Dynamic base-age invariant site index models based on generalized algebraic difference approach for mixed Scots pine (Pinus sylvestris L.) and Oriental beech (Fagus orientalis Lipsky) stands

Abstract

Data from stem analysis of 397 dominant trees (198 Scots pine and 199 Oriental beech) were used to evaluate dynamic base-age invariant site index models derived from Bertalanffy–Richards, Hossfeld, and Lundqvist–Korf functions with the generalized algebraic difference approach (GADA) for mixed Scots pine (Pinus sylvestris L.) and Oriental beech (Fagus orientalis Lipsky) stands. These functions were compared with respect to residuals of these models; specifically, the evaluation criteria were bias (– root mean

square error (RMSE), absolute mean error (AME), adjusted coefficient of determination (R2

), and Akaike information criterion.

The best results were obtained with generalized algebraic difference equations derived from the base models of Bertalanffy–Richards for Oriental beech and Hossfeld for Scots pine. These selected models accounted for 95%–96% of the total variance in height–age relationships in dominant trees with bias of 0.049841 and 0.00171, RMSE of 1.55624 and 1.353736, AME of 0.940128 and 0.884034, and AIC of 723.55 and 1250.78 for Scots pine and Oriental beech, respectively. These dynamic base-age invariant site index models for the 2 tree species presented more effective and accurate polymorphic site index curves with multiple asymptotes than earlier site index models for Oriental beech and Scots pine. The important differences of height growth trend between developed base-age invariant site index models in this study and earlier site curves were determined by graphical comparisons for site index predictions. Therefore, the new dynamic base-age invariant site index models developed based on GADA methodology can be recommended for dominant height prediction and forest site quality evaluations in the mixed stands of these 2 species.