map the WUI in these hotspots based on high-resolution satellite data. Our 3rd objective is to analyze multiyear WorldView data with semantic segmentation to map building footprints in 10 of the hotspots from objective 2. Our 4th objective is to map WUI growth in these 10 hotspots based on the building footprints (obj. 3), and wildland vegetation (obj. 2),. Based on the results from objective 4, we and identify the causes of WUI growth, and we will measure impacts, such as the number of houses in the WUI, and rates of WUI growth.
Our proposal is highly responsive to the call, meets all requirements, and is well aligned with the goals and emphases of the solicitation. We will analyze multi-source satellite imagery, including high-resolution imagery from NASA’s archives, using recently developed methods to map building footprints from very-high-resolution satellite imagery (WorldView), together with medium-resolution satellite imagery (Landsat/Sentinel-2) to map wildland vegetation and imperviousness, and VIIRS nighttime lights for a global assessment of potential WUI areas. We will focus on WUI, a land use change with high societal impacts due to the escalating rate ofwildfires found there. And lastly, we will demonstrate that our methods are scalable by mapping 20 potential WUI hotspots with coarse- and medium-resolution satellite data, and 10 of them in detail with high-resolution imagery. In doing so, we will combine multi-source imagery so that their individual strengths contribute towards an integrated analysis that no single dataset could provide.