A long history of fire suppression by federal land management agencies has interrupted fire regimes in much of the western United States. Many forest types that historically burned frequently have undergone significant changes in species composition and have heavy accumulations of surface and canopy fuels. Fuel quantity and flammability are important local predictors of fire severity. The climate system operates at both broad and fine spatial and temporal scales to favor conditions that increase fuel loading through biomass accumulation and accelerate drying of fuels; and maintain active fires under favorable concurrent atmospheric conditions. Observed increases in large fire occurrence and area burned in recent decades are explained by warmer, drier, and longer growing season conditions in the West. There has not yet been a large-scale study that examines patterns and controls of high severity fire in the western US. We use a 30 year record of fire severity to identify the controls of high severity fire across the western US, develop statistical probability models for high severity fire occurrence and area burned, and examine the impacts of climate change on high severity fire risk. In examining topography, vegetation and fire-year climate as predictors we found that inclusion of both vegetation and fire-year climate predictors was critical for identifying fires with high fractional fire severity and capturing inter-annual variation in high severity fire occurrence. While a single, west-wide model was able to predict high severity fire occurrence with some accuracy, it was necessary to develop regional models to accurately predict high severity area burned for forests in extreme fire years. A simple generalized Pareto distribution model with maximum temperature the month of fire, annual normalized moisture deficit and location explains forest high severity area burned in a west-wide model, with the exception of years with especially large areas burned with high severity fire: 1988, 2002. With respect to mitigation or management of high severity fire, understanding what drives extreme fire years is critical. For the Northern Rocky Mountains, Sierra Nevada Mountains, and Southwest forests, topography, spring temperature and snowpack condition, and vegetation condition class variables improved our prediction of high severity burned area in extreme fire years. We used the models developed for the Northern Rocky Mountains to examine how fractional area of high severity fire will change with climate. Application of output from global circulation models to large fire occurrence and size models in the Greater Yellowstone Ecosystem indicates that climate conditions by mid-century will result in an increase in the frequency of large fire events and area burned. We applied GCM output to a set of probabilistic models for high severity fire occurrence and burned area for the Greater Yellowstone Ecosystem. We found that fraction of high severity burned area increases to levels by mid-century that are three times greater than a 1961-1990 reference period. These potential changes in high severity area burned and frequency of occurrence may result in changes to species composition in these high elevation forests. If a goal of management is to mitigate extreme fire events in terms of fire severity, we would conclude that knowledge of fire year climate is essential. All of the models we developed predict high severity fire occurrence and area burned with reasonable accuracy in all years when fire year climate and vegetation predictors are included. The inclusion of fire-year climate variables allows these models to forecast inter-annual variability in areas at future risk of high severity fire, beyond what slower-changing fuel conditions alone can accomplish. This allows for more targeted land management, including resource allocation for fuels reduction treatments to decrease the risk of high severity fire. Models like this will be important tools for assessing interactions between changing climate and fuel profiles under a diverse menu of future climate and management scenarios.

The escalating frequency and severity of wildfires in California have precipitated substantial economic losses and social strains, underscoring the imperative to comprehend the dimensions of wildfire management. This dissertation amalgamates three pivotal research endeavors focusing on different dimensions of wildfire risk: perceptions among wildland-urban interface residents, predictions of wildfire burn severity, and visualization of uncertainty. The surge in wildfires, coupled with the increasing population moving to Wildland-Urban Interface (WUI) areas, highlights the urgency of understanding both the physical and human dimensions of wildfire risk management. While various management practices involving communities have emerged as favored solutions, barriers to their implementation persist. Understanding public attitudes and perceptions regarding these practices is essential for successful fire management efforts.Furthermore, the warming climate and increasing fuel loads due to fire exclusion, compounded by climate change and drought, have led to more frequent, extensive, and severe wildfires. Burn severity, a metric that measures the ecological impact of fire on vegetation, is crucial for post-fire management. The Composite Burn Index (CBI) emerges as a preferred method of characterizing fire effects due to its comprehensive approach, offering a systematic and visually intuitive estimation of ecological impacts following a fire. Effective wildfire severity and risk information to stakeholders is paramount for enhancing understanding and promoting resilience. However, conveying complex information poses significant challenges. Visualization techniques play a vital role in conveying risk information and aiding comprehension of complex wildfire-related information. This research introduces a scalable visualization model for a use in predicting and managing the complex dynamics of wildfire occurrences. This dissertation advances our understanding of wildfire management by elucidating the complex interplay between public perceptions, burn severity estimation, and risk visualization. By integrating social perspectives with empirical modeling and visualization techniques, it offers multifaceted approach to addressing the challenges posed by wildfires in California. The insights garnered from this dissertation are crucial for informing policy decisions, guiding mitigation efforts, and fostering community resilience in the face of escalating wildfire threats.