Generalização da prescrição de manejo por classe diamétrica para floresta heterogênea multiânea

Abstract

The practice of sustainable forest management involves many challenges. Since the management prescription is based on the modeling of the forest's diameter structure, it is expected that the higher the accuracy of the function fitting, the greater the accuracy of the management prescriptions. Thus, the objectives of this work were (i) to test the performance of nonlinear models in the fitting to the diametric distribution of an unequal forest; and (ii) propose a generalization of the BDq method aiming at a better management prescription, which guarantees the sustainability of the activity. Five models were tested, among them the Meyer model (most commonly used) and the Biexponential model. The comparison of the fittings was performed using the residual standard error, Akaike information criterion, chi-squared test and scatter plots of residuals distribution. The Biexponential model presented better performance, with a mean relative error of 7.13%, while the Meyer model presented a 70.08% relative error. The cut prescriptions of the two models presented differences, since the Meyer model allows greater exploration of the larger diameter classes while the Biexponential model allows for a greater exploration of the initial classes. Excessive exploitation of larger diameter classes may impair continuous production and sustainability of the activity.