Search

This paper presents a model to predict annual wildfire risk in pure and evenaged eucalypt stands in Portugal. Emphasis was in developing a management- oriented model, i.e., a model that might both: (a) help assess wildfire occurrence probability as a function of readily available forest inventory data; and (b) help predict the effects of management options (e.g., silvicultural treatments) on the risk of fire in eucalypt stands. Data from both the 1995/1998 and the 2005/2006 Portuguese National Forest Inventories as well as wildfire perimeters' data were used for modeling purposes. Specifically, this research considered 1122 inventory plots with approximately 1.2 million trees and 85 wildfire perimeters. The model to predict the probability of wildfire occurrence is a logistic function of measurable and controllable biometric and environmental variables. Results showed that wildfire occurrence probability in a stand increases with the ratio basal area/quadratic mean diameter and with the shrubs biomass load, while it decreases with stand dominant height. They further showed that the probability of wildfire occurrence is higher in stands that are over 1 Km distant from roads. These results are instrumental for assessing the impact of forest management options on wildfire risk levels thus helping forest managers develop plans that may mitigate wildfire impacts. [ABSTRACT FROM AUTHOR]