Modeling Long-Term Fire-Caused Mortality of Douglas-Fir

Abstract

Mortality was determined in a stand of Douglas-fir 8 years after 20 plots were treated with light surface fires. Logistic regression was used to model long-term mortality as functions of morphological variables measured shortly after burning. Independent variables were diameter at breast height, height of needle scorch, percentage of the prefire crown volume scorched, season of burn, and the number of quadrants with dead cambium at 1.4 m bole height. Mortality increased with increasing scorch height, percent crown scorch, and dead cambium. It decreased with larger diameter. The best predictor of mortality was the number of quadrants with dead cambium. Percentage of crown volume scorched was a better predictor than lethal scorch height. For a given level of damage, mortality following fall season fires was slightly higher than following spring fires. Models may be used in planning prescribed fires and for salvaging fire-damaged Douglas-fir. For. Sci. 34(1):190-199.