Modelagem não linear mista para descrever o afilamento de árvores clonais de Eucalyptus sp

Abstract

The aim of this work was to evaluate the quality of the logistic model related to taper in order to estimate the volume of the individual trees of Eucalyptus in a clonal plantation with different classes of age and diameter. For testing purposes, we used the non-linear mixed effects regression model, and the classes of diameter and age were considered as random effects of this process. The model considered three different scenarios: (i) fit of the model in its original format, considering diameter’s class within different age classes with random effect; (ii) fit of the model in its original format, considering diameter’s class within different age classes with random effect plus the addition of total height as covariate; and (iii) relativization of the original model for diameter and total height and by considering age as a random effect. In short, comparing the first two scenarios, the standard error was reduced by 40 %, ranging from 1.4727 to 0.8970 cm, and when comparisons were made between the second and third scenarios for volume estimation purposes, an increase of the model accuracy to 1.3 % with the application of the third scenario was observed; furthermore, improvements in the scatter plot of the observed versus predicted values were observed.