Development Scenarios and Wildfire Risk in Central Oregon

Risk from natural hazards occurs where areas vulnerable to natural hazards and development intersect. Development in vulnerable areas is a societal choice that has a lasting impact as our built environment is relatively permanent. Development patterns that contribute to increased wildfire risk mostly occur within the wildland urban interface, a land use type where development is at the fringes or intercept of wildland areas and is more flammable due to surrounding vegetation, slope, local climate, and other factors. To understand how past and future development potentials may impact fire risk in wildfire prone areas, a case study of Deschutes County, Oregon was conducted. In Chapter 1, past and future development trajectories were evaluating using data sets of land use and cover (NLCD), wildland urban interface (WUI; University of Wisconsin-Madison's SILVIS Lab), and surveys and interviews with local land use planners in Deschutes County, Oregon. Results from land cover and WUI data indicate the footprint of low-density, sprawling WUI grew more than 20% from 2000 to 2020 and as of 2020, over 80% of the development in the county was considered WUI. Interviews outlined three main mechanisms that would allow for further sprawling growth in the region, including demand for rural resorts, agricultural land conversion to developable land, and rural ADU additions to the landscape. Interviewees indicated that Oregon's land use planning program, which implements urban growth containment strategies as a way to preserve resource lands, was the main factor standing in the way of sprawling rural development, while other major constraints included water availability, wildfire risk, and insurance availability. The results suggested two major alternative scenarios are plausible depending on societal choices. The first scenario was characterized by compact development constrained by urban growth boundaries, indicative of Oregon's land use planning system; the other scenario characterized by unconstrained development where the three main mechanisms mentioned in the interviews would allow for more frequent low-density rural development. In fire-prone Deschutes County, the findings of Chapter 1 beg the question of whether different societal choices about development patterns are likely to lead to differences in wildfire risk. In Chapter 2, this question was addressed using a pilot version of the landscape simulation modeling platform called Envision. The agent-based simulation modeling software was used to model two potential future scenarios called compact and sprawling development, along with two wildfire regime scenarios, one representing historic fire patterns and another representing a more extreme fire future. Modeling experiments were used to learn if development patterns are likely to impact wildfire risk. Statistical analysis determined that the spatial pattern of development was not a significant indicator of risk; however, the extreme wildfire regime scenario that increased wildfires to 2-3 times more area than the historical regime significantly influenced the amount of risk to development indicating future risk is highly influenced by climate factors. The pilot version of Envision included a relatively new feature for wildfire simulation models -- the development of urban wildfire potential where urban areas can carry wildfires rather than be represented as unburnable. Although counterintuitive that development patterns would not influence the extent of risk, visualization and evaluation of simulation runs highlighted several spatial patterns of wildfire risk that help explain the result: (1) given that the majority of urban development didn't vary between the two scenarios, when wildfire entered densely populated areas the destruction across scenarios was similar and extreme; (2) defensible space and wildfire resilient construction materials may have limited wildfire spread and losses. Additionally, rural wildfires were more destructive under the sprawling scenario which had more rural development and the inclusion of prescribed fire as a mitigation strategy across scenarios in the model may have limited wildfire spread and losses in both scenarios. This research was used to inform updates to the pilot version of Envision for the National Science Foundation funded FireNET project and could help inform decisions by policy makers and land use planners in fire prone regions. Policies relating to land use planning and natural hazards can be used as a tool to address how we develop in vulnerable areas and may reduce the risk communities experience in similar areas.