Modeling postfire mortality of ponderosa pine following a mixed-severity wildfire in the Black Hills: The role of tree morphology and direct fire effects

We examined the relationship among tree size, crown and stem damage, and 5 years of postfire mortality of ponderosa pine (Pinus ponderosa Dougl. ex. P. & C. Laws.) in the Black Hills following a large, mixed-severity wildfire. We measured tree morphology and direct fire effects on 963 trees and assessed individual tree mortality annually from 2001 to 2005. We used logistic regression to model tree mortality as a function of tree morphology (dbh) and bark thickness (BARK) and direct fire effects [percentage of the live crown scorched (PSCOR) and basal char measured as the percentage of the bole charred below 30 cm (CHAR)]. Models using dbh and BARK were modeled separately due to correlation between the variables. In all models, mortality decreased with increasing dbh and BARK and increased with increasing PSCOR and CHAR. Basal char contributed to the mortality of trees less than 40 cm but became less influential as dbh and BARK increased. Overall, probability of mortality modeled as a function of dbh, PSCOR, and CHAR correctly predicted the status of 78% of trees, whereas the model predicting mortality as a function of BARK, PSCOR, and CHAR had an increase in prediction accuracy of only 1%.