Evaluating multi-scaled patterns, trends, and drivers of fire severity to inform adaptive management of western North American forests

Abstract

In the last several decades, profound changes in wildfire activity influenced by climate change, exclusion of Indigenous stewardship, and changes in land management have had enormous impacts on forest ecosystems across much of western North America. Of particular interest to managers and ecologists are observed shifts in fire severity (i.e., the magnitude of ecological change caused by fire, typically measured by vegetation killed), which have important implications for carbon storage and sequestration, wildlife habitat, economic and cultural resources, and post-fire forest regeneration under a warming climate. To develop management strategies aimed at adapting forests to changing wildfire regimes under climate change, information about both the causes and consequences of shifting fire severity patterns is critical. In this dissertation, I used remote sensing datasets to evaluate drivers, trends, and spatial patterns of fire severity from the scale of individual burn patches to ecoregions that encompass hundreds of recent wildfires in California, Oregon, and Washington. In Chapter 1, I evaluated trends, spatial patterns, and cumulative impacts of fire severity across a range of unburned, low-moderate, and high severity burn patches to evaluate implications for post-fire management in recent California wildfires. I found that a small number of large fires disproportionately contributed to increases in high severity effects, but simultaneously presented an opportunity for managers to use areas burned with low and moderate severities as an initial “treatment” to be followed by additional post-fire fuel treatments. Chapter 2 of this dissertation examined the drivers of changing severity patterns, with a focus on understanding how existing management practices influence fire effects. I found that land ownership designation as a proxy for general management practices was a strong predictor of fire severity across the study area along with other ‘bottom-up’ controls such as pre-fire forest structure. Finally, in Chapter 3, I evaluated patterns and drivers of fire severity to inform adaptive management strategies within the Northwest Forest Plan, which directs the management of federal forests across nearly 10 million ha within the study area. I found that historically frequent-fire dry forest types – particularly mixed conifer forests, mixed evergreen forests, oak woodlands, and pine-oak woodlands – have faced the greatest impacts in terms of stand-replacing fire. Although fire weather was the most important predictor of fire severity at the regional scale, factors such as pre-fire forest cover and topography exerted strong local controls. Collectively, this dissertation work contributes to enhanced understanding of adaptive forest management strategies for both pre-fire and post-fire forest conditions in increasingly fire-prone landscapes of western North America.