مجله جنگل ایران (Feb 2017)
Determining the most important environmental variables affecting on oriental beech (Fagus orientalis Lipsky.) site productivity using random forest technique in Khayroud forest, Nowshar
Abstract
Studying the relationship between site productivity, as a key indicator in forest management, and environmental variables is an important issue in forestry. In the past, studying such a relationship was done using linear regression model, but nowadays ensemble techniques like random forest are used. A stratified random sampling method based on landform was used to locate 114 0.1 ha circular sample plots in beech dominated forests. The height of five largest diameter beech trees within each plot was recorded as dominant height along with elevation, azimuth, and slope of the ground. In addition, at the center of plot, soil samples from first layer (0-10 cm) were taken for analyzing several soil variables. Evaluation of site productivity by using random forest technique showed that about 90% of variability in oriental beech productivity could be justified using environmental variables. Phosphorous and percentage of nitrogen, as edaphic variables, and slope and altitude, as physiographic variables, are among the most important variables affecting the beech forest productivity. Using criteria including R squared and adjusted R squared showed random forest technique had very good performance in assessing beech forest site productivity. These analyses underscore the importance of ensemble techniques in forest ecology studies and furnish framework for future modeling of beech forest productivity in relation to environmental predictors.
