Site index curves in thinned and non-thinned eucalyptus stands

The aim of this study is to compare different non-linear models in order to classify the productive capacity of thinned and non-thinned eucalyptus stands by using the guide curve and longitudinal measurement modeling methods of the height of dominant and co-dominant trees, in different sample units and site potential. The Schumacher, Chapman-Richards, logical, Gompertz, and Bailey-Clutter models are tested in the guide curve method. These models are compared using statistics for residual standard error (Sy.x) and percentage (Sy.x%), Akaike information criteria, Bayesian information criteria, and the consistency observed between site index curves as well as the observed dispersion of the dominant heights. In addition, at the base without thinning, conformation of the data is observed using the Schumacher model based on dominant height groups at the reference age (84 months). Moreover, we conduct a comparison of the parameters based on the Z test. The Chapman-Richards model most effectively adjusts to the two data sets (i.e., those with and without thinning) and produces curves that best reflect the behavior of the dominant height variable. Thinning at the base does not affect dominant height growth, and no polymorphism is observed between the curves of the groups. The range of the site classes defined by the proposed method may be more suitable for differentiating the productive classes.