Evaluation and development of first order fire effects mortality model predictions for eastern hardwood forests

Abstract

The First Order Fire Effects Model (FOFEM) is a useful tool for predicting fire-induced mortality in western coniferous forests. However, it lacks robust models for eastern hardwood species. Managers in the eastern United States that are interested in using fire as a tool for forest and woodland management would benefit from the addition of eastern hardwood models to this software for planning purposes. This study determined the accuracy of the current FOFEM hardwood equations at predicting mortality at the stand and individual tree levels by comparing outputs from FOFEM 6.4 to mortality observed at 5 prescribed burn studies across the eastern US. FOFEM was found to overestimate mortality for all sites and all species based on predicted crown scorch and bark thickness. These same study datasets were then used to develop new mortality equations for certain species based on terrain, fire effects, and tree-level characteristics. New models for several species were developed. DBH was found to be a strong predictor for all three measurement periods, while slope and aspect were significant in the One-Year and Three-Year periods. The overall best models for all three periods included species and DBH as the best predictors. Receiver Operating Characteristic curve analysis demonstrated better discrimination of live versus dead trees according to the probability assigned by the new models than the current FOFEM model. DBH was found to be a better predictor than bark thickness, suggesting a need for new bark thickness models. These results show that there is room for improvement in the FOFEM model and identify potential variables that may provide for more accurate mortality modeling across eastern forests.