Simultaneous modelling of stand volume yield, dominant height and basal area growth models

2001

S. 285-293

5 Tab., 27 Lit. Ang.

Unselbständiges Werk

200087278

Forest growth and yield models are often composed of a system of compatible, interdependent and analytically related equations rather than of a single equation. Most forest researchers have recognized that the forest stand dynamics system should be described by a system of simultaneous and interdependent equations. The two- and three-stage least square (2SLS and 3SLS) and seemingly unrelated regression (SUR) techniques from econometrics have been widely used to estimate the coefficients of forest growth and yield model system. This study analyzes a system of three interdependent equations for predicting the future stand volume yield, the future stand dominant height and the future stand basal area for eucalyptus plantations as an integrated system given by initial stand conditions. The coefficients of the system are estimated from Eucalyptus globulus Labill. stands in the Central inland of Portugal by using 3SLS, SUR and OLS, respectively. The three methods are evaluated and compared on some statistic indicators. The results indicate that there is a small difference between the three approaches in the particular case, but the system estimation methods perform better for a system of simultaneously interdependent equations in theory. Therefore, the appropriate system estimation approaches are recommended for estimating coefficients in simultaneously interdependent systems of forestry equations.