Very large wildfires in the western contiguous United States: probabilistic models for historical and future conditions
Abstract
Wildfires, especially the largest ones, can have lasting ecological and social effects both directly on the landscape and indirectly on the atmosphere and climate. Both climate and fire regimes are expected to change into the future while air quality, the composition of the atmosphere, continues to be regulated. It is necessary to understand how climate, wildfire, and air quality interact to mitigate air quality. Existing studies, however, span spatial and temporal scales necessary for only linking two components at a time (e.g. climate and wildfire or wildfire and air quality). Appropriate scales of data and modeling are required to integrate all three components and understand the system as a whole. To lay the foundation for studying interactions among these three components, I investigated the relationship between climate and very large wildfires, here defined as megafires (≥ 50,000 ac ~ 20,234 ha), at spatial and temporal scales appropriate for future work to bridge results into air quality modeling. In this dissertation, I demonstrated, using a systematic approach, that broad spatial and fine temporal resolutions are the best scales by which to understand how climate, wildfire, and air quality interact. Thus, using broad wildfire data aggregated to the spatial scale of eight US National Interagency Fire Center Geographic Area Coordination Centers (GACCs) across the western contiguous US, and daily and monthly climate data, I developed logistic regression models to predict the probability that a megafire will occur in a given week. Significant climate predictors of megafires vary by GACC and are similar to those found by other studies for aggregate annual area burned. Thus megafires may influence the analysis of aggregate statistics substantially. For all eight GACCs, projecting these models showed a significant (p≤0.05) difference between the historical period from 1979 to 2010 and Intergovernmental Panel on Climate Change future scenarios, representative concentration pathways (RCPs) 4.5 and 8.5, during 2031 to 2060. Generally, with the exception of the Southwest and Northern California, megafires will be more likely both throughout the fire season and from year to year, with more pronounced patterns under RCP 8.5 than RCP 4.5. This work investigates the effects of a changing climate on megafires at scales that can aid policy and management to mitigate their effects. It also provides a foundation by which to improve understanding of the climate and carbon systems. Lastly, it illuminates the need to investigate how fire statistics are aggregated and how this affects climate associations.
Collections
- Forestry [391]