Wildfires in the Western US Rangelands: Spatial and Temporal Patterns and Impacts on Vegetation Structure

Abstract

Wildfires have become more frequent and larger across the western United States over the last few decades. Fire is a significant component of many rangeland ecosystems, and burn severity, which indicates the level of ecological change caused by fires, has received less attention, especially in large scale rangeland wildfire studies. The research of this dissertation is four-fold and it (1) examined the temporal patterns of large wildfires and their burn severity to assess wildfire trends across the western US rangelands, (2) evaluated the prefire vegetation structure of high severity wildfires to identify characteristics of vegetation at risk of high severity wildfires, (3) assessed the impacts of wildfires with different severities on post-fire vegetation structure to examine vegetation response to wildfires, and (4) explored the landscape-scale estimation and mapping of rangeland herbaceous biomass to assess fine fuel loads and predict fire behavior.

The results showed that the total land area burned in western US rangelands had significant increasing trends, while generally no significant trends in the proportion of area burned at different levels of burn severity (low, medium, and high). High woody plant cover was a key characteristic of prefire vegetation structure associated with high severity wildfires. Examinations of post-fire vegetation structure found that the impacts of wildfires on fractional cover were greater for woody than for herbaceous functional types. Moisture condition in the months prior to the fire led to larger decreases in herbaceous cover after the wildfires, likely due to increased fuel loads, while drier prefire conditions resulted in a greater decrease in woody cover. In estimating rangeland herbaceous biomass, the use of field data and high spatial resolution imagery in a machine learning model (random forest) resulted in more than 80% of the variance in biomass being explained. Fuel type and NDVI were the top predictors contributing to biomass estimation. The findings of these studies contributed to a better understanding of the trends and impacts of rangeland wildfires across the western US and provided insights for improving policy and management strategies to reduce the risk and impact of rangeland wildfires in the western US.