Developing effective wildfire risk mitigation plans for the wildland urban interface

•Demonstrate that wildfire mitigation planning is challenging in the WUI.•Important and interesting tradeoff solutions are invaluable for decision making processes.•Social vulnerability represents an important yet neglected aspect of wildfire risk.•Case study results for Santa Barbara, California, USA offer rich insights for mitigating wildfire risk. The wildland urban interface (WUI) is a transition zone between mostly undeveloped, vegetated lands and more densely populated areas with buildings and other infrastructure. This interface is of great interest and concern in wildfire prone areas like California, as significant increases in scale, frequency and intensity of fires have resulted in devastating impacts to life-safety, property and other high value assets. Ensuring that people, property and infrastructure are safe is a major priority, but will not happen by chance. Analytical tools, risk reduction strategies and operational plans to enhance wildfire resilience, particularly in the WUI, are vital. Examples of various fire resiliency approaches involve vegetation treatment such as prescribed burns, strategic fuel breaks, fuel thinning, grazing and mastication, while other risk mitigations may include structural hardening, creation of defensible space and other measures. A challenge is selecting the best areas for wildfire risk mitigation within the WUI, among many considerations, with viable projects required to be contiguous and manageable in size. Spatial optimization has much potential for informing planning and policy efforts, enabling the formalization of goals as well as offering approaches for identifying the best solutions possible. However, underlying geographical structure and spatio-temporal characteristics are formidable obstacles in problem solution. This paper highlights geographic analytics to support mitigation initiatives within the WUI, including the use of remote sensing, topography, climate, weather, wildfire behavior simulation, parcel data, structure and infrastructure information integrated using GeoAI along with a spatial optimization model that reflects the intent to identify the best project areas. The analysis area is comprised of millions of spatial raster cells, resulting in a model with decision variables corresponding to over 26 thousand land parcels. Application results for the Santa Barbara region are detailed, demonstrating the importance of spatial optimization combined with GeoAI in strategic coordination of scarce