Fire-induced tree mortality in the mixed conifer forests of the Sierra Nevada, California.

摘要

This work explores how fire-caused damage to trees can be measured, assessed and used to predict future tree mortality. Chapter one deals with the use of long-term monitoring data for fire effects, with special attention given to problems of pseudoreplication. Chapter two describes a series of experiments on bark heat resistance in small-sized trees. Fire simulations demonstrated that bark thickness was the primary determinant of cambial heat resistance. We found only slight, but statistically significant, among species differences in bark thermal properties. We also determined small-sized trees to be more resistant to heating than expected from analytical models.; Chapter three details an experiment where we manipulated the importance of crown and bole damage in determining post-fire mortality. The treatments consisted of groups where (1) the basal bark was artificially thinned, (2) bole damage was inhibited by raking fuels away from the base of the stem, and (3) an untreated control. Mortality rates and mortality risk factors were different among the treatments. Trees that received the raking treatment had almost zero post-fire mortality. Trees with thinned bark suffered relatively high rates of mortality, and death was best predicted by the severity of bole scorch, bole diameter, and crown scorch. Untreated control trees suffered high rates of mortality, but death was associated only with crown scorch.; Chapter four describes a new approach to modeling tree mortality. Death in unburned stands is usually modeled as a function of growth rate, while death in burned stands is modeled as a function of crown scorch. We sought to narrow this conceptual gap by determining (1) whether growth rate, in addition to crown scorch, is a significant predictor of tree death in burned stands, and (2) whether a single, simple model could predict tree death in both burned and unburned stands. For white fir (Abies concolor ) we found that radial growth rate is a significant predictor of post-fire survival, particularly for trees sustaining low to moderate amounts of crown scorch. We applied our model of tree death in burned stands, with crown scorch set to zero, but were only moderately successful in predicting unburned tree mortality.