Monday, 17 November 2003
Analysis of algorithms for predicting canopy fuel
We compared observed canopy fuel characteristics with those predicted by existing biomass algorithms. We specifically examined the accuracy of the biomass equations developed by Brown. The following variables were used to calculate canopy fuel characteristics: DBH, height of tree, crown ratio and crown length. We examined the accuracy of the algorithms by computing crown fuel biomass from 5 different study areas by destructively sampling each study site. In each of the 5 study areas the biomass was computed by destructively sampling selected branches. The biomass of individual size classes was calculated and standard regression techniques were used to predict the biomass for each size class for the entire stand. These measurements were used to obtain estimates of crown fuel characteristics of the entire stand. We also calculated crown fuel estimates using Brown’s equations for the entire stand. Summary statistics were used to compare the accuracy of Brown’s equations to our predictions. Comparisons were made for individual species in our study sites such as ponderosa pine, Douglas fir, and lodgepole pine. In addition, we observed the appropriateness of using a similar species to predict canopy fuel characteristics when the actual species is not accounted for using Brown’s equations. For example, we used western red cedar in place of incense cedar and grand fir instead of white fir. We also evaluated the importance of tree dominance as a predictor variable of crown biomass. Adjustments were made to Brown’s equations in order to improve the predictability of the equations for future use.