Relationship between particular dendrobiometrical indicators of natural European beech (Fagus sylvatica L.) dendrocenoses in Central Balkan Range

In parallel studies, different regression models were tested to identify relationships between particular dendrobiometrical indicators on two sample plots representing forests dominated by the European beech in the Central Balkan Range (Bulgaria). The presence of incomplete multicollinearity was studied through correlation matrix for factor variables. To avoid multicollinear negative impact, step multiple regression was applied and adequate regression equations of the relationships under consideration were formulated. The results of statistical analysis confirmed that the link between the investigated indicators is strong and that the ’cloud‘ data show some ’sphericity‘ and distribution close to normal. In one of the sample plots, one major volume-forming factor – height does not participate in the obtained regression equation, so it is not possible to estimate its influence. By testing linear and several nonlinear regression dependencies and by mediating widely used statistical criterions for model selection, the optimal linear model of the considered link was chosen.