A multi‐scale assessment of interior Douglas‐fir tree mortality for hazard and risk assessments

Abstract Land managers use hazard (susceptibility) and risk rating systems to guide the application of forest management treatments that aim to reduce future damages to forests. Rating systems are typically designed for individual damage agents, but tree mortality often results from multiple agents without a clear proximate cause. In interior Douglas‐fir (DF, Pseudotsuga menziesii var. glauca ) forests in the Northern Rocky Mountains, USA, multiple damage agents are commonly associated with DF tree mortality, such as insects, disease, weather, and fire. We investigated how recent DF tree mortality from insects and diseases (excluding fire and harvest) shifted stand structure and composition in DF forests and was influenced by susceptibility (e.g., stand structure and composition, topography, and spatial variability in climate) and risk (biotic agent pressure). Our multi‐scale analysis used 884 plots remeasured after 10 years from the USDA Forest Inventory and Analysis program with support from spatial datasets. Across a large, forested landscape, 60% of the plots had no new DF mortality, and most plots (80%) experienced mortality 25% loss of DF basal area, occurred in 6% of plots. Most of the dead DF trees (68%) were smaller diameter (12.7–29 cm at breast height) and mortality rates of smaller trees were significantly greater than those of larger diameter trees (>29 cm), a finding consistent with natural stand development processes. During the remeasurement period, average DF tree size increased in most plots (80%) and <4% of DF‐dominated plots with severe mortality shifted in dominance from DF to another tree species. Greater DF mortality (percentage of initial DF basal area) was associated with lower tree growth rates, larger average tree sizes, greater availability of DF tree hosts for biotic agents, cooler and wetter topo‐climatic locations, and higher Douglas‐fir beetle population pressure. The relative importance of each variable differed west and east of the Continental Divide. By identifying thresholds in susceptibility and risk variables associated with higher DF tree mortality, our results support adaptive forest management for multiple damage agents, when the goal is to reduce tree mortality of a widespread and abundant conifer.