Delayed conifer mortality after fuel reduction treatments: interactive effects of fuel, fire intensity, and bark beetles

Abstract

Abstract. Many low-elevation dry forests of the western United States contain more small trees and fewer large trees, more down woody debris, and less diverse and vigorous understory plant communities compared to conditions under historical fire regimes. These altered structural conditions may contribute to increased probability of unnaturally severe wildfires, susceptibility to uncharacteristic insect outbreaks, and drought-related mortality. Broad-scale fuel reduction and restoration treatments are proposed to promote stand development on trajectories toward more sustainable structures. Little research to date, however, has quantified the effects of these treatments on the ecosystem, especially delayed and latent tree mortality resulting directly or indirectly from treatments. In this paper, we explore complex hypotheses relating to the cascade of effects that influence ponderosa pine (Pinus ponderosa) and Douglas-fir (Pseudotsuga menziesii) mortality using structural equation modeling (SEM). We used annual census and plot data through six growing seasons after thinning and four growing seasons after burning from a replicated, operational-scale, completely randomized experiment conducted in northeastern Oregon, USA, as part of the national Fire and Fire Surrogate study. Treatments included thin, burn, thin followed by burn (thin þ burn), and control. Burn and thin þ burn treatments increased the proportion of dead trees while the proportion of dead trees declined or remained constant in thin and control units, although the density of dead trees was essentially unchanged with treatment. Most of the new mortality (96%) occurred within two years of treatment and was attributed to bark beetles. Bark beetle-caused tree mortality, while low overall, was greatest in thin þ burn treatments. SEM results indicate that the probability of mortality of large-diameter ponderosa pine from bark beetles and wood borers was directly related to surface fire severity and bole charring, which in turn depended on fire intensity, which was greater in units where thinning increased large woody fuels. These results have implications when deciding among management options for restoring ecosystem health in similar ponderosa pine and Douglas-fir forests.

Development of Vegetation and Surface Fuels Following Fire Hazard Reduction Treatment

“…Thirty-three percent of sites were treated with either pile or broadcast burning after harvesting in order to address natural and activity fuels. Burning reduces surface fuels but also produces direct and indirect tree mortality [44][45][46]; over time, fire-killed snags enter the surface fuel pool. Indeed, in burned sites, the number of snags in the most recently treated stands was 1.4-6.7 times that of the next age class, indicating that some trees killed during burning and still standing 2-4 years after treatment will become part of the surface fuel pool in the coming years.…”

Structural Equation Modeling: Theory and Applications in Forest Management

“…Laughlin and Abella [21] and Laughlin et al [22] applied SEM to an observational study of abiotic and biotic factors influencing plant community composition and species richness in a ponderosa pine (Pinus ponderosa P. and C. Lawson) forest ecosystem. Youngblood et al [23] studied the effects of experimental thinning and prescribed burning treatments on mortality of ponderosa pine with ANOVA and SEM. They found that SEM was able to represent and detect cascading effects of several firerelated factors on mortality, a process that the ANOVA model could not represent.…”

Modelling natural disturbances in forest ecosystems: a review

“…Models of such interactions mainly focused on descriptive, statistical approaches, including various logistic regression models (Bebi et al, 2003;Bigler et al, 2005;Breece et al, 2008), generalized linear models with different link functions (Peltonen, 1999;Eriksson et al, 2005;Hood and Bentz, 2007) and classification tree models (Kulakowski and Veblen, 2007). As an alternative approach for evaluating hypotheses and conceptual understanding about fire-bark beetle interactions, Youngblood et al (2009) demonstrated the utility of structural equation modelling. In a more process-oriented approach Seidl et al (2007) used a dynamically calculated estimate of drought-induced host tree stress to account for increasing tree susceptibility to I. typographus attack.…”

“…In this regard statistical modelling can provide insights on quantitative relationships for exploratory research questions. For example, structural equation modelling (e.g., Youngblood et al, 2009) or hierarchical Bayesian methods (e.g., McMahon et al, 2009) are particularly suitable for such tasks, allowing the consideration of simultaneous (and interacting) drivers as well as of non-Gaussian, nested and random effects. Furthermore, recent methodological advances have improved our inference abilities in working with the highly variable, incomplete and noisy characteristics of most disturbance datasets (e.g., machine learning algorithms such as random forests, genetic algorithms, and neural networks).…”