Heterogeneity in forest structure prior to restoration by fire: Multi-scale methods, measurements, and fire modeling from extensive field plots in the mixed conifer forest of the southern Sierra Nevada

Heterogeneity plays a key role in ecosystem function and quantifying it can be critical for understanding ecosystem processes. During a century of fire suppression, however, forests may have become more homogeneous. In many western forests the natural disturbance of fire is being reintroduced yet the effects are not well understood. As a result, the best method of reintroducing fire to forested ecosystems has been debated. Should fire be introduced directly, with little pre-treatment of fuels, or should mechanical treatment precede the reintroduction of fire? To assess the effects of disturbance on structure we must quantify heterogeneity in the forest before and after disturbance occurs.; In this dissertation, I explored means to assess heterogeneity and measured it in the mixed conifer forest and in the litter base that vectors fire. In chapter two, I present a new statistical metric for describing the distribution of forest structure---Menning's Departure Index (MeDI). MeDI can be used to compare any test distribution with any similarly measured reference distribution to determine the direction, magnitude and location of departure from one to the other.; The metric also can be used to assess forest structural heterogeneity at different spatial scales. The standard deviation of MeDI from a cluster of plots is used to measure variability in forest structure. This method is used in chapter three to test differences between two dominant slope aspects in the watershed. The distribution of forest structure between the aspects was not found to be significantly different. This led to revised hypotheses that insolation differences between dominant aspects might affect structural variability at the water-stressed, lower-elevation margins of the forest or create patchiness in community distribution.; Ground and surface fuels are essential for vectoring a contagious disturbance like fire. Fire models often assume uniform fuel characteristics for a forest community such as the mixed conifer forest. In chapter four, I explored how a single factor, litter bulk density, could vary by dominant tree species. I modeled how variability in litter bulk density could affect fire behavior and found significant differences in fireline intensity and rate of spread among different dominant tree species.