Prescribed fire mortality of Sierra Nevada mixed conifer tree species: effects of crown damage and forest floor combustion

Logistic regression equations of prescribed fire mortality were developed for white fir (Abies concolor [Gord. and Glend.] Lindl.), sugar pine (Pinus lambertiana Dougl.), ponderosa pine (Pinus ponderosa Laws.), incense-cedar (Calocedrus decurrens [Torr.] Floren.), and giant sequoia (Sequoiadendron giganteum [Lindley] Buchholz) in the southern Sierra Nevada, California. A total of 1025 trees were analyzed in this study. Variables included in the mortality equations are diameter at breast height, percent crown volume scorched, crown scorch height, and local forest floor consumption. The likelihood ratio 7,2 was highly significant (P < 0.0001) for all models developed and the receiver operating curve statistic ranged from 0.736 to 0.997 indicating good overall model performance. None of the logistic regression models developed for California black oak (Quercus kelloggii Newb.) produced a maximum likelihood estimate indicating that the variables measured were not significant predictors of mortality. Probability of death is lower for giant sequoia, incense-cedar, and ponderosa pine at high levels of percent crown volume scorched when compared to sugar pine and white fir. Forest floor consumption was a significant factor in the majority of the models developed indicating that mortality is not solely a function of crown damage. These models may be useful to forest managers planning prescribed fires and to ecologists interested in modeling the effects of fire on mixed conifer forests. (C) 2002 Elsevier Science B.V. All rights reserved.